Abbott SG Wtms

COPYRIGHT AND CITATION CONSIDERATIONS FOR THIS THESIS/ DISSERTATION
o Attribution — You must give appropriate credit, provide a link to the license, and indicate if
changes were made. You may do so in any reasonable manner, but not in any way that
suggests the licensor endorses you or your use.
o NonCommercial — You may not use the material for commercial purposes.
o ShareAlike — If you remix, transform, or build upon the material, you must distribute your
contributions under the same license as the original.
How to cite this thesis
Surname, Initial(s). (2012). Title of the thesis or dissertation (Doctoral Thesis / Master’s
Dissertation). Johannesburg: University of Johannesburg. Available from:
http://hdl.handle.net/102000/0002 (Accessed: 22 August 2017).
South African Healthcare and the
Fourth Industrial Revolution:
New Applications of Technology
Simon George Abbott
A thesis submitted for the degree of
Doctor of Philosophy
in
Information Technology Management
Department of Applied Information Systems

College of Business and Economics
University of Johannesburg
Supervisor: Professor Abejide Ade-Ibijola
March, 2022
Simon Abbott. 2022.
South African Healthcare and the Fourth Industrial Revolution: New Appli-
cations of Technology
Copyright © University of Johannesburg, Johannesburg, South Africa
All rights reserved. No part of this publication may be stored in a

retrieval system, transmitted, or reproduced, in any form or by any
means, including but not limited to photocopy, photograph, magnetic
or other record, without prior agreement and written permission of
the copyright holder.
supervisor:
Professor Abejide Ade-Ibijola
supported by:
The Supervisor-Linked Bursary (2018 to 2022), University of the Jo-
hannesburg.
D E D I C AT I O N
This work is dedicated to my sister for her continuous support.
iii
D E C L A R AT I O N
I, Simon G Abbott, hereby declare the contents of this doctoral the-

sis to be my own work. This thesis is submitted for the degree of
Doctor of Philosophy in Information Technology Management at the
University of Johannesburg, Johannesburg. This work has not been
submitted for any other degree.
Author’s signature.
March 2022
ABSTRACT
T
he advancements of the Fourth Industrial Revolution within
Healthcare are ever present, impacting the South African (SA)
healthcare system.
Introduction
Health technology and the Fourth Industrial Revolution (4IR), are
advancing globally, being propagated by the World Health Organi-
sation (WHO), global eHealth strategy, recently launched. This has
resulted in the global transformation of healthcare, towards Digital
Health 4.0. Healthcare is currently experiencing an unprecedented
modernisation through the development of digital technologies. The
South African National Department of Health (NDoH) has recently
launched its own local version of an ‘eHealth Strategy’, engaging the
Fourth Industrial Revolution (4IR). We endeavour to present a few
of these technologies that are, being designed for the healthcare sys-
tem thus advancing digital healthcare through Artificial Intelligence
(AI) technologies, and solving some clinical value based healthcare,
through these designs.
Problem statement
We present the challenges faced by the South African Healthcare land-
scape, by identifying the need for ICT technological solutions, within
the healthcare industry. We further include the adoption of modern
software applications and information technology in addressing the
4IR, eHealth and the application of AI technologies in solving cer-
tain healthcare problems, within the SA healthcare sectors. This in-
cludes the convergence of health and technology in producing such
improved value based health outcomes for all, within the SA health-
care system.
Methodology
This thesis presents five artefacts, as 4IR technological contributions to
the South African healthcare system, namely: The MoonBoot, Med-
Bot, Decryption tool, Anonymisation tool and Trend Analysis tool.
The design of these artefacts is based on the Design Science Research
(DSR) methodology — which is a structured sequential order, address-
ing a specific problem. It follows an agile method of structured or-
der with each iteration. Thus, modifying the problem formulation,
through each cycle, within the DSR cycle, in reaching the final out-
comes based design of the artefact. The designs of these tools were
validated through evaluations, directed at a specialist focus groups of
practitioners, within different specialised fields of healthcare.
vii
Results
The designs of the five healthcare tools are presented including the
corresponding survey results. The inferences and validations of these
five tools are presented. The results reveal the modern day practi-
tioner acknowledging the presented 4IR technologies, within the SA
healthcare system.
Conclusion
This thesis presented five 4IR technological artefacts that is expected
to make a difference in the South African healthcare system. These
artefacts were designed using the DSR methodology. The resulting
products were evaluated by experts/practitioners in the different health-
care sub-domains. Results from evaluations show that these tools will
be useful in advancing the healthcare system in South Africa in the
4IR.
viii
P U B L I C AT I O N S
S
ome ideas in this thesis — including verbatim images, equa-
tions, and concept descriptions have been featured in the fol-
lowing articles that have been published and others completed,
submitted, or published at the time of submitting this dissertation:
[1] Abbott, S.G. and Ade-Ibijola, A. (2018). Architectural analysis of

the South African public healthcare industry. In Third Interna-
tional Conference on the Internet, Cybersecurity and Informa-
tion Systems (ICICIS 2018), pp. 39–43, Gaborone, Botswana.
[2] Abbott, S.G. and Ade-Ibijola, A. (2019). Trend analysis: a deci-

sion tool in SA healthcare. In the proceedings of the IEEE In-
ternational Multidisciplinary Information Technology and En-
gineering Conference (IMITEC 2019), pp 155–161, ISBN: 978-1-
7281-0040-1, November 21st to 22nd, Johannesburg.
URL: https://ieeexplore.ieee.org/document/9015863.
[3] Abbott, S.G. and Ade-Ibijola, A. (2019). Algorithms and a tool

for automatic decryption of clinical notes. In the proceedings
of the 6th IEEE International Conference on Soft Computing
and Machine Intelligence (ISCMI 2019), pp 137–143, ISBN: 978-
1-7281-4576-1, November 19th to 20th, Johannesburg.
URL: https://ieeexplore.ieee.org/document/9004426.
[4] Abbott, S.G., Ade-Ibijola, A. and Roestenburg, W. (2021). Big

data in social research. In Book: Research at Grassroots, 5th Edi-
tion, Van Schaik Publishers (Subsidiary of Media24), pp 181–
196, Pretoria.
URL: https://www.vanschaiknet.com/book/view/509.
[5] Abbott, S.G. and Ade-Ibijola, A. (2022). An Architectural Anal-

ysis of the South African Healthcare Landscape. Journal article
submitted for review.
[6] Abbott, S.G. and Ade-Ibijola, A. (2022). Big Data in Healthcare.

Journal article submitted for review.
ix
ACKNOWLEDGMENTS
T
his Doctoral programme would not have been successful with-
out the help of a few people. My supervisor, Prof Ade-Ibijola
for whom I am thankful and grateful for giving me the oppor-
tunity to work with him. His insight and support gave me direction
towards the completion of this thesis.
For their support during this Doctoral programme, I give thanks to

the following people (in no particular order).
• Formal Structures Lab — Nikita Patel, Lloyd Radebe, Safwaan Pa-

tel, Jakote Lejaha, Bathini Mkhaliphi, Keagan Young, Dr Chinedu
Okonkwo and Sonny Kabaso, for their continued assistance.
I would also like to thank them all for their support. My sincere apolo-
gies for anyone I have forgotten to mention.
Sponsors. This research was supported by the following grants and

scholarships.
• University of Johannesburg Supervisor-Linked Bursary (2018 to

2022).
• Prof. Abejide Ade-Ibijola’s research funds (for other expenses

such as conference travel, registrations, language editing, and
partial tuition support).
xi
P R E FA C E
I
n this thesis, we present some applied AI solutions to the South
African Healthcare sector, drawing reference to and emanating
from the impacts of the Fourth Industrial Revolution (4IR). This
section highlights the key contributions of the research and the organ-
isation of this dissertation. It also includes a list of domains that are
related to, and non-academic talks that were presented on this work.
Key contributions.
Throughout this thesis, the following key contributions were made:
1. Architectural analysis of the South African healthcare landscape: we

have presented an in depth analysis of the economics and dy-
namics of the South African Healthcare (SAHC) landscape. It
includes the private and more specifically the public healthcare
sector. It outlines an understanding of the current challenges
facing the South African eHealth strategy and the potential im-
pacts of the 4iR technologies [Abbott and Ade-Ibijola, 2018].
2. Healthcare robotic designs: we have presented the design of a mod-

ified medical MoonBoot is presented in aiding the monitoring
and the management of a chronic Diabetic Foot Syndrone (DFS).
3. Trend analysis: we have presented a case scenario of searching

medical disease patterns and possible trends in gaining insights
into healthcare conditions. The analysis of certain medical con-
ditions is shown as per region [Abbott and Ade-Ibijola, 2019b],
thus creating a method for improved clinical decisions, an ana-
lytical tool for possible healthcare interventions, per region or
country.
4. Clinical note decryption tool: we have presented an algorithm for

the decryption of complex medical specialist reports, clinical
narratives, reports, logs and pathological notes, by translating
them into plain language, for the whole medical team [Abbott
and Ade-Ibijola, 2019a].
5. Clinical chatbot: we have presented a clinical chatbot application

within the South African healthcare context, in delivering pa-
tient clinical information. The chatbot aids the patient in the
management of a medical specific condition, such as a diabetic
foot.
xiii
6. Clinical big data anonymisation: we have presented an algorithm
for the anonymisation of a patients personal medical details, in
clinical big data sets. These data sets are highly sensitive data,
subscribing to certain governance rules and laws. We have also
validated the perceptions and opinions of the aforementioned
AI tools with a focus group of healthcare specialists by present-
ing them with evaluation survey results and inferences.
Dissertation organisation
This dissertation is organised into parts, each consisting of two or
more chapters. Part i presents the introduction, background and prob-
lem definition, research aim and objectives, an overview of the method-
ology applied in this work, literature review, and definition of terms.
It further analyses the current healthcare architectural domain includ-
ing the challenges, within the South African healthcare landscape en-
gaging the eHealth policy. This includes the Fourth Industrial Revolu-
tion (4IR), which compares the applications of AI within the defined
domains. Part ii presents the theoretical background to this study and
the research methodology in details. The major contributions of this
work are in Part iii, and Part iv. Part v evaluates the developed tools
and discusses conclusions and future work.
Domain of research
This research falls under the following categories.
1. Medical healthcare sciences, an in-depth insight into the eco-

nomics and epidemiological ramifications, pertaining to specific
medical conditions with an understanding of the current chal-
lenges facing the strategic intent of eHealth and the impacts of
the Fourth Industrial Revolution technologies [Abbott and Ade-
Ibijola, 2018].
2. Computing methodologies, artificial intelligence, natural lan-

guage processing [Abbott and Ade-Ibijola, 2019b].
Academic presentations
Some of the ideas in this dissertation were presented at the following
academic events.
1. Presented The Impacts of 4iR on the South African Healthcare

market at the South African Podiatry Medical Conference, Sand-
ton, Johannesburg, South Africa — Podiatry Association of South
Africa (PASA) Aug, 2017.
2. Presented An Architectural Analysis of the South African

Healthcare Landscape at the ICICIS Conference proceedings,
xiv
Sandton, Johannesburg, South Africa Aug, 2017 (Hosted by the
University of Botswana).
3. Presented An Architectural Analysis of the SA Healthcare

Sector at the ICICIS Conference proceedings, Gaborone, Botswana
Sep, 2018 (Hosted by the University of Botswana).
4. Presented a public lecture on the analysis of the South African

healthcare sector encompassing the integration challenges fac-
ing the National Health Insurance at the University of Johan-
nesburg (2017).
xv
CONTENTS
i introduction, literature and related work 1

1 introduction and background 2
1.1 Problem definition . . . . . . . . . . . . . . . . . . . . . 3
1.1.1 The Fourth Industrial Revolution (4IR) within
the South African healthcare landscape . . . . . 3
1.1.2 New applications of technologies in South African
healthcare . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Research context . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Aim and objectives . . . . . . . . . . . . . . . . . 4
1.2.2 The purpose of the research . . . . . . . . . . . . 5
1.2.3 The benefits of the research . . . . . . . . . . . 5
1.2.4 Scope . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.5 Research questions . . . . . . . . . . . . . . . . . 6
1.3 Methodology: research design . . . . . . . . . . . . . . . 8
1.4 Key contributions . . . . . . . . . . . . . . . . . . . . . . 8
1.4.1 Applications . . . . . . . . . . . . . . . . . . . . . 8
1.4.2 Evaluation . . . . . . . . . . . . . . . . . . . . . . 9
1.5 Dissertation organisation . . . . . . . . . . . . . . . . . . 9
2 preliminaries 11
2.1 General definitions . . . . . . . . . . . . . . . . . . . . . 12
2.2 Artificial Intelligence definitions . . . . . . . . . . . . . 15
2.3 Medical definitions . . . . . . . . . . . . . . . . . . . . . 16
2.4 Summary of chapter . . . . . . . . . . . . . . . . . . . . 18
3 literature review and related work 19
3.1 The South African Healthcare Landscape . . . . . . . . 19
3.1.1 Introduction . . . . . . . . . . . . . . . . . . . . . 19
3.1.2 The SA healthcare economic background . . . . 20
3.1.3 The healthcare sector in South Africa . . . . . . 21
3.1.4 The South African public healthcare sector . . . 23
3.1.5 SA e-Health national digital strategy 2019 -2024 25
3.1.6 SA public health projects . . . . . . . . . . . . . 26
3.1.7 Healthcare technology in South Africa . . . . . 28
3.1.8 Current challenges within the South African health-
care Industry . . . . . . . . . . . . . . . . . . . . 30
3.1.9 The National Health Insurance (NHI) . . . . . . 31
3.1.10 The South African private healthcare technol-
ogy . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.1.11 Discussion and analysis . . . . . . . . . . . . . . 32
3.1.12 Data analysis within public health . . . . . . . 36
3.1.13 Possible outcomes . . . . . . . . . . . . . . . . . 37
xvii
xviii contents
3.1.14 Summary of SA Healthcare Landscape . . . . . 39

3.2 Fourth industrial revolution (4IR) . . . . . . . . . . . . . 40
3.2.1 The 4IR industries . . . . . . . . . . . . . . . . . 42
3.2.2 Digital healthcare 4.0 . . . . . . . . . . . . . . . . 43
3.2.3 Conclusion on 4IR . . . . . . . . . . . . . . . . . 52
3.2.4 Summary of 4IR . . . . . . . . . . . . . . . . . . 52
3.3 Interoperability: A basis for 4IR . . . . . . . . . . . . . . 52
3.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . 53
3.3.2 Healthcare knowledge management . . . . . . . 54
3.3.3 The South African public healthcare challenges 55
3.3.4 The South African public healthcare architecture 59
3.3.5 The challenges of interoperability standards in
Africa . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3.6 Knowledge management in healthcare . . . . . 67
3.3.7 A new era in patient care . . . . . . . . . . . . . 71
3.3.8 The era of information-based medicine . . . . . 73
3.3.9 Architectural recommendations . . . . . . . . . 76
3.3.10 Interoperability considerations for the SA health-
care landscape . . . . . . . . . . . . . . . . . . . 78
3.3.11 Conclusion on Interoperabilty . . . . . . . . . . 79
3.3.12 Summary on Interoperability . . . . . . . . . . . 80
3.4 Electronic Health Record . . . . . . . . . . . . . . . . . . 81
3.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . 81
3.4.2 Electronic Health Record (eHR) . . . . . . . . . 82
3.4.3 The challenges facing an eHR . . . . . . . . . . . 85
3.4.4 The basic fields required within an eHR . . . . 87
3.4.5 Conclusion on eHR . . . . . . . . . . . . . . . . . 88
3.4.6 Summary on eHR . . . . . . . . . . . . . . . . . 89
ii theoretical background and research method-

ology 91
4 theoretical background to the study 92
4.1 Theoretical approach to the study . . . . . . . . . . . . 93
4.2 An overview of the Positivist and Critical approaches . 93
4.2.1 The Positivist Approach . . . . . . . . . . . . . . 93
4.2.2 The Critical Approach . . . . . . . . . . . . . . . 94
4.3 Key constructs . . . . . . . . . . . . . . . . . . . . . . . . 95
4.3.1 Key constructs: Positivist Approach . . . . . . . 95
4.3.2 Key constructs: Critical Approach . . . . . . . . 95
4.4 Rationale for using Positivist and Critical approaches . 96
4.5 Critics and limitations of the chosen theories . . . . . . 98
5 research methodology 100
5.1 Philosophical foundations of the study . . . . . . . . . 100
5.2 The Design Science Research (DSR) methodology . . . 101
5.2.1 An overview of the DSR approach . . . . . . . . 102
contents xix
5.2.2 DSR in Information Systems or Technology . . 103

5.2.3 The benefits of DSR methodology . . . . . . . . 103
5.2.4 The risks of the DSR methodology . . . . . . . . 104
5.3 The rationale for selecting artefacts . . . . . . . . . . . . 104
5.4 Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.4.1 Study population . . . . . . . . . . . . . . . . . . 105
5.4.2 Sampling procedures . . . . . . . . . . . . . . . . 105
5.4.3 Expert sampling . . . . . . . . . . . . . . . . . . 105
5.4.4 Inclusive criteria for respondents . . . . . . . . . 106
5.4.5 Sample size . . . . . . . . . . . . . . . . . . . . . 106
5.4.6 Data collection . . . . . . . . . . . . . . . . . . . 106
5.4.7 Designing the questionnaires . . . . . . . . . . . 106
5.5 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . 107
5.6 Ethical considerations and procedures . . . . . . . . . . 107
iiicontributions of 4ir technologies to south african

healthcare 109
6 big data in healthcare 110
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 110
6.2 Big data concepts . . . . . . . . . . . . . . . . . . . . . . 112
6.2.1 Big data . . . . . . . . . . . . . . . . . . . . . . . 112
6.2.2 Big data analytics . . . . . . . . . . . . . . . . . . 115
6.3 Big data architecture . . . . . . . . . . . . . . . . . . . . 117
6.4 Big data analysis . . . . . . . . . . . . . . . . . . . . . . 119
6.4.1 Big data analysis tools . . . . . . . . . . . . . . . 119
6.4.2 Cloud technology . . . . . . . . . . . . . . . . . . 121
6.5 Big data in healthcare . . . . . . . . . . . . . . . . . . . 121
6.5.1 The nature of big data in healthcare . . . . . . . 121
6.5.2 Big data value for healthcare . . . . . . . . . . . 122
6.5.3 Big data in healthcare . . . . . . . . . . . . . . . 123
6.5.4 Big data challenges for healthcare . . . . . . . . 124
6.6 Anonymisation of healthcare data . . . . . . . . . . . . 125
6.6.1 Algorithms for health data anonymisation . . . 125
6.6.2 Implementation . . . . . . . . . . . . . . . . . . . 127
6.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 128
6.7.1 Big data for research . . . . . . . . . . . . . . . . 129
6.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 130
7 trend analysis: a decision tool in sa healthcare 132
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 132
7.2 Related work . . . . . . . . . . . . . . . . . . . . . . . . . 135
7.2.1 Google trends for cancer screening . . . . . . . 135
7.3 Methodology . . . . . . . . . . . . . . . . . . . . . . . . 135
7.3.1 Google flu trends development . . . . . . . . . . 135
7.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
xx contents
Google trends scenarios - South African health-

7.4.1
care . . . . . . . . . . . . . . . . . . . . . . . . . . 137
7.4.2 Tobacco smoking can result in an increase in
lung cancers . . . . . . . . . . . . . . . . . . . . . 137
7.4.3 Diabetic neuropathy and lower limb amputations 138
7.4.4 Breast cancer screening . . . . . . . . . . . . . . 140
7.4.5 The prevalence of diabetes and cardiac arrest . 141
7.4.6 The most prevalent diseases in South Africa are
cancers, diabetes, HIV and Tuberculosis (TB) . . 143
7.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 145
7.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 146
8 natural language processing(nlp): decrypting clin-
ical notes 148
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 148
8.2 Background and related Work . . . . . . . . . . . . . . . 150
8.2.1 NLP in healthcare . . . . . . . . . . . . . . . . . 150
8.2.2 Basic structure . . . . . . . . . . . . . . . . . . . 151
8.2.3 Parsing . . . . . . . . . . . . . . . . . . . . . . . . 152
8.3 The methodology . . . . . . . . . . . . . . . . . . . . . . 152
8.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
8.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 155
8.5.1 Current NLP challenges . . . . . . . . . . . . . . 158
8.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 160
iv robotics and mobile applications 163

9 robotics in healthcare: theraboot 164
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 164
9.2 Robotics in healthcare . . . . . . . . . . . . . . . . . . . 166
9.2.1 Medical robotics . . . . . . . . . . . . . . . . . . 166
9.3 Robotic surgery . . . . . . . . . . . . . . . . . . . . . . . 169
9.3.1 Minimally Invasive Surgery (MIS) . . . . . . . . 169
9.3.2 Telerobotic systems . . . . . . . . . . . . . . . . . 171
9.4 Internet of Things (IoT) . . . . . . . . . . . . . . . . . . 174
9.5 The TheraBoot . . . . . . . . . . . . . . . . . . . . . . . . 175
9.5.1 The diabetic smart boot- TheraBoot . . . . . . . 178
9.5.2 The new therapeutic boot - TheraBoot . . . . . . 179
9.5.3 Applying AI to other pressure injuries . . . . . 182
9.6 Medical ethics and medical robots . . . . . . . . . . . . 187
9.7 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . 188
9.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 189
10 mobile application: healthcare chatbot 192
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 192
10.2 Background to mobile applications . . . . . . . . . . . . 194
contents xxi
10.2.1 mHealth . . . . . . . . . . . . . . . . . . . . . . . 194

10.2.2 mHealth platforms . . . . . . . . . . . . . . . . . 196
10.3 Healthcare chatbots . . . . . . . . . . . . . . . . . . . . . 196
10.4 Methodology . . . . . . . . . . . . . . . . . . . . . . . . 198
10.5 The Design of the MedBot ver 3.0 - HTML . . . . . . . 198
10.6 The Implementation and results of the MedBot . . . . . 199
10.7 The Logic of a MedBot for a diabetic foot . . . . . . . . 202
10.8 Evaluation and application of the MedBot . . . . . . . . 208
10.8.1 The MedBot for diabetic foot syndrome . . . . . 208
10.8.2 The evaluation of the MedBot . . . . . . . . . . 208
10.8.3 The application of the MedBot . . . . . . . . . . 208
10.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 209
10.10Summary of chapter . . . . . . . . . . . . . . . . . . . . 209
v evaluation and conclusion 212

11 evaluation 213
11.1 MoonBoot . . . . . . . . . . . . . . . . . . . . . . . . . . 213
11.1.1 Practitioners suggestions . . . . . . . . . . . . . 214
11.2 MedBot . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
11.2.1 Practitioner perceptions . . . . . . . . . . . . . . 217
11.3 Clinical note decryption algorithm . . . . . . . . . . . . 219
11.4 Anonymisation of personal medical details algorithm . 221
11.4.2 Practitioners Suggestions . . . . . . . . . . . . . 222
11.5 Disease Trend Analysis . . . . . . . . . . . . . . . . . . . 223
11.5.2 Practitioners Suggestions . . . . . . . . . . . . . 225
12 conclusion and future work 228
12.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 228
12.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . . 229
12.2.1 South African public health (NDOH) . . . . . . 229
12.2.2 Big data in healthcare . . . . . . . . . . . . . . . 232
12.2.3 Robotic technology and the TheraBoot . . . . . 232
12.2.4 Trend analysis in healthcare . . . . . . . . . . . . 233
12.2.5 Natural language processing - decryption of clin-
ical notes . . . . . . . . . . . . . . . . . . . . . . . 233
12.2.6 New AI health tools . . . . . . . . . . . . . . . . 234
12.2.7 Healthcare and data science . . . . . . . . . . . . 234
12.2.8 Machine-learning tools . . . . . . . . . . . . . . 234
12.2.9 Policy and governance frameworks . . . . . . . 235
References 268
LIST OF FIGURES
Figure 1 The economic overview of the SA healthcare

landscape. Adapted from:[Abbott and Ade-Ibijola,
2018] . . . . . . . . . . . . . . . . . . . . . . . . 22
Figure 2 Current private SA Healthcare switching topolo-
gies. Adapted from:[BCX, 2016; NDOH, 2014a] 62
Figure 3 Current SA public healthcare systems. Adapted
from: [Botha et al., 2016; Katuu, 2018; NDOH,
2014a] . . . . . . . . . . . . . . . . . . . . . . . . 63
Figure 4 Four stage process in creating commonality of
standards. Source: IHE process:[Carr and Moore,
2003; IHE, 2016] . . . . . . . . . . . . . . . . . . 66
Figure 5 The knowledge management flow cycle. Source: [Bali,
2013; Haslinda and Sarinah, 2009; Klimko, 2001;
Wiig, 2004] . . . . . . . . . . . . . . . . . . . . . 71
Figure 6 A new architectural Cloud solution for SA pub-
lic health. . . . . . . . . . . . . . . . . . . . . . . 77
Figure 7 A possible architectural conceptual design for
SA public health. . . . . . . . . . . . . . . . . . 78
Figure 8 Anonymisation algorithm for patient details . 126
Figure 9 Synthesised hospital names . . . . . . . . . . . 126
Figure 10 Synthesised medical aid names . . . . . . . . . 127
Figure 11 Synthesised patient identification numbers . . 127
Figure 12 Smoking versus lung cancer within South Africa138
Figure 13 Smoking versus lung cancer within regions in
South Africa . . . . . . . . . . . . . . . . . . . . 138
Figure 14 Trend analysis of diabetic neuropathy versus
lower limb amputation in South Africa . . . . 139
Figure 15 Trend analysis of diabetic neuropathy versus
lower limb amputation as per region in South
Africa . . . . . . . . . . . . . . . . . . . . . . . . 139
Figure 16 Trend analysis of breast cancer screening and
breast cancer within South Africa . . . . . . . . 140
Figure 17 Breast cancer screening versus breast cancer as
per region within South Africa . . . . . . . . . 141
Figure 18 Trend analysis of diabetic prevalence and heart
failure within South Africa . . . . . . . . . . . . 142
Figure 19 Trend analysis of diabetic prevalence and heart
failure within South Africa as per region . . . 142
Figure 20 Trend analysis of cancer, diabetes, HIV and tu-
berculosis(TB) within South Africa . . . . . . . 144
xxii
List of Figures xxiii
Figure 21 Trend analysis of cancer, diabetes, HIV and tu-

berculosis(TB) within South Africa per region 144
Figure 22 The process of deciphering clinical notes into
plain text. . . . . . . . . . . . . . . . . . . . . . . 154
Figure 23 Clinical notes translator(the deciphering of com-
plex medical notes notes into readable text). . 155
Figure 24 The Robo Nurse aiding patients in Italy. Source:
www.pri.org [Kumar, 2018; Riek, 2017]. . . . . 165
Figure 25 The Robo Nurse aiding patients in Italy. Source:
www.pri.org [Kumar, 2018; Riek, 2017]. . . . . 167
Figure 26 The TeleRobotic daVinci surgical system. Source:
www.ResearchGate.com and daVinci surgical
systems [Avgousti et al., 2016; Levi Sandri et
al., 2017]. . . . . . . . . . . . . . . . . . . . . . . 171
Figure 27 The Telerobotic surgical team. Courtesy red-
dit.com [Avgousti et al., 2016]. . . . . . . . . . . 172
Figure 28 The TeleRobotic daVinci surgical system. Source:
www.ResearchGate.com and daVinci surgical
systems [Avgousti et al., 2016; Iroju and Olaleke,
2015]. . . . . . . . . . . . . . . . . . . . . . . . . 174
Figure 29 The standard MoonBoot. Source: www.Google-
scholar.com and Amazon.com. [Armstrong et
al., 2017, 2005]. . . . . . . . . . . . . . . . . . . . 176
Figure 30 The LAN layout as presented for the Smart In-
sole. Source [Jegede et al., 2015]. . . . . . . . . . 177
Figure 31 The Smart Insole with adapted sensors. Source
[Jegede et al., 2015]. . . . . . . . . . . . . . . . . 178
Figure 32 The modified concept of a MoonBoot or TheraBoot
with zones of demarcated sensors. . . . . . . . 180
Figure 33 The padded hospital bed: note the pressure
pads. Source: www.Medicalexpo.com . . . . . 184
Figure 34 The MAP sensor mat for a hospital bed: A-
Remote monitor device; B-Sensor mat under
sheet; C-Monitoring of pressures; D-Sensor Data.
Source:https://www.XSensor.com hospital sys-
tems . . . . . . . . . . . . . . . . . . . . . . . . . 185
Figure 35 A wheelchairs padded seat - note the pressure
cushion. Source: https://www.drivemedical.com186
Figure 36 The smart hospital bed - note the smart pres-
sure mat. Source: https://www.Boditrak.com 187
Figure 37 The MedBot conversational screening desktop.
(1/4) . . . . . . . . . . . . . . . . . . . . . . . . 200
(2/4) . . . . . . . . . . . . . . . . . . . . . . . . 200
(3/4) . . . . . . . . . . . . . . . . . . . . . . . . 201
xxiv List of Figures

(4/4) . . . . . . . . . . . . . . . . . . . . . . . . 201
Figure 41 The MedBot conversational screening on a smart-
phone. (1/6) . . . . . . . . . . . . . . . . . . . . 202
phone. (2/6) . . . . . . . . . . . . . . . . . . . . 203
phone. (3/6) . . . . . . . . . . . . . . . . . . . . 204
phone. (4/6) . . . . . . . . . . . . . . . . . . . . 205
phone. (5/6) . . . . . . . . . . . . . . . . . . . . 206
phone. (6/6) . . . . . . . . . . . . . . . . . . . . 207
Figure 47 Specialty groupings. . . . . . . . . . . . . . . . 214
Figure 48 The potential benefits of such a design on the
improved off-loading of pressure points in the
modified AI Moonboot. . . . . . . . . . . . . . 215
Figure 49 Survey: opinions on the benefits of the Moon-
Boot/ TheraBoot design. . . . . . . . . . . . . . 215
Figure 50 Survey: Expert opinions on the potential bene-
fits of the MoonBoot/ TheraBoot design. . . . 216
Figure 51 Survey: The potential benefits - design of the MedBot.218
Figure 52 Survey: The potential benefits - design of the MedBot.218
Figure 53 Survey: The potential benefits - design of the
medical notes decryption tool. . . . . . . . . . 220
Figure 54 Survey: The potential benefits - design of the
medical notes decryption tool. . . . . . . . . . 221
Figure 55 Survey: the potential benefits - design of the
anonymisation of patient details tool. . . . . . 223
Figure 56 Survey: the potential benefits - design of the
anonymisation of patient details tool. . . . . . 224
Figure 57 Survey: the potential medical benefits - disease
trend analysis tool. . . . . . . . . . . . . . . . . 226
L I S T O F TA B L E S
Table 2 Practitioner perception of AI and digital tools

in 4iR . . . . . . . . . . . . . . . . . . . . . . . . 227
xxv
ACRONYMS
Acronym Description
4iR Fourth Industrial Revolution (ie. Industry 4.0)
AI Artificial Intelligence
ANSI American Naming Standards of Industries
BRFSS Behavioural Risk Factor Surveillance System
CA Cardiac Arrest
CCD Continuity of Care Document for HL7 V_3
CDA Clinic Document Architecture for HL7 V_3
CDC Center for Disease Council (USA)
CNT Clinical Note Translator
CRM Customer Relationship Manager
DFS Diabetic Foot Syndrome
DFU Diabetic Foot Ulcers
Diab Diabetic
DMBOK Data Management Body of Knowledge
DRU Direct Robotic User
DU Diabetic Ulcers
Gov Government of South Africa
GT Google Trends
HDFS Hadoop Distributed Filing System
Health 4.0 Digitalisation of Healthcare i.e. 4th Paradigm
HIC High Income Countries
HIE Health Integration Enterprise
HINTS Hypertension Intervention Nurse Telemedicine Study
HISA Health Information Systems Architecture
HiSP Health Information Systems Processes
HL7 Open Source Healthcare Software
HST Health Systems Trust of South Africa
HTTPS Hyper Text Transfer Protocol
ICD 10 International Codes of Diagnoses V_10
ICHI International Classification of Health Interventions
ICT Information Communications Technology
IDDB Insulin Dependent Diabetic
IoT Internet of Things
IR Information Retrieval
xxvi
acronyms xxvii
LIC Low Income Countries

LOINC Logical Observation Identifiers Names and Codes
MD Master Data
MDG Millennium Development Goals
MDT Multi-Disciplinary Team (MDT)
MHR Medical Health Record
MIC Middle Income Countries
MS Microsoft Systems or Products
NCD Non Insulin Communicable Diseases
NDoH National Department of Health
NDOH National Department of Health
NDP National Development Plan
NER Name Entity Recognition
NGO Non Government Organisation
NHI National Health Insurance
NHIS National Health Information System
NHS National Health System of UK
NIDDB Non-Insulin Dependent Diabetic
NIDS National Indicator Data Sets
NLP Natural Language Programming
OLAP Online Analytical Processing
OLTP Online Transaction Processing
OPE Out of Pocket Expenses
PMP Project Management
POPI Protection of Private Information Act (2013)
POS Part of Speech tagging
PU Pressure Ulcer
RCW Removable Cast Walker
RDBMS Relational Database Management Systems
RS Robotic Surgery
SA South Africa (Zuid Afrikaans)
SAHC South African Healthcare
SAP Systems Application Products
SCM Supply Chain Management
SDG Strategic Development Goals
SMM Social Media Marketing
SNOMED Systematized Nomenclature of Medicine
SOFIE Surgical Operating Force Feedback Interface Eindhoven
SPA Single Part Access
SQL Structured Query Language, for databases
xxviii acronyms
SRL Semantic Role Labelling

TOGAF The Open Group Architecture Framework
UI User Interface
UIA User Interface Architecture
UN United Nations
WHO World Health Organisation
XML Extra Markup Language
ZAR South African Rands
Part I
I N T R O D U C T I O N , L I T E R AT U R E A N D R E L AT E D
WORK
The South African healthcare industry is currently under-

going further transformation, against the backdrop of po-
litical and economic transformations that are, dominating
the landscape. In order to understand the future of its
course, or discourse, we need to understand the economic
dynamics of its origins, and current mechanics. The in-
coming 1994 regime for South Africa, has brought in fur-
ther change and expansion of healthcare, for the whole
population, encountering its inherent inequities and per-
sistent disproportionate budget spending.
In addition, the WHO is prescribing eHealth policies, glob-

ally, including the digitisation of healthcare through the
advancement of the Fourth Industrial Revolution (4IR).
South Africa is facing additional challenges in adapting
to the modernisation of healthcare. The first part of this
thesis explores the background and historical origins to
these consequent challenges, which lays the basis for the
advancement of healthcare from a technology point of
view. It further sets the scene for opportunities of creative
innovation, desperately needed within the SA healthcare
industry.
1 INTRODUCTION AND BACKGROUND
T
he SA healthcare system in general consists of the private sec-
tor (privately funded) and the public health sector (state funded)
addressing vastly differing populations with different fund-
ing mechanisms. The health care industry has many facets, or layers
of “knowledge”, which comprises of patient data, transactional data,
analytical data (meta data), business intelligence data, spatial or Ge-
ographic Information Systems (GIS) management information data
and patient data trends or information [Telkom, 2015].
The construct of the vision of the impact of the Fourth Industrial

Revolution(4IR), within healthcare, is indirectly promulgated via the
eHealth policies and tools of the World Health Organisation (WHO),
and includes the South African National Department of Health poli-
cies and legislated polices, according to the National Development
Plan and the SA eHealth policy of recent development [Herselman
and Botha, 2016b].
The United Nations (UN) Millennium Development Goals (MDG’s)

were agreed upon, and signed by 189 UN Member States in 2000,
consisting of eight major goals [Nations, 2015]. They are all commit-
ted to eradicating or alleviating poverty, disease, hunger, environmen-
tal degradation, illiteracy, and discrimination against women. These
MDGs have subsequently been superseded by the Sustainable Devel-
opment Goals by 17 integrated objectives or goals based on the MDGs
but they appear to be broader in scope. The SDGs were signed by all
its member countries in 2015 [Mayosi, 2012; Ruxwana, 2010].
The South African National Department of Health (NDoH) has

subsequently adopted these goals for public healthcare strategically
within all their future strategies in developing national strategic ob-
jectives and plans accordingly [Ruxwana, 2010]. The universal access
to healthcare (UHC), for all South African citizens, is the most signif-
icant of the Strategic Development Goals (SDGs) where the National
Health Insurance (NHI) plan, is at its core [Haywood, 2016; Mayosi,
2012].
Further to this initiative, is the World Health Organisation (WHO)

eHealth strategy which has also placed itself into the South African
National Department of Health (NDoH), strategic plan of develop-
ment, whereby the SA eHealth strategy will take precedence across
2
1.1 problem definition 3
all future developments. This eHealth strategy has been adopted and
applied by the South African Department of Health (NDoH), in con-
junction with the South African National Development Plan (NDP) as
a strategic plan of governance currently being implemented [Katuu,
2016a].
South Africa is in dire need to upgrade and modernise its pub-

lic healthcare system: by aligning with these global trends through
the newly introduced National Healthcare Insurance (NHI) model of
2011, which is currently under development in a pilot phase. The NHI
programme is envisaged to be fully operational by 2030 and projected
to be developed in phases over the next 10 to 14 years [Archer, 2016].
1.1 problem definition
In this section, we define the problem addressed in this thesis.
1.1.1 The Fourth Industrial Revolution (4IR) within the South African
healthcare landscape
By presenting the challenges faced by the South African healthcare

landscape, we can identify the need for ICT technological solutions,
within the healthcare industry.
The South African public healthcare sector is lacking in imple-

menting and adopting modern software applications and information
technology in addressing the Fourth Industrial Revolution (4IR) and
eHealth. These are currently being prescribed by the World Health
Organisation (WHO), in delivering improved healthcare services to
the greater population of South Africa [Botha et al., 2016; Herselman
and Botha, 2016b].
the application of artificial intelligence (ai): To ex-

plore new applications of Artificial Intelligence (AI) within
healthcare (The major driver of 4IR) [Botha et al., 2016;
Herselman and Botha, 2016b].
the resultant impacts of such ai applications: To imple-

ment certain AI technologies to show the positive impact on
the South African healthcare sector [Belle et al., 2015; Chen and
Asch, 2017].
the application of such ai applications: To evaluate addi-

tional applications of technologies and their future impacts on
healthcare outcomes, nationally and globally.
4 introduction and background
Further, this includes the convergence of health and technology

in producing such improved value-based health outcomes for
all, within South Africa [Seara et al., 2016].
1.1.2 New applications of technologies in South African healthcare
A literature research was conducted to gain an in-depth understand-

ing of subject matter from experts in the field of healthcare and tech-
nology. It proposes a strategy to further enhance the objective of
bringing quality public healthcare to the South African population.
The literature research gives impetus to an interpretive approach in
addressing the SA public healthcare technology challenges within the
framework of the Fourth Industrial Revolution.
1.2 research context
1.2.1 Aim and objectives
The objectives of this study are to design such technologies for the
healthcare, by identifying certain technologies that can provide a prac-
tical technological solution for a current healthcare problem.
These solutions have been investigated in aiding the eHealth strat-

egy and digital health in its primary objective of providing universal
access to healthcare, aimed at the majority of the population who
does not have access to any healthcare. This thesis aims to provide an
understanding of the technological innovations and the development
thereof, to engage modern technology in the context of the Fourth
Industrial Revolution. The common thread is the creative innovation
and application of AI technologies selected for immediate impact dur-
ing this process.
1.2.1.1 Sub-objectives
The objectives of this thesis are further broken down into sub-objectives
as follows:
1. to explore, and compare the implementation and impact of known

emerging AI technologies that have previously not been applied
entirely in healthcare such as emerging technologies,
2. to investigate the impacts of adopting such technological appli-

cations within the context of (4IR) healthcare innovation,
1.2 research context 5
3. to create AI artefacts that can be deployed to improve the South

African healthcare system, and
4. to evaluate created artefacts and show that they indeed will be

useful in the targeted system.
1.2.2 The purpose of the research
The purpose of this research project details a transformation by outlin-

ing the application of certain software technologies, within the South
African healthcare sector, directly related to the Fourth Industrial Rev-
olution (4IR). The research will highlight some technology initiatives
to aid and assist the common objectives of the 4IR within the health-
care sector:
1. To explore new applications of Artificial Intelligence (AI) within

healthcare - the major driver of 4IR.
2. To implement certain AI technologies to show that they have a

positive impact on the South African healthcare sector.
3. To evaluate additional applications for the clinical environment

and provide a design within the clinical environment.
Further, this includes the convergence of health and technology in

producing improved health outcomes for the South African popula-
tion.
1.2.3 The benefits of the research
A multitude of benefits could emanate from the study which will

enhance the quality of healthcare outcomes through the: healthcare
analytics and AI tools described. This will validate the tools in pro-
moting more efficient health processes with the convergence of tech-
nologies providing a more efficient predictability of healthcare out-
comes based on improving data efficiencies through the impacts of
the Fourth Industrial Revolution (AI-tools).
1.2.4 Scope
We base this research on the assumptions that:
1. the scope is confined to the Fourth Industrial Revolution within

South African healthcare encompassing the public and private
healthcare sectors,
2. we only focus on the application of selected technologies ad-

dressing the inherent challenges within the SA healthcare land-
scape. It further relates to the current global challenges, in many
instances, in applying such technologies, and
3. we assume the SA eHealth strategy is the foundation of all fu-

ture technological developments within the SA healthcare con-
text, thus forming the common platform for the development
of the National Health Insurance plan, currently legislated, ad-
dressing all the infrastructural technological Internet and Com-
munication Technologies (ICT) challenges.
1.2.5 Research questions
Currently, there is a lack of evidence of a strategic approach in ad-

dressing the design and adoption of specific solutions in 4IR, pertain-
ing to South African healthcare.
The primary question is: How can we design 4IR artefacts to support the
South African healthcare sector?
Five tools were carefully chosen specifically that can answer this
question. However, there are many more solutions which remain too
exhaustive for this thesis. These specific solutions were selected based
on their immediate practical design, application and adoption, for
specific healthcare problems being experienced.
During the course of this work, we have addressed a number of

sub-questions from the primary research question stated above.
1.2.5.1 The sub questions from the primary research questions
1. What efforts/initiatives are present within the SA healthcare

industry addressing the Fourth Industrial Revolution (4IR)?,
2. How are the efforts of eHealth supporting or aiding the advent

of 4IR?,
3. What systems and platforms are in place in ensuring improved

interoperability across all health information systems – provin-
cial and national, in support of 4IR?,
4. Which AI software applications need to be applied and made

available in order to produce improved health outcomes i.e. big
data, NLP, robotics, chatbots and trend analysis?
1.2 research context 7
1.2.5.2 The medical MoonBoot
5. What AI designs/initiatives can be harnessed in addressing the

management of a Diabetic Foot Syndrome (DFS) in preventing
lower leg amputations?,
6. How can we improve the preventative methods and outcomes

of these patients facing foot ulceratons and potentially a foot
amputation?,
1.2.5.3 The Medical Chatbot
7. What systems and platforms are in place in ensuring improved

patient assistance in the communication of basic medical infor-
mation, without consulting a medical professional?,
8. What medical chatbot technologies can be utilised in aiding the

patient along well constructed information pathways aiding a
patient in the management of a diabetic foot, as an example?
1.2.5.4 A decision analysis tool
9. What medical analytical tool is widely available on the Internet,

in supporting the public, patient and practitioner, in answering
simple medical questions relating to medical conditions, viral
spreading or medical interventions, from a basic statistical point
of view?
1.2.5.5 A clinical note decryption tool
10. Which AI software applications need to be applied and made

available in order to decrypt a medical report and made avail-
able for the whole medical team to understand, being presented
in a simple format?
1.2.5.6 An anonymisation tool for personal medical details
11. What technologies can be applied in aiding a healthcare spe-

cialist in improving the big data analytics in healthcare through
the anonymisation of personal medical details, in providing vi-
tal medical insights through analysis and pattern recognition?
and,
12. how will the Industry apply these software technologies in mod-
ern healthcare?
1.3 methodology: research design
In Part ii, this work followed the Design Science Research method
(DSR) in the design and development of the artefacts presented in
this thesis. The evaluation of these artefacts is based on a quantitative
research design. In the evaluation sections, the questionnaires were
targeted at specific experts in the field of healthcare in South Africa.
The developed tools are: the decryption of medical notes, medical
chatbots, disease trend analysis, anonymisation of personal medical
details, and the new MoonBoot design, named the TheraBoot.
1.4 key contributions
The key contributions are presented in five artefacts:
1. Big data anonymisation algorithm,
2. Algorithms for decryption of medical notes,
3. Disease trend analysis,
4. Applied robotics — TheraBoot, and
5. mHealth chatbots — MedBot.
1.4.1 Applications
In Part iii, we have presented the first three applications, namely:
1. a tool for anonymising complex clinical healthcare datasets,
2. a chatbot in aiding a practitioner in the management of patients,

and
3. a method of applying healthcare trend analysis in aiding a prac-

titioner within their clinical environment.
In Part iv, we present two more applications, namely:
1. Robotics in healthcare: AI tools that perform a medical pro-

cedure remotely, with precision similar to a surgeon in another
location, and
1.5 dissertation organisation 9
2. Medical chatbots: an AI tool through a mobile application that

is able to simulate a medical discussion or consultation in seek-
ing medical knowledge remotely.
1.4.2 Evaluation
In Part v, we have:
1. evaluated how useful and effective the proposed designs of the

AI healthcare tools are for healthcare experts, and
2. presented an evaluation of the five different designs across an

array of healthcare experts in these subject fields, in a series
of data visuals generated by a Google Form questionnaire. The
data presented is to corroborate our findings and designs of
such tools to be of value to the healthcare practitioner, and the
healthcare industry of South Africa.
1.5 dissertation organisation
This dissertation is organised in parts, each consisting of one or more

chapters addressing the components of this research. The parts are as
follows:
Part i presents the introduction, background and definition of terms.

It outlines the literature and related work. It describes the South
African Healthcare sector from an economic and an ICT per-
spective, concerning the technology platforms and the health-
care applications and current models from the SA Healthcare
perspective, focusing on the SA Public healthcare sector. In ad-
dition, the literature and the related work concerning the Fourth
Industrial Revolution (4IR), focusing on the healthcare environ-
ment, technological applications and the benefits thereof.
Part ii outlines the theoretical background and research methodol-

ogy. It further describes the approach and its suitability for this
thesis whereby the rationale of the methodology is put forward.
This included the Design Science Research (DSR) methodology
as the adopted technology for this study. In addition to the ra-
tionale for selecting the artefacts making up the contributions
of this study.
Part iii outlines the healthcare system interoperability historical

background and the challenges currently facing the SA health-
care landscape, whilst adopting an eHealth strategy and the 4IR
transformation. This will include the background and inherent
challenges within the possible adoption of an electronic medical

or health record for healthcare within the SA public healthcare
sector
Part iv presents the five chapters on the contributions towards a

modernised healthcare platform incorporating the eHealth strat-
egy and the 4IR transformation. The technological contributions
will add value within the following five chapters covering: the
anonymisation of healthcare big data; the decryption of clinical
notes in the healthcare organisation; trend analysis of impend-
ing endemics or pandemics; robotic technology in healthcare
and the mobile applications applying the technology of Chat-
bots, within healthcare institutes.
Part v evaluates the developed tools, draws conclusions and pro-

poses future work.
PRELIMINARIES
2
T
he definitions of terms used in this dissertation are derived
from the fields of modern day healthcare and applied tech-
nologies, emanating from the Fourth Industrial Revolution
(4iR) technologies. 4iR is also termed as the ’Fourth Wave’ giving
impetus to the digitalisation of healthcare. It is the convergence of
digital and genomic technological revolutionary science, fusing with
healthcare and society in general, primarily centred on empowering
the patient, to manage his or her personal health [Mesko, 2017; Reddy
and Sharma, 2016; Sonnier, 2016].
The field of healthcare varies from the supportive, nursing and

aiding, to the modern day cutting edge technologies of telemedicine,
being applied to many specialist branches of medicine. One such spe-
cialist field of medicine relates to diabetes, commonly referred to as
diabetology. Diabetology is the study of a metabolic disorder called,
diabetes, in a broad context, where it carries significance for many
other applications of modern medicine [Soley-Bori, 2014].
The podiatrist, specialises in providing foot treatment and clinical

support to the diabetic patient, with other healthcare workers or pro-
fessionals. The podiatrist, focuses in the medical field of foot patholo-
gies, treatment and the management of foot pathologies. Diabetes is
one of many health conditions a podiatrist will treat and manage,
in collaboration with other medical team members [Armstrong et al.,
2017].
Digital health technologies can be applied through the applica-

tions of Artificial Intelligence (AI) within the ambit of the 4iR, thus
creating exponential value to the field of healthcare [Mesko, 2017;
Reddy and Sharma, 2016; Sonnier, 2016].
The application of Artificial Intelligence (AI) is not only relevant

to the field of diabetes within this study, but to other indirect fields
of nursing care and the collection of data through high end data
analytics and modern era technologies. This creates valuable medi-
cal insights, ultimately creating value-based outcomes of clinical care
which are currently in high demand from the healthcare industry [Cabestany
et al., 2018; Pang et al., 2018].
11
12 preliminaries
We have endeavoured to shed light on the application of these

4iR technologies, through AI, in the public and private healthcare
industry.
2.1 general definitions
Definition 2.1.1 (Healthcare). Healthcare is the maintenance and im-

provement of ones health through effective diagnosis, prevention mea-
sures, the recovery and treatment of a disease, illness, injury or any
other mental and physical anomalies managed by a health profes-
sional, across all health fields of medicine. It can also include the
effective and efficient management of such recovery plans of a pa-
tient through different interventions of effective management strate-
gies with the aid of technology tools [Bauer, 2018; Pang et al., 2018;
Porter and Teisberg, 2008].
Definition 2.1.2 (The Fourth Industrial Revolution (4iR)). The revolu-

tion or a new era of technology innovation that will vastly improve
the relationship between human and machine interactions, that will
unlock new products and markets, thereby fueling growth within the
global economy. Many industry sectors are making efforts in this
global vision, such as healthcare, amongst others, in general. 4iR is
largely based on the revolution of new technologies such as the ush-
ering in of 3D printing, the internet of things (IoT), the internet, ar-
tificial intelligence and genetic engineering with biomedical sciences
and healthcare [Chou, 2018; Schwab, 2017].
Definition 2.1.3 (Digital health). Digital health is the convergence of

digital technologies and genomics within healthcare, health, living,
and society [Sonnier, 2016]. Digital health encompasses the follow-
ing essential technologies: wireless wearable devices; sensors; micro-
processors; the internet and internet of things (IOT); social media;
mobile applications or mHealth; pharmacogenomics; telehealth or
telemedicine; robotics; big data analytics and personal genetic infor-
mation. These technologies are utilised, in an engineered manner, in
order to prevent and predict the future health, or the well being, of
an individual or population [Morrison, 2016; Pang et al., 2018; Son-
nier, 2016]. Digital health can be used as a general overarching term
encompassing all of the above terms that essentially engages digital,
mobile and wireless technologies in achieving a specific clinical out-
come [Fatehi et al., 2018].
Definition 2.1.4 (Electronic Health Records - eHR). An eHR, also

electronic patient record (ePR) or computerised patient record. It is
an evolving concept defined as a systematic collection of electronic
2.1 general definitions 13
health information, about individual patients or populations. It is a

record in electronic format that is capable of being shared across dif-
ferent health care settings, embedded in network-connected enterprise-
wide information systems. An eHR record may include a whole range
of data in comprehensive, or summary form, including demograph-
ics, medical history, referrals, medication and allergies, immunization
status, laboratory test results, radiology images, and billing informa-
tion” [WHO, 2012].
An eHR is based on specific dimensions surrounding the com-

pleteness of a patient’s health record over an entire lifespan, encom-
passing all dimensions of data for use by physicians, in addition to
other authorised hospital personnel and the patient. It can be termed
a longitudinal health record that resides across multiple health care
settings, meaning the record is a person centric health record [Kotzé
and Alberts, 2017].
Definition 2.1.5 (Electronic Medical Health Record - MHR). An elec-

tronic medical health record (MHR), is a type of in-house static record
within a single institute, clinic, or hospital. MHRs contains all relevant
medical data under the custodianship of the institute itself, and not
for public access, including by the patient. It is not to be confused
with the eHR. An eHR is different in nature, being a more complete
or “fluid” document over the full lifetime of a patient [Tan, 2005].
The industry focus is on the eHR, as its durability of time, with a
medical history of multiple dimensions, is of great value to any bio-
informatics statistician, or data scientist [Duwe, 2004].
Definition 2.1.6 (Health Level Seven Protocol(HL7)). The Health Level

Seven (HL7) is an international standards development organisation
(SDO). It maintains and develops certain standards of exchange and
the integration of electronic healthcare information, consisting of ad-
ministrative and clinical functions. It is currently the most widely
used standard and, is headquartered in USA. It is based on corpo-
rate membership, unlike other SDO’s, which is American National
Standards Institute accredited (ANSI). The level seven refers to the
seventh layer of the Open Systems Interconnection (OSI) model [Ben-
son, 2013; Katuu, 2018; Kotzé and Alberts, 2017].
Definition 2.1.7 (Clinical Document Architecture (CDA)). Clinical Doc-

ument Architecture supports medical records which are a standard
specification for the semantics of clinical documents and is widely
adapted globally. This is primarily a mark-up standard that speci-
fies the semantics and structure of clinical documents for data ex-
change between providers of healthcare. These clinical documents
would include discharge summaries, reports, lab reports, procedures
and pathology findings. CDA clinical documents are coded in Exten-
14 preliminaries
sible Markup Language (XML). This would include a primary detail

section with a main body section. The current version of CDA is re-
lease3 which serves as the foundation for all current implementations
which are ANSI approved HL7 standards [Benson, 2013].
Definition 2.1.8 (Continuity of Care Document (CCD)). The CCD en-

ables interoperability of clinical data enabling doctors to send elec-
tronic medical information to other regions or provinces without los-
ing the meaning of such records. CCD is a joint effort between HL7
and the American Society for Testing and Materials (ASTM). These
are the files that carries the patient medical history which comprises
a set of templates of summary medical records. These templates con-
tain the medical history with treatment plans that can be used in
other CDA types. It also serves as a basis for the interoperability in
the US Health Information Technology Standards Panel (HITSP), inte-
grating the health enterprise(IHE). It is an XML based standard which
specifies the coding and structure of a patient summary clinical doc-
ument [Benson, 2013; Mead, 2006].
Definition 2.1.9 (CEN/ISO13606 Standards- OpeneHR ). The Com-

munication from the European Committee for Standardisation/In-
ternational Standard Organisation is an approved international ISO
standard which was designed to achieve semantic interoperability
for eHR. The CEN/ISO13606 is a standard defining a data archi-
tecture for the communication of medical records between health-
care systems. This includes the communication between electronic
health records (eHR) systems or a centralised repository and or mid-
dle ware [Mead, 2006].
Definition 2.1.10 (Integration of Healthcare Enterprises (IHE)). IHE

utilise the coordinated use of established standards such as ISO, HL7,
W3C, OASIS, DICOM and IETF, underpinning patient care records.
IHE provide standards-based frameworks that share information be-
tween disparate systems that address critical interoperability issues
between service providers and patients, security, workflow and ad-
ministration of the workflow [Carr and Moore, 2003].
Definition 2.1.11 (Interoperability of Healthcare Systems). Interoper-

ability is fundamentally the ability of healthcare systems to share and
exchange clinical information, in order to improve health outcomes.
The patient remains the central subject, in the process of integration.
For this to be seamless and effective, the systems involved need stan-
dards, to operate within their respective technologies, regionally and
nationally. The global eHealth programmes are striving to achieve
a common set of global and local standards under the guidance of
the WHO, enabling multidisciplinary teams of health care providers,
2.2 artificial intelligence definitions 15
to share information and coordinate interventions effectively, and to

eradicate inefficiencies [Ardebesin, 2013; Foster, 2013; Nat, 2012].
Definition 2.1.12 (Internet of Things). Referred to as the Internet of

everything, or rather IoT, meaning the connecting of most devices
through the Internet. The concept allows the connection of many
work pieces together in facilitating the capturing and movement of
surveilled or monitoring data, to a central location that can be anal-
ysed and mined. The mined data provides patterns and insights that
could never have occurred without the constant monitoring of such
data inputs. The new insights provide improved decision and man-
agement processes [Chou, 2018; Pang et al., 2018; Porter and Teisberg,
2008].
Definition 2.1.13 (Healthcare Data). The intelligent health data, re-

lated to health conditions, patient details, population, reproductive
outcomes, causes of death, quality of life, clinical metrics, socioeco-
nomic factors and, behavioural information related to health and well-
ness. The data can include the financial status of a patient regarding
public or private healthcare status. The sources of such data can be
obtained from medical health records, electronic health records and
sensory or monitoring devices within a clinical setting. The data can
be aggregated in a structured or unstructured format. Its value is un-
locked in presenting those patterns and relationships in aiding the
diagnosis and the healing process. Healthcare data is proving to be
increasingly more critical for a favourable clinical outcome of the pa-
tient than ever before, in general [Mesko, 2017; Reddy and Sharma,
2016; Sonnier, 2016].
2.2 artificial intelligence definitions
Definition 2.2.1 (Artificial Intelligence (AI)). AI is a classification of

computer systems and processes that can perform intelligent tasks or
actions, as performed by a human in performing remedial repetitive
tasks over a longer sustained period of time than a human can accom-
plish [Chou, 2018; Chang, 2020; Pang et al., 2018]. The original intent
of AI was to enhance and aid and, not to replace the human role.
Definition 2.2.2 (Artificial Intelligence in Diabetes). Digital health is

the modern era of technological advances within healthcare, such as,
wearable devices and remote monitoring, telemedicine and other di-
agnostic technologies that are optimising the quality of healthcare
outcomes. The main entity within digital health is the use of AI
analysing vast data sets from these devices, to facilitate improved
acute and chronic disease management protocols. AI will enhance
16 preliminaries
the clinical decision support mechanisms in preventing pathologies

from occurring such as the complications of diabetes, hypertension
and heart failure [Chang, 2020; Chou, 2018; Pang et al., 2018].
Definition 2.2.3 (Machine Learning). Machine Learning is a scientific

discipline in AI that focuses on learning capabilities and automated
improvement of a computer system based on previous experiences
and empirical data to execute a given task [Schwab, 2017].
Definition 2.2.4 (Deep Learning). Deep Learning is a sub domain

of Machine Learning (ML) that utilises a multi-layer neural network
with two or more hidden layers capable of learning from large amounts
of unstructured data [Ajayi et al., 2019; Schwab and Davis, 2018].
Definition 2.2.5 (Natural Language). Natural Language is human lan-

guage that has evolved naturally using basic rules and representa-
tions that do not relate to technological evolution which suggests that
using Natural Language is based on human biology.
Definition 2.2.6 (Natural Language Processing). Natural Language

Processing (NLP) is the computational approach to analysing text,
and more recently the spoken linguistic analysis, based on sets of the-
ories and computational technologies, for the purpose of producing
simple user friendly records [Liddy, 2001; Pang et al., 2018; Reddy
and Sharma, 2016]
2.3 medical definitions
Definition 2.3.1 (Diabetes). Diabetes Mellitus (DM), is a chronic metabolic

disorder that exhibits abnormal elevations of blood sugar, and which
requires effective sustained healthcare, to avoid or delay the onset of
chronic complications. Common to this disorder are Type 1 and Type
2 diabetes. Type 1 is a familial autoimmune disorder and manifests
early in life or at birth. In both types 1 and 2, the pancreas are defi-
cient in producing vital insulin, and/or the body muscle fails in the
absorption of insulin [Afzal et al., 2018; Fatehi et al., 2018].
Type 2 diabetes mellitus (T2DM), is caused by insufficient secre-

tion of insulin which is generally diet related. Type 1 diabetes mellitus
(T1DM) is a genetically inherited autoimmune disorder involving the
pancreas cells. Both these types of diabetes require a large degree of
self-monitoring and self-education, with life style management, that
can be complex in nature [Afzal et al., 2018; Fatehi et al., 2018].
The estimated number of diabetics world wide is set to increase

from 415 million in 2015, to over 640 million in 2040, which repre-
2.3 medical definitions 17
sents a global pandemic [Fatehi et al., 2018]. The two biggest com-
plications of diabetes are the neurological loss of sensation or pain
perception, in both feet, leading to a deficiency in micro circulation
in limbs, and feet in particular. These events are stabilised with effec-
tive treatment [Afzal et al., 2018; Fatehi et al., 2018].
The management of diabetes requires a large amount of self mon-

itoring and data analysis, making it ideal for any ICT web based
tools and technology intervention, such as telemedicine, telehealth,
mHealth, eHealth, digital health, robotics and ICT applications in as-
sisting the diabetic patient, from the cradle to the grave [Fatehi et al.,
2018].
Definition 2.3.2 (Diabetic Foot Condition or Syndrome (DFS)). The di-

abetic foot is experiencing a chronic loss of pain perception or nerve
sensation, which ultimately leads to an end-stage gangrenous (septic)
foot if untreated, which is commonly termed a “Diabetic Foot Condi-
tion/Syndrome” [Armstrong et al., 2018; Fatehi et al., 2018].
According to the World Health Organization, this conditon in-

cludes all foot complications within the term of diabetic foot syn-
drome (DFS) that has been defined as “ulceration of the foot (distally
from the ankle and including the ankle) associated with neuropa-
thy and different grades of ischemia and infection” [Armstrong et
al., 2018; Fremmelevholm and Soegaard, 2019].
The eventuality of a surgically amputated foot then becomes a

diagnostic code, under the International Classification of Diagnosis
(ICD) coding system. The aim is to promote the prevention of such
a condition eventually occurring which is now paving the way for
predictive analytical medicine. This will ensure better value-based
‘outcomes medicine’. The objective is to prevent such a devastating
outcome or “condition”. AI and NLP can create such an intervention,
by means of an effective algorithm to detect this event in due course
over time [Afzal et al., 2018; Armstrong et al., 2018; Jegede et al., 2015].
One of the most common causes of such an event is a diabetic

foot ulcer which becomes septic and gangrenous (chronically septic
and cellulitic). The emphasis is now placed on the value derived from
the improved ‘outcomes based medicine’ or ‘favourable clinical out-
come’. This proves that the prevention or preventative measures of
treatment for a diabetic foot, should ultimately prevent a lower leg
or foot amputation, and will always precede the diagnosis (ICD code)
of an amputated foot. We have devised an algorithm that is based on
the preventative value-based outcomes of such an event [Afzal et al.,
2018; Fatehi et al., 2018; Jegede et al., 2015].
The pre-empting of such a condition occurring, by a focused anal-

ysis of the signs and symptoms, found in clinical notes of the health
18 preliminaries
team is key to the prevention of a diabetic foot ulcer. The objective of

the algorithm is to pre-empt the fatal outcome, or at least raise aware-
ness of the common precursors within all the medical teams’ notes,
reports and narratives, to such an extent that a definitive decision or
intervention can be incorporated into the prevention of a diabetic foot
ulcer [Afzal et al., 2018; Fatehi et al., 2018; Jegede et al., 2015].
Definition 2.3.3 (The Medical Disciplinary Team (MDT)). The clinical

team will be any healthcare worker or provider within the team, man-
aging the clinical treatment plan of the diabetic patient. Each member
is a specialist in his/her own field of expertise and will enter narra-
tives or notes. Their respective clinical notes will all differ, with a
common theme or consensus, on signs and symptoms in order to pre-
vent a traumatic event, such as an amputation [Armstrong et al., 2018;
Fremmelevholm and Soegaard, 2019].
Definition 2.3.4 (Diabetic Foot Ulcer Signs). Foot ulceration and com-
plications are very common amongst diabetics where ulcers are the ul-
timate eventuality of such foot conditions and are difficult to treat and
resolve. They can take months to heal even with diligent treatment in-
terventions. If these diabetic foot ulcers are not treated properly they
will lead to soft tissue infections, then gangrene, which requires am-
putation [Afzal et al., 2018; Armstrong et al., 2018; Fremmelevholm
and Soegaard, 2019].
The final clinical diagnoses of such a chronic foot ulcer is based on

signs and symptoms: I. Presence of pathogens on culture of wound
specimen or swab such as debriding of the wound whereby staphy-
lococcus aureus (gram positive pathogen or coccus) is the most com-
mon organism found on culture. II. Wounds or ulcers are purulent
and oozing fluid which are mostly found on the plantar aspect of
foot or the balls and heels of the feet of a patient. III. Pedal pulses
are non-palpable, weak or absent. IV. Soft tissue swelling and red-
ness with streaking are suggestive of a surgical intervention such as
debridement or amputation [Armstrong et al., 2018; Fremmelevholm
and Soegaard, 2019].
2.4 summary of chapter
In this chapter, we have defined the terms used in the context of

the Fourth Industrial Revolution (4iR) technologies within modern
healthcare, bearing reference to a future concept of digital health or
healthcare 4.0. In Section 3.1, we review and discuss the literature
on previous work in the fields of healthcare economics and health
informatics, within the South African context.
L I T E R AT U R E R E V I E W A N D R E L AT E D W O R K
3
3.1 the south african healthcare landscape
S
outh Africa has nine provinces that consist of 52 district mu-
nicipalities, 237 local municipalities, and eight metropolitan
municipalities [Commission and others, 2013b; Telkom, 2015].
3.1.1 Introduction
South Africa has a three way parallel healthcare system: the National
Health System, the Provincial Health System, and the District Health
System. The three systems are centrally governed by the SA National
Department of Health with little autonomy given across the board
which includes financial budgeting which is managed from the SA
Treasury and the Central SA National Department of Health [Botha
et al., 2016; Herselman and Botha, 2016b; Commission and others,
2013b].
In order for these public healthcare systems to function seamlessly,

the administrative, financial and support services, planning and hu-
man resources, are provided through negotiated agreements or Ser-
vice Level Agreements. These integration points lie between the vari-
ous public healthcare divisions, from District Health to the Provincial
and National Governing bodies, including professional medical bod-
ies, across the country. It creates a complex system to manage, in an
efficient and streamlined manner [Harrison, 2010; Corp, 2013].
The challenges within the Public Healthcare System of the Na-

tional Department of Health remains an ongoing problem, as they re-
quire skilled managers and efficient business processes which are rel-
atively archaic and manual in nature. However, this is changing and
about to change drastically with the forthcoming national eHealth
strategy which will bring with it the digitalisation of healthcare or
Healthcare 4.0, commonly termed in Europe and the USA. The pub-
lic healthcare system is servicing the greater part of 85% percent of
the SA population of roughly 58 million inhabitants [Abbott and Ade-
Ibijola, 2018; Harrison, 2010; Ruxwana, 2014].
19
20 literature review and related work
The private healthcare system covers approximately 15% of the

population through private funders and insurers who enjoy the lions
share of the healthcare in the country. Simply put, public healthcare is
under-funded and under-serviced, whereas the private sector is over-
serviced and over-funded with private funding mechanisms. It is this
inequity in basic access to health services, for all South Africans, that
needs addressing by political transformation [Coovadia et al., 2009;
Harrison, 2010; Corp, 2013].
Government is addressing this imbalance by intending to promul-

gate the National Health Insurance plan in a country wide transfor-
mation drive of the public healthcare sector. However, the adoption of
the digitalisation of the healthcare sector, through the Fourth Indus-
trial Revolution and the national eHealth strategy, is critical to this
initiative [Harrison, 2010; Commission and others, 2013b].
3.1.2 The SA healthcare economic background
The SA healthcare system primarily consists of the private sector

(privately funded) and the public health sector (state funded), servic-
ing vastly differing populations with different budgets and funding
mechanisms [Coovadia et al., 2009; Mayosi, 2012]. The healthcare in-
dustry has many nodes of data or layers of “knowledge” comprising
of patient data, transactional data, analytical data (metadata), Busi-
ness Intelligence (BI) data, spatial or Geographic Information Systems
(GIS) management information data and patient data trends or health-
care information [Coovadia et al., 2009; Telkom, 2015]. The construct
of this vision in healthcare knowledge management is widely pro-
mulgated through the eHealth policies and tools of the World Health
Organisation (WHO) [Botha et al., 2016; Herselman and Botha, 2016b;
Theron, 2016].
The United Nations (UN) Millennium Development Goals (MDG)

were agreed upon, and signed by 189 UN Member States in 2000,
consisting of eight major goals [Botha et al., 2016]. They all commit-
ted to eradicating or alleviating poverty, disease, hunger, environmen-
tal degradation, illiteracy, and discrimination against women. These
MDG’s have subsequently been superseded by the Sustainable De-
velopment Goals of 17 integrated objectives or goals based on the
MDG’s, but broader in scope. The SDG’s were signed by all member
countries in 2015 [Chopra et al., 2013].
The South African National Department of Health (NDOH) has

subsequently adopted these goals for public healthcare within all
their future strategies in developing national strategic objectives and
plans accordingly [Botha et al., 2016; Ruxwana, 2014]. Universal ac-
3.1 the south african healthcare landscape 21
cess to healthcare (UHC), for all South African citizens, is the most
significant of the Strategic Development Goals (SDG’s) where the Na-
tional Health Insurance (NHI) plan, is at it’s core [HST, 2016; Moyo,
2012].
Further to this initiative is the World Health Organisation (WHO)

eHealth strategy, which has also placed itself on the SA (NDoH)
strategic plan of development, whereby the eHealth strategy will take
precedence across all future developments [HST, 2016].
This eHealth strategy has been adopted and applied by the South
African Department of Health (NDoH) in conjunction with the South
African National Development Plan (NDP) as a strategic plan of gov-
ernance, currently being implemented [Mayosi, 2012]. South Africa
needs to upgrade and modernise its public healthcare system: by
aligning with these global trends through the newly introduced Na-
tional Healthcare Insurance (NHI) model of 2011, which is currently
under development in its pilot phase. The NHI programme is envis-
aged to be fully operational by 2030 and projected to be developed in
phases over the next ten to 14 years [NDOH, 2017; HST, 2016].
3.1.3 The healthcare sector in South Africa
The South African healthcare system comprises of a dual system: the

public healthcare system; and the private healthcare system. The pri-
vate healthcare system has undergone an investigation, by the Com-
petition Board of the SA government, investigating the reasons for
the rampant escalation of medical costs outstripping the accepted in-
dustrial sector indices [Archer, 2016; Serfontein, 2016; Theron, 2016].
This makes healthcare practically unaffordable for the majority of the
SA population. Currently, only the higher income brackets can afford
private healthcare, thus healthcare is perceived as a luxury for a priv-
ileged few.
The private health sector originated from the mining houses in

the late 1800’s and early 1900’s where the mining companies built
their own in-house healthcare financing schemes and hospital sys-
tems, purely as a supportive function. The mining hospitals served
the black and white workforce alike, which later created their own
non-profit medical aid structures supporting their own in-house med-
ical and financing structures. These medical aids later departed from
the mines and proliferated into mainstream business organisations
which gave rise to modern day medical aid schemes [Harrison, 2010;
HST, 2016].
Today the medical aid industry is a multi-billion-rand industry,

serving approximately 30%, of the population. The private health-
care system addresses approximately 14% of the population, with a

budget of R189 billion plus per annum, which is self-funded by a net-
work of 215 private hospitals and 83 medical aids, servicing approxi-
mately 8.3 million people out of approximately 60 million [Coovadia
et al., 2009; HiSP, 2012]. The public healthcare sector services approx-
imately 84% of the population with a budget of R183 billion per an-
num through a network of approximately 4200 public hospitals and
clinics as shown in figure 1. [Serfontein, 2016; Telkom, 2015].
Figure 1: The economic overview of the SA healthcare landscape. Adapted

from:[Abbott and Ade-Ibijola, 2018]
The “Out of Pocket Costs” (OPC) spend, approximately R22 bil-

lion in the 2017 financial year, amounts to private money being spent
on unpaid benefits, from medical aids shown in Figure 1. This figure
is increasing yearly [Archer, 2016; Jeffery, 2016]. The OPC figure is in-
creasing as costs escalate and the corresponding benefits are similarly
curtailed by medical aids. Those who cannot afford medical costs are
dropping out into the public system which inevitably has to be car-
ried by the public healthcare budget, placing it under strain [Hay-
wood, 2016; Serfontein, 2016].
Drastic measures are called upon across the board as the figures
of the stated healthcare costs are now in question, The private Health-
care budget spend is not covering 14% of the population but a much
higher percentage, than the original figure stated. ‘Healthman Con-
sultants’ (2015) place this figure around the 38%, where the budget-
spend is much higher, as a result [Archer, 2016; Jeffery, 2016; Ser-
fontein, 2016].
Further, the analysts state that approximately 37% of the SA popu-

lation utilise private doctors. Approximate 29% of the SA population
make use of private hospitals, due to the rise of OPC costs. Simply
stated: private health funding utilisation is increasing due to the inad-
equacies of the medical aids in meeting all the members’ healthcare
needs due to decreasing benefits and increasing OPC costs [Haywood,
2016; Serfontein, 2016].
The imbalance is evident from the figures which are widely pub-
lished by the National Department of Health (NDOH) and Health
Systems Trust (HST) of SA. This remains a topic of great debate
amongst officials and analysts. This divide is growing, much to the
despair of the Ministry of Health [Haywood, 2016; Serfontein, 2016;
HST, 2016]. While huge improvements have been made in the access
to healthcare for all, and equitable health management, within the
public sector, since the 1994 elections, those very efforts have largely
been eroded by the burden of disease related to HIV/AIDS, weak
health systems management, bad management and low staff morale.
The result is a dire picture of poor health outcomes relative to the
total health expenditure in the SA public health sector [Coovadia et
al., 2009; Harrison, 2010].
It is now of utmost strategic importance for government to estab-

lish clear service structures, or mechanisms to retain medical person-
nel and make better use of private sector personnel, academics and
Non-Governmental Organisations (NGO). This includes the establish-
ments of private-public partnerships in the service delivery of public
health. This is yet to be utilised on a wider scale [Serfontein, 2016;
Harrison, 2010; HST, 2016].
The public healthcare financial model termed the National Health-

care Insurance (NHI) model, will ultimately give universal access to
public healthcare for all its citizens. The key objective of the SA NHI
is to provide access to Universal Healthcare (UHC) across all sectors
of the South African population [NDOH, 2015; Harrison, 2010; Hay-
wood, 2016].
3.1.4 The South African public healthcare sector
The NHI is primarily envisaged as a government-proposed financial

model, designed to pool and fund medical services for all, irrespec-
tive of socio-economic class. It will be a single master fund, publicly
managed and administered by central government. This fund will op-
erate on a pre-payment system ensuring cross subsidisation across all
the services within the healthcare system. The implementation will be
performed in various phases over the coming years i.e. in a 14-year
plan with a projected budget of R256 billion which has subsequently

been reviewed [NDOH, 2015; Mayosi, 2012].
Further, the NHI is aligned with the Bill of Rights, embedded in

the South African National Constitution: Section 27 envisages univer-
sal access to healthcare for the entire SA population [NDOH, 2015;
Mayosi, 2012]. "Everyone has the right to have access to healthcare
services including reproductive healthcare. The State must take rea-
sonable legislative and other measures within its available resources,
to achieve the progressive realisation of each of these rights. No one
shall be refused emergency medical treatment" (Section 27 of the SA
Constitution 1995) [Haywood, 2016; NDOH, 2015].
This is stated in the National Health Insurance (NHI) white paper

which was released in December 2015. It is published on the (NDOH)
website for the public to read. The high healthcare spend, or health
costs in South Africa, is listed as being one of the highest in the world,
and is currently being debated widely [Coovadia et al., 2009; HST,
2016].
According to the National Department of Health (NDOH), the

primary reasons are commonly known as the ‘quadruple burden of
disease’:
1. communicable diseases – Human Immune Virus (HIV), Sexu-

ally Transmitted Infections (STI’s), Tuberculosis (TB), Tropical
diseases (Malaria) and others,
2. non-communicable diseases – (“Western diseases” sweeping Africa)

cardiovascular diseases detailing high blood pressure, diabetes,
cardiac failures and cancer,
3. maternal pediatric or mother child diseases - Mortality rates are

abnormally high compared to other countries, and
4. trauma and injury related conditions - all forms of violence in-

cluding rape and motor accidents [Theron, 2016; HST, 2016].
The above burdens of disease are the primary causes of rising

costs, within the public and the private healthcare sectors, according
to the Minister of Health [Serfontein, 2016]. To address the burdens
of these diseases the NDoH has developed a five-year macro plan
for public health which is detailed in the “Outcome no 2” of the
Negotiated Service Delivery Agreement (NSDA). This is all encom-
passing regarding their vision: “A long and healthy life for all South
Africans”. Included in the NSDA document is the implementation
of the National Health Insurance (NHI) [Coovadia et al., 2009; HST,
2016].
3.1.5 SA e-Health national digital strategy 2019 -2024
The National eHealth strategy is of importance to the future of South

African healthcare, incorporating the laying of the foundations for the
NHI. This includes the conception of Healthcare 4.0 and the Fourth
Industrial Revolution technologies that will be of strategic importance
for the public health sector. The original eHealth strategy was drafted
for the period preceding the current strategy. The global eHealth strat-
egy was established by the WHO where all African countries and
member countries have a strategic plan in process, to address the dig-
italisation of health globally [WHO, 2012].
The initial eHealth strategy of 2012 has subsequently been re-

viewed and updated with new intentions of a fresher approach, ac-
cording to the SA government. This revised approach is to address
and provide opportunities in strengthening all its major health sys-
tems and technologies, nationally, with its origins in the National De-
velopment Plan (NDP) 2030 and vision for the foreseeable future [Com-
mission and others, 2013b].
The eHealth strategy document is concisely laid out with all crit-
ical interventions which must take precedence in the foreseeable fu-
ture. The overarching vision is to "achieve better health for all South
Africans being enabled by person-centred digital health". The strat-
egy is founded upon five principles of the eHealth strategy of access
to all, person or patient centered orientation, innovation, and digital
workforce with an all-encompassing government approach [Commis-
sion and others, 2013b; HST, 2016].
This vision can only be achieved by following nine strategic objec-

tives:
1. Leadership
2. Stakeholder engagement
3. Strategy and investment
4. Governance
5. Architecture and standards
6. Infrastructure and connectivity
7. Legislation, policy and compliance
8. Capacity and workforce [Commission and others, 2013b]
The first eHealth strategy which was written in 2012 produced

some significant improvements in the deficient health system regard-
ing technologies. In addition, the mHealth policy framework certainly
strengthened the strategic objectives since its enactment [Commission

and others, 2013a; Wolmarans et al., 2014].
3.1.6 SA public health projects
Over the past years various projects have been executed or currently
underway through policies or strategic documents (White Papers)
managed by the SA Government Information Technology (IT) agen-
cies and their partners [Harrison, 2010; Telkom, 2015]
1. The Department of Health Strategic Plan Document (2014).
2. The Department of Health ‘eHealth Strategy’ document adopted

from the World Health Organisation (WHO) eHealth global strat-
egy and toolkit (2012).
3. The National Health Information Systems strategic document

(NHIS).
4. The National Development Plan program document (NDP).
5. The National Department of Home Affairs with CSIR and oth-

ers in creating a national Health Patient Registration System
(HPRS) currently being rolled out.
6. The National Health Insurance financial model (NHI) White Pa-

per. Section 27: the ‘Health Bill of Rights’ ensuring universal
access to healthcare across South Africa.
7. The Government Wide Enterprise Architecture (GWEA) forum

and strategy document.De
8. The World Health Organisation (WHO) six blocks of strategy

document.
9. The United Nations (UN) consisting of 17 Strategic Develop-

ment Goals (SDGs), subscribed to by South Africa over a thirty
year plus period.
10. The Negotiated Service Delivery Agreement (NSDA) document [Com-

mission and others, 2013b; HST, 2016].
The policies or strategic white papers listed above are not all en-
tirely focused on healthcare but do overlap into the public healthcare
domains, besides the WHO, UN, NHI, NHIS and eHealth strategies
which are purely health related [Mayosi, 2012; HST, 2016; HiSP, 2012].
In accordance with the WHO eHealth strategy the SA Government

has endeavored to drive the following solution-based projects with
its own eHealth strategy through its partners [NDOH, 2015; HST,
2016]namely:
1. a national Patient Master Index (PMI) with a unique identifier –

a system currently being rolled out as a medical indexing card
system i.e. Health Patient Registration System (HPRS), as an
initiative with Home Affairs,
2. the IACT project in the Eastern Cape Province. HISP-SA has

been contracted to develop an Anti-Retroviral Treatment (ART)
module to capture the adult clinical record called Debo,
3. National Health Information Repository and Data Warehouse

Project (NHIRD) is supporting the National Department of Health
(NDoH) data quality across the country where many workshops
and data clean ups, are currently underway,
4. CDC contract to support the development of the National Health

Information Information System (NHIS) development in South
Africa (2011-2016), which is a project in collaboration with Jembi,
Health Partners SA and the, University of Pretoria. (The aim of
the project is to assist the Department of Health in strengthen-
ing the national health information system at national, provin-
cial, district, sub-district, facility and community levels),
5. DHER – District Health Expenditure Review. HISP-SA is de-

signing the District Health Information System (DHIS) being a
primary source of data for this review [Mayosi, 2012; HST, 2016;
HiSP, 2012],
6. web based surveillance facility management systems such as

the TB Electronic Disease Registers (EDR) which are facilitated
through Tier.Net software (HISP-SA has partnered with WamTech-
nology and the University of Cape Town Center of Infectious
Diseases and Epidemiology Research (CIDER) regarding this
initiative),
7. facility-based HIV management system involving data collec-

tion and analysis software (Tier.Net) which is the data exchange
between Tier.Net and third-party tools using a modified version
of the HIV Cohorts Data Exchange Protocol (HICDEP) format
(Tier.Net imports the data from the “Electronic Medical Record”
(EMR) systems. HISP-SA has partnered with WamTechnology
and CIDER regarding this initiative) [HiSP, 2012],
8. the vital registration of deaths and births across all areas includ-
ing rural areas,
9. health knowledge management guidelines such as best prac-

tices that are currently non-existent except for a few isolated
areas of expertise within the landscape,
10. mHealth comprising the use of mobile devices where there are
currently a few projects in operation in certain rural areas in
conjunction with the NGO (HISP) [Mayosi, 2012; HST, 2016],
11. Telemedicine in providing care at a distance,
12. the National Health Laboratory Services (NHLS) with its ‘Gate-
way’ project in progress,
13. virtual healthcare through collaborative health information and

education,
14. health research such as gaining vast coverage while utilising

large data sets,
15. real-time workflow of scheduling and dispatch, and
16. remote monitoring, diagnostics and environmental monitoring [Telkom,

2015].
Certain governing bodies have made progress in applying effec-

tive IT strategies, in the form of designing and developing new in-
formation structures called the National Health Information Systems
(NHIS). This has been planned in conjunction with the Health Infor-
mation Systems Program (HISP). The National Department of Health’s
(NDoH) National Health Information System (NHIS), has initiated
some programmes such as the Electronic Health Records (EHR) and
the Electronic Medical Record (EMR) with marginal success to date [Com-
mission and others, 2013b; HST, 2016]. Only a third of SA’s provincial
hospitals have embarked on some sort of electronic patient file to date.
According to HISP there is a pilot study underway in the Eastern
Cape, regarding EMHR operating on the district public health (DHiS)
system, operating on the protocol HL 7 version 2.4/3.00, which is a
global standard on healthcare informatics protocols [Botha et al., 2016;
Herselman and Botha, 2016b; HiSP, 2012].
3.1.7 Healthcare technology in South Africa
The private sector is largely siloed in its own respective environ-

ment’s with many different systems addressing the private health
sector, mostly operating on the Edifact protocol [NDOH, 2015; HST,
2016]. Moreover, the private sector fails to integrate or share the same
protocol with the public-sector ICT environment. This is a techno-
logical constraint of the private healthcare system. The Council for
Medical Schemes (CMS) is addressing this integration problem with
many projects to be initiated soon [Commission and others, 2013b;

HST, 2016].
Public healthcare utilises approximately 35 different systems across

nine provinces [Commission and others, 2013b; HST, 2016]. SA is a
member of the ISO/TC215 health informatics for data standards pro-
moting more efficient data interchange. South Africa subscribes to
the ICD-10 coding data system which is now the national diagnos-
tic coding standard. The HL7 ver 2.4/3.00 is the national messag-
ing standard in the public health sector, but has not yet been fully
adopted [Coovadia et al., 2009].
Mobile technologies have penetrated and proliferated across the

business and social sector within South Africa, giving rise to mo-
bile applications or mHealth technologies. The penetration of mo-
bile handsets is estimated to be between 80% and 90%. Analysts pre-
dict that this mHealth technology, will further expand rapidly [HST,
2016].
Telemedicine, medical informatics, marketing, surveillance and

healthcare education will experience similar growth. While these ini-
tiatives are being closely aligned to the eHealth strategy, within gov-
ernment, they are facing challenges in the implementation phase. Fur-
thermore, a great need is arising to build capacity within health infor-
matics, at all levels [Abbott and Ade-Ibijola, 2018; Telkom, 2015].
The eHealth strategy inherently has its own challenges and con-
straints, which are summed up by the Negotiated Service Delivery
Agreement (NSDA) [HST, 2016]. While there has been substantial
spending to procure ICT and NHIS in the past, it has largely failed
in meeting the basic requirements to support the business processes
within the healthcare system. Thus, rendering the public healthcare
system incapable of producing quality data information services for
the effective management and evaluation of performance in general,
largely due to the lack of governance and policies.
Besides these eHealth strategies, constraints and limitations, there

are challenges in the system that currently restrict the effective im-
plementation of the eHealth strategy rollout which needs reviewing
[Coovadia et al., 2009; Mayosi, 2012]. The eHealth strategy will form
the main framework to the roll out of the NHI model, as stated by the
Minister of Health, where the model will dictate the eHealth strategy
as the roadmap in achieving an effective national health information
system revolving around the patient [Commission and others, 2013b;
HST, 2016]
3.1.8 Current challenges within the South African healthcare Industry
The main challenges within the public healthcare system are the fol-
lowing:
1. healthcare data islands of data across various divisions with no

integration or interoperability,
2. hospitals and clinics lack interconnectivity and interoperability

structures on a common platform,
3. the centralisation of the Relational Database Management Sys-

tems (RDBMS) for effective Business Intelligence (BI) and Data
Mining techniques (there are projects in early phases)
4. a lack of standardisation across heterogeneous systems as the

main obstacle to interoperability,
5. a huge gap in skills development in support of these systems

where key skills are being lost to the private sector,
6. differing stages of system maturity regarding eHealth environ-

ments,
7. absence of a National Patient Master Index (PMI) primarily fo-

cusing on unique patient identifiers in the system,
8. broadband uptake in the public sector is too slow and expen-

sive,
9. minimal to completely absent inter-departmental collaboration

or basic interoperability across major public health and govern-
mental departments,
10. differing national and provisional budgetary spending,
11. too many islands of uncoordinated projects with no national or

global project management at a much higher level,
12. differing levels of organisational structures requiring standardi-

sation to create more efficient roll-up hierarchies, and
13. the lack of high-level coordination between the national eHealth

strategy and corresponding Enterprise Architecture (EA). This
should be based upon common standards evolving towards
standardisation on a local and a global template which requires
a high level of management commitment [Abbott and Ade-Ibijola,
2018; Archer, 2016; BCX, 2016; Mars and Seebregts, 2008; Ouma,
3.1.9 The National Health Insurance (NHI)
The following questions regarding the latest white paper of the NHI
(2015/2016), need to be considered for the implementation of such
technology programmes [Matsoso and Fryatt, 2013; NDOH, 2015; Wol-
marans et al., 2014].
1. Will the proposed NHI model function within the current state
of the Public Healthcare infrastructure?
2. Which governing body is overseeing the big picture of imple-

mentation and coordination of these respective programmes an-
swering to the overarching strategic objective of the Department
of Health (NDOH) and the wider National Health Information
System (NHIS) and,
3. Has the SA government made provision for these portfolios of

programs to be managed under a “Mega Project” banner re-
quiring specialist skills where South Africa is lacking [Flyvberg,
2014; WHO, 2012].
The NHI propagates access for all, to Universal Healthcare (UHC),

which has been legislated under ‘section 27’ of the SA National Con-
stitution which also in part applies to the private sector in part [Hay-
wood, 2016; NDOH, 2017]. The private healthcare sector has an obli-
gation to fulfil this right to the SA government and the people of
South Africa [Theron, 2016] by:
1. the implementation of an effective technological strategy that

will enable the architectural changes to facilitate the structural
improvements in the current public health model, ultimately fa-
cilitating access for all to universal healthcare as explicitly stated
in the National Health Insurance [Mayosi, 2012; HST, 2016], and
2. employing instruments and methods to measure disease out-

comes and management protocols, which will continue to be
critical to the understanding of the effectiveness of future knowl-
edge related public health strategies.
The ability to engineer a culture of change and, learning with in-

novation will play a major role in health care. The knowledge related
strategies will dominate the public and private healthcare domains
for many years, adding to the knowledge revolution [Theron, 2016].
3.1.10 The South African private healthcare technology
To facilitate an effective integration of the private sector with the pub-

lic sector, varying degrees of entities called Public Private Partner-
ships (PPP’s) need to be negotiated. The technological expertise and
platforms of the private sector need to be shared, by means of close
collaboration, with public health which will enhance the public health
sector, over the long-term [BCX, 2016; Telkom, 2015; Theron, 2016].
The technology platforms consist primarily of network topology

houses, facilitating the switching of transactions. The private health
sector network architectures follow the Edifact or EDI protocol, in
switching such data, which is a different protocol to the public health
protocol of HL 7 version 2/3 and the Health Information System (HIS)
2.00, called HISP-SA and currently supported the NGO [BCX, 2016;
Telkom, 2015; Theron, 2016].
These main private network switching houses are: HealthBridge,

MediSwitch, DH Switch, MEDiPrax, MEDIKREDIT, Mediscor, Inter-
Pharm, Med EDI, iTrack, Panacea and DataMax. These are the main
contenders in the private healthcare switching space, for patient and
transactional data, from service provider to funder, and vice versa.
However, there are many more smaller network providers, actively
involved in the private health sector [Abbott and Ade-Ibijola, 2018;
Telkom, 2015; Theron, 2016].
3.1.11 Discussion and analysis
The SA public healthcare sector is in dire need of transformation,

both legislatively and technologically. This can possibly be achieved
through the development of a solid platform by implementing, basic
technologies within the public sector with well-structured ICT tech-
nologies, laying a solid foundation for effective collaborative knowl-
edge management [Abbott and Ade-Ibijola, 2018; Botha et al., 2016].
This approach will encompass the public hospital/clinic networks
and data architectures that need to be re-engineered comprising over
4000 hospitals and clinics, laying the foundation for Primary Health-
care – community patient data structures [Wolmarans et al., 2014; Gov-
ernment, 2012b; Ruxwana, 2010].
In order for the healthcare system to evolve into an effective and

efficient system, healthcare risk management companies must align
the needs of the three major actors in the market. They are the con-
sumer, the healthcare provider, and the healthcare funder. It must
have a well-structured risk model, based on accurate historical clini-
cal and financial data [Prather et al., 1997; Ruxwana, 2010]. The new
NHI healthcare system can succeed if the necessary infrastructures

is in place, and government legislature is revised, in support of this
initiative [Botha et al., 2016; NDOH, 2017; Jeffery, 2016; Serfontein,
2016].
The South African Minister of Health has produced a structured

plan where these legislated initiatives are currently in mainstream
political thought. The most notable of these strategic initiatives is the
National Development Plan (NDP) [Botha et al., 2016; Commission
and others, 2013b; WHO, 2012]. It is a blue-print of development,
where the focus is, Universal Healthcare for all. The National Devel-
opment Plan (NDP) ‘2030 Vision document’, produced “Nine Health
Goals”, namely:
1. life expectancy to be aimed at 70 years old (currently its 63),
2. Tuberculosis (TB) prevention and cure,
3. to decrease and prevent Non-Communicable Diseases (NCD),
4. to decrease and prevent Maternal child mortality rates,
5. to reduce and prevent violent trauma,
6. to complete health systems reforms,
7. to improve Primary Health Care (PHC),
8. Universal Access to Health Care (UAC), and
9. human resources management to be improved drastically [Gov-

ernment, 2012b; NDOH, 2015].
From these nine strategic health goals the NDOH has subsequently
produced six important ‘Strategic Programmes’:
1. administration and patient files,
2. health Systems – encompassing the NHI,
3. HIV/TB – maternal child infections,
4. improving Primary Health Care (PHC),
5. developing hospital services workforce, and
6. health regulation and compliance [Government, 2012b; NDOH,

2015].
From these programmes a “Five-year strategic plan” was drawn

up and set out by the NDOH, namely:
1. decreasing the burden of disease and its prevention,
2. Universal Health Care (UHC) through the NHI,

3. re-engineering of the Primary Healthcare (PHC),
4. improvement of health facilities,
5. improvement of capacity planning, revenue collections, Human

Resources (HR) capacity within the supply chain of public health,
6. develop an efficient health management system for improved

decision making,
7. improve the quality of care regarding norms and standards, and
8. improve human resources through effective training and account-

ability measures [Government, 2012b; NDOH, 2015].
Special attention has already been given to the second programme,

(the enhancing of the health systems platform and the NHI to be
developed, in pilot phase) [Abbott and Ade-Ibijola, 2018; Botha et al.,
2016; Theron, 2016]. Legislative issues are increasing the uncertainty
in the private sector with new changes being promulgated.
In response to these challenges, a proposed re-engineering of health

industry legislation in South Africa is required, with regard to pri-
vate hospitals countrywide. Anthea Jefferey (2016), from the Institute
or Race Relations (IRR) and a strong advocate for access to Univer-
sal Health Care (UHC) suggested a review of the current healthcare
legislation, at a recent seminar held at the Free Market Foundation
(FMF). Dr. Jeffery (2016) suggested the following changes to current
legislature [Jeffery, 2016]:
1. public hospitals are in desperate need of experienced and qual-

ified managers,
2. private hospitals need to be allowed to employ doctors as salaried

workers. This is not currently not allowed, which completely
negates the “peer review” process within private hospitals –
many overseas hospital groups employ ‘resident doctors’ as for-
mal employees of the hospital and not as independent practi-
tioners,
3. presently private hospitals are prevented from training local

doctors where the local medical schools are far behind in meet-
ing the demand of the population - despite the current arrange-
ment of the Cuban student exchange programmes - this needs
to be further reviewed by government,
4. the concept of ‘low cost medical schemes’ need to be reconsid-

ered to cover the Prescribed Minimum Benefits (PMB’s) which
are currently throttling all private medical aids locally,
5. the government ban on medical insurance needs to be lifted and

addressed accordingly,
6. the Set Exit Pricing (SEP) of drugs which disallows bulk dis-
count negotiation must be reviewed as this could possibly bring
down the cost of medicines,
7. the poor standards of public healthcare and hospitals are shift-

ing patients into private hospitals from desperation and this
trend is increasing monthly,
8. the Medicines Control Council (MCC) needs to review its lengthy

process (5years) for drug approvals - re-engineering of these
processes are currently in place, transforming the registration
process down to less than a year, and
9. Government does not pay or takes months to pay its suppli-

ers. The supply chain needs to be seriously re-engineered as it
currently experiences ‘stock outs’ leaving patients with no med-
ication [Haywood, 2016; Jeffery, 2016; Serfontein, 2016].
The proposed new NHI model is loosely based on the American

“capitation” model. It is tailored to suit the unique South African
environment. The NHI is fundamentally a financial model or frame-
work, for collecting or pooling all medical funds and giving access
to Universal Health Care (UHC) [Archer, 2016; Africa, 2012; Theron,
2016].
The financing of the NHI remains a questionable subject, given

projections of R256 billion over the next 14 years, based on a projected
growth of 3.5%, Economists fail to see how this will be achieved as
the current growth rate is below 1%. Analysts are showing their own
calculated projections in excess of R550 billion [Archer, 2016; NDOH,
2015].
Moreover, it remains unclear as to who will ultimately fund this

huge figure, which is loosely based on a local tax base of less than
10%. Many challenges and unanswered questions lie ahead, concern-
ing the actual support and financing of the NHI. The fact remains,
from the studies conducted that the necessary information technol-
ogy platform, in addition to the collaborative sharing of informa-
tion within the ambit of data management, is imperative for health-
care efficiency. It remains key to the success of the future NHI sys-
tem [Archer, 2016; Haywood, 2016; Serfontein, 2016].
The analysis has a bearing on the technology alignment process

and specifically on the topics of alignment of IT with strategy, the use
of IT in a re-engineering exercise, IT advances in the health industry
worldwide and the transformation of the health industry in South
Africa by effective collaborative knowledge management [Prather et
al., 1997; Westphal and Biaxten, 1998]. It is therefore most important
that a strategy for the public health care sector is formulated and
positioned at ministerial or executive level [Haywood, 2016; Jeffery,

A few building blocks of the envisaged foundation need to be

laid down, however, such as a robust and solid national technolog-
ical (ICT) platform, within the SA public healthcare network. It is
currently operating below capacity for the demand [Abbott and Ade-
Ibijola, 2018; Coovadia et al., 2009; Ouma et al., 2011; Passchier, 2017].
3.1.12 Data analysis within public health
In addressing the data analysis capabilities within the SA public health-

care one cannot ignore the current deficiency of technology platforms
across the public healthcare landscape. Healthcare systems must inte-
grate seamlessly to facilitate a patient identity or hospital ID, across
the landscape, from the community clinic to the hospital. Currently,
this is not possible at a regional or national level. The data capturing
is done on paper based data sets, starting at the clinical visit. Futher-
more, this process is compounded by inferior systems and poor inter-
operability [Botha et al., 2016; Herselman and Botha, 2016b].
A plethora of disparate systems currently exist across the pub-

lic healthcare landscape, with poor to no integration. The NDOH
has laid down certain criteria addressing the possible stacks of stan-
dards, across these systems, namely the HL7 V_3 incorporating the
CDA and CCD, the ISO 13606, MIOS v4.1 and the IHE suite [Katuu,
2018]. These are essentially inter operable standards currently resid-
ing within the public healthcare architecture of SA. In addition to the
ICD-10 codes, procedure codes are also used in a variety of healthcare
systems, adding to the complexity [Katuu, 2016a, 2018; HST, 2016].
Data analysis requires a seamless integrated platform in order to

produce reliable data structures from a well-structured stack of stan-
dards across the landscape. Many see this as lost opportunities, to ex-
ploit critical clinical value from clinical data analysis; from improved
clinical outcomes to basic managerial operational data. The NDOH
is currently addressing this issue with the advance of the eHealth
strategy [Katuu, 2016a,b].
The SA government is laying down these foundations for effec-

tive interoperability, across all health systems, within the SA public
healthcare landscape. It is currently being tested by the implementa-
tion of certain clinical scenario’s being performed across clinics where
the sharing of healthcare information is monitored, with actual de-
mographic data [Commission and others, 2013b; Katuu, 2019; HST,
2016].
It will, however, require the NDOH to address the following:
1. a national eHealth standards board - instilling governance poli-

cies and processes,
2. review current policy decisions on security, data management

and privacy of clinical data within the clinical environment,
3. architectural standards, policies and processes to be developed

and deployed, and
4. a data dictionary for the eHealth and data analytical capabili-

ties within the context of SA public healthcare. A data dictio-
nary creates the uniformity, validity, availability, reliability and
consistency of such data.
Furthermore, it is critical that this collaborative infrastructure be

developed with high priority. It paves the way for effective data man-
agement and eventually data mining: from the basic retrieval of a
patient record, to the updating of such files in the system and then
retrieving such data sets from an analytical perspective [Mayosi, 2012;
Herselman et al., 2016].
3.1.13 Possible outcomes
Vast geographical coverage offers affordable convergence across fixed,

mobile, data and cloud. This would fill the immediate need for high
speed inter-connectivity between public hospitals, as a primary ini-
tiative in the form of a Wide Area Network (WAN) and Virtual Pri-
vate Networks (VPNS) providing high speed fibre optics coupled to
data centre hosting services as a possible solution. Cloud computing
can potentially provide an alternate solution, hosted by an enterprise
platform, such as Amazon services, as a suggestion [BCX, 2016; Mars
and Seebregts, 2008; Telkom, 2015]. The expected benefits from imple-
menting these measures:
1. synergistic platforms with Private Public Partnerships (PPP),

necessary for collaborative knowledge sharing,
2. robust transactional platform across the public-sector network

in support of all relevant business processes of the proposed
NHI model i.e. the primary healthcare network being the back-
bone,
3. standardising of data structures and protocols across such plat-

forms and including the supply chain transactional hubs, this
will include a massive re-engineering of its core processes and
data analytical capabilities,
4. the further development and enhancement of the Electronic Med-

ical Record (EMR) within a clinic or Hospital [Khalifa, 2013].
Medical/health record is centralised within architecture of dig-
italisation and centralisation of data, reading to and from the
national Health Patient Registration System (HPRS), with a na-
tional patient index record data base structure enabling effective
Business Intelligence (BI) tools to be deployed [HST, 2016],
5. the centralisation of patient databases within the public sector

creating a central repository for a national patient index sys-
tem (ID index file currently being rolled out with the Home
Affairs National Identity System (HANIS)), with unique iden-
tifiers that could be inter-operable with an Electronic Medical
Record (EMR), for a patient history from the primary clinic to
the hospital [Coleman et al., 2011; HST, 2016],
6. collaborative interoperability of all relevant transactional plat-

forms within the public healthcare sector,
7. an all-inclusive “pervasive connectivity” would include the SMAC

technologies, commonly referred to as the “Third Platform” where
Gartner calls it the “Nexus of Forces”, but “SMAC” is more
commonly used to represent this technology stack [Mgudiwa
and Iyamu, 2018 04; Weisinger, 2016],
8. cloud computing, supports big data technology which caters

for large volumes of data to be analysed and mined for trends,
and the tracking of patient performances and statuses regarding
the patient, care [Grossman and Mazzucco, 2002; Westphal and
Biaxten, 1998],
9. the importance of the creation of ‘data stewards’ or ‘Knowledge

Workers’, within the public sector. The importance of assigning
qualified and accountable people to manage systems and its
data structures,
10. the high-level coordination of such large projects and programs,

under the banner of programme or portfolio management, which
needs to be effectively managed on a large scale, and
11. the establishment of a project practice within the NDOH public

health structure for large project, or mega project management,
as listed above [Abbott and Ade-Ibijola, 2018].
The ongoing NHI project will require judicious high-level pro-

gramme and portfolio management at the highest degree regarding
national, district and primary healthcare strategic plans to be fulfilled,
on time and on budget [Abbott and Ade-Ibijola, 2018; Flyvberg, 2014].
Further, to this concept of mega project methodology, it must

be noted that the NHI architectural/technological platform require-
ments must blend or merge with the National Healthcare Informa-
tion System (NHIS) wider strategy, and national implementation, on
a grand scale.
3.1.14 Summary of SA Healthcare Landscape
The South African healthcare industry is rich in data, but poor in

information, or the interpretation of such knowledge from the data.
This information can be transformed into better disease management
practices and trends, ultimately reducing the ever-expanding burden
of the healthcare budget in this country. The focus must shift to pre-
ventative medicine, as opposed to curative medicine.
The SA public health sector should strive to formulate a design or

framework to establish an efficient technology ICT platform, paving
the way for an effective NHI healthcare model. It will require a re-
view of the current status quo of the public hospital networks and
data structures. Primary healthcare is currently at the focal point of
government, setting the stage for patient data architecture. The trans-
actional data lies within the procurement and finance structure which
need to be aligned with all future developments, within the public
healthcare sector.
The impact of the knowledge evolution will instil a culture of

change, learning and innovation which are party to this evolution
of knowledge. The objectives of the SA government lies in improving
the quality of and access to public healthcare, whilst reducing costs. It
needs to be matched in applying technologies through improving the
capabilities of healthcare information systems, overall efficiencies and
processes. Access to universal health care will be realised through the
collaborative sharing of patient and biomedical information. Thus,
creating a larger reach in lessening the spiralling costs of healthcare.
This complex public health initiative can only be managed as a ded-
icated mega project, which will ultimately increase the efficiency of
the future NHI.
In this section, we have analysed the South African Healthcare

(SAHC) system and presented the two differing worlds of the public
and the private healthcare sectors. This divide brings with it an imbal-
ance in the economic and political landscape. It skews the healthcare
spend towards the private sector serving a fraction of the popula-
tion. These investments have not produced the intended outcomes.
Health technology is advancing globally at a rapid rate. The Fourth
Industrial Revolution (4IR) is advancing rapidly across all industries.
Healthcare will be affected by this advancing technology. This can

only be achieved by the re-engineering the current state of the public
healthcare technological infrastructure.
In the next section, we present the technological advances of the

4IR, regarding the digitalisation of healthcare, and the potential it
holds for the South African healthcare sector.
3.2 fourth industrial revolution (4ir)
The Fourth Industrial Revolution (4IR) is undeniably in our midst.

We should not deny this fact. No other industry will be as profoundly
impacted as healthcare, considering the seismic shifts about to occur
within the South African public and private healthcare sectors collec-
tively. Embracing the 4IR means a committed attitude to not being left
behind and eager to face change. The private healthcare role in the
Fourth Industrial Revolution (4IR) will be pivotal, if not critical, for
the future development of the South African public healthcare sector.
The paradigm shifts within mainstream healthcare concerning 4IR
are manifold, such as: Ambient Assisted Living technologies (AAL),
Internet of Services, eHealth technologies, Internet of Things (IOT)
and telemedicine [Javaid and Haleem, 2019; Lin et al., 2018; Singh et
al., 2017].
The first industrial revolution originated in the 18th century with

the invention of steam engines which was brought about by the dis-
covery of coal. This gave rapid rise to the large scale industrial man-
ufacturing of textiles through the very first stages of industrialisa-
tion [Prisecaru, 2016; Schwab and Davis, 2018].
The second industrial revolution emanated in the early 1900s, where

the main energy precursor, was oil and electricity thus creating the in-
genious invention of the internal combustion engine. This led to the
production of the train carriage and first automobiles [Kumar et al.,
2020; Javaid and Haleem, 2019; Prisecaru, 2016; Schwab, 2017].
The third industrial revolution started around the 1960s where

more discoveries of nuclear and natural gas energy led to the inven-
tion of high tech machines such as the first computers and robotics
[Prisecaru, 2016].
The fourth industrial revolution is upon us today where fossil

fuels are now being replaced by alternative energies, referred to as
renewable’s [Prisecaru, 2016]. The 4IR is based largely on the revolu-
tion of new technologies such as the ushering in of 3D printing, the
internet of things (IoT), the internet, artificial intelligence and genetic
engineering with biomedical sciences, supporting industrial produc-
3.2 fourth industrial revolution (4ir) 41
tion and manufacturing [Chou, 2018; Kumar et al., 2020; Gröger, 2018;
Ghobakhloo, 2020; Masood and Sonntag, 2020; Schwab, 2017].
Many top industrial specialists are expounding on the potential

business benefits and technological leaps that 4IR will bring about
within their respective environments and other fields [Pang et al.,
2018; Schwab and Davis, 2018; Schwab, 2017]. The German commu-
nity of networks has identified that 3D printing and genetic engineer-
ing will have the most profound impact on industrial development,
however it will not be without its risks. This is based largely on the
advancement of technology connectivity and the Internet of Things
(IOT), to facilitate these, developments [Ghobakhloo, 2020; Gröger,
2018; Pang et al., 2018; Schwab and Davis, 2018; Schwab, 2017].
Deloitte Consulting has termed it, the digital transformation of the

manufacturing field which incorporates sensors, nano-technologies,
artificial intelligence (AI), robotics and quantum computing which is
becoming more accessible and affordable [Ajayi et al., 2019; Kraftová
et al., 2018; Martin, 2016]. This transformation is contributing towards
paradigm shifts in business models and outputs. Ultimately the fourth
industrial revolution is going to marginalise current jobs and produce
new skill sets and job portfolios in the near future [Prisecaru, 2016;
Ślusarczyk, 2018; Zheng et al., 2021].
The CEO of General Motors has predicted the inter-connectivity

of motor vehicles which will be smarter and greener, with efficiency
like never seen before. These interconnected cars will be electronically
self driven and powered by multiple energy sources. The future cars
will become wi-fi hot spots of connectivity to many devices and other
cars via the vehicle to vehicle (V2V) concept. The next phase will
be to connect to the infrastructure which increases safety and the
complete avoidance of hazards on the roads [Prisecaru, 2016; Schwab
and Davis, 2018; Zheng et al., 2021].
Economists are fearing that these sweeping technological changes

of digital, big data, robotics and connectivity technologies will have
profound impacts on labour markets which will eventually under-
mine fiscal polices through changing revenue and taxation models [Prise-
caru, 2016].
The modern medical landscape or architecture is changing with

major shifts towards integrative models, all based on digital platforms
or medical eco-systems, from the manufacturing of drugs to patient
care, and supply chains underpinned by these digital ecosystems. The
new ecosystems are emerging as the new era of patient care 4.0 or In-
dustry 4.0 capabilities [Sonnier, 2016; Schwab and Davis, 2018]. The
challenges are within the extensions of these medical ecosystems, fo-
cused on trusted integrated networks. This will apply pressure to
policy and strategy designers, who need to be the drivers of transfor-

mation [Chute and French, 2019; Javaid and Haleem, 2019].
The chapter endeavours to highlight the distinct advances and

the derived applications of the fourth industrial revolution regarding
digital health and patient care 4.0. It covers the advantages, disad-
vantages, risks and medical gaps with the inherent challenges facing
modern medical science, primarily within the South African health-
care sector.
3.2.1 The 4IR industries
The confluence of data and vast computational storage with smart

extraction technologies will transform most modern industries to a
level or degree that was never imagined before. This sweeping trans-
formation is approaching and the industries most affected will be
healthcare, engineering and medical sciences. Quantum computing
will defy "Moores Law" through sheer computational power that will
solve the most pressing problems of the world at speeds never expe-
rienced before [Ajayi et al., 2019; Prisecaru, 2016; Schwab and Davis,
2018].
The primary industries of the world to be directly impacted upon

by the fourth industrial revolution will be the following (according to
the World Economic Forum) [Schwab, 2017; Schwab and Davis, 2018].
1. Artificial intelligence and robotic technology
2. Manufacturing
3. Nanotechnologies
4. Healthcare and bio-technologies
5. Virtual augmented reality
6. Energy technologies
7. Material technologies
8. Computer scientific technologies
9. Geo-engineering
10. Graphic arts and culture industry
11. Space technologies
12. Drone technologies - regional surveillance technologies

3.2.2 Digital healthcare 4.0
The definition or description of digital health is complex, as it in-

volves a complexity of technologies building this landscape of con-
vergence, as a simple explanation. Each country has its own defini-
tion: from electronically or digitally, connecting all the points of care
in order to share medical information, in a secure manner, to the
application of digital tools, processes and services in healthcare deliv-
ery [Morrison, 2016; Pang et al., 2018].
According to Paul Sonnier (2016): digital health is the “the con-

vergence of the digital and genomics revolutions within healthcare,
health, living, and society” [Sonnier, 2016]. It is clear that digital
health remains in flux and is being reshaped by the thought leaders.
Digital health applications remain inclusive of the following essential
technologies: wireless devices, sensors, microprocessors, the internet
and internet of things (IOT), social media, mobile applications and
personal genetic information [Morrison, 2016; Pang et al., 2018; Son-
nier, 2016].
The fourth industrial revolution and digital health 4.0 are striving
and endeavouring to close the inequality gap within modern health-
care, being driven by the World Health Organisation (WHO). This is
driven by a set of 17 sustainable development Goals (SDGs), adopted
by many member states in 2015, striving to end poverty, starvation,
poor education, little access to healthcare, and improve energy and
clean drinking water. These are known as the global goals which
requires extensive re-engineering of policies and healthcare strate-
gies, underpinned by the global and South African eHealth strate-
gies [Coovadia et al., 2009; Prisecaru, 2016; WHO, 2012].
These eHealth strategies are currently driving the technological

advances in healthcare which requires innovation and creative appli-
cations, often driven by the private sector. The South African public
healthcare sector has adopted, and subscribes to, the eHealth global
strategy thus making efforts in aligning and implementing such strate-
gies. Various organisations and NGO’s have been established to steer
these eHealth strategies and policies within the SA public health
sector. The challenge lies in aligning its current digital architectural
structures to support this strategy for the advent of the National
Health Insurance plan, currently underway [Sonnier, 2016; Telkom,
2015; Theron, 2016; Wolmarans et al., 2014].
3.2.2.1 Digital health trends
The digital revolution is changing the landscape and creating smart

systems and eco-environments, setting the trend for an emerging dig-
ital transformation of health, and business [Javaid and Haleem, 2019;

Schwab, 2017]. Various trends will thus emerge:
1. National health strategies.
The emerging eHealth strategies and 4IR will place pressure on

governments and policy makers is adopting such technologies
for the evolution of their own digital health systems from local
wards to regions and then centrally to regions. The impetus is
adaptability and adoption of such emerging technologies is key
to its success [Coovadia et al., 2009; WHO, 2013, 2012].
2. eHealth.
eHealth strategies have been put forth by the respective gov-

erning bodies of the World Health Organisation (WHO) global
health and local health legislature of the South African National
Department of Healthcare (NDOH) [Ajayi et al., 2019; WHO,
2013, 2012]. In other words this is the digitalisation of public
health which is currently prescriptive in nature for all regions
to adopt. eHealth benefits are manifold for the patient, and they
are described as emergent healthcare technological transforma-
tion. It is the intersection of medical informatics and public
health whereby the public health services are enhanced through
the internet of services and all its encompassing digital tech-
nologies [Cabestany et al., 2018; HST, 2016].
It is the transformation of healthcare to a new modernised ver-

sion never seen before at local, regional and national levels. This
is largely a national strategy but also an important foundational
trend which is currently being adopted by all countries, in par-
ticular the five major emerging developing economies called
BRICS (Brazil, India, China and South Africa) [Cabestany et al.,
2018].
The digitalisation of public health being prescribed by the SA

Government for all regions have the purpose to: create superior
efficiencies, personalise the approach to all patients, universal
access to healthcare services and access to their own data. These
goals are to be achieved through a transformation process of em-
powering the patient. One of the strategic service deliverables
within this grand scope of eHealth, is the electronic health file or
record (eHR) [Cabestany et al., 2018; Reddy and Sharma, 2016;
Sonnier, 2016].
It will be the building block for all future digital efficiencies,

including access to a clean data record with advanced AI ca-
pabilities facilitating big data predictive analysis and advanced
analysis [Cabestany et al., 2018].
3. Borderless healthcare.
The future of modern medicine lies in the interconnectedness

of these medical eco-communities whereby healthcare will be-
come borderless, through digital healthcare tools. Patients will
be transferred or remotely treated from other regions or coun-
tries [Cabestany et al., 2018; Pang et al., 2018; Reddy and Sharma,
2016].
4. Chatbots.
Healthcare Chatbots are being developed for patients calling

in to medical aid and websites, seeking more information. The
next level of medical Chatbots lies in the aiding of fellow doc-
tors and health providers, in offering informed decisions of cal-
culated inputs by the provider. This trend is new and expanding
rapidly, with added benefits to patient and doctor [Cabestany et
al., 2018; Pang et al., 2018; Sonnier, 2016].
5. Motor vehicles.
The family vehicle is the most ’intimate’ or closely used piece

of technology that is utilised, on a daily basis. Car manufactur-
ers see this as their next market whereby they are now arming
all their latest models with bio-metric sensors for driver touch
points in the vehicle. The drivers vital signs, temperature and
even moods are monitored constantly for any deviation or ab-
normality detected by the system and an alarm raised to driver
and system [Cabestany et al., 2018; Pang et al., 2018; Schwab and
Davis, 2018].
6. Bio-printed 3D tissues.
This is a relatively new branch of medical science currently in

the development stage with tissue printing and regeneration.
Medical biological start-up companies are getting involved in
this emerging market and will be part of the future where bio-
materials, with growth factors, are combined in creating tissue
like structures. The material used is a type of bio-ink that creates
layer by layer, of these structures. It imitates or mimics natural
tissue. This technology is used to repair damaged organs by
depositing cells into a wound, or damaged bone or tissue and
can even fabricate heart valves or nerve tissue [Cabestany et al.,
2018; Pang et al., 2018; Reddy and Sharma, 2016; Sonnier, 2016].
7. Smart sleep alarms.
Our sleep patterns are symbolic of our circadian rhythms, or

patterns of sleep, which go through cycles of light and deep
sleep with more in-depth scientific relevance. Waking up to a
set alarm can be in the wrong cycle to wake, causing mild dis-
turbances in bodily function. The monitoring of these patterns

through sensor technology and AI will wake us, in the perfect
moment. It is gaining more focused attention by sleep clinics,
and the general population [Cabestany et al., 2018; Pang et al.,
2018; Reddy and Sharma, 2016].
8. Artificial Intelligence (AI).
Artificial intelligence based algorithms and services approved

by the FDA, will gain increased momentum and acceptance
into mainstream healthcare. The access to medical big data in
healthcare through the help of AI will bear valuable insights
never seen before in medical science. AI is improving efficien-
cies regarding research and development for the registrations
of certain drugs with pharmaceutical companies. AI based algo-
rithms are aiding oncologists in giving them options for treat-
ment plans. Deep learning and machine learning are currently
being adopted by mainstream healthcare with the successes of
approvals, coming from the governing bodies [Cabestany et al.,
2018; Gröger, 2018; Pang et al., 2018; Reddy and Sharma, 2016;
Schwab, 2017; Sonnier, 2016].
9. Internet of Things (IoT).
In recent years, the services of the internet have been profound

and pervasive across all areas of business and social life in gen-
eral. It has given rise to developing advanced biomedical sen-
sors and sensing technology that enables the physical and the
psychological monitoring of patients and people [Pang et al.,
2018; Schwab, 2017; Sonnier, 2016].
The IoT is seen as an ever expanding ecosystem that integrates,

hardware, software, devices, objects and people over an inte-
grated network which enables all these objects to communicate
by the collection and exchange of data in the monitoring of
the patient. The objective is to enable a smart decision support
mechanism with additional options for the patient [Cabestany
et al., 2018; HST, 2016].
Trends for these sensing devices are becoming highly customised

and personalised to the patient requirements of being sleek, natural
looking, unencumbered, long lasting, unobtrusive, non visible, hands
free and networked in the natural living environments of a person or
patient. These sensing devices are often embedded in clothing with
direct contact with the persons body or be embedded in a system or
device worn by the patient. They must be bio compatible with the ca-
pability of collecting large health data sets for the effective consump-
tion for big data analysis through smart algorithms in order to extract
derived meaningful information and insights for the patient [Chou,

2018; Pang et al., 2018; Sonnier, 2016; Schwab, 2017].
The data is generated across an intelligent, yet simple multi lay-

ered architecture that consists of four layers [Cabestany et al., 2018]:
1. First layer: sensory in the form of a patch or wearable.
2. Second layer: network layer.
3. Services layer: meeting patients requirements.
4. Interface layer: interactions with other users or applications.
3.2.2.2 Big tech in big pharma
The big pharmaceutical companies are now collaborating with the

high tech companies, in finding futuristic healthcare solutions. Big
tech companies such as Google, Apple, Amazon and Microsoft have
committed to healthcare technologies as their next big market: Apple
Inc. is addressing monitoring and sensing devices; Amazon is refin-
ing its logistics and distribution technologies and expertise; Google
specialising in analysis and big data mining with encompassing AI
and the taxi services such as Uber are, currently redesigning their
operational strategies to replace ambulances, amongst many others
[Gröger, 2018; Cabestany et al., 2018; Chou, 2018; Schwab, 2017].
The healthcare industry, in the USA, is a trillion dollar indus-

try and ripe for transformation and disruption. Only the big tech
companies can cause such disruptions from a technological perspec-
tive, as opposed to policies and governance changes. These envisaged
changes will be, and need to be, primarily to the benefit of the patient
and the practitioner [Chou, 2018; Schwab, 2017].
3.2.2.3 The future 4IR healthcare technologies
The manufacturing industry went through major changes in attaining

industry 4.0 status which paved the way for smarter manufacturing. It
soon spilled over into many other industries such as agriculture, com-
merce, transport, tourism, construction and now healthcare [Chou,
2018; Javaid and Haleem, 2019; Pang et al., 2018]. The influence of
the 4IR is pervasive and hence the inter-connectivity of healthcare
services is now key to the success of the future digital healthcare. It
will consist mainly of the following technologies [Chou, 2018; Mesko,
2017; Pang et al., 2018]:
1. Augmented reality - is an interactive experience where real

world objects are enhanced by electronically generated graphics
which can include certain modalities such as visual, auditory,

somatosensory, and even olfactory. The visual reality fields or
projected graphics are placed (transposed) over anatomical ar-
eas, locating an accurate picture of anatomical landmarks for
the sake of the surgeon or practitioners visual of a patient.
2. Virtual reality - the advancement of 3D virtual reality of certain

processes and procedures will prove critical in future training of
professionals on a simulation protocol. This includes the simula-
tion of surgery and, clinical skills training to train professionals
or refresh their skills. Healthcare has adopted this technology
widely as its proven to be most beneficial in a practical clinical
sense.
3. Robotics - the evolution of robotic systems in aiding patients

and providers is ongoing with the advancement of wi-fi and
other technologies. Telerobotic surgery locally and remotely is
being developed at immense rates of adoptability in mainstream
medicine.
4. Artificial Intelligence (AI) - AI is under development within the

healthcare clinical setting with smart technologies such as Nat-
ural Language Processing (NLP) with algorithmic driven pro-
cesses and outcomes supporting evidence based medical op-
tions for the doctor. AI encompasses smart sensing (IoT) and
Big Data analysis.
5. 3D printing - artificial and prosthetic limb manufacturing is be-

coming a home based practice as 3D printing technology is be-
ing developed to new levels of application across many tech-
nologies. Health device technology is being transformed with
the development of 3D printing technology. Not only can one
design and print modified organic tissue components with bio-
ink, but also the 3D prosthetic structures such as functional legs
and hands, with all their bio-functional joints and components.
6. Tissue engineering - a new class of bio-engineering where biomed-

ical tissue is engineered and enhanced with components to im-
prove healing with skin grafts for diabetic ulcers and cutaneous
implants for burn wounds etc. The component of this technol-
ogy is basically a bio-ink print used to create tissue structures
in a laboratory.
7. Portable diagnostics - sensor pads or electronic tattoos are devel-

oped for the monitoring of skin temperatures to heart beat sens-
ing in detecting cardiac abnormalities. There are many other
applications for such technology in the monitoring and sens-
ing of abnormal patterns of temperature, movement and bodily
functions in a live networked environment.
8. Digital tattoos - are similar to sensory pads or monitors but are

sleeker and take the shape of the anatomical area and are hardly
noticeable to the patient. They are easily clipped or stuck to the
patient body.
9. Nutrigenomics - biomarkers within genomic engineering have

the capacity to depict a pattern of DNA or DNA profile. From
this a suitable diet plan is engineered for the optimal function-
ing of the patient’s genetic profile - type of customised diet plan
applicable to an accepted genetic profile. The purpose is to im-
prove or enhance a clinical outcome for that patient.
10. Medical chatbots - commerce has been using Chatbots for many
years - healthcare companies are finally understanding the clin-
ical benefits. Medical Chatbots are finding application with doc-
tors, where critical information is often desperately needed. Doc-
tors with a heavy workload cannot keep up with the load of
mainstream information required; the Chatbot can fulfil this
need.
11. Cloud computing for healthcare - more healthcare companies

and national healthcare departments are considering, or have
already adopted the concept of the cloud computing offering,
with all its encompassing services and benefits offered. It is an
internet shared pooling of services and resources by virtue of an
outsourced basis of rich pools of resources on demand. Only the
users of this cloud technology are billed for the use of such in-
frastructure, as opposed to laying out large capital investments,
in purchasing and managing such technology.
12. 5G technologies - the advancement of these technologies will

depend largely, on 5G technology. Despite all the health warn-
ings it brings with it. 5G will vastly accelerate the future digi-
tal technologies within healthcare, amongst others. The speeds
of 5G bandwidth (100 times faster than 4G with a capacity to
support a million devices per square kilometre as opposed to
4G supporting a maximum of 4000 devices per square kilome-
tre). 5G brings unlimited digital opportunities within healthcare.
The largest impact is the success of telerobotics and surgical
robotic technology [Ajayi et al., 2019; Chou, 2018; Javaid and
Haleem, 2019; Mesko, 2017; Pang et al., 2018; Schwab and Davis,
2018].
Underpinning these technologies is a requirement for high avail-

ability of bandwidth, especially 4G and 5G. Within the literature it
states that 5G alone is destined to inject $12 trillion into the global
economy by 2035 where the USA will receive up to 22 million direct
jobs as a result . It is critical for the forthcoming healthcare technolo-
gies which will have a large dependency on fast stable high frequency
bandwidth and low latency period within. The high frequency band-
width of 5G will require more towers over shorter distances, as op-
posed to 4G in its current state [Ajayi et al., 2019; Mesko, 2017].
3.2.2.4 Robotic systems
The evolution of robotic systems has played an important role within

the second and third industrial revolutions, mainly applicable to the
manufacturing and automation production processes. It has been rev-
olutionary in its entirety merging into the service industry and, com-
ing from the manufacturing industry. Robotic systems are now pro-
viding automating nursing care to surgical procedures with far greater
efficiency, with less risk than in traditional approaches. However, tele-
surgery is still in its infancy stage. It is mostly dependent on a com-
munication layer, which is also evolving at rapid speeds [Martin, 2016;
Wiederhold, 2017].
The Gartner Group predicts that robotic health monitoring is en-

tering a profiteering stage. Technologies such as ‘biochips’, within
robotic technology, are currently under development [Gartner, 2018;
Martin, 2016]. The healthcare trend in Japan is the implementing of
such telerobotic and robotic systems, across the board. The devel-
opment of such technologies is becoming increasingly cheaper and
faster with time [Gkegkes et al., 2017; Martin, 2016]. Investors need
assurances that the ongoing challenges of ethical standards will be
overcome, which will pave the way for a plethora of new profitable
products, within the healthcare sector.
The advantages of robotic systems are manifold, with respect to

lowering costs, reduction of labour costs, increased independence of
the patient, increased quality of care and doing daily mundane tasks
or humanly impossible tasks. At this stage, it remains a challenge
to apply medical ethics, as we know it. Cost measuring structures
are changing rapidly with the improvement of the advancing tech-
nology in this field of science and engineering. It further applies to
measuring the valued outcomes, of any healthcare protocol or pro-
cess [Cabestany et al., 2018; Pang et al., 2018].
3.2.2.5 The challenges of the 4IR
Despite the effects of the fourth industrial revolution within health-

care and its potential for many groundbreaking digital solutions, for
future digital healthcare 4.0, it presents many challenges. The ar-
rival of 4IR will bring new operating models while the importance
of collaboration will be pivotal in the success of these 4IR technolo-
gies. Governments and politicians, including healthcare governors,
will need to address and drive new policies in adapting and adopt-
ing these pending technologies. It will involve new systems and ar-
chitectures driving these services and systems, for improved health-
care [Pang et al., 2018; Ślusarczyk, 2018].
Within the general SA healthcare landscape, the concept of 4G

WiFi availability , is currently sparse in rural areas. It is mainly con-
centrated in islands around cities of South Africa. The availability of
consistent bandwidth is a major challenge for Government. It requires
a drastic review of infrastructure and data costs, the most expensive
in the world. Neighbouring countries such as Angola and Botswana
have cheaper data rates, on average, and it is available in most rural
areas. It must change for the future success and benefit of the fourth
industrial revolution, within South Africa [Herselman et al., 2016;
HST, 2016; Mayosi, 2012; Ruxwana, 2014].
The issue on privacy and compliance to regulations will deter-

mine the success or failure of 4IR, across many countries. New stan-
dards and compliance’s for integration will need to be established
in addressing the new digital healthcare landscape, requiring new
system architectural landscapes addressing medical, legal, manage-
ment, financial and privacy issues. The integration of social media
to banking systems, supply chain logistics institutes, healthcare ser-
vice providers, hospitals and healthcare providers, will take on a new
level of interaction. It will require more cohesive engagement and
collaboration of these entities [Chou, 2018; Mahomed, 2018; Schwab,
2017].
The demand for advanced skills and different skill sets will grow
exponentially where those institutes will need to adapt to the grow-
ing demand of skills required. University programmes will change
with new ones introduced, and many old programmes being eradi-
cated. These new programs will be based on biomedical data engi-
neering, health data informatics, eHealth/digital health technologies,
smart system engineering, healthcare system engineering, security
specialist courses and data science development courses, amongst
others and many more down the line [Ghobakhloo, 2020; Gröger,
2018; Pang et al., 2018; Schwab and Davis, 2018; Schwab, 2017; Son-
nier, 2016].
4IR can, potentially, widen the gap of social inequality in SA, as

many will not adapt and up-skill where those opportunities arise. It
has the potential of creating an ever widening divide between these
two groups. The SA government is now under pressure to address
the skills gap in formulating policies and governance laws accord-
ingly [Herselman et al., 2016; Mahomed, 2018; Mayosi, 2012; Ruxwana,
2014].
3.2.3 Conclusion on 4IR
The 4IR has arrived and we all need to swim or sink in the process,
as described by many thought leaders and change agents. Political
and social policies have to change which must be implemented and
enforced to assist the transformations in a successful and meaning-
ful manner. This will prevent the displacement of jobs, currently the
concern. New technologies and skill sets must be embraced in order
to reap the benefits. Learning objectives need to be redefined and
government needs to be ahead in preparing the healthcare landscape
with an aggressive digital adoption and readiness. These developing
digital foundations should be observed from other countries, such as
Brazil and China, who have higher IT proficiency within their own
4IR efforts. It is both a global and a country wide effort towards this
digital revolution. The manufacturing sectors went through a similar
revolution. Now the healthcare sector is experiencing a similar rev-
olution with overwhelming benefits coming from 4IR. It cannot be
forced or pushed solely by legislative strategies, but rather by the
evolutionary forces of data driven technologies, and service delivery
driven efforts, focusing primarily on the patient, at the core of the
future digital healthcare 4.0.
3.2.4 Summary of 4IR
In this section, we discussed the evolutionary path of the four indus-

trial revolutions, over the preceding century. This included the devel-
opment of the associated technologies spilling over into the health-
care industry where analysts foresee this as having the most impact
of all the industries. We presented an overview of the 4IR technology
impacts on the healthcare industry. We discussed the advances of the
pending Healthcare 4.0 with its transforming attributes, including the
available technologies in creating these changes. Further, the merging
of the Fourth Industrial Revolution (4IR) and healthcare in creating
an evolved Healthcare 4.0 platform, including the patient-centric ap-
proach being enabled by the said technologies. In the following chap-
ters, we will discuss the challenges of healthcare interoperability of
healthcare systems, and the pending electronic healthcare record in a
clinical environment.
3.3 interoperability: a basis for 4ir
For many years, the eHealth strategy has been trying to gain trac-
tion according to the South African eHealth strategy, which was pro-
3.3 interoperability: a basis for 4ir 53
mulgated in 2012 by the South African Public Health Department

(NDOH) [Foster, 2013; Nat, 2012].
3.3.1 Introduction
The South African National Department of Health (NDOH) is adopt-

ing the accepted ‘eHealth strategies’ and objectives with all their as-
sociated health objectives and policies. The eHealth strategy is depen-
dent upon a robust platform, which is lacking, in order to facilitate
the system integrations and the fourth industrial revolution encom-
passing the digitalisation of healthcare(Healthcare 4.0) [Pang et al.,
2018; WHO, 2012].
In addition it includes the envisaged National Health Insurance

model which will require a relatively sophisticated national digital
platform in transforming the South African public healthcare model.
This is applicable when an electronic health record (eHR) appears
high on the NHI agenda. The South African public healthcare sys-
tems do not integrate at varying levels, proving to be a barrier to the
effective implementation of the eHealth strategy needed in the public
healthcare domain [Botha et al., 2016; Foster, 2013].
The SA health sector is facing three major transformation obsta-

cles: inequity, legacy of fragmented systems, and a service delivery
structure, which is slanted towards a curative, and not a preventative
health care structure. In South Africa, a problem remains, of hetero-
geneous systems with approximately 40 differing systems scattered
across the SA public health care landscape. Most of these public and
private healthcare systems do not have any interoperability capabili-
ties [Coovadia et al., 2009; Nat, 2012; Moyo, 2012].
There are no local or international standard practices in place that

have the capabilities to integrate externally or internally. This essen-
tially signifies an inherent lack of interoperability of systems within
the public sector, across the SA public health domain [Ardebesin,
2013; Passchier, 2017; Ruxwana, 2010; Telkom, 2015].
Interoperability is the ability for healthcare systems to share and

exchange clinical information, improving the clinical outcome. It should
be achieved by a process of seamless integration, but the systems in-
volved need standards, to operate, within their respective technolo-
gies. The eHealth strategy has a strategic objective in achieving seam-
less integration, through standardisation of technologies [Ardebesin,
2013; Foster, 2013; Nat, 2012].
The interoperability of systems forms the core of any effective

Knowledge Management (KM) platform, for any transactional envi-
ronment. Interoperability is critical in fostering a robust knowledge-

based platform with effective and efficient structures in place for fast
and easy accessibility [Benson, 2013; Moyo, 2012; Rupali, 2017].
3.3.2 Healthcare knowledge management
Knowledge management is a critical factor in global health which

bears significance for public health, by improving health outcomes
through merely connecting social systems and development systems.
This knowledge system will create certain dependencies ensuring a
reliance on the accessibility of the latest research, delivering quality
care. However, this requires well designed systems with the capabili-
ties to integrate seamlessly and smoothly. Currently there are gaps in
this domain across the SA public healthcare sector [Bali, 2013].
Within the context of this dilemma, the technology needed in

the health sector needs to be increased globally, to be effective. This
ever-pervasive healthcare knowledge management can only flourish
within a well-structured technology landscape i.e. a robust platform
[Bali, 2013].
Further, good knowledge management practices pave the way for

a robust platform to improve healthcare generally. Within the context
of global healthcare delivery, the challenges are across these wide and
diverse platforms, within healthcare environments. Privacy and secu-
rity will always remain contentious issues and a major hindrance, to
the overall progress of developing best practices within global health-
care platforms [Haslinda and Sarinah, 2009; Rupali, 2017].
To meet this healthcare information requirements, we need all

practitioners to collaborate in creating these strategic connections, for
the ability to tap into vital networks and bodies of knowledge, to
reap the benefits. However, it remains a global challenge as primary
healthcare is where the major information requirements emerge. Suf-
fers from poor infrastructure and scarce skills to perform these tasks.
A great deal of focus is shifting towards the social aspect of healthcare
needs: social media is destined to fill this gap [Haslinda and Sarinah,
2009; Rupali, 2017].
Interoperability is the ability of health systems and software ap-

plications to communicate and exchange data, to make meaningful
use of the information. It is the ability of health systems to commu-
nicate within and across varying organisational boundaries. Thus, it
advanced the status of an individual’s health by the effective delivery
of healthcare [Higman et al., 2019; HST, 2016].
There are essentially three levels of IT interoperability:

1. Foundational- allowing data to read from one system to another

without any interpretation.
2. Structural- defines the structure or format of data exchange to

be interpreted at the data field level.
3. Semantic- is interoperability at the highest level which takes ad-

vantage of both the structuring of the data exchange and the
coding of data, including the vocabulary, making it easier to in-
terpret the data. Thus making it interoperable across disparate
systems, such as an electronic health record system [Ardebesin,
2013; Higman et al., 2019].
3.3.3 The South African public healthcare challenges
In keeping with the foundations of the South African eHealth strategy

of building a robust electronic platform enabling the interoperability
of many disparate systems, the SA government has endeavoured to
lay down certain frameworks and norms. The SA healthcare system
has been plagued for many years by fragmentation with a distinct
lack of coordination and interoperability as stated by the Minister
of Health, which in turn has produced very little return on these
collective investments [Botha et al., 2016; Passchier, 2017].
The major factors are: widely differing levels of maturity of tech-

nology environments, islands of information not shared, broadband
connectivity too expensive, absence of a national master patient index
and a lack of convergence of these encompassing technologies to de-
liver much needed value [Ardebesin, 2013; Kotzé and Alberts, 2017;
Mayosi, 2012; Moyo, 2012].
The SA National Department of Health (NDOH) has laid out its

strategic intentions to address the issues and problems with infras-
tructure, connectivity, basic ICT literacy, financial planning and hu-
man resource planning [Ardebesin, 2013; Kotzé and Alberts, 2017;
Mayosi, 2012; Moyo, 2012; Mudaly et al., 2013].
3.3.3.1 The focus of interoperability in SA public healthcare
The lack of interoperability is evident thus far with a large variety of

standards of technology to consider. The SA government has spent
much effort and money trying to establish a clean set of standards
to bridge the misalignment across disparate systems [Nat, 2012]. The
standards of communication under consideration must facilitate four
main types of interoperability, in general:
1. technical interoperability: concerns connecting the systems and

services through interfaces and protocols etc. encompassing hard
core system engineering enabling machine to machine commu-
nication,
2. syntactical interoperability: concerns the messaging component

of communication protocols that must have a well-defined syn-
tax and encoding,
3. semantic interoperability: concerns data formats for the human

messaging content be clearly understood, by sender and re-
ceiver alike, rather than any technical component or protocol,
and
4. organisational interoperability: concerns the definition of the

business goal, processes, and collaboration across varying or-
ganisations regionally. This is dependent on the above aspects [Ben-
son, 2013; Katuu, 2018].
The eHealth strategy is creating the way for enhanced interoper-

ability as it is focusing on the following data technologies:
1. diagnostic coding (ICD 10),
2. diagnostic groupings,
3. procedural coding,
4. laboratory/pathology coding (LOINC),
5. standards for clinical information, and
6. pharmaceutical coding (NAPPI Codes) [Benson, 2013].
In addition to this, the NDOH will focus on creating bodies of

standards addressing these technologies and establishing a minimum
set of interoperability standards (MIOS) across the landscape. The
MIOS is a technical standard defined by the government for achiev-
ing interoperability and the compatibility of these differing systems,
across the public healthcare sector [Kotzé and Alberts, 2017; NDOH,
2014a; Nat, 2012].
Government has established a forum of senior level ministers of

departments called the ‘eConnectivity Forum’ which aims to deliver
the eHealth strategy in a cost-effective manner which answers to the
National Service Delivery Act (NSDA) ensuring a long and healthy
life for all South Africans. This led to the establishing of the National
Health Normative Framework for Interoperability (HNSF) which set
the foundations for interoperability, as demonstrated in the eHealth
Strategy South Africa (2012-2016) [Commission and others, 2013b;
HiSP, 2012; Kotzé and Alberts, 2017; Katuu, 2016b; Mayosi, 2012].
3.3.3.2 Health Normative Standards Framework (HNSF)
The HNSF was formalised by the Meraka Institute of the Council

for Scientific Research (CSIR) who subsequently collaborated with
the Nelson Mandela Metropolitan University (NMMU). This was the
development of the enterprise architecture within a standards based
framework and when fully developed it will represent an integrated
eHealth solution in both the public and private health systems [Katuu,
2016b; Kotzé and Alberts, 2017].
This exercise started out with an intensive study or review of the

current landscape and the current standards in place, which included
use case scenarios, reflecting the quadruple burden of disease:
1. HIV/ AIDS/ Tuberculosis management,
2. Mother and child health,
3. Diabetes disease management, and
4. Violence and injury management.
These use case scenarios were the basis to an information flow

requirement giving rise to the defining of standards for these data
sets. This resulted in a set of base standards or cohesive ‘stacks’ of
standards, with three ‘stacks’ of standards identified [Katuu, 2016b;
Kotzé and Alberts, 2017; Schabetsberger et al., 2010; Wolmarans et al.,
2014]:
1. standards based HL7 Version 3. Reference Information Model

(RIM),
2. standards based on the ISO 13606 / OpenEHR Reference Model

(RM),and
3. the standards base profiles developed by the global organisa-

tion – Integrating Health Enterprises (IHE) [Katuu, 2016a; Wol-
The standards based profiles and standards were further assessed

and developed into National Indicator Data Sets (NIDS) which gen-
erated additional requirements and recommendations:
1. NIDS should aggregate person-centric data,
2. an eHealth standards board should be established,
3. gaps are to be identified,
4. the standards of HL7, Clinical Document Architecture (CDA)

and the Continuity of Care Document (CCD) are to be localised
to ensure shared information within an electronic patient record

infrastructure,
5. data dictionaries are to be set up,
6. review the enterprise architecture,
7. cloud-based shared national eHealth platform to be established,
8. policies and legislation to be reviewed in support of the above,
9. monitoring, governance and testing mechanisms are to be in

place, and
10. standards are to be applied for a possible Electronic Medical

Record (EMR) system to exchange content data with internal
and external systems alike in a smooth manner [NDOH, 2014a;
Benson, 2013; Katuu, 2018].
The HNSF is mainly concerned with the semantic, syntactic, and

organisational interoperability of all relevant patient information. The
technical interoperability is not within the HNSF domain as it is pri-
marily concerned with data integration or messaging in a shared in-
frastructure [Ngwenya, 2018; Wolmarans et al., 2014].
The HNSF refers to three types of health records:
1. Electronic Medical Record (eMR): is an electronic record of an

event or a medical intervention within a single institute such
as a doctor’s visit. The service provider holds partial custodian-
ship of the content and it is termed a provider-centric record,
2. Electronic Health Record (eHR): is a longitudinal complete record

across many care settings over the patient’s lifetime, and
3. Personal Health Record (PHR): is a complete health record held

primarily by the patient or guardian of the patient which is a
person centric health record [Wolmarans et al., 2014].
The technical interoperability is outlined within the set of IT stan-

dards which is compatible with the Minimum Interoperability Stan-
dards for Government Information Systems (MIOS V_5). MIOS V_5
remains under the governance of the State Information Technology
Agency (SITA) that was established in 1999 [Katuu, 2018; Ngwenya,
2018; Wolmarans et al., 2014].
MIOS is integral to the foundation of the Government Wide Enter-

prise Architecture (GWEA) framework as MIOS prescribes the min-
imum architecture model and standards in achieving interoperabil-
ity. GWEA also prescribes the adherence to MIOS, in developing ICT
plans in government, from a pure technical perspective which does
not include standards to business process or functionality or security

practice standards [Katuu, 2016a; Wolmarans et al., 2014].
MIOS technical standards are applied to national, provincial, and

local levels in government in endeavouring to enhance the interoper-
ability of information systems across the public sector. It is frequently
reviewed and advanced. All eGovernment systems must adhere to
a certification process in adherence with MIOS. It includes all exist-
ing Supply Chain Management (SCM) and encompassing solution
development and, middleware with integration processes in meet-
ing the requirements, which are closely monitored and governed by
SITA [Katuu, 2018; Commission and others, 2013b; Ngwenya, 2018;
Wolmarans et al., 2014].
3.3.4 The South African public healthcare architecture
Enterprise Architecture (EA) is a comprehensive framework which

is used to align the organisations IT technological assets, processes,
people and projects underpinned by its operational processes. The
EA is aligned in such a design as to aid the support of the health-
care landscape in an effective and efficient manner thus providing
added benefits i.e. seamless blending of operational systems and pro-
cesses [Mudaly et al., 2013].
EA currently does not exist in the South African public healthcare

sector and remains a prerequisite in building the foundations of an
interoperability (ICT) strategy. The EA structure is of utmost impor-
tance in building the foundations for effective interoperability and in
keeping with the eHealth strategy and objectives [Katuu, 2018; Com-
mission and others, 2013b; Ngwenya, 2018; Wolmarans et al., 2014].
The building of this architecture can take on different forms and

strategies according to Coiera (2009) cited in Mudaly et al, (2013),
where the EA structure in Low/Middle Income Countries (LMICs)
and High-Income Countries (HICs) can be a ‘top down’, ‘bottom up’
or a ‘middle out’ approach in designing the EA technicalities of inter-
operability [Mudaly et al., 2013].
The middle-out approach creates a set of technical goals which

connects health providers, IT and government. This approach is geared
towards developing IT frameworks of standards and guidelines to im-
prove the Interoperability of their Health Information Systems (HIS) [Mu-
daly et al., 2013].
The HICs tend to adopt the top down approach as they have mas-
sive capital injections or budgets and the resources to advance this
approach. The UK has achieved great success in its public health ser-
vice. Unfortunately, the LMICs such as those in Africa do not have the
means and capital with skills and the much-needed resources, cou-
pled with immature system environments rendering them to adopt a
‘siloed’ approach in their respective healthcare technology landscapes
and Health Information Systems (HIS) [Mudaly et al., 2013].
The LMIC environments tend to depend on donors aiding them

in an ad hoc fashion which deters from a systematic approach to
health information system design, at a national level. Countries such
as Ghana and Rwanda have adopted the Open Medical Record Sys-
tem (OpenMRS), in addition to the Open Health Information Ex-
change (OpenHIE). Ghana, South Africa and Rwanda are currently
collaborating on common enterprise architecture on certain artefacts
and standards such as EA, interoperability, national registries, data
warehouses and data set exchanges [Mudaly et al., 2013].
The systematic evolution of a Health Information System (HIS)

and District Health Information System (DHIS) is imperative to the
success of enhancing the interoperability of any public healthcare
management system and policy making system. There are ISO stan-
dards of eHealth architectures available that can be adopted such
as the ISO DTR14639 which describes the components therein. The
Health Informatics Service Architecture (HISA) is a European stan-
dard that defines an open and interoperable architecture based on a
middleware layer [Mudaly et al., 2013].
Central to this theme of efficient interoperability is a centralised

National Health Information System (NHIS) that can develop and
promote collaboration between systems, applications and processes
(which is the primary aim of the NHIS). The eHealth strategy of the
SA government must be attuned to the national government Wide
Enterprise Architecture (GWEA) which uses the TOGAF and Zach-
man framework. TOGAF is used within many corporations and gov-
ernment departments throughout the world. TOGAF is described as
’The Open Group architectural Framework’ V_9.2 which is a business
architectural methodology [Katuu, 2016a; Mudaly et al., 2013].
These architectural frameworks with their encompassing key re-

quirements must support National Health Information Systems (NHIS)
and all respective local Health Information Systems (HIS) providing
effective interoperability across all systems within the healthcare plat-
form [Mudaly et al., 2013].
As a result of the internal challenges in interoperability, there are

more complexities on the external integrations into other third-party
systems, such as the private sector. Therefore, the nurturing of exter-
nal public private partnerships (PPP) is the latest trend in healthcare,
which is proving difficult with private public sectors, community out-
reach and vertical programmes [Mayosi, 2012; Moyo, 2012].
The integration of the public and private sector with all its associ-
ated services, pertaining to tuberculosis, HIV and non-communicable
diseases requires a review and revitalisation, including surveillance
and health information systems.
The following section will highlight these technical incompatibili-

ties across the public and private health sectors.
3.3.4.1 Current architectures in SA healthcare domains
The SA private healthcare landscape
The private sector is under the scrutiny of the Minister of Health,

who is trying to curb the rampant expenditure and cost ratios. Pri-
vate healthcare is almost unattainable by the common man in South
Africa. It is a reality where costs and inflation are outstripping most
other industry sectors which are well supported in the local media,
including health economists.
The private health care sector is a competitive siloe’d environment

with many profiteering corporations. Each with competitive strate-
gies and platforms of technologies built to outperform each other in
the market place. Patient-centricity is weak and the data structures
are found in patient market data, and financial data statistics, primar-
ily catering for payments and financial data.
The opposite is true for the public sector where it is more focused
on patient centric data structures and clinical data, devoid of any
financial data structures for the datamining functions of financial and
payment data systems. It remains a fundamental difference in data
systems, having an impact on the interoperability between the public
and private health sectors. The private sector has the capability to
adopt any technology, being competitive and profit driven [Archer,
Integration technologies are driven mainly by the leading indus-

tries such as Discovery Health utilising Healthbridge and the likes of
Pharmacare, Medikredit and DH Switch, with many smaller health-
care transactional switching houses, within the private sector Figure 2.
They dominate the market space within the private health sector and
are displaying advanced and sophisticated data mining analytical
tools. Their switching technologies operate on an xml-based protocol
called ‘Edifact’ [BCX, 2016; Telkom, 2015].
The medical aid houses have advanced technologies for private

medical systems and hospital group systems. They have well-developed
back offices supporting sophisticated data mining analytical systems.
It is based on being at the cutting edge of technology.
Organisation System protocol

HealthBridge XML/Edifact
MediKredit XML/Edifact
Pharmacare XML/Edifact
DH Switch XML/Edifact
Mediswitch XML/Edifact
e-MD XML/Edifact
MedPrax XML/Edifact
Net Pharma XML/Edifact
Mediscor XML/Edifact
Figure 2: Current private SA Healthcare switching topologies. Adapted

from:[BCX, 2016; NDOH, 2014a]
The SA Public healthcare systems
South Africa’s healthcare system sector or landscape is largely

fragmented and siloed in nature, with disparate systems across provinces
and districts alike. This renders it in operable with little integration,
primarily caused by the decentralisation of the National Healthcare
Management Information System (NHC/MIS) which was established
in 1996 [Botha et al., 2016].
Each province procured its own systems with little consideration

for interoperability standards at a national governance level, or any
provincial standards including district standards. Since then there has
been a concerted effort through central governance and structures to
comply to certain standards, through the eHealth strategy and others.
Hence, the SA healthcare landscape has many heterogeneous systems
that are unable to communicate with each other, at a provincial and
national level. Currently, there is a drive to create structures of inter-
operability, to standardise the sharing and exchange of information
and knowledge. The following Figure 3 depicts the disparate systems
across the provinces of South Africa:
Province Patient hospital system Protocol

Eastern Cape Delta 9 TCP/IP v_4
Free State Meditech, PADS TCP/IP v_4
Gauteng Medicom, MedSuite, PharmAssit, PAAB TCP/IP v_4
Kwazulu-Natal Medicom, Meditech, PALS, ProClin, ReMed TCP/IP v_4
Limpopo Medicom TCP/IP v_4
Mpumalanga PAAB TCP/IP v_4
North West PAAB TCP/IP v_4
Northern Cape Nootroclin System TCP/IP v_4

Western Cape Clinicom, Delta 9, PHICS, JAC Pharmacy TCP/IP v_4
Figure 3: Current SA public healthcare systems. Adapted from: [Botha et al.,

2016; Katuu, 2018; NDOH, 2014a]
At the primary level of care in South Africa it remains a paper

based file kept within the community health centre or the hospital.
The two paper records are disjointed with no link between the two
files, although they contain same information. There remains a lo-
cal paper-based patient master index (PMI). It is held in a local pa-
tient registry and, not shared with any other medical facility. The
system comprises of a facility and provider register, which may not
hold a common file for any individual patient. This model of op-
eration is found across most clinics, rural hospitals and provincial
hospitals [Botha et al., 2016; Katuu, 2018].
The challenge within government is to align and enable the health

information systems for integration to be based on a national shared
electronic health record (eHR). The paper patient records are aligned
with the structure of the eHR by means of a standardised form for
all medical records and results with pathology and laboratory notes,
laid down by the HNSF adhering to content standards and to align
the patient registries to the national population registry, or patient
master index file (PMI) [Botha et al., 2016; Katuu, 2018].
Integration is achieved through the application of interoperability

protocols by selecting a common set of standards across the land-
scape. Much research and analysis have already been undertaken.
Three main groups or stacks of standards to be considered from a
basket of 800 standards of interoperability, have been established. It is
worth noting that these ‘stacks’ of standards, are to be considered by
the SA NDOH, for the basis of seamless integration across the health-
care landscape. Including the community clinic to a provincial hospi-
tal at national level of governance [Botha et al., 2016; Katuu, 2018].
3.3.4.2 The HL7 V_3 protocol
Health Level Seven (HL7) is an international Standards Development

Organisation (SDO). It maintains and develops certain standards of
exchange and the integration of electronic healthcare information,
consisting of administrative and clinical functions. It is currently the
most widely used, and is headquartered in USA. It is based on cor-
porate membership, unlike other SDO’s, that are American National
Standards Institute accredited (ANSI). Level seven refers to the sev-
enth layer of the Open Systems Interconnection (OSI) model [Benson,

2013; Katuu, 2018; Kotzé and Alberts, 2017].
Currently HL7 V_2 is widespread which underpins the exchange

of admission information, discharges and transfer including results
and billing information. It has undergone several upgrades to V_2.7
and now V_3 giving support to interfaces and detailed clinical data
files with reference to a Reference Information Model (RIM). The RIM
is the clinical dictionary that supports the grammar and semantic
level of the information within the message. However, HL7 V_3 is not
reverse compatible due to the fact that it utilises two widely used com-
ponents namely the Clinical Document Architecture (CDA) and the
Continuous Care Documentation (CCD) [Benson, 2013; Katuu, 2018;
Kotzé and Alberts, 2017]. These two components are further widely
used in conjunction with the earlier versions of HL7 V_2x:
1. CDA – the Clinical Document Architecture supports medical

records which are a standard specification for the semantics
of clinical documents and is widely adapted globally. This is
primarily a mark-up standard that specifies the semantics and
structure of clinical documents for data exchange between providers
of healthcare. These clinical documents would include discharge
summaries, reports, laboratory reports, procedures and pathol-
ogy findings. CDA clinical documents are coded in Extensible
Markup Language (XML). This would include a primary detail
section with a main body section. The current version of CDA
is release 3 which serves as the foundation for all current imple-
mentations which is ANSI approved HL7 standard (National
Department of Health/CSIR, 2014).
2. CCD – the Continuity of Care Document enables interoperabil-

ity of clinical data, capacitating doctors to send electronic medi-
cal information to other regions or provinces without losing the
meaning of such records. CCD is a joint effort between HL7 and
the American Society for Testing and Materials (ASTM). These
are the files that carry the patient’s medical history which com-
prises of a set of templates of summated medical records. These
templates contain the medical history with treatment plans that
can be used in other CDA types. They also serve as a basis
for the interoperability in the US Health Informaton Technol-
ogy Standards Panel (HITSP), integrating the health enterprise
(IHE). It is an XML based standard which specifies the coding
and structure of a patient summary clinical document [Benson,
2013; Higman et al., 2019; Mead, 2006].
3.3.4.3 The CEN/ISO 13606 standards- OpeneHR
The Communication from the European Committee for Standardi-

sation/International Standard Organisation is an approved interna-
tional ISO standard which was designated to achieve semantic inter-
operability for eHR. The CEN/ISO 13606 is a standard defining a
data architecture for the communication of medical records between
healthcare systems. This includes the communication between elec-
tronic health records (eHR) systems or a centralised repository and
or middleware. Currently, a few central African countries are utilising
this protocol for their public eHR system [Mead, 2006].
3.3.4.4 Integration of Health Enterprises- IHE
There are many organisations and NGO’s (Non-Governmental Organ-

isations) that endeavour to create common standards of interoper-
ability across medical domains. One of them is the IHE (Integrated
Health Enterprise). IHE’s products apply the coordinated use of es-
tablished standards such as ISO, HL7, W3C, OASIS, DICOM and
IETF, underpinning patient care records [Katuu, 2018].
IHE claims to be international and has established many clinical

domain committees in this manner, among others: pathology, den-
tistry, cardiology, endoscopy, opthalmoscopy, surgery, pharmaceuti-
cal, radiology, oncology, laboratory work and public health. IHE pro-
vides standards-based frameworks for sharing information between
disparate systems that address critical interoperability issues between
service providers and patients, security, workflow and administration
of the workflow. IHE has more than 250 vendors globally and is gain-
ing momentum in the health market [IHE, 2016].
IHE follows a simple four stage process as shown in Figure 4

which illustrates the process of transformation and integration of
differing system domains. They provide the foundations, globally,
for the electronic health records and medical health records, by way
of common standards of interoperability and health information ex-
changes [IHE, 2016]. The functioning of IHE is internationally man-
aged by a board that requires membership alongside 200 stakeholder
organisations. IHE holds a liaison category A, alongside the WHO, in-
cluding the ISO Health Informatics Technical Committee [IHE, 2016].
IHE is primarily found in the private health sector, however, the

public domain is of greater importance as it affects more patients
in any given country. There remains no such integration platform
within the public sector that could be utilised in a similar manner. It
is forward thinking to integrate private platforms with public sector
platforms [Carr and Moore, 2003; IHE, 2016].
Figure 4: Four stage process in creating commonality of standards. Source:

IHE process:[Carr and Moore, 2003; IHE, 2016]
3.3.5 The challenges of interoperability standards in Africa
From the literature, it is evident that a plethora of standards exists,

which must be explored and analysed accordingly. Considering a set
of 800 standards, one must identify a stack of standards that will
support one’s electronic exchange of medical records for the future.
It remains an academic decision as to which strategic objective de-
termines the standards, within the clinical environment. However,
the standards in question can be difficult in acquiring, as there are
distinct boundaries in becoming a member of the Standards Devel-
opment Organisation (SDO) [Ardebesin, 2013; Schabetsberger et al.,
2010].
The adoption of eHealth is the buzzword as-is, is clearly carv-

ing a path within the digitalisation of healthcare, locally and glob-
ally. The transformation of healthcare is ever-present and needs to be
addressed. Multidisciplinary teams within healthcare will have cer-
tain medical information that needs to be shared, where the benefits
outweigh the disadvantages of electronic healthcare. The standard-
isation of protocols is the cornerstone in achieving interoperability.
When considering a common standard, the following protocols must
be analysed, debated and explored at length [Ardebesin, 2013; Cole-
man et al., 2011]:
1. identifier standards such as the ISO standards for the identifica-

tion of entities,
2. Messaging standards such as HL7 V_3 with CDA and CCD,

3. structure and content standards such as Logical Observation

Identifiers Names and Codes (LOINC) for lab tests. Including
the International Coding of Diseases ICD 10 diagnostic cod-
ing and the Systematised Nomenclature of Medicine – Clinical
Terms (SNOMED-CT),
4. electronic health record standards for EMRs and MHRs,
5. security and access control standards for such medical records,
6. however, in Africa and developing countries there are defined

barriers of entry and challenges in gaining access to these global
standards of interoperability which include: limited to no in-
volvement in the development of these standards; too little in-
volvement of diverse users of standards development; lack of
commitment to standardisation by government at national level,
7. the lack of foundational infrastructure,
8. the lack of guidelines for the implementation of such standards,

and
9. the lack of skilled resources and the capacity for standards de-
velopment [Ardebesin, 2013; Coleman et al., 2011].
To overcome these barriers at a national level there are certain

defined enabling environment components and two ICT components:
1. governance and leadership at a national level,
2. investment and Strategy plan or framework,
3. legislation, policy and compliance,
4. workforce and resources,
5. standards and interoperability, and
6. the ICT components are: infrastructural platform forming the

foundation with services and applications for systems.
3.3.6 Knowledge management in healthcare
Knowledge Management (KM) is an old term. It has been used in the

healthcare industry for many years with little progress in producing
real value. Aristotle was one of the first philosophers to consider the
abstract nature of metaphysical reasoning. He insisted that one must
think not only about knowledge per se, but also about the processes,
in seeking out the truth [Limaye et al., 2017; Metagroup, 2005].
Since the turn of the century, economies have changed from a

labour-oriented, production valued system to an intellectual and skill-
valued system. Technology makes up the strongest contribution, through
the means of interconnected systems [Metagroup, 2005].
3.3.6.1 The evolution of knowledge management in health
Knowledge management brings its own challenges of integration and

extraction of data driven value, within healthcare. It essentially comes
from many millennia of religion, philosophy, psychology, history and
economics and, more recently, business theory. It is only over the
last twenty years that it has become a dedicated discipline, or busi-
ness practice for the sole purpose of profit. The practice integrates
processes, systems and technology in a collaborative business man-
ner [Limaye et al., 2017].
Two types of knowledge exist: tacit and explicit. Tacit knowledge,

resident in individuals’ minds, is typically developed from experi-
ence. It is unstructured and informal, making it difficult to communi-
cate. Explicit knowledge is formal, structured, and well documented.
Thus, it can easily be communicated. Discrete elements of explicit
knowledge and information can be combined and analysed for gain-
ing perspective. While both tacit and explicit knowledge can con-
tribute to organisational knowledge creation, it is the interaction be-
tween these two concepts that creates powerful knowledge patterns [Li-
maye et al., 2017; Metagroup, 2005; Rupali, 2017].
The first generation of KM primarily focused on extraction of cod-

ified knowledge, mostly in the private sector [Limaye et al., 2017].
The second generation of KM then decided to focus on human learn-
ing and interaction capability with the use of technology. This led to
a proliferation of many KM strategies and methodologies. In other
words, there was an innate shift away from the actual technologies of
databases and data set technology to the more human discipline of ac-
quiring this valuable knowledge, within the organisation for its own
profiteering. A knowledge management culture was born [Limaye et
al., 2017].
The third generation of KM was soon developing deeper into the

people collaborative and cooperation nature, in acquiring this knowl-
edge. It is today’s founding practices of KM whereby these practices
are entrenched by way of people sharing knowledge. It is different to
the previous two generations that primarily focused on the technol-
ogy capabilities of collating data. Currently it is more focused on the
interaction of persons and organisation, in acquiring such data [Li-
maye et al., 2017].
To this end, a large proportion of knowledge management practice

draws upon a unique set of KM theories or frameworks for develop-
ing and implementing a KM system. These widely used frameworks
include the following [Limaye et al., 2017] :
1. The DIKW pyramid - functional relationships between data, in-

formation, knowledge, and wisdom in a pyramid model,
2. Demerest’s model- identifies the production of knowledge within

the organisation which is not exclusive to scientific inputs but
inclusive of social construction of knowledge where this knowl-
edge is then embodied within the organisation through explicit
programmes and social interchange,
3. Fred’s knowledge management model- according to Fred’s frame-

work of knowledge management maturity assessment levels, a
KM implementation consists of five levels:
First level – knowledge chaotic which is the vision at this early

stage.
Second level – knowledge awareness is the divisional knowl-

edge awareness.
Third level – knowledge focus covers the implantation aspects

of embedding process engineering, KM infrastructure, services
and training, support early adopters and finally monitoring.
Fourth level – this is a phase of adoption of these activities men-

tioned in level one, two and three.
Fifth level – is where the company becomes knowledge centric

placing a high value on knowledge assets and has reached the
highest level of maturity regarding the institutionalisation of
intellectual assets within the organisation.
Nonaka’s SECI model of knowledge dimensions – a model of

knowledge creation which explains how tacit and explicit knowledge
are converted into organisational knowledge. This process consists of
a ‘spiral effect’ of the accumulation of knowledge and growth birthing
organisational innovation and learning.
Skandia intellectual capital model of knowledge management –

the Swedish firm Skandia developed this model of intellectual knowl-
edge. The model focuses on the significance of equity, customer, hu-
man and innovation in measuring the flow of knowledge across its
networks of partners. It further assumes all intellectual property can
be transformed into valuable assets of the organisation [Haslinda and
Sarinah, 2009; Nonaka, 1994].
Each of these have widely differing contributions to KM within

global health, but they remain limited. This limitation has led to the
next era of generation of KM which is the social media aspect that can
buffer or fill this gap of KM within healthcare. One of the modern-
day tools in extracting this knowledge from mass volumes of data
within social media platforms is known as big data analytics [Limaye
et al., 2017].
Knowledge Management (KM) systems are integral in getting the

right information to the right people, by providing them with the
tools for analysing that information, which gives them the ability to
respond to the insight gained from that information. Knowledge man-
agement is a defined framework of disciplines within an organisation
that views and understands the organisation’s intellectual capital as a
managed asset. It is a about gaining insight into a business via many
avenues [Bali, 2013; Haslinda and Sarinah, 2009; Nonaki, 1991].
The old ways of doing business with TQM, re-engineering con-

cepts and activity based costing is outdated. According to Microsoft,
companies now understand that managing knowledge is key in re-
maining competitive in today’s business landscape [Metagroup, 2005;
Nonaka, 1994; Nonaki, 1991].
Knowledge Management (KM) has special reference to the health-

care industry in that it is the one of the most knowledge and data
intensive, of most industries with patient data and the unstructured
information of processes, diagnostics and preferred specialist knowl-
edge of commonly accepted practices. Knowledge-management is vir-
tually unexplored within the South African public and private health-
care sector. The private health sector is intensely isolated and compet-
itive in nature, with industry secrecy abounding, for obvious compet-
itive and profiteering reasons - juxtaposed to the public sector. The
two sectors hardly collaborate for the sake of improved health out-
comes within the South African context. However, great strides are
being made towards improvement [Limaye et al., 2017; Metagroup,
2005; Nonaka, 1994; Nonaki, 1991].
3.3.6.2 Information is the lifeblood
For hospital management purposes, access to information is critical.

Information is the lifeblood of the hospital, providing management
data that assists in the effective utilisation of available resources. For
example, using historical data one can evaluate the cost of procedures
in accordance with average time spent in operating theatres, bed util-
isation, materials used, etc. In terms of scheduling, having an online
system that allows one to instantly see what facilities are available at
any given time or what revenue has been derived from emergency
room inpatients, among others [Calhoun, 1997].
Information management and productivity have been boosted at

Afrox Healthcare with an investment in Microsoft’s net offerings. Wire-
less devices are now used to transact at the bedside, reducing time
spent by medical staff on administration and increasing the quality of
patient care. It has had a knock-on effect in terms of being able to bill
more rapidly. The mechanics of government and SA’s national wel-
fare, depend on information gathered from the health system. While
the government has taken some important steps on its way to the
delivery of an integrated national healthcare management informa-
tion system, there remain many challenges in the SA healthcare land-
scape [Calhoun, 1997; Haslinda and Sarinah, 2009].
In Figure 5 is an adapted knowledge flow diagram. It is generic

to most organisations and gives an account of the flow of knowledge
and information through an institute or industry. The flow diagram
was adapted from various thought leaders and authors.
Figure 5: The knowledge management flow cycle. Source: [Bali, 2013;

Haslinda and Sarinah, 2009; Klimko, 2001; Wiig, 2004]
3.3.7 A new era in patient care
In 2003, the news broke that scientists had successfully decoded the
human genome. This proved to be one of the most significant sci-
entific breakthroughs ever, providing the road map for the scientific
community to search for knowledge about life. The metaphor that
appropriately described the triumph was “a race to the starting line.”

The map of the human genome triggered a race to understand the
origins of diseases and how to combat them, as well as how genes
and proteins can influence a person’s well-being [Singh et al., 2012].
Information technology is an important driver when DNA se-

quencing could take place fast enough by using powerful systems
and high-throughput sequencing techniques, when the international
scientific community was able to decipher the entire human genome.
Today, it is virtually impossible to complete any biological research
without high-velocity computing, storage systems and software to
manage volumes of big data [Roche, 2006; Singh et al., 2012].
The phenomenal progress researchers are making to understand

Severe Acute Respiratory Syndrome (SARS) offers a case study in the
convergence of information technology and emerging biology. Sci-
entists at the Michael Smith Genome Sciences Centre in Vancouver
recently completed the world’s first genetic sequence of the corona
virus, believed to be responsible for SARS, just months after the dis-
covery of the new deadly virus. Their research tools included an
IBM computing system and high-throughput DNA-sequencing de-
vices [Jones et al., 2014].
3.3.7.1 Future implications for patient care
The bridges between information science and biology, are strength-

ening. The major challenge ahead is the improvement of the human
condition through better patient care. The average human life span,
in developed countries, has risen by 20 years. Much of this increase
in life span is the result of scientific breakthroughs, in the 1920s and
1930s, attributable to the discovery of antibiotics and vaccines. Infec-
tious diseases have declined dramatically and in developed countries
and various diseases have disappeared entirely [Ahmed et al., 2017;
Singh et al., 2012].
Influenza, pneumonia and tuberculosis no longer top the list of

leading causes of death. The top of the list now belongs to heart dis-
ease, cancer and stroke. We are on the brink of discovering the mech-
anisms of such diseases. If the research brings another 20-year leap in
human life span, our children and grandchildren may routinely live
into their 100’s [Ahmed et al., 2017; Singh et al., 2012].
The implications of genomic research in medicine are profound,

also for major medical research institutions, such as the Mayo Clinic
and Johns Hopkins University, including pharmaceutical and biotech-
nology companies. Aventis and “deCode” genetics, are currently un-
dertaking large research projects. Their research focuses on under-
standing the basic mechanisms of disease and the likeli¬hood that
those mechanisms operate in individual patients [Ahmed et al., 2017;

Singh et al., 2012].
3.3.8 The era of information-based medicine
Many research efforts are catalysts for a new era of medicine, where
doctors will have more information at their fingertips, to make better
diagnostic and treatment decisions, signifying the era of “information-
based medicine.” [Jones et al., 2014; Singh et al., 2012].
Doctors in the future, through information-based medicine, will

be able to diagnose and treat patients as individuals, not as statistics.
They will be enabled to make better assessments and prescribe better
preventive care long before any symptoms appear. By performing
simple diagnostic tests, doctors will be able to rule out drugs, likely
to have adverse or no effects on their patients. Doctors will select
the drugs designed to be most effective with minimal side effects.
They will be able to identify potential clinical trial participants more
effectively [Jones et al., 2014; Porter, 2009].
Today, most drugs on the market have been developed using a

“one size fits all” approach. Consequently, only one in three patients
taking prescribed drugs improves. What if appropriate markers could
be developed, ideally based on knowledge of mechanisms of action of
the drug candidate’? Suppose we could identify those patients who
will be high responders? It is called precision medicine and gaining
momentum [Howie et al., 2019]. The best example is a drug called
Herceptin, developed by Genentech, to treat women with aggressive
breast cancer. It is helping patients who desperately need and can
benefit from it — women identified through genetic testing, as high
responders [Roche, 2006].
The Mayo Clinic is among the pioneers of information - based

medicine, in collaboration with IBM, has archived more than 4 million
patient records — data collected from informed, consenting patients
into an advanced data management system. The Mayo Clinic Life
Sciences Warehouse can perform complex cross-patient correlations
across patient demographics, diagnostics and laboratory results. The
system can complete certain types of medical searches that once took
months, in a matter of min¬utes. The system will allow the Mayo
Clinic to conduct population based research more quickly on a broad
range of diseases and translate those research results into improved
patient care [Ahmed et al., 2017; BCX, 2016].
Another leader in information-based medicine is deCode genetics,

whose work in population genetics is creating a new paradigm for
healthcare. deCode has collected population data going back more
than a thousand years. The company is mining this data to isolate key
genes and drug targets, in nine common diseases, including arthritis,
schizophrenia and spinal muscular atrophy [Mesko, 2017].
Researchers at iCAPTURE in Vancouver are investigating links be-

tween genetic and environmental influences on the leading causes
of death in North America: heart, lung and blood-vessel diseases.
Through collaboration with IBM, this publicly funded research in-
stitute is building an information system to help researchers relate
genetic susceptibility of patients with cardio-vascular and respiratory
diseases to environmental influences, such as culture and socioeco-
nomic status, educational backgrounds, inhaled cigarette smoke, pol-
lutants, viruses, allergens, diet and obesity. It is only the beginning of
the revolution sparked by the convergence of information technology
and basic medicine. Through information-based medicine, we have
much to look forward to, including longer, healthier lives [Ahmed et
al., 2017; Jones et al., 2014].
The public healthcare sector is responsible for more than fifty Mil-
lion people of the SA population with little funding, bad manage-
ment practices and virtually no access to quality information. With
more effective use of interconnectivity and data management, a more
informed approach towards patient knowledge and the management
thereof, is possible to create a broader reach for the greater popula-
tion. This will address the burden on the South African public health
sector [HST, 2016; Mayosi, 2012].
Implementation of the new NHI model has the potential to save

the private and the public health industry in SA. Many more South
Africans will have access to public health care, or better healthcare,
which will relieve the pressure on the public health system. The vi-
sion cannot progress without a supporting IT model, which bears sig-
nificant implications for the health industry in South Africa. The col-
laborative sharing of information, such as efficient medical processes
and, data and disease management protocols, will offer affordable
and efficient healthcare to all citizens [Coovadia et al., 2009; Katuu,
2016a].
The consensus is that an IT infrastructure platform is essential,

for the success of the proposed implementation of a new model, for
the South African Public Healthcare System. Cloud technology could
be the elixir, for South Africa’s ailing hospital information technol-
ogy applications. Hospitals can use internet technology to integrate
older systems, in a phased approach, to improve patient care manage-
ment [Ajayi et al., 2019; Ochian et al., 2014; Weisinger, 2016].
Cloud technology can aid the many cash-strapped public health

state hospitals facing the dilemma of investing in new technology to
make their administrations more efficient. The approach should be
based on having a simple-to-use system that allows hospitals to link

up to the Cloud for accessing large data sets through client server
technology or an on-premise Cloud technology [Ajayi et al., 2019;
Ochian et al., 2014; Weisinger, 2016].
Cloud technologies now facilitate fast and cost-effective develop-

ment of new applications, that take advantage of information from
multiple information sources across the healthcare organisation. By
using existing Cloud technology to integrate critical applications, with
web technologies, it creates highly customizable front-ends, which
allow access to the information through the web [Ajayi et al., 2019;
Ochian et al., 2014; Weisinger, 2016].
A high degree of integration and interoperability is possible. Ap-

plications in the Cloud are relatively easy to prototype. Doctors and
other hospital staff can collect and define specific requirements, be-
fore implementation. Clinical IT technologists will continue to ad-
vance thus avoiding a capital investment into a new system. This
forms the substrate for the knowledge sharing platform [Ajayi et al.,
2019; Ochian et al., 2014; Weisinger, 2016].
3.3.8.1 Access to information
In the context of healthcare provision, a borderless environment saves

lives. Whether it be between different nations, regions, or hospital
wards. It is essential that administrators put in place systems that
facilitate the rapid transfer of relevant, accurate information across
secure networks.
3.3.8.2 Security
Health care providers are especially protective of data regarding their

patients health, as it can directly impact their career - they have to
be reassured that this information is being held in a safe place, far
from prying eyes. Governmental regulation has been put in place to
protect sensitive patient information and many have specified severe
penalties for organisations that misuse or mismanage this data. It is
therefore essential that the environment must be secured against theft
or hacking attacks [Ajayi et al., 2019; Ochian et al., 2014; Weisinger,
2016].
3.3.8.3 The Social Welfare Project (SOCPEN)
This is a successful cloud solution that the SA government has imple-

mented and utilised extensively, in distributing disability funds. It has
since been upgraded a few times but the Cloud concept has proven
to be successful. The logistics of maintaining and updating databases,
managing payrolls and keeping track of the thousands of pension-
ers, who rely solely on this income for survival are demanding. In
response to the dilemma faced by beneficiaries who are refused pay-
ment, including the widespread abuse of social security benefits, each
province has implemented an automated payment system which is
cloud based [Petersen et al., 2015].
The automated system makes use of smart card technology and

biometrics to manage and secure the payment process. Smart cards
operate in a similar way to ATM cards, but have an embedded chip
that can hold data such as the holder’s identification number, pass-
port number and bank account details. Biometrics enables fingerprint
verification, facial recognition or iris scanning, for the unique identi-
fication of individuals [Petersen et al., 2015].
On enrollment, the beneficiary’s information is downloaded to a

central database that is connected to a Wide Area Network (WAN).
Regional offices use mobile systems for enrollment and download
the demographic data to file servers connected to the WAN via Cloud
technology. Photographs and fingerprints are obtained using live scan-
ners so that the quality of the information can be checked immedi-
ately and updated easily. The card production system prints and ini-
tialises the smart cards, which contain encrypted data on the name
and ID number of the beneficiary, as well as fingerprints and a pho-
tograph [Petersen et al., 2015].
The secure encryption of this information, coupled with biometric

technology that enables fingerprints, to be recorded and verified, pre-
vents the abuse of social security benefits. The occurrence of fraud is
also avoided by careful management of the database and a payment
system which cannot be overridden. Details of all transactions and
withdrawals are instantly recorded and retained by the smart-card
to avoid over, or under payments. The government with the State
Provincial Authority has access via a secure intranet to the central
repository [Petersen et al., 2015].
3.3.9 Architectural recommendations
3.3.9.1 Conceptual design for the SA public health
The connectivity of over 4200 public hospitals and clinics, through

wireless connectivity and Cloud based technology, will enhance the
integration of disparate systems. It will bring together healthcare
data, from disparate systems, into a common dashboard. Thus, Cloud
technology remains an option which will cost far less than a full clin-
ical system solution in every province. The data is analysed from a
central repository, by the National Department of Health. They will
own and manage the data for health statistics. The data will be of
clinical and statistical value, as shown in Figure 6 [Herselman et al.,
2016; Kotzé and Alberts, 2017; Makovhololo, 2018; Schabetsberger et
al., 2010].
Figure 6: A new architectural Cloud solution for SA public health.
The population-based surveys give a measure of equity in terms

of development and demographic indicators. The health service util-
isation statistics provide us with measures of inequity of access to
health care access on a geographical basis. Very little integration of
information across the domains exist, i.e. quality, disease monitoring,
patient volumes, finance. Information from corporate information sys-
tems are not integrated with service and utilisation of the data avail-
able in systems, such as the Department of the Health Information
System (DHIS). Without the integration and interoperability of data
across the various domains, information will always be of limited use.
This is a lost opportunity for public health resource planners. Further,
management can be entrusted to make effective decisions within their

own environments at local government level [Makovhololo, 2018; Hersel-
man et al., 2016; HiSP, 2012; Kotzé and Alberts, 2017; Schabetsberger
et al., 2010].
Figure 7: A possible architectural conceptual design for SA public health.
A conceptual design is proposed in Figure 7, to harness lost data

and create a secure, efficient and cost-effective means to connect these
data sources, for effective retrieval and management. Thus, creating
value for the greater healthcare population of South Africa, industry,
government and healthcare planners.
3.3.10 Interoperability considerations for the SA healthcare landscape
The South African public health care environment has many chal-
lenges to negotiate in order to reach a level of value driven knowl-
edge management, within the sector. The first major consideration is
the integration and connectivity between hospitals coupled with the
enablement of these systems to communicate with each other via in-
tegration and interoperability of such systems. At a high level, they
are:
1. human resources or skills and capacity planning of such skills

- scarce to nonexistent, especially in outlying rural areas, with
only community clinics,
2. lack of training programmes or incentives to study and acquire

skills,
3. bandwidth problems are a constant headache with costs being

too high and services generally scarce in the rural areas,
4. lack of commitment and direction from government with well

defined strategies and policies in place,
5. general lack of large project planning or coordination of such

program/project plans at a district, provincial and national level,
6. lack of commitment from government as it might not have to-

tal decisive power over the provinces, besides the Minister of
Health,
7. lack of architectural foresight and the commitment to transfor-

mation of such a national initiative, and
8. the lack of funding - programme requires large funding budgets

which the country cannot afford at this stage [Coleman et al.,
2011].
3.3.11 Conclusion on Interoperabilty
‘Standardisation’ of healthcare system protocols is key, in ensuring

the exchange of healthcare information for the continuity of seamless
interoperability, within all public primary, provincial and national
healthcare services. It forms the basis for any medical knowledge
sharing or collaboration within the clinical environment – integration
and interoperability are key to building a knowledge management
culture. However, there are distinct barriers which need to be over-
come to facilitate the wide spread adoption of such standards, across
the landscape. Therefore, governments need to be proactive in the
adoption for such user - centered approaches. This can be achieved
by means of changing policies and legislature with a committed atti-
tude towards healthcare integration from the community clinic to the
Provincial Hospital. This includes the centralisation of the databases
of the Department of Health (NDOH) and other government divi-
sions enabling a seamless integration of services for the patient and
citizen. This could lead to ultimately unlocking further value from the
data, via a sound strategic policy based on a knowledge management
culture.
3.3.12 Summary on Interoperability
In this section, we discussed the importance of systems integrating

with each other, across the healthcare domain, referred to as interoper-
ability. Interoperability is critical for any seamless intra and cross bor-
der integration, of the public eHealth system. This integration is cur-
rently being promoted by the NDOH and the WHO’s, eHealth strat-
egy. South Africa is experiencing poor integration of public healthcare
systems. Moreover, interoperability is indicative of improved health
quality outcomes, influencing all members of the healthcare teams
and specialists, alike.
The technologies of such an integration were highlighted as well

as the challenges inherent to the interoperability of healthcare sys-
tems. This includes the challenges facing the Health Normative Stan-
dards Framework (HNSF) and the benefits of standardisation of terms
and technologies, across all systems and provinces in order to facil-
itate seamless integration i.e. the benefits of using a standardised
topology, such as HLV_3 for hospitals, widely used in Africa and
Europe. Further, the Integrated Health Exchange (IHE) and ISO stan-
dards, currently gaining momentum in public healthcare as an ac-
cepted standard.
The SA public healthcare architectures of the Department of Health

Information Systems (DHIS) were described with the accompany-
ing governing bodies aiding these architectures, such as Government
Wide Enterprise Architecture (GWEA). It included the breakdown of
the private healthcare architecture. The healthcare domain is in des-
perate need of knowledge management principles, and information
is the lifeblood of any healthcare system.
The era of information and quality based healthcare outcomes, is

upon us and South Africa needs to adopt such a culture and re-align
its future strategies accordingly. A new architectural design of the
SA public healthcare structure was proposed, as possible solution to
the integration and interoperability challenges, faced by SA public
healthcare. The private and public healthcare domains have the po-
tential for future integration efforts: unlocking numerous benefits for
the South African patient.
Finally, considerations for the interoperability of healthcare sys-

tems were highlighted, emphasising the value of a knowledge man-
agement culture, within the healthcare sector.
3.4 electronic health record 81
3.4 electronic health record
Fourth industrial revolution impacts several industries. Healthcare is

going to be impacted the most, reaping more benefits than any other
industry. At the core of the new digitalisation (Healthcare 4.0) of the
healthcare sector, central to the patient, is a medical electronic health
record (eHR). An eHR underpins the full collaboration of data points,
data objects, clinical services and processes which holistically under-
pins the full digitilisation of modern healthcare [Kotzé and Alberts,
2017; Kleynhans, 2011a].
3.4.1 Introduction
The global eHealth strategy, set out by World Health Organisation

(WHO), is pressing upon all countries to adopt the prescribed ‘eHealth
Strategies’. The global health objectives and policies includes the elec-
tronic health record (eHR) as a key objective. The WHO describes the
eHR as a digital record being shared across different healthcare land-
scapes, recording all the respective events data in a comprehensive
or summarised format. It must include demographics, allergies, med-
ical history, medications, status, lab tests, scans, immunizations and
billing information [Foster, 2013; Kotzé and Alberts, 2017; Kleynhans,
2011a; WHO, 2012].
The South African eHealth strategy, promulgated in 2012, expresses

the need for an electronic patient record, against the backdrop of a
specific governance framework. This strategy for eHealth within the
SA public healthcare segment of the economy has paved the way for
the development of a standards based eHealth enterprise architec-
ture [Wolmarans et al., 2014].
The eHealth strategy provides a roadmap for a future state digi-

tal public healthcare system, where the patient becomes centre-stage.
Within the eHealth strategy white paper, the implementation of an
Electronic Health Record (eHR) is implemented into a national pa-
tient registry with a Patient Master Index (PMI), currently underway
as a national strategic project within the South African public health
sector. At the core of such an initiative is the eHR [Geldenhuys and
Botha, 2015; Katuu, 2019; Masilela et al., 2013; Wolmarans et al., 2014].
Studies have been conducted over the last few years in the devel-
opment of such an eHR. However, it requires a transformation of the
public healthcare system. The upgrade of the health systems needs to
address the community clinic, district clinic and the provincial hospi-
tals across the full domain of public health [Kotzé and Alberts, 2017;
Katuu, 2016a].
The envisaged transformation calls for an upgrade to the current

inadequate health care system/architectural platforms underpinning
the electronic health record across the entire healthcare compendium.
The key structure is the primary health care system which is in dire
need of such an upgrade; aiding the support of the envisaged effi-
ciencies and advantages of such an electronic record [Katuu, 2016a,b,
2019; Mayosi, 2012].
The eHR is primarily designed to create a paperless environment

with a high degree of quality and standardisation across many health
systems and domains ensuring continuity of effective treatment. Most
countries are moving towards this trend of eHR which aligns with the
WHO strategic health goals known as the “Millennium Development
Goals” (MDGs) and “Strategic Development Goals” (SDGs) [Commis-
sion and others, 2013b; Kleynhans, 2011a; Kotzé and Alberts, 2017;
WHO, 2012].
3.4.2 Electronic Health Record (eHR)
3.4.2.1 eHR
Electronic Health Records are sensitive topics within any healthcare

environment as they demand focused attention, inter alia the security
of information, authorised accessibility or restrictions concerning the
patient record data, amongst others. Commonly reference is made to
an electronic medical record, versus the electronic health record. At
face value, both appear to describe the same thing or object, but they
are different in nature.
An eHR, also electronic Patient Record (ePR) or computerised pa-

tient record, is in digital format and includes a whole range of data in
comprehensive, or summated form, including demographics, medical
history, referrals, medication and allergies, immunization status, labo-
ratory test results, radiology images, and billing information” [WHO,
2012].
An electronic Medical Health Record (MHR), is a type of static

record within an institute or hospital. It is not to be confused with the
electronic health record (eHR) which is markedly different in nature,
being a more complete and a “fluid” document over the full lifetime
of a patient [Tan, 2005].
The ‘eHealth Strategy’ of 2012 incorporates ten strategic priori-

ties as the roadmap in guiding the public health sector to a well-
integrated and functioning digital health care platform, as set out by
the Minister of Health in numerous papers [Foster, 2013; Geldenhuys
and Botha, 2015; Kotzé and Alberts, 2017; Commission and others,
2013b].
3.4.2.2 The origins of the eHR
The electronic health record originated in the late 1960’s with many
software problems and challenges such as: no data dictionaries, sys-
tem interfaces and connectivity issues. These issues persist today. Typ-
ically, the content information in an eHR will include a unique iden-
tifier [Kleynhans, 2011a]. The eHR must be managed as a secure and
confidential file, meeting all the legal requirements, surrounding a
patients medical details. However, the true value of such a file lies at
the primary healthcare clinic [Ruxwana, 2014].
The SA public health ICT infrastructure must be upgraded. in

support of the eHR. Further, an efficient platform for an eHR file, in
the rural areas, will require a unique patient identifier, as many adults
and children do not have ID documents [Geldenhuys and Botha, 2015;
Kleynhans, 2011a; Ruxwana, 2010].
3.4.2.3 The benefits of eHR
The eHR will eventually find its way into mainstream digital health.
It is the foundation of all future data mining and data informatics
determinants, within the SA public health care sector. It forms the
basis to the future eHealth Strategy of the National Department of
Health. However, a substantial amount of strategic thinking and de-
signing with the current health care landscape is needed in building
the required ICT architectures [Kleynhans, 2011a; Kotzé and Alberts,
2017].
The concept of an eHR brings many benefit’s in the form of added

features, of the artefact i.e. (non-functional):
1. high degree of accessibility and flexibility,
2. portability across many domains and environments,
3. simplification of healthcare processes,
4. high degrees of security and confidentiality,
5. rich clinical data collated over long periods (life span) with a
high degree of continuity,
6. the archiving of medical data in large repositories, and
7. enabling the functions of sorting, searching and sharing: prepar-

ing for big data analytics and data mining.
Kleynhans (2011), in his thesis, asks the pertinent question if South

Africa is actually ready for such an eHR [Kleynhans, 2011a]. Many
healthcare analysts do share the opinion that the SA healthcare land-
scape has a long journey ahead. The first ever published strategy for
an eHR dates back to 2007, when a few advantages were tabled:
1. the ability to enhance the efficiency of the delivery of health

services across the board,
2. to achieve privacy and confidentiality,
3. to integrate the health record systems in the country through

the holistic integration of such systems within provinces and
across the provinces,
4. to improve the governance, management and planning of such

systems at local and national level,
5. the ability to monitor health trends and manage them accord-

ingly, and
6. the outcomes of such advantages were the following: reduce

human and medical errors,tracking the patients health over a
lifetime; recording all events into a single record; improve the
referral system with surveillance methods.
3.4.2.4 The eHR platform
The foundations for such an electronic health record form one of the
ten strategic directives, or priorities according to the ‘eHealth Strat-
egy’, drafted by the Minister of Health. It includes a requirement
for a central patient register, or repository for all the health informa-
tion systems. The eHR is central to the national ‘eHealth Strategy’ of
2012 [Geldenhuys and Botha, 2015; Kleynhans, 2011a].
In order to develop an eHR, with full interoperability with other

software systems in the landscape, certain foundational technologies
are required. Thus, transforming the environment into a more mature
environment, enabling seamless integration [Kleynhans, 2011a; Kotzé
and Alberts, 2017]. The key areas of government’s ‘eHealth Strategy’,
addressing these building blocks are the following:
1. infrastructure development is key across all facilities,
2. connectivity of such infrastructure,
3. registration of patients in a central database according to clin-

ic/hospital/facility and providers (approximately 4200 such fa-
cilities), and
4. a basic national eHR.

Currently there is no such technology in any public institute be-

sides a few pilot projects and interventions, attempting to build this
foundation. It excludes technologies such as an eHR record system, e-
training, e-prescriptions, e-referrals and an electronic booking system,
in rural and most urban hospitals. Accessing files, electronically, and
recording pertinent clinical information, will offer improved efficien-
cies by means of the elimination of duplication of paper prescriptions.
It adds to the security and identification of rightful patients.
The referral of patients between hospitals is also cumbersome,

with paper notes as referrals being lost en route, to the designated
hospital where the e-referral system will secure the process. The eHR
will keep a record of patient movements and their contact details will
be translated into alerts, for the patient to report to a specific hospital
or clinic to fulfil this referral effort. These referrals must be followed
through for the continuation of treatment, thus ultimately reducing
costs in the long term [BCX, 2016; Telkom, 2015; Wolmarans et al.,
2014].
3.4.3 The challenges facing an eHR
The eHR comes with its own set of challenges, which are well docu-
mented in other first world countries, such as the UK, including the
Middle East hospitals. These barriers for an eHR system appear to
be generic. Khalifa (2013) conducted one such field study (question-
naire to medical personnel) of two major hospitals in the region. The
barriers to an electronic healthcare record system were the follow-
ing [Khalifa, 2013]:
1. human beliefs and barriers to change are more apparent than

anticipated(a major barrier where personnel are inherently averse
to change and especially where it will require more effort ini-
tially before any benefits are realised),
2. professional barriers pose a change risk, seen as an intrusion

into the nature and scope of job descriptions and might pose as
a delegation of duties to others(implies clear and concise visibil-
ity of other practitioner’s notes and management protocols),
3. technical barriers regarding information technology encumbrances

will impose its limitations on staff and personnel to varying de-
grees,
4. financial management barriers where exact costing of such a

system is often sketchy and non-descript, and
5. the legal and regulatory barriers go beyond local boundaries to

possibly cross over into legislative and governance policies. It
applies when transferring patients inter continentally, or cross

border. It includes the eternal medical argument as to who owns
the patient data and ultimately has jurisdiction of such data [Khal-
ifa, 2013].
Notably, Ruxwana (2014) further states that the implementation of

an electronic health record continues to fail - mainly due to poor pro-
ject/programme management of such projects, according to a survey
of fifteen hospitals in the region of the Eastern Cape SA. The sur-
vey included various weaknesses of quality assurance programmes
in these hospitals [Ruxwana, 2010, 2014].
On a wider economic scale further dimensions of such impedi-

ments were noted which reinforces Khalifa’s (2013) findings. These
dimensions that, originate from a first world environment in the re-
gion of Serbia Russia, appear similar in nature, which are:
1. human resource issues,
2. the training of such resources,
3. the resistance of users due to changes in working practices and

duties,
4. technical issues such as the lack of integration with other sys-

tems and applications,
5. financial issues will always be a factor such as maintenance and

capital/operational and software license costs, and
6. certification, ethics and security and privacy with confidential-

ity issues.
Within the South African context it is worth considering the re-

strictions posed by the following policies and regulations which are
industry standards and policies, promulgated by the South African
government. This places further restrictions and challenges in rolling
out such a system.
According to two pilot sites or community clinics in Gauteng and

the Eastern Cape the manual input of patient data on spread sheets,
or patient ledger journal ticking was proving to be archaic. Many
patients could not be positively identified on the system as many did
not produce ID cards. A centralised patient master index system will
positively identify all patients at these clinics [Ruxwana, 2010].
The functional requirements have been processed through coun-

try wide Joint Application Development (JAD) workshops and doc-
umented accordingly. The overall requirement incorporates the abil-
ity to remotely identify, look up, review, manage, update and, record
medical findings with treatments. This includes managing the system-
wide settings and access levels. Moreover it will enable the managers
of such clinics to extract reports and, manage the super users and
settings [Geldenhuys and Botha, 2015].
3.4.4 The basic fields required within an eHR
The South African public healthcare patient requires a patient record

that can integrate, with other health care systems. The architecture
needs to be robust, well designed and well structured. It must inte-
grate this artefact with encompassing databases and interfaces, into a
centralised patient master indexed database such as the Home Affairs
systems for ID and details. It will require a high level of digital matu-
rity, ensuring the efficiency of a record system with the required gov-
ernance policies firmly in place [Ngwenya, 2018; Kleynhans, 2011a;
Geldenhuys and Botha, 2015].
The following is a basic list of required data fields when devel-

oping an electronic health record system. It does not include the full
stakeholder, solution and non-functional requirements for such an
electronic patient record system.
The eHR must provide sufficient patient data for identification

purposes of the correct patient, through a cross reference of name, ID,
address and patient file with a cell number and a telephone number,
available at any medical facility within the public health domain. The
following fields are the basic requirements at a high level of the fields
inherent within a basic eHR.
3.4.4.1 Demographic and analytics field
Name; initials; surname; ID; next of kin; guardian details; insurer or

medical aid; medical aid no.; employment; highest education; reli-
gion; marital status; no of children; unique identifier (UID); blood
groups; any current chronic conditions?; any current medication?;
current practitioner; disability status; pregnancy status; smoker sta-
tus; drinking status; Date of Birth (DOB); gender; nationality; family
linkage; telephone no; cell phone no; address lines or fields; postal
code; HPRS number (Health Patient Record System); allergies to list
in field; Referral yes / no; pre - conditions to list in field and Admis-
sion Diagnosis field.
3.4.4.2 Major medical events fields
Parity / gravidity; genetic markers; pre-disposition to illness; current

treatment; blood group; allergies; donor status and episode History.
3.4.4.3 Health clinic/hospital events fields
facility or clinic; healthcare provider; ICD-10 coding for diagnoses;

procedures (CPT-4); discharge summary (copied from file notes); med-
ication prescribed and dispensed and Lab results
3.4.4.4 Past medical history fields
Diagnoses: Treatment and procedures; medications; free text field;

hospitals / clinics attended; practitioners; dates of treatment; out-
comes; previous blood results – history and updates; Test results; vac-
cinations and confidentiality indicator.
The above fields complete the format, typically found in any basic
electronic health file or eHR. These fields consist of the demographic
and high level medical event register of fields. Moreover, this would
entail a high degree of interoperability.
The clinical medical fields would need to be further broken down

into more detailed fields, creating a high degree of integration and
complexity, across the whole patient profile. However, the eHR needs
to attach medical reports and scans, to be uploaded and viewed by
the attending healthcare professional.
3.4.5 Conclusion on eHR
The development of an electronic health record is more complex than

anticipated, due to various degrees of integration and interoperabil-
ity capabilities, needed. The environment must be mature enough
to support such a record system, including an encompassing data
model and sound architectural platforms in support of such an initia-
tive. The key objective with an electronic health record system is to
accurately and consistently identify the correct patient, through a cen-
tralised patient master index file, then to record the medical events
accurately and unambiguously against the patient’s unique identi-
fier. The eHR is viewed as a major advancement in the public health
care sector, ultimately leading the SA public healthcare transforma-
tion process. However, there remains many barriers and complexities
to navigate.
Apprehension to change, confidentiality, and security issues within

the greater context of the healthcare sector are concerns. More work
is needed in the public health sector, such as setting the foundational
technologies and enabling interoperability across such platforms. The
future eHRs will eventually re-engineer the healthcare processes, cre-
ating efficiencies which will translate into more efficient healthcare,

for all South Africans in the long run.
3.4.6 Summary on eHR
In this section, we have described the complexities of the technolo-

gies encompassing an electronic health record (eHR) in the SA pub-
lic healthcare system, in addition to defining the benefits and value
based health outcomes of the development of the eHR. The digitalisa-
tion of the health sector, under the stipulations of the South African
National Department of Health (NDOH) and the World Health Or-
ganisation (WHO), as a global initiative. We have endeavoured to
highlight the challenges, at the core of such an initiative namely the
concept of a patient Electronic Health Record (eHR), giving impetus
to easier instant access to quality driven patient healthcare records,
through eHealth strategies. The eHR underpins the full collaboration
of data points, data objects, clinical services and processes which un-
derpin the full digitilisation of modern healthcare. We further high-
lighted the need to upgrade the South African healthcare system and
architectural platforms. It remains a pre-requisite for implementation
of an eHR, across the entire healthcare compendium. The key struc-
ture is the primary health care system, in dire need of this upgrade.
An improved ICT infrastructure will enable the transformation of the
public health care system to a large degree.
Part II
THEORETICAL BACKGROUND AND

RESEARCH METHODOLOGY
In this part we have presented the theoretical background

to this study and the research methodology. In Chapter 4
we stated the theoretical approach to this study with an
overview of the critical and positivist approaches. There-
after we presented the key constructs of these approaches
that made them suitable for this study. This was followed
by clear rationale of these theoretical approaches. The chap-
ter concludes with critics and limitations of the theoreti-
cal approaches. In Chapter 5, we have presented the De-
sign Science Research (DSR) methodology as the adopted
technology in this study. This is followed by the rationale
for selecting the artefacts that make up the contributions
of this study. This chapter ends with details of sampling
methods that was adopted to evaluate all artefacts pre-
sented in this work, additional notes on data analysis, and
research ethics.
4 THEORETICAL BACKGROUND TO THE STUDY
T
his chapter outlines the theoretical basis and approach for this
study. Most theories are developed in order to understand,
explain, predict and then to challenge existing knowledge
within the limits of critical thinking, leading one to assumptions and
new knowledge [Bishop, 2015; Bhattacherjee, 2012]. The theoretical
constructs set the scene and defines the approach, from which the
study is viewed. The theoretical approach or framework outlines the
theory that defines the research problem statement and as to why the
study exists.
There are multiple complex paradigms of research, within many

domains of research. However, the choice of paradigm is key to the
successful outcome of the research conducted i.e 4IR technology ap-
plications in healthcare. The process we have followed in selecting
the appropriate research paradigms is informed by literature from
the works of [Brown and Dueñas, 2020; Venable et al., 2017].
The basis to our approach, is to establish a consensus of a ”posi-

tivist” outcome by means of solutions, or solution artefacts. This was
established by the critical analysis of the existing healthcare literature,
within the South African context. The literature review was extensive,
establishing a thorough body of knowledge that was critically anal-
ysed using the Critical Approach (CA). The critical analysis of the
literature produced many shortcomings which highlighted a few crit-
ical gaps in the technology domain. These gaps can be resolved by
designing and applying a few 4IR-driven solutions. The solutions are
designs of practical artefacts, designed and applied in the healthcare
domain, producing a ’positivist’ outcome.
This chapter is organised as follows. Section 4.1 explains the the-

oretical approach to this study. Section 4.2 provides an overview of
the positivist and critical approaches, while Section 4.3 highlights the
key constructs of these approaches. In Section 5.3 we presented the
rationale behind the selection of these approaches and listed the limi-
tations of these approaches in Section 4.5. The last section, Section 4.6
provides the summary to this chapter.
92
4.1 theoretical approach to the study 93
4.1 theoretical approach to the study
There are multiple complex approaches to research, within many sub-

jects areas of research. However, the choice of theoretical approach
and paradigm is key to the successful outcome of the research con-
ducted i.e 4IR applications in South African healthcare. The approach
taken in this is epistemological in nature, as it refers to the theory of
knowledge and the manner in which we obtain knowledge Alvesson
and Deetz [2000]; Hirschheim [1985]. Hence, the acquired knowledge
(from the results presented in this thesis) aligns with with a specific
research approach. We decided on Critical Approach (CA) for the rig-
orous review of the literature of the South African healthcare, com-
bined with a "positivist" approach based on our proposed new tech-
nological solutions that is aimed to address the issues and/or gaps
from the literature. The gaps identified with the CA presented oppor-
tunities to design 4IR artefacts to address the problems.
Further and specific to this thesis we adapted the Design Science

Research methodology (DSRM) specific to the development of a num-
ber of solutions or artefacts, in solving real-world healthcare prob-
lems. The DSRM is discussed in details in the next chapter.
4.2 an overview of the positivist and critical approaches
Positivist research principles are based on, and give meaning to the
creation of new knowledge that brings about positive change to the
world and improves the well-being of mankind Ryan [2006]. The ap-
proach has a firm belief of a full understanding based purely on ex-
periment and observation. These concepts of knowledge are the direct
products of experience which is further interpreted through rational
deduction according to Ryan Ryan [2006]. In addition these views
are subject to scrutiny. This modernistic model arose to as positivist-
empiricism which places a lot of value in quantifiable data or findings
which can predict correct answers Ryan [2006].
4.2.1 The Positivist Approach
Positivism is a theoretical approach based on genuine knowledge

which is positive in nature and is a natural phenomena derived from
ones senses and experiences [Kaboub, 2008]. It is primarily derived
and interpreted through reason and logic related to validated empiri-
cal data. The Positivist Approach (PA) further asserts that real events
can be observed in a scientific manner and further explained through
logical analysis. This can be validated at the micro level experimenta-
94 theoretical background to the study
tion generating scientific knowledge about reality [Hirschheim, 1985;

Kaboub, 2008].
The results are relatively precise and further generalised, based

on empirical findings. This knowledge gained is tested, measurable
and validated in a lab-like environment [Hirschheim, 1985]. This au-
thenticity of knowledge through observation and logical reason have
an iterative association [Bhattacherjee, 2012].
Positivist theory, in the information systems technology domain,

further assert that:
1. semantic expressions or words will remain inadequate in artic-

ulating what actually transpires in the world, and
2. the implementation of technology is performed by human ac-

tors which results in only one or a series of common experiences
for all [Bhattacherjee, 2012; Hirschheim, 1985; Kaboub, 2008].
Based on the premise of this Positivist theory approach, the logi-

cal application of software artefacts to address real shortcomings in
the healthcare domain is the motivation for the adoption of this ap-
proach. In addition, this approach was suited and based upon the CA
approach where these gaps were directly observed through logical
reasoning and analysis, we managed to establish the theoretical basis
to this research thesis — namely, the CA and PA. CA and Positivists
theory leads one directly into quantitative analysis. It consists of the
collection of data which is based on a measurement tool (i.e. vali-
dation tool such as a questionnaire for experts reflecting their own
evaluation) [Brown and Dueñas, 2020].
4.2.2 The Critical Approach
The critical research is fundamentally defined by the intent to change

society and improve our social standing and this is furthermore based
on critical research paradigms and methodologies Stahl [2008]. This
is further argued by the fact that critical research is motivated and
supported by morals and ethics in nature Stahl [2008]. The CA is
becoming an important stream in information system research which
has the ability to enrich our understanding of the current technology
and how it can be improved Myers and Klein [2011].
The critical research enhances our understanding and knowledge

by selective communication concerning shortfalls in the design of in-
formation systems and technology in general [Kvasny and Richard-
son, 2006]. The main objective is to force one to evaluate these short-
comings and openly criticise or challenge them Brooke [2002]. Fur-
thermore, critical research approach tends to focus on the experiences
4.3 key constructs 95
with certain IT artefacts addressing the shortfalls rather than on what

is working well Walsham [2005]. Premised on CA lies three impor-
tant elements namely insight, critical analysis, and the redefining of
transformation Alvesson and Deetz [2000].
From the CA review of the background literature of the South

African healthcare domain and applying a critical analysis of the cur-
rent healthcare technologies, including various shortcomings, led us
to understand the deficits and opportunities that modern technology
could provide a viable solution by means of solution artefacts. Based
on our analysis and experience within the healthcare industry we se-
lected five applied artefacts.
The solutions are practical artefacts, designed and applied in the

healthcare domain, producing a ’positivist’ outcome. In the next sec-
tion, we state the key constructs of the CA and PA approaches that
make these suitable for this study.
4.3 key constructs
4.3.1 Key constructs: Positivist Approach
Positivism is a philosophical theory which determines that true knowl-

edge is positive, been derived from experience and observation of
natural phenomena, and is true in nature [Kaboub, 2008; Howcroft
and Trauth, 2004]. This knowledge is derived through reason and
logic which is validated, and verifiable as positive facts. The physical
knowledge has and will always function according to defined laws
of nature. The acquired knowledge, through sensory experience and
observation, is known as empirical evidence. True to this deduction,
positivism is based on empiricism [Bhattacherjee, 2012; Brown and
Dueñas, 2020; Howcroft and Trauth, 2004; Kaboub, 2008].
In addition it rejects further metaphysics and even intuitive knowl-

edge as this knowledge is not verified through natural senses. The
knowledge of science is guided by three main categories; technical,
practical and emancipatory - inherent with the human factor [Haber-
mas, 1976]. This is also referred to as the habermasian factor in trans-
mitting new knowledge.
4.3.2 Key constructs: Critical Approach
The critical research pathway to any research, commonly referred to

as critical discourse, is an approach that analyses relationships be-
tween differing texts that can reinforce an imbalance or inequality in
the system, or a shortcoming [Howcroft and Trauth, 2004]. This can

only be achieved through defining, discovering and designing new
ways or concepts of knowledge.
There are essentially five ways, or themes, outlining the constructs

of a critical epistemology [Howcroft and Trauth, 2004]:
1. emancipation where it involves freeing up certain steadfast be-

liefs,
2. the elimination of those causes of domination [Alvesson and

Deetz, 2000; Myers and Klein, 2011],
3. further developing knowledge on the premise of minimum in-

put for maximum output, thus assisting managerial efficiency [Howcroft
and Trauth, 2004],
4. disrupting the status quo by deconstructing firmly held assump-

tions, and
5. the critique of traditional technologies and highlighting its inad-

equacies in furthering technical change [Howcroft and Trauth,
2004; Myers and Klein, 2011].
Critical theory forms a basis for modern thought or modernism. It

is also widely used in the social and health information sciences [Howcroft
and Trauth, 2004].
4.4 rationale for using positivist and critical approaches
The Positivist approach was incorporated in the research by under-

taking a thorough cross analysis of the literature, in developing solu-
tions or artefacts in a scientific manner. It would effectively address
the gaps or shortcomings in the South African healthcare domain,
through the application of 4IR artefacts, producing a positive impact
— albeit disruptive. The rationale for the Positivist approach were the
following [Brown and Dueñas, 2020; Kaboub, 2008; Ryan, 2006]:
1. the research is largely broad based where many concepts are

interrelated,
2. theory and practice is directly related regarding technology and

healthcare,
3. the commitment to solving real world human problems with

technology,
4. positivism remains a dominant public model for research,

4.4 rationale for using positivist and critical approaches 97
5. positivist research methods and quantitative method are related

to each other,
6. the data/knowledge collected mirrors reality,
7. we believe in reaching a full understanding of the solutions

based on scientific method of applying technology to human
problems within healthcare,
8. the commitment to research is central and critical to the well

being of mankind in general,
9. the research is based on scientific method and empirical data

collection for the validation of the designs,
10. the data collected is valid and rational which validates the de-
sign of the solution whereby the data is quantified, and
11. new knowledge is acquired which makes for further contribu-

tions to a body of science and technology [Brown and Dueñas,
2020; Hirschheim, 1985; Ryan, 2006].
From the Positivist approach, we decided on the CA. The main

overarching rationale was a critical analysis of a body of healthcare
and technology literature spanning many decades. This literature re-
search was analysed and summarised. The cross analysis was specific
to identifying the gaps in the South African healthcare domain. These
gaps presented opportunities of technological artefacts that can be de-
veloped by the direct application of 4IR technologies. To this end, we
sought the design of a few selected 4IR artefacts. The following are
the steps taken from our CA and PA theories to arrive at the selections
and designs of the new 4IR artefacts presented in this thesis:
1. An extensive systematic critique of literature specific to exist-

ing healthcare literature, within the South African context, un-
earthing critical gaps in the technology domain,
2. the identified gaps produced opportunities for solutions — by

way of 4IR technological artefacts,
3. the critical research is defined by the intent to change society, in

order to improve our social and ethical standing Stahl [2008],
4. our research has the ability to enrich and improve our under-
standing of the current healthcare technology and how it can
be improved through the application of 4IR technologies/arte-
facts Myers and Klein [2011],
5. improves our knowledge in the design of 4IR artefacts [Kvasny

and Richardson, 2006],
6. the objective is to force one to evaluate these shortcomings and

openly criticize and challenge the status quo [Brooke, 2002],
7. a critical approach focuses on the experiences with certain arte-

facts addressing the gaps, rather than on what is currently work-
ing well Walsham [2005], and
8. our research held three important elements namely insight, crit-

ical analysis, and the redefining of transformation Alvesson and
Deetz [2000].
In the next section, we present critics of the chosen theories.
4.5 critics and limitations of the chosen theories
Historically one of the main criticisms of positivism has been reduc-

tionism, where all scientific processes are reduced to physical or bi-
ological events [Horkheimer, 1972]. In addition, positivism ignored
the role of the observer in constituting the social reality. Positivism
is inherently too conservative and at times supported the status quo
without, challenging it [Horkheimer, 1972].
The main critics of positivist epistemology are of the belief that

no knowledge is neutral. In addition, this knowledge being black or
white is inadequate where the ethics of such research is always ques-
tionable. Including the fact that ’Positivist’ approach dominates in re-
search, where one depends entirely on aggregated data in revealing
an overall truth of knowledge [Ryan, 2006].
Similarly, in critical theory, it has been widely criticised by the-

orists who claim that a critical approach can be too subjective, and
too limited by the observers knowledge and communication skills. In
addition it has been widely criticised for not offering any future con-
sensus to a possible solution in solving a real world problem [Kvasny
and Richardson, 2006; Walsham, 2005]. In other words it can be limit-
ing in expanding the mind of the researcher, in developing a sustain-
able solution. This is not the case in this study where we combined
this with the PA, and consequentially designed solutions in form of
artefacts for the South African healthcare industry.
This chapter focused on the theoretical approach of the research, in-

cluding its theoretical positioning with some background information
regarding positivist and critical theory. The overview of the paradigms
of positivist and critical theory were further described, giving the in-
4.6 summary of chapter 99
dicative style selected for this thesis, namely the Positivist and Critical
Approach.
The chapter further expanded on the merits for such an approach

being well suited for this research thesis, in seeking solutions within
information systems and healthcare. The rationale for the selected
artefacts, consolidated the approach.
We further defined the critics and limitations to the selected re-

search approaches, including the benefits of its application. We be-
lieve that both CA and PA are complimentary in style, hence, suitable
for this research. Positivism and critical theory has evolved over the
decades, into a scientifically accepted methodology in information
sciences, including the designing of applied 4IR healthcare solutions
or artefacts.
5 RESEARCH METHODOLOGY
I
n this chapter, we present the research methodology adopted for
this work — Design Science Research (DSR) methodology. DSR
was used in the design, development, and evaluation of the arte-
facts and/or 4IR technologies presented in this thesis. These arte-
facts supports the positivist theory discussed in the previous chap-
ter. Paradigms are the ground rules for the research project that pro-
vides information in order to best evaluate emerging concepts or
artefacts [Brown and Dueñas, 2020]. The most important aspect is
to make the correct selection of the research paradigm in the very
early phases of one’s research. The areas of value must be identi-
fied and aligned with the research question that allows for logical
interpretation of the results, and then communicated effectively to
one’s audience [Brown and Dueñas, 2020]. This chapter highlights
the philosophical foundations of the thesis including the methodol-
ogy used.
This chapter is organised as follows. Section 5.1 outlines the philo-

sophical foundations for the Design Science Research (DSR). Section 5.2
further outlines the components of DSR as applied in this thesis, in-
cluding the risks and benefits of the DSR methodology. Section 5.3
supports the rationale for using DSR. Section 5.4 describes the study
population engaged in the research including the sampling meth-
ods incorporated and why we focused on expert sampling — in-
cluding the sampling size, and criteria used for these respondents.
Section 5.4.6 explains the process of collecting (i.e. data collection in-
strument) the respondents data with the aid of questionnaires, while
Section 5.4.7 presents the questionnaires used for all artefacts created
with the DSR in this thesis. Section 5.5 details the method of scoring,
while Section 5.6 outlines the ethical considerations. The last section,
Section 5.7 provides the summary to this chapter.
5.1 philosophical foundations of the study
Philosophy originated from the Greeks, which essentially means the

‘love of wisdom’ and is viewed as a specific applied thinking process
[Partington, 2002]. It is build upon the manner in which knowledge
is acquired and created across many fields of intellectual endeavours.
It forms a coherent body of knowledge over time through refining
100
5.2 the design science research (dsr) methodology 101
the knowledge and identification, selective abstraction, and recombi-

nation [Partington, 2002].
The four pillars of philosophical thinking pivot around metaphysics,

epistemology, logic, and ethics. The process of logic is more con-
cerned with the methods in extracting and capturing broad gener-
alisations as to how things interact in reality [Partington, 2002]. The
ontology of a subject is a knowledge repository which is essentially a
conceptualisation based on terms and definitions of concepts giving a
realist and relativist ontology [Bishop, 2015]. Ontology is based on the
pure nature of reality, epistemology is focused on knowing [Bishop,
2015; Partington, 2002]. These epistemological studies involves the
reflection of standards and methods in order to produce reliable out-
come [Partington, 2002].
In the traditional sense, quantitative research approach is based

on post-positivist inclined epistemology. Qualitative research is based
on the constructionist or interpretive epistemology [Bishop, 2015]. It
leaves the researcher to adopt a positivist paradigm or a construction-
ist paradigm which will in all probability have an influence on the
design and evaluation process [Bishop, 2015].
Pragmatism will always view scientific truth with conditions, gained

through cycles of experimentation and experience. The knowledge
acquired is grounded in our physical world [Bishop, 2015]. Episte-
mological, design science researchers are fundamentally pragmatists,
underpinned by predictable and functional artefacts which are evalu-
ated against a knowledge base of information [Radebe, 2021]. In the
next section, we present details of the DSR methodology.
5.2 the design science research (dsr) methodology
The adopted methodology is the Design Science Research methodol-

ogy (DSR). The DSR allows for the construction of a novel artefact
being created from a list of requirements and theory. An acceptable
design is created through an iterative process of various design cy-
cles, until a list of design sequences is established [Gacenga et al.,
2012; Hevner et al., 2004; Hevner, 2007].
DSR is a relatively recent paradigm originating in the domain of

information science [Gregor and Hevner, 2013; Hevner, 2007]. DSR
is defined and known for the creation of system artefacts in address-
ing a research problem [Peffers et al., 2007]. Researchers mainly se-
lect design-theory or pragmatic-design methods [Gregor and Hevner,
2013; Hevner, 2007]. DSR makes a well-defined contribution that is
pragmatic in nature that clearly involves relevance [Hevner, 2007].
102 research methodology
5.2.1 An overview of the DSR approach
The framework of DSR follows the specific phases of design cycles

of understanding the problems, firstly, then applying practical solu-
tions through an iterative process in a pragmatic and relevant man-
ner [Hevner, 2007,?].
The design cycles are namely:
1. the relevance cycle — consisting of requirements gathering or

understanding through a literature research in establishing a
conceptual model amongst specialists and experts,
2. the design cycle — the actual design of the artefact and pro-
cesses involved in the “build” process, and
3. the rigour cycle — the evaluation and grounding of the concept

in the selection and reliability of the methods, then adding to
the knowledge base [Hevner, 2007].
The DSR approach is primarily the construction of artefacts based

on specific requirements and the underlying concept. The defining
of relevant requirements helps the review process of previous ap-
proaches. A sequence or structure of activities emerges which are
commonly referred to as the DSR activities. These are the design sci-
ence phases, methods and tools, activities, and the final outputs [Radebe,
2021]. The Knowledge base includes the literature review of various
studies which explore the identification and relevance of the prob-
lem [Peffers et al., 2007]. The problem is identified and well-defined
in order to develop an optimised solution. This is in turn justifies the
artefacts value in the final solution.
The definition of activities or objectives relating to the knowledge

base will determine the final output of the solution design. The collec-
tive requirements that are created should address the research ques-
tion. The design and development of the solution will inevitably cre-
ate an artefact that addresses the problem [Radebe, 2021]. The DSR
cycles are followed by an evaluation of the artefact designs which
contribute to design science and the creation of a design, followed
by an evaluation process. The rigour cycle then makes a meaningful
contribution through the evaluation of artefacts and concepts in the
knowledge domain.
The DSR methodology would typically follow these stages:
1. problem identification,
2. objectives of the solution in the requirements,
3. design and development of artefact — producing the artefact(s),

5.2 the design science research (dsr) methodology 103
4. implementation of artefact(s),
5. evaluation of the artefact(s), and
6. the conclusion and results of the designed artefact(s).
Every stage of the DSR cycle methodology follows an iterative

process towards the final design outcome [Hevner et al., 2004].
5.2.2 DSR in Information Systems or Technology
Further debates in research methodologies as relating to information

systems and computer science, led to DSR being widely published as
an accepted paradigm [Venable et al., 2017]. Variation of the DSR has
been proposed in recent years. These methodologies are:
1. System Development Research Methodology (SDRM) — con-

structing frameworks for systems design,
2. DSR Process Model (DSRPM) — act of a design approach with

emphasis on knowledge flows,
3. Action Design Research (ADR) — this methodology combines

Action Research (AR) and DSR with four stages of activities,
4. Soft Design Science Methodology (SDSM) — consists of eight

activities of a design cycle in solving a specific problem, and
5. Participatory Action Design Research (PADR) — for developing

solutions to problems facing a large population or organisation.
These five variations of DSR methodologies all follow similar stages

of design. They differ in nature regarding applicability and its own
evolution from the system artefact design, with very little client en-
gagement, to practices with a high level of client involvement, which
combines AR and DSR [Venable et al., 2017].
5.2.3 The benefits of DSR methodology
The benefits are focused around solving human problems with scale-
able solutions, across system architectures, locally or globally [Pef-
fers et al., 2007]. The overarching benefit is its strong dependence on
function — grounded in functional requirement that relate to pre-
scribed components of the artefact or solution. It is focused on out-
comes based methodology building on new solutions or artefacts in
creating new realities for mankind [Bisandu, 2016]. It builds upon the
acquisition of knowledge and thus producing new knowledge which
is further enhanced over time. It adds to any body of knowledge in

designing and devising new concepts.
5.2.4 The risks of the DSR methodology
Any methodology will present its own risks and limitations to a de-
gree depending on many factors such as the nature of the research
problem and the solution being designed. These problems are re-
ferred to as “wicked problems” [Hevner et al., 2004]. These problems
seem to arise in complex socio-technical systems whereby the out-
comes solutions become unpredictable which are the following [Bisandu,
2016]:
1. the requirement definition is unstable in an ill-defined environ-

ment,
2. too much dependence on human cognition in producing an ef-

fective solution,
3. the interactions between the sub-components of the problem

and the solution become too complex, and
4. the processes of solution design become too malleable and vague.
In this work, the risks of the DSR did not present any concern as
the problem identified in the South African healthcare sector, were
clearly defined. Hence, we applied DSR hitch-free.
5.3 the rationale for selecting artefacts
In the light of current 4IR technologies being rolled out to many di-
verse industries, we have selected a number of 4IR solutions that can
be applied in the mainstream South African healthcare sector; based
on the reasons given in this section. The reasons for selecting the arte-
facts presented in this thesis are as follows:
1. we identified gaps within the healthcare sector based on the ex-

tensive critical analysis undertaken on the South African health-
care sector,
2. strong evidence of a few selected feasible 4IR solutions that can

be applied immediately,
3. the availability and accessibility of healthcare domains within

the private and public sectors is always a challenge. Our se-
lected artefacts are in healthcare domains where we have been
5.4 sampling 105
granted access to consult the experts and for which we have

received ethical clearance, and
4. the artefacts selected could be developed swiftly within the pe-

riod of the doctoral study and applied in addressing real health-
care problems.
In the next section, we present the sampling techniques that we

have adopted for analysing the data collected from the evaluation of
the new 4IR artefacts presented in this work.
5.4 sampling
5.4.1 Study population
A survey was developed for a focal group of medical experts and

healthcare practitioners in order to validate the artefacts designed in
this work. The sample population of the respondents were selected
based on their expertise in their respective domains with a total of
38 respondents. Due to them being a select focal group of experts,
the sample size was confined to their respective fields of expertise in
healthcare and healthcare data analysis. However, approximately 20
specialist experts from the selected focal group, diligently responded.
It was a good response considering the small population of experts
in the selected domain of expertise.
5.4.2 Sampling procedures
The sampling method leaned towards non-probability sampling which

involved a non-random selection of experts in healthcare, based on
specific inclusion criteria. The questionnaires were anonymous and,
presented through ‘Google Forms’. The questionnaire supported the
evaluation of the five developed tools.
5.4.3 Expert sampling
The selection of the sample of respondents was critical in the final

evaluation of the designs of the artefacts. The sample was drawn
from a select pool of experts within the field of clinical healthcare and
healthcare data analysis. This sampling method is used in seeking
expert opinions with: a high degree of expertise within the study do-
main, respective clinical experience, authors in the field, professionals
registered in the field. Expert sampling is more of a simple sub-type

of Purposive Sampling [Partington, 2002; Peffers et al., 2007].
5.4.4 Inclusive criteria for respondents
The targeted subjects were based on the non-probability sampling

method where the respondents are personally selected on the basis
that they should at least have the following attributes:
1. healthcare professionals registered with the South African Health

Professionals Council of South Africa, who is progressive and
active within the industry and familiar with the modern trends
of medicine encompassing modern technology. This includes
the management of the diabetic foot, for the MoonBoot ques-
tionnaire.
2. the focal group of experts was extended to healthcare data ana-

lysts, for the anonymisation of medical details, only.
5.4.5 Sample size
The targeted sampling size of the respondents was initially at 38, con-
sidering its a select focal group of experts. However, approximately
20 specialist experts from the selected focal group, responded effec-
tively.
5.4.6 Data collection
Data collection was done with questionnaires, electronically using

Google Forms. The data collected was stored in the cloud. Design
of questions took the form of multiple choices, short texts and nu-
merical scales.
5.4.7 Designing the questionnaires
The questionnaire was self-designed and administered. The expert

respondent was requested to fill out the questionnaire in his or her
own time and remained anonymous. The five newly designed arte-
facts had separate questionnaires, one for each. All links to question-
naires were presented to each expert via an introductory email. The
questionnaires are available at:
5.5 data analysis 107
1. The Moonboot: https://tinyurl.com/y56txf4l
2. The MedBot: https://tinyurl.com/y5bmoxf3
3. Medical note decryption: https://tinyurl.com/y567f7v9
4. Anonymisation algorithm: https://forms.gle/Y5rezSw5N2C3dtzc8
5. Trend analysis: https://tinyurl.com/y23jezf9
These links are for the developed 4IR tools, namely: the decryp-
tion of medical notes; medical chatbots; disease trend analysis; anonymi-
sation of personal medical details and the MoonBoot design.
5.5 data analysis
The scoring of each questionnaire was stored in the Google forms

cloud storage. The respondents scores were presented in performance
based scoring as a percentage, for each respondents questionnaire,
in support of their respective validations and motivations, for the
respective artefact designs. The scores and findings were downloaded
and analysed, and presented in the later chapters of this thesis.
5.6 ethical considerations and procedures
This research thesis was accepted and approved by the University

of Johannesburg ethics committee. The developed tools are original,
with no third party involvement of persons or subjects, which was
presented with a formal ethical introductory letter, on email, high-
lighting the ethical committee acceptance and procedure. This in-
cluded a brief instruction letter in order to proceed. The respondent
then returned an acceptance email — essentially an acceptance from
the respondent.
If the respondent accepted they proceeded to the Google form

link, on the email. On entering the Google form the respondent is
then presented with more detailed component of the design proposal
and a formal letter presented by the University of Johannesburg ethi-
cal committee to acknowledge and sign.
Once the acceptance and understanding of the introductory letter

and proposal letter, was accepted and signed (via a digital checkbox),
including the ethical letter of acceptance, the respondent could pro-
ceed to answering the questions as per artefact design. There were
approximately twelve questions per artefact design presented in mul-
tiple choice format, based on a simple scoring mechanism.
The data collected was stored on Google Drive (Google Forms)

adhering to normal security protocols. The data is anonymous and
cannot be traced to any respondent in the questionnaire. The data
remains in safe-keeping within the Google Cloud. Finally, the results
of these surveys, including the inferences, are presented in this thesis.
In this chapter we have presented the philosophical foundations per-

taining to this thesis’ research methodology. An overview and mo-
tivation for the DSR method was outlined. We explained how DSR
was used in the design of five 4IR artefacts, within the South African
healthcare sector. We further defined the theoretical background of
the methodology of DSR, including the advantages and limitations of
its application.
The risks and rationale for selecting the five artefacts was further
outlined. The procedures of the sample selection of 38 specialists were
discussed. The expert sampling was based on certain inclusive crite-
ria in their respective domains, such as clinical medicine or healthcare
data analysis. This included the method of design of the question-
naires (including the links), relating to the expert respondents.
In addition, the data analysis section highlights the assimilation

and scoring of the data, within the Google platform as per the de-
signs. Finally, the ethical considerations and procedures and, the data
collection procedures were presented. The data collection and storage,
by means of a Google form questionnaire, was defined.
Part III
CONTRIBUTIONS OF 4IR TECHNOLOGIES TO

S O U T H A F R I C A N H E A LT H C A R E
This part consists of Chapter 6, which presents the dy-

namic impacts of big data within healthcare. A simple
search tool is presented in detecting disease trends or pat-
terns of pathologies in Chapter 7. Trend analysis can be
utilised by patients and practitioners alike being Open-
Source on the Internet. We present an algorithmic based
solution in anonymising large personal healthcare data
sets. Natural Language Processing (NLP) in Chapter 8
presents the design of an AI tool in deciphering clinical
reports or notes, for all practitioners to easily understand.
Robotics in healthcare is gaining momentum where we
present an AI tool designed in Chapter 9, as an added fea-
ture to a MoonBoot, in the management of a diabetic foot.
Finally, in Chapter 10 we present a mobile application de-
sign in the form of a MedBot, aiding the practitioner in
defining a possible outcome for a diabetic patient.
6 B I G D ATA I N H E A LT H C A R E
B
ig data analysis in healthcare is becoming a prerequisite for
the digitalisation of healthcare 4.0, and a formidable building
block of the Fourth Industrial Revolution (4iR), across all in-
dustries. The largest impacts yet to be reaped from 4iR in healthcare,
will no doubt be forged from the development of big data real time
analysis.
6.1 introduction
The big data revolution has impacted all facets of business including
the biomedical and health sciences. Primarily facilitated by genetic
sequencing, imagery, the development of vast patient data sets from
wearable sensory devices, the ever expanding knowledge pool of clin-
ical data through the electronic health file (eHR) and digital trend
plotting [Amiriam et al., 2017; Kaur et al., 2019].
Healthcare data consists of patient clinical data, transactional data

and the healthcare funder data, or healthcare service payer financial
data. It goes further into healthcare resource data and other asso-
ciated business data streams. Conventional data bases and systems,
overburdened with an array of medical data streams, showed an in-
ability to manage such quantities of big data sets [Amiriam et al., 2017;
Gartner, 2015].
The South African public healthcare system is undergoing trans-

formation and a major re-organisation. Drastic measures are called
upon, for transformation whereby data analytics and, data mining
will eventually become the primary focus and ultimately the centre
stage. Currently there are numerous disparate siloe’d systems across
the South African public and private healthcare landscape. All dis-
playing a unique set of architectures and tools, attempting to address
data mining and business intelligence capabilities, by means of its
current National Health Information Systems (NHIS). The National
Department of Health (NDOH) is aware of these architectural shifts
and skills needed, in addressing the proliferation of big data analyt-
ics, within the public healthcare sector, in order to produce valuable
clinical insights for public health [Abbott and Ade-Ibijola, 2018; Wol-
110
6.1 introduction 111
While these specific NHIS systems are isolated and, heterogeneous

in nature, they strive to derive an assemblage of data analytical value,
albeit sparse. The analytics process in healthcare is critical in deriving
knowledge from data or information that is collected from the clinical
environment, in the form of big data from different sources. Primarily,
the benefits of the big data revolution will generate value through de-
scriptive, predictive and prescriptive techniques which must address
the four major pillars of healthcare namely patient care, patient mon-
itoring, predictive analytics and treatment management [Chen et al.,
2014; Kaur et al., 2019].
A commonly defined process of data analytics is referred to as the

methods and tools used with the interpretation, by the practitioner,
or healthcare provider in providing meaningful insights. Moreover,
the definitions of big data analytics vary widely. The literature refers
to healthcare data analysis, as the discovery of meaningful and in-
sightful patterns of data which ultimately produce common value,
through a defined life cycle of data collection, ex-traction, transfor-
mation, analysis, interpretation and reporting thereof. Thus, taking
the healthcare professional from hindsight to insight, then towards
foresight, from the analysed clinical data sets. Primarily, big data
analytics will improve the operational capabilities or processes; and
the patient interaction – similar to cus-tomer experience - forming
the basis to economic transformation across the industry [Belle et al.,
2015; Chen et al., 2014; Gartner, 2015; Grossman and Mazzucco, 2002;
Schmarzo, 2016].
The investments made by South African National Health Depart-

ment (NDOH) over the last decade, have not created the intended
value. The outcome has been one of too many missed opportunities
with poor healthcare analysis, in comparison to the total spend by the
NDOH, according to the Minister of Health [Nat, 2012]. The structure
of the SA healthcare industry primarily consists of a dual system,
private and public healthcare. These two different economic indus-
try landscapes have differing data architectures of data content and
data flow. The public sector is still serving the majority of the South
African population [Abbott and Ade-Ibijola, 2018; Government, 2016].
The private healthcare data landscape is more focused on the fi-

nancial transactional data sets of benefits, payments with the provider,
payer and patient scenario of financial (transactional) data. Focusing
less attention on clinical data, as compared to public healthcare. Con-
trary to private healthcare, the public healthcare data landscape, is
more focused on the patient statistical data, or clinical notes and pa-
tient organisational static data. These are two markedly different sce-
narios and data flow processes due to the nature of the industry focus,
in spite of the private and public healthcare sectors, both serving the
SA population [HiSP, 2012; NDOH, 2014b].
112 big data in healthcare
Big data analytics is upon us. The governments eHealth strategy

is being adopted, in order to fulfil this strategy. It depends on the
building blocks and foundations being correctly adapted, in the ar-
chitectural IT landscape. Modern healthcare, globally, has failed to
exploit the benefits and opportunities, offered by the evolving big
data analytical world, but it is changing. However, the NDOH is mak-
ing strides in laying solid foundations of transformation across the
landscape, in order to derive value from the untapped clinical big
data with its foundations, spanning from the NHIS system landscape
to skills development. The proposed solution of anonymisation of
medical health data sets, addresses the ethical issues of the privacy
of medical health data records, in medical healthcare research [Nat,
2012; NDOH, 2014b].
Electronic devices and wearable surveillance devices, capturing

vast data sets, have become common practice, in transmitting this
daily data. Thereby, enabling the data collection for millions of peo-
ple across the globe, essentially revolutionising healthcare data analy-
sis [Chen et al., 2014, 2012]. It can include the monitoring of household
appliances through sensor technology and, even the weather patterns
affecting crops, triggering harvesting times. Healthcare is catching up
with wearable sensors, through smart applications of such technology.
At a basic level it includes wrist watches monitoring vital bodily func-
tions to the monitoring of steps taken and recording this daily data
to the Cloud, for data analysis [Chen et al., 2014, 2012].
This chapter seeks to present an overview of big data in medical

healthcare as a cornerstone of the Fourth Industrial Revolution (4iR).
It will include an understanding of the concepts of big data analysis,
platforms and its tools with Cloud technology. In addition, it is a pos-
sible solution for the growing demand for healthcare data analytics
being confronted by compliance and ethical issues, within the health-
care industry. This chapter endeavours to present the background
and related work of big data in general, the design of an algorithm
solution as a proposed solution, for the anonymisation of personal
medical details, enabling and enhancing healthcare analysis.
6.2 big data concepts
6.2.1 Big data
Big data refers to large data sets originating from various informa-
tion system sources, which are then aggregated for business value.
Big data comprises of large quantitative and qualitative data sets, col-
lected from various sources, in increasing volumes and velocity [Schmarzo,
6.2 big data concepts 113
2016]. Big data is analysed in order to extract knowledge to reveal

trends, patterns, relationships, and interactions for decision makers.
Conventional analytical software was unable to manage big data sets,
therefore new methods were developed to manage large datasets [Chen
et al., 2012; Gartner, 2015; Lin et al., 2018; Corp, 2013].
Data anonymisation: is a process whereby all personal health data

is sanitised where no reference or relation can be traced back to a sin-
gle person or entity that identifies the patient. The data is not only
sanitised but new data is synthesised and replaced with fictitious
replacement data, thus preserving all privacy aspects of the actual
identification of a patient or entity. The patient always remains anony-
mous [Gartner, 2015].
A data lake is a system or a large data repository that holds raw

data in its native format for further requirements where the data is
extracted with specific tools for further analysis. A typical data lake
would reside in the Microsoft Azure Cloud, as an example [Belle et
al., 2015; Chen et al., 2014].
Conventional IT architectures, or ‘legacy systems’, use online an-

alytical processing tools (OLAP) and static data bases - relational
databases (RDBMS), commonly found across commercial landscapes.
Naturally, they became inadequate with the advancing mainframe, in
the 80’s. The parallel database system was developed to meet the in-
creased demand [Piateski and Frawley, 1991; Schmarzo, 2016]. The
clustered storage system generated the first ever Teradata parallel
database. The K-Mart retail stores group, acquired its first Teradata
database system in 1980’s. It was seen as a pioneering platform for
large data sets. The exponential increase in demand for Internet ser-
vices and the resulting search engines, paved the way for the creation
of Google, and many other social media sites [Chen et al., 2014; Ellis,
2014; Wamba et al., 2015].
Big data was born and soon became a technological concept, de-
scribed as the ‘next revolution in business’, commonly referred to as
the ‘4th paradigm’ [Chen et al., 2014]. The new data lake, is an exam-
ple of a large data repository that houses native data, extracted with
specific tools for analysis. The Microsoft Azure Cloud would typi-
cally host such a data lake. Big data primarily possess the following
attributes, known as the 4V’s [Amiriam et al., 2017; Chen et al., 2014;
Ellis, 2014; Wamba et al., 2015]. The 4 V’s of Big Data comprises:
1. volume: large data sets with pictures and video streams,
2. variety: media data, streaming data - structured, unstructured

or semi-structured data,
3. velocity: live streaming data sets, and

4. veracity: authentic real time data.
Big data is being referred to as the “fourth paradigm of science”

or “the next frontier”, bringing a revolution of science across various
industries. The McKinsey report on big data, predicts a saving of $300
billion across the board, with 8% of that saving coming from health-
care, within the United States of America [Belle et al., 2015; Chen et
al., 2014; Cawsey, 1997; Corp, 2013; Sharma et al., 2018; Wamba et al.,
2015].
According to the Industrial Data Corporation (IDC), recently stated

that global data could be in the region of 1.8 Zetabytes, which will
double every two years [Chen et al., 2014; Gartner, 2018]. Google
can produce hundreds of Petabytes per month. Facebook produces
around 10 Petabytes per month. YouTube, uploads 72 hours of video
every minute. The rapid growth of data regarding the Internet of
Things (IOT), has given rise to Cloud computing platforms on a large
scale. The scalability of storage, for these data structures, has literally
expanded beyond the structures of traditional data and enterprise ar-
chitectures. Space for such data volumes, ran out, which inevitably
led to the creation of Cloud technology [Chen et al., 2014; Gartner,
2018; Kaur et al., 2019].
According to Gartner (2015), there are four basic types of analyt-

ics:
1. descriptive - which is essentially Business Intelligence (BI) and

data mining,
2. predictive - is the future view of such data sets,
3. prescriptive - incorporates the optimising the data sets and what

needs to be done, and
4. diagnostic - analyses data and provides answers as to why events

happened.
According to McKinsey (2013), there are primarily five industries,

which produce needed value and benefit from big data, adding to the
greater US economy. The primary industries are healthcare, public
sector administration, retail, manufacturing and global location data.
Further, McKinsey sees this concept of Big Data as the next frontier
for innovation, competition and productivity. Akin to the concept of
Big data is the growth and management of such data structures [Chen
et al., 2014; Corp, 2013].
The big data concept is currently producing hidden values and

potential business opportunities, never realised before, purely by ef-
fectively managing these large data sets. The healthcare industry is
slowly catching on to this ‘missed opportunity’, previously stated.
6.2 big data concepts 115
Global healthcare is currently poised to exploit this opportunity in

dealing with vast transactional and clinical data sets, originating from
various sources including sensors, satellites, photos, GPS signals, so-
cial media feeds, financial systems, twitter feeds, video, scans and cell
phones [Belle et al., 2015; Chen et al., 2014; Wamba et al., 2015].
Only 42% of healthcare companies are employing big data tool

sets that will support their decision processes whereby only about
16% are deriving some form of value from big data mining processes
and tools [Sharma et al., 2018; Wamba et al., 2015]. Through the evolu-
tion of technology, the concept of Business Intelligence (BI) and data
mining, within its formal structures, definitely laid the foundations
for modern big data concepts and structures. This is similar to a new
era of business exploration, where fresh insights into patients’ needs
are discovered. It can be further refined through the use of predictive
models anticipating such needs. Enter ‘Data Science’, which provides
for these predictive tools, consisting of machine learning algorithms
which can automate those complex time-consuming activities [Belle
et al., 2015; Chen et al., 2014; Sharma et al., 2018; Wamba et al., 2015].
It is evident that big data is to a large degree linked to the birth of

data science, to a large degree. Data science is a relatively new field
of advanced analytical skills in deriving business value from these
large data sets. While the value of data can only be extracted and
presented to the change agents, data remains a data store [Wang et
al., 2018a; Wamba et al., 2015].
The Data Management Body of Knowledge (DMBOK), framework

of concepts goes further to say that big data is linked to data science.
Big data can also appear vague in many respects, but data science
and big data remain interrelated. It allows data scientists or business
people to store, generate and analyse larger sets of data. Data analysis
can predict and influence patient behaviour to a large degree which
is in high demand [Belle et al., 2015; Cawsey, 1997; Cupoli et al., 2014;
Murdoch and Detsky, 2013].
Further, the Internet of Things data (IOT), monitors and conducts

surveillance of real-time health bio-data, from sensors and streamed
to a platform through wifi [Abinaya, 2015; Cupoli et al., 2014].
6.2.2 Big data analytics
Big data analytics is a process utilising specific methods and tools for
the interpretation of big data to provide meaningful insights. Amidst
various definitions of big data analytics the central idea is the discov-
ery of meaningful and insightful patterns in data which ultimately
produce common value, through a defined life cycle of data collec-
tion, extraction, transformation, analysis, interpretation and report-

ing [Kaur et al., 2019; Kankanhalli et al., 2005].
Data analytics takes the user of data from hindsight to insight,

then towards foresight, illustrating the retrospective and prospective
view on information. Analytics transforms raw data into meaning-
ful information that is useful for decision-making [Kankanhalli et al.,
2005]. Primarily, big data analytics becomes a feedback mechanism
that improves the efficiency of operational processes of an organisa-
tion, the service user or customer’s experience of the services pro-
vided by that organization, and eventually the services itself [Kaur et
al., 2019; Sharma et al., 2018].
There are four basic types of analytics:
1. Descriptive - which are essentially business intelligence and

data mining and focuses on describing populations, users and
their profiles.
2. Predictive - which is the future view in terms of what may be

likely to happen.
3. Prescriptive - which incorporates the optimising of data sets

and decisions on what needs to be done.
4. Diagnostic - which provides answers as to why certain things

happened [Gartner, 2018; Metagroup, 2005].
Big data analytics differ from static data analysis in that it focuses
on real time, ongoing analysis, which can be automated. In this re-
gard, Big Data is referred to as the ‘fourth paradigm of science’ or
‘the next frontier’, which is bringing a revolution of science due to
its strategic potential across various industries. The McKinsey report
on big data predicts a saving of $300 billion, across the board with
eight percent of that focused on healthcare within the United States
of America [Belle et al., 2015; Corp, 2013; Wamba et al., 2015].
The Data Management Body of Knowledge (DMBOK), framework

of concepts, states that big data is linked to the concept of data science
(the science of analytics), which is gaining more recognition. The pro-
fessionalisation of data science enables industries to have more rigor-
ous data analysis capabilities. In terms of business development, the
big data concept is producing value never seen before from data, and
potential business opportunities never realised before, purely by ef-
fectively managing and drawing useful information from these large
data sets. The healthcare industry is gradually catching on to the ben-
efits of the big data wave [Belle et al., 2015; Cupoli et al., 2014; Mur-
doch and Detsky, 2013].
6.3 big data architecture 117
6.3 big data architecture
Big data structures remain generic across most industries, until called
upon for specific analysis, where it will then differ in composition.
The basis or founding structures will incorporate the primary gener-
ation of datasets, then their ingestion of such data sets, and finally
their transformation of such data sets. The platform analytical tools
will process these large data sets where, finally, the analytical tools are
engaged to extract the value of data for business use. It is generally, an
accepted framework for big data, across most industries [Ragupathi
and Ragupathi, 2014; Sharma et al., 2018; Schmarzo, 2016].
The large data sets originate from an array of sources within the
healthcare sector, which is decentralised, and then processed in par-
allel at a multi nodal level. The data sets are vast, and of a hetero-
geneous nature. These platforms are mostly open source and consist
mainly of the following platform technologies, which are freely avail-
able in the Cloud [Amiriam et al., 2017; Ragupathi and Ragupathi,
2014; Sharma et al., 2018; Schmarzo, 2016].
The sources of big data within healthcare originates from various

sources internally within the clinical/medical environment, and ex-
ternally from third party sources such as laboratories, or government
agencies. In addition the data comes in multiple formats such as flat
files, csv, excel, ASCII and text files alike which originate from trans-
actional data and various other data types such as:
1. machine to machine data - originating from remote sensors and

vital sign monitors,
2. web based social media data- twitter and social media interac-
tive data with additional monitoring such as on “Clickstream”,
Facebook, Twitter, blogs and smartphones application data,
3. transactional data - transactional medical data in the form of

billing records in, structured and unstructured formats,
4. biometric data - finger print, medical ID data, scans, medical

records, X-rays and specific medical details on the patient, and
5. human generated data - electronic medical records written by

the practitioner on the patient.
The big data analytical platform consists of five major architec-

tural layers. They enable healthcare managers to transform and inter-
pret the data into meaningful clinical information by means of a big
data implementation [Chen et al., 2014; Schmarzo, 2016; Vayena and
Blasimme, 2018; Wang et al., 2018a]. The five architectural layers are
the following:
1. data layer- the sources of data to provide the insights for the
business decisions scenario. The data presents in a traditional
structured format such as data from an Electronic Health Record
(eHR), in an unstructured format like monitoring logs from de-
vices on patients and clinical images where the format will de-
termine its target data base,
2. data aggregation layer – the sourced data is processed by the fol-

lowing means; data acquisition, data transformation and data
storage. Data collation, in any big data project, is when the
sources are varied in nature and format. The transformation
phase is primarily responsible for moving, cleaning, splitting,
translating, merging, sorting and validating data. The data is
loaded into target databases, such as the Hadoop Distributed
Filing System (HDFS). Hadoop cloud performs further analysis
in batch or real time process,
3. analytical layer – consists of two major components; the Hadoop

Map/Reduce; stream computing and data base analytics. Map
reduce is the de facto programming tool, having the capability
of processing large volumes of data the most efficiently. Stream
computing is performed best in healthcare real-time environ-
ments, especially for detecting an emergency. The in-house high
speed parallel processing, is effective in predicting healthcare
analytical outcomes. This component supports the evidence based
medical practices by virtue of eHR analysis, patient histories,
patterns of care and other therapies involved,
4. information exploration layer – the visualisation or reporting

layer of business insights, a feature layer in big data analytics.
The reporting produces the business value where strategic deci-
sions are made. It is imperative within healthcare, especially in
streaming real-time insights, to medical problems. Further, op-
erational issues and Key Performance Indicators (KPI’s) at the
operational level through dashboards, and
5. data and governance layer – governance policies, tools, stan-

dards, and procedures relating to the Master Data Management
(MDM). The master data has to be standardised to create imme-
diacy, availability and accuracy of master data sets, in support
of effective decision making.
Another important component is the data life cycle management

scenario. The data is managed from the archiving of data, data-warehousing,
testing, and disposing of such data, to the cost efficiencies of manag-
ing the in-house data.
Further, an important component is the security and privacy of

such data management. Security provides the platform for enterprise
6.4 big data analysis 119
level data services by means of discovery, configuration, monitoring

and auditing functions. The healthcare industry has special considera-
tions in this regard which are far reaching such as medico legal, med-
ical ethics in handling sensitive clinical data, the ownership or autho-
risations for medical data and the privacy of medical data in general.
Relevant and suitable policies must be well implemented and well
managed, in accordance with national healthcare regulations [Chen
et al., 2014; Schmarzo, 2016; Vayena and Blasimme, 2018; Wang et al.,
2018a].
This architectural framework has produced a wide variety of skills

needed in order to extract clinical data by means of applied mathe-
matics/statistics, economics, programming skills and, computer sci-
ence in order to assimilate, aggregate, manipulate, analyse and ex-
tract the value of such healthcare data. The most significant plat-
form is the Hadoop/Map Reduce platform, as an open-source dis-
tributed data processing platform. It belongs to the “NoSQL” tech-
nology where other technologies include Mongo DB and Couch DB.
Hadoop has the ability to process large data sets, and assign them
to different servers or nodes, which solve different parts of the prob-
lem then integrates the final answer on a massive scale [Chen et al.,
2014; Ragupathi and Ragupathi, 2014; Shinde, 2016; Schmarzo, 2016;
Vayena and Blasimme, 2018; Wang et al., 2018a].
6.4 big data analysis
6.4.1 Big data analysis tools
Big data analysis consists of two categories, namely real-time analysis

and off-line analysis, or retrospective data analysis. This is key when
it comes to the structure of the architecture of the platform.
1. Real-time analysis – is mostly found in corporate retail systems

or in e-commerce and financial systems. The functions involved
are rapid analysis and fast turnaround times. The technology
platform will include: parallel processing clusters using tradi-
tional RDBMS database structures, and in-memory based com-
puting platforms, such as the SAP Hana platform, as an exam-
ple.
2. Offline analysis – commonly used within environments not re-

quiring high response times having performed the data anal-
ysis retrospectively: machine learning, statistical analysis and
automated algorithms, as examples. The required architecture
would typically consist of a Hadoop platform which will dras-
tically reduce the cost of data acquisition. And other examples
would be Facebook’s open source tool called Scribe. In addition

Linked-In has a tool called Kafka and the, Taobaos open tool
Timetunnel [Chen and Asch, 2017; Chen et al., 2014].
There are many data analytical tools used within big data which
are either open source or commercially available tools [Shinde, 2016].
The five most commonly used tools in big data software in today’s
business environments, namely:
1. ‘R’ language: an open-sourced based language designed for data

visualisation. It was designed by Bell Laboratories and used for
statistical analysis. ’R’ is highly respected within the software
engineering community and currently outranks Java and SQL.
Oracle and Terradata have recently produced a few products in
R.
2. MS Excel: a key component within the Microsoft office business

software which provides statistical analysis capabilities. The plug-
in utilities,‘Toolpak’ and ‘Add-in Solver’, further enhance its an-
alytical and statistical capabilities.
3. RapidMiner: open-source tool for data mining, predictive analy-

sis and machine learning. RapidMiner also provides rich visual-
isation, modelling, data pre-processing and evaluation capabili-
ties. The data-mining component is written in Java.
4. KNMINE (Konstanz Information Miner): an open source ana-

lytical platform, creates data flows in a visual format, providing
analysis and data models. Written in Java and Eclipse.
5. Weka/Pentaho: an open source machine learning application,

written in Java. It provides functionalities, regression, clustering,
data processing, and association rule with visualisation capabil-
ities. Further, it includes charting and integration.
The disadvantages of these technologies are that they require spe-

cialist programming skills and analytical knowledge, which are cur-
rently proving difficult to attract and retain. This is evident within the
public health and social care sectors, where these skills are relatively
scarce [Coovadia et al., 2009].
In addition, there are numerous commercially available data an-

alytical tools, commonly found across most industries, namely: Mi-
crosoft Azure Analytics and Power BI; SAP Business Objects; Qlikview;
Tableau and Google Analytics. Many of these popular data analytical
tools are part open source, and part fee for service driven. The ma-
jority of these tools are making strategic marketing entries into the
health sciences landscape, producing value driven based medicines
outcomes objectives [Chen et al., 2014].
6.5 big data in healthcare 121
6.4.2 Cloud technology
The Cloud is a term describing large data storage platforms. It houses

its own computational technologies, with distributed storage capacity
within Cloud computing. Thus, managing extremely large data sets
or pools of data [Belle et al., 2015; Chen et al., 2014; Young and Zhang,
2017].
Cloud computing is closely aligned to big data technology, Inter-

net of things (IOT), the data centre and Hadoop technologies. Cloud
computing is a component on its own, with its own technology and
architectural structure. It consists of various components within its
own structure. The largest dependency or technology requirement
from big data is the capacity to store data, where the Cloud is opti-
mised for this function. Cloud computing provides for the capacity
and solutions for the storage and processing power of Big Data [Chen
et al., 2014; Gartner, 2015].
The two concepts, Cloud and Big Data, are essentially different
but complement each other. Cloud computing is transforming main-
stream business which has evolved from the Virtual Private Network
(VPN) scenario, while the big data technology is influencing and de-
riving more effective strategic business decisions [Belle et al., 2015;
Chen et al., 2014]. Cloud computing has become a service/function
offering for big data.
6.5 big data in healthcare
6.5.1 The nature of big data in healthcare
Healthcare data is multifaceted, it originates from various sources,

adding to its complexity and equally to its richness. The data is col-
lated as clinical or transactional data, depending on the environment.
The nodes of digital health data originates in the following forms:
clinical outcomes reports, insurance detail data, consumer data, socio-
economic data, claims data, biometric data, epigenetic data, and ge-
nomic data. Due to this massive increase in data fields, conventional
databases cannot cope with the streaming variety and quantity of
data. Therefore, storage facilities had to be upgraded [Amiriam et al.,
2017; Kankanhalli et al., 2005].
Globally, healthcare has lagged behind in exploiting the many op-

portunities within the big data world, although it is changing. The
National Health Information System (NHIS) in South Africa, appears
to be fragmented across the public and private healthcare sectors [Ab-
bott and Ade-Ibijola, 2018]. It is scattered with different architectures

and systems, with limited capability or simply nothing in place for
any advanced analytics.
The private healthcare landscape is more focused on benefits, funds

and financial data, with little focus on clinical data. The South African
public healthcare landscape is more concerned with patient clinical
and statistical data, relating to its organisational structural data, ana-
lytically retrospective in nature [Abbott and Ade-Ibijola, 2018,?; Reddy
and Sharma, 2016].
The SA eHealth strategy aims to address the big data evolution.

The building blocks of such a venture are being applied in the public
healthcare landscape. The underlying theme is transformation through
the adoption of the eHealth strategy [Botha et al., 2016; Sharma et al.,
2018; Wang et al., 2018a].
Clinical and social data are vastly different to transactional data

across the system, across the landscape and, it requires a high degree
of integration. The researcher or analyst will face new challenges in
deriving patterns and trends from these varied data sets, via vari-
ous devices, from the Cloud, thus enabling epidemiological analy-
sis [Belle et al., 2015; Chawla and Davis, 2013; Kannel and McGee,
1979; Wang and Alexander, 2016].
Evidence-based health outcomes data, based on scientific statis-

tical principles, is poised to displace ‘expert’ opinion. Big data can
enhance the digital analysis, or economics, of healthcare and improve
efficiencies, beyond any other industry [Nadkarni et al., 2011; Mur-
doch and Detsky, 2013; Sharma et al., 2018].
6.5.2 Big data value for healthcare
The value and benefits of big data will change the healthcare environ-
ment [Belle et al., 2015]. The patient and healthcare provider will both
benefit from these contributions:
1. personalised applications to individual patients through alerts,
2. real-time synopsis of sign and symptoms,
3. improved treatment outcomes through treatment interventions,

and
4. immediate response to an emergency coming from a sensor or

data findings.
Big data value creations are numerous, ranging from predictive

risk modelling and, preventative healthcare, to the implementation
6.5 big data in healthcare 123
of healthcare interventions and rapid decision making [Chen et al.,

2014; Vigilante et al., 2019]. Traditional research data, referred to as ret-
rospective predictions are primarily based on historical data. Future
predictive models are part of a big data science capability making it
a powerful analytical tool [Chawla and Davis, 2013; Chapman et al.,
2000; Piateski and Frawley, 1991].
It creates more transparency, reduced time spent on analysis, more

experimentation and discovery of variables, for different analytical
scenarios. Patients can further be segmented according to specific
criteria for specialised interventions, enabling higher order decision-
making capabilities for the innovation of new products. Big data
remains new in healthcare, but will bring in evolutionary improve-
ments [Chawla and Davis, 2013; Chapman et al., 2000; Pang et al.,
2018].
6.5.3 Big data in healthcare
Reflecting on the multiple facets of data within modern healthcare, it

makes sense to dissect three major clinical big data nodes, commonly
found in any healthcare organisation, or hospital [Belle et al., 2015;
Chapman et al., 2000; Chawla and Davis, 2013; Kannel and McGee,
1979]. It bears reference to the value it brings for improved medical
outcomes:
1. medical imaging – image processing is a big component of the

treatment programme, always used for diagnostic purposes and
therapy assessments. The sources of such data images come in
the form of Computed Tomography (CT) scans, X-rays, sonar
scans, fluoroscopy, Magnetic Resonance Imaging (MRI), molecu-
lar imaging, ultrasound, mammography, and Positron Emission
Tomography Computed Tomography (PET-CT). These are all
imaging techniques, within a clinical environment. The data or
bandwidth varies from megabytes to gigabytes when it comes
to thin slice CT scans. It requires massive storage capacity, in-
cluding fast algorithms for extraction and presentation capabili-
ties, for special investigations by a data scientist, for a specialist,
2. signal processing - is similar to image processing by way of

bandwidth demand and storage capacity. Signal processing comes
in volumes and with velocity, constant and in real-time. It con-
sists of critical monitored physiological data, and analysed. It
is in addition to situational clinical data which is embedded
within the continuous monitoring and predictive or decision
support systems in place, and
3. genomics – in line with the decrease in cost of technology, over

time, with increased technological capacity, the cost in mapping
the human genome has also decreased considerably, over the
past years. This is attributed to the high throughput sequencing
technology, in the field of computational biology. This partic-
ular field of medical science is still growing at a rapid rate in
alignment with the technology which enhances the outcomes of
such a medical science. However, it is becoming a focal point at
national level within the public health sector in order to build
data repositories of these genome scale data sets of different
pathologies. It will eventually lead to the modelling of a patient
profile. SA public healthcare is currently nowhere near the ma-
turity level of such a technology to be of any value. This will
naturally gain momentum in the near future.
6.5.4 Big data challenges for healthcare
While big data has many advantages to offer healthcare it does come
with challenges and operational constraints presenting to the health-
care specialist or analyst. These challenges include:
1. anonymisation of patient privacy of medical data - within any

health organisation, creates a challenge for personal identifiers
and health files,
2. redundancy - obtaining clean, compressed data sets and for-

mats, remains an ongoing challenge,
3. collaboration - across multiple professional domains where im-

proved interaction is preferable,
4. data representation - heterogenous data sets - source data will

need to be accurately represented throughout the lifecycle of
the data analytical process, which may hinder analysis,
5. technology constraints - no room for legacy systems which es-

sentially means the upgrading of all digital platforms, includ-
ing databases, tools and networks in order to collate, aggregate,
analyse, visualise and digest big data,
6. organisational change - will mean the changing of workflows

and even operational processes,
7. access to data - critical that researchers have immediate univer-

sal access to various data modes for the purposes of analysis,
and
8. industry structure - public sector generally lacks a competitive

incentive, it narrows the scope for creative productivity and ef-
6.6 anonymisation of healthcare data 125
ficiency [Chen et al., 2014; Chen and Asch, 2017; Wamba et al.,
2015].
Another challenge for big data is the ethical question – mainly

for medical research purposes. Many ethical risks face the healthcare
specialist using big data tools and concepts [Mittalstadt and Floridi,
2015]. The analyst or data scientist will have to obtain informed con-
sent for research purposes which may prove difficult with big data
sets. Research data has to be anonymised by using algorithms. In-
cluded within anonymisation of data is data ownership, coming from
multiple data sources [Vayena and Blasimme, 2018]. The data scien-
tist or analyst will need to investigate the complete ethical impact,
before beginning any large data analysis exercise.
Other challenges, such as the analytical capabilities of the analyst,

who would need to collaborate with other scientists and, specialists in
their respective fields [Vayena and Blasimme, 2018]. As the process is
mainly dynamic monitoring and digital surveillance, he or she would
have to perform a static replication of such analysis, before-hand, in
order to verify the hypothesis [Mittalstadt and Floridi, 2015; Vayena
and Blasimme, 2018].
6.6 anonymisation of healthcare data
6.6.1 Algorithms for health data anonymisation
Large data records, encapsulate numerous records of a single patient,

commonly found in databases or spreadsheets and, are classically
termed big data sets. These records incorporate highly confidential
and private personal identification and medical data. These medi-
cal data sets are strictly governed by various privacy laws, such as
the protection of personal information legislation (POPI Act of 2013)
and the National Health Act of 2005, amongst others. The POPI Act
is being enforced globally, by all governments and corporate institu-
tions [Cupoli et al., 2014; Nat, 2012; NDOH, 2005].
Many ethical risks are found within healthcare big data, as it

always requires ethical clearance or consent, for any access to any
biomedical data. There is also always the question of data ownership.
These are, only a few of the major ethical issues in high-end Big Data
analysis [Machado et al., 2012; Mittalstadt and Floridi, 2015; Young
and Zhang, 2017].
There is an ethically moral obligation and operational procedure

for any healthcare data specialist in designing such an algorithm
for the identification, or anonymisation of personal medical details
within these large health records. On a small scale, the following is

the design of an algorithm for such an exercise, on a big data health-
care repository Figure 8:
Figure 8: Anonymisation algorithm for patient details

.
Figure 9: Synthesised hospital names

.
Production 1 in Figure 8, starts by defining a date of birth as that

of the patient’s. In this production, yy is a constant extracted from
the plain data. <mm> and <dd> are synthesised with the random
context free grammar rules shown in Production 5. Production 5 also
describes rules for describing valid dates.
In Production 3, patient ID’s are synthesized using Global Unique

Identifiers or GUIDs.
6.6 anonymisation of healthcare data 127
Figure 10: Synthesised medical aid names

.
Figure 11: Synthesised patient identification numbers

.
Production 6, describes the formulation of date of birth – these

are created with a similar rule as Production 1. Names for medi-
cal aid schemes are synthesized by using a data bank of names as
prefixes and a lookup table of common sub-strings of medical aid
schemes Figure 8.
As an example, if the random name “John” is chosen by the Con-

text Free Grammar (CFG), a suffix such as “Med” will produce the
medical aid name: “JohnMed”. This is considered realistic in this
study, as a potential substitution of the real record. The definition of
these rules are in Production 8 and 9. Productions 10 to 12 follow the
previous productions, defining medical aid numbers (using GUIDs),
and fictitious hospital names Figure 8.
6.6.2 Implementation
Large healthcare data sets, biomedical data or healthcare records, re-

quire strict ethical clearance and compliance criteria as a medico legal
requirement, to be adhered to at all times. The demand for big data
analytics is increasing. The case for big data analysis becomes difficult
with such constraints and laws governing medical data. The anonymi-
sation of large medical data sets has become a focal point in gaining
access for Big Data analysis. The first step is to gain ethical clearance
regarding personal medical data. The next step is the anonymisation
of medical data records [Machado et al., 2012; Mittalstadt and Floridi,
2015].
The following example is a solution to the anonymisation of large

healthcare medical data. It consists of 244 000 patient health records
in a spreadsheet, within a rehabilitative health environment. The spread-
sheet has approximately thirty columns of category data. This partic-
ular data repository qualifies as big data, judging by the magnitude
of lines of records.
The objective is to anonymise, by way of an algorithm Figure 8,

all the relevant and related personal identification data of the patient,
in the columns as:
1. The name of the patient,
2. address of the patient,
3. patient identification number,
4. medical aid name,
5. patient medical aid number,
6. date of birth, and
7. hospital/clinic name.
The patient identifiers will require a specifically written algorithm

to anonymise certain columns of data whereby it anonymises and ran-
domises any attachment of a patient ID, name to a diagnosis, place or
the name of a clinic. In other words the biometric data is completely
disguised or camouflaged in such a way that the privacy of any pa-
tient is completely retained in the sharing of such data in Figure 11.
6.7 discussion
The tables in Figure 11, depict a library of fictitious names are gener-
ated and randomised within the data set according to the rules of the
algorithm. The algorithm performed the anonymisation of the names
of hospitals in Figure 9, patient IDs and medical schemes in Fig-
ure 10 and generated fictitious names replacing all private medical
patient identifiers thus ensuring the preservation of privacy of per-
sonal health information as seen in Figure 10. This protects the per-
sonal identification and the relationship to the identity of a patient or
entity.
6.7 discussion 129
We made extensive reference to the Design Science Research (DSR)

methodology in designing this artefact, through an iterative process
following through the cycle of stages [Gacenga et al., 2012; Hevner et
al., 2004], eventually arriving at the artefact addressing the problem
in a logical manner. The algorithm provides a solution for the preserv-
ing of any personal identification of a patient and entity, thus paving
the way for big data healthcare analytics by addressing any ethical
issues surrounding the privacy of personal health information.
6.7.1 Big data for research
The potential of multiple sources of real-time data flows of large data

files produces new opportunities for investigating the finer links be-
tween clinical health, demographics, transactional and financial data.
The explosive growth of these data sets, creates major challenges for
the healthcare analyst, never seen before.
The role of the healthcare analyst in big data analysis, differs from
that of the traditional industry data analyst, who intends to automate
analytics for improving business processes. The healthcare analyst
intends to identify patterns for the sake of discovery in a field of
clinical study, testing hypotheses that make theoretical, rather than
business sense, then to publish these in relevant journals.
Big data analytics in the healthcare sciences will deepen our appre-
ciation of disease patterns and trends, within different healthcare out-
comes, individuals, regions and communities, including the impacts
of healthcare interventions. The clinical value of big data science lies
in the feedback processes, enriching the valued outcomes. It requires
familiarity with the data fields, in gaining insights into the relation-
ships of those variables. It would include pattern recognition, such as
the understanding of a practitioners prescribing patterns, regarding
a certain medical condition, for a medical aid company [Mittalstadt
and Floridi, 2015; Vayena and Blasimme, 2018].
An important aspect of big data analysis, is to establish the anonymi-

sation of medical personal detail data. Large data sets containing de-
mographic information have existed for ages and are available. These
data sets would be classified as big data sets. Healthcare sciences data,
may still have a long journey, but the transformation of live data col-
lection, is already underway [Mittalstadt and Floridi, 2015; Vayena
and Blasimme, 2018].
6.8 conclusion
Big data continually evolves with modern technologies emerging. These

technologies leverage vast data sets, from disparate sources, either
structured or unstructured. The spectrum of analytical capability is
growing rapidly. The focus was on healthcare data in general, cover-
ing only three healthcare data components. The exponential growth
of such data sets places pressure on healthcare analysts, to innovate
new solutions, in keeping abreast with the requirements of the de-
mand. The NDOH is aware of this healthcare industry requirement,
and is currently building the foundational platform, for public health-
care. All medical professionals in conventional medicine understand
the growing demand to upskill, in the field of data science.
Clinical medicine is now merging into clinical analytical scenar-

ios, where understanding big data will become a required skill set
for medical professionals. The integration of biomedical data science
and high-end healthcare analysis, has become mainstream practice in
order to deliver effective clinical recommendation’s, based on solid
data findings. It will become a prominent skill set, for the health-
care professional, healthcare worker, government official, administra-
tor, executive, CEO, CIO and industry alike. Pursuant to this is a
critical component, in finding a solution, is the anonymisation of a
patient’s medical details, for ethical and privacy of data reasons, as
recently legislated.
In this chapter, big data was covered, originating from multiple nodes
data flows. The data is either stored in the Cloud or on the premises
of an organisation The data flow originates from a streaming data
set produced by a patients medical device, or wearable device in the
form of IoT, reports, demographics, eHR and even from transactional
data or retrospective medical data.
The nature of big data is predominantly real-time data flow, re-

quiring a whole new skillset. These skills include the automation of
certain operational functions, in order to improve the quality of value
added services.
While the traditional functions of storage and the management

of data has proven inadequate, big data has effectively changed the
approach to managing data, across the board. Conventional storage
of data, primarily focused on well structured, unstructured and semi
structured data sets, lacked the technical ability to manage streaming
real-time data, with expanding volumes.
It is at this juncture that the big data concept was born. Big data
further developed its own capabilities of high-end data analytics. It in-
cluded, valued insights - through predictive rather than interpretive,
descriptive and diagnostic capabilities, never seen before.
The five layers of a big data architectural platform was detailed

originating from the source layer, aggregation, analytical, the visuali-
sation layer and finally ending at the data governance layer.
An algorithm was designed and developed, as one of many Big

Data tools, used for the analysis of healthcare big data sets. It was
designed in order to anonymise a patients personal medical and de-
mographic details, in compliance to the modern day privacy laws
and governance policies, noted in the literature. The algorithm was
presented with a feasibility study of the design, written up as a con-
tribution to healthcare analytics.
Many big data tools exist in the market where the most relevant
in modern healthcare were highlighted. These included MongoDB,
RapidMiner, Excel, Tableaux and others. Cloud technology houses
these modern big data frameworks and platforms. Various benefits
and challenges, including the ethical debates and policies, regarding
big data was covered. The planning and approaches to implementing
big data for healthcare were drawn from mainstream literature.
7 T R E N D A N A LY S I S : A D E C I S I O N T O O L I N S A
H E A LT H C A R E
T
he use of medical surveillance and predictive analysis, is now
more readily available through the intelligent use of modern
technology, with the use of software tools. A few of those
tools are simple open source tools available to the general public,
which produce accurate and effective results, having been tweaked
and improved upon over the years. One such tool is Google Trends,
which has made an impact on the public whilst impressing the sci-
entific community spanning a range of medical subjects and regions,
globally [Abbott and Ade-Ibijola, 2019b].
7.1 introduction
The true essence of epidemiological trend analysis is the early detec-

tion of a pattern, or trend of a specific disease, making use of vari-
ous clinical methods in detecting those trends per country, state or
region [Carneiro, 2009; Schootman et al., 2015]. This is commonly re-
ferred to as medical surveillance or monitoring of such clinical pat-
terns in order to provide a solution before it becomes an epidemic dis-
aster. It has been performed through traditional medical surveillance
methods deployed within communities and regions, albeit often slow
in nature. The use of medical surveillance of information seeking and
predictive analysis has been elevated to a higher level, with the intel-
ligent use of modern technology such as the use of internet software
tools [Carneiro, 2009; Schootman et al., 2015].
Over the last 20 years the internet has become integral to public
health surveillance, by reducing the time to the actual recognition of
a disease outbreak or trend [Carneiro, 2009; Schootman et al., 2015].
Google Trends (GT) was launched in 2008, as an open-source based in-
ternet research tool for the surveillance of flu, in the USA. It is primar-
ily used for the early detection and ongoing monitoring of epidemics
through the use of monitoring internet portal message search data,
which is collated over a time period, per region or country [Carneiro,
2009; Nuti et al., 2014; Olson et al., 2013].
GT primarily focused on detecting an outbreak of various strains

of influenza (flu) infecting different regions, sporadically or endemi-
132
cally through information seeking by internet users. It is an adjunct to

the Google search engine criteria with more directed algorithms tar-
geting web search volumes of an outbreak of the flu, within a region
or country, by virtue of tracking the internet messaging movements or
volumes of these tweets or messages, over the internet [Brownstein et
al., 2009; Carneiro, 2009; Lotto et al., 2017]. The results of such a trend
are plotted over time, and presented in a graph. This makes it easy
to read and interpret, for anyone gaining access to GT. The causal
inferences are self-explanatory through logical interpretation.
Various medical professional bodies are turning to GT in monitor-

ing the portal traffic of their respective specialities. The World Dental
Federation monitors the incidences of toothache patterns and then
make new seasonal discoveries for the benefit of public interest [Lotto
et al., 2017]. The medical profession in Australia is making valuable
comparisons, regarding the series volume indexes (SVIs) for colon,
prostrate, breast and skin cancer against the Age Standardised Rela-
tive index searches (ASRs) [Huang et al., 2017]. ASR is the product
of sophisticated predictive modelling techniques used for age index
searches [Huang et al., 2017]. They are relying on GT tool to explain
certain changes and fluctuations, with qualified accuracies.
The data used in this study has been downloaded directly from
GT, applicable to the South African regions and demographics, which
is open to the public and allowing any internet user to access such in-
formation seeking trends(www.google.com/trends). Furthermore, no
ethical research requirement is needed, from any governing commit-
tee or body, to access such information, as it is anonymised [Schoot-
man et al., 2015]. The reason is that the data is not directly related to
any research subjects or clinical field studies requiring ethical clear-
ance, from an Institutional Review Board (IRB) [Lotto et al., 2017].
The data is analysed over a time period and extrapolated, in order

to estimate the volume of searches; a graph is mapped where these
spikes in volumes are noted as causal inferences by the observer. The
data is primarily scaled by way of relative and fixed scaling [Carneiro,
2009; Schootman et al., 2015]. The outcome produces a seasonal trend,
or pattern, which is presented in a graphical display, as a time-series
graph, which is similar to a periodogram [Brownstein et al., 2009;
Huang et al., 2017].
The relative data is scaled by averaging the search volume of a

time period whereas the fixed scaling data is normalised using the
extrapolated search volume at a fixed time point i.e. the data is nor-
malised such as the flu data index graph [Nuti et al., 2014]. Essen-
tially this means that an increase in volume of an internet search will
increase its own average and denominator over time [Brownstein et
al., 2009; Carneiro, 2009; Nuti et al., 2014]. These web user searches
134 trend analysis: a decision tool in sa healthcare
originate from the Google search engine and affiliated search engines
where users can download the results of such searches [Nuti et al.,
2014]. The user may request to indicate a geographic region by dis-
trict, region, country or global region. The portal will measure the
volumes of searches as per location, identified by the user over a spec-
ified time frame which then gives impetus to comparative analysis
across a myriad of research questions and social events.
A study was conducted in Kenya where it found GT to be ex-

panding and increasing annually, as an adopted research tool within
the medical college of medicine [Rotich and Onyancha, 2017]. It also
holds true for GT in the near future and brings a renewed surge in
such a research tool. GT users are utilising the tool for healthcare
searches; this is now becoming evident in the increase of global cita-
tions and publications [Nuti et al., 2014; Schootman et al., 2015]. It is
all too evident in the large proportion of cited literature of infectious
diseases where GT is extensively used for causal inference, in order
to answer many unanswered questions [Carneiro, 2009; Schootman et
al., 2015].
One such event is the measuring of global awareness campaigns

concerning breast cancer awareness, whereby the interest and aware-
ness can be retrospectively analysed, measured and proven to be ac-
curate over the last decade [Benson, 2013; Sugrue et al., 2016]. In addi-
tion, the occurrence of breast cancer is in decline as a result of these
positive campaigns where GT is proving the value of such internet
monitoring of campaigns [Benson, 2013; Sugrue et al., 2016].
The contribution of this chapter includes the benefits of a tool

utilised in the early detection of disease trends, from the web portal
searches on the web. Further, it serves to propose an approach to solv-
ing the problem of improving the early detection of a disease trend
in addition to traditional clinical surveillance methods deployed. In
addition, various disease scenarios are applied within the general do-
main of South Africa and its regions producing clinical inferences in
adapting possible solutions within the healthcare domain of South
Africa.
The layout of this chapter is organised in the following manner.

Current GT work related to this medical trend research is presented
including the methodology and development of the GT tool and its
method of application. The results of different scenarios, or diseases,
as applied in the South African healthcare landscape are discussed,
with the outcomes of such results and its implications by utilising GT.
Finally, the conclusions are stated.
7.2 related work 135
7.2 related work
7.2.1 Google trends for cancer screening
Cancer screening remains integral to public health, by virtue of di-

recting immediate healthcare to an early diagnosis and possible pre-
vention. There are various non-medical behavioural and risk method-
ologies developed, which mainly comprise of traditional surveillance
systems or mechanisms such as questionnaires and telephonic inter-
views [Carneiro, 2009; Lotto et al., 2017; Schootman et al., 2015].
The traditional surveillance systems deployed are the Behavioural

Risk Factor Surveillance System (BRFSS), National Health Surveil-
lance (NHS), and Hypertension Intervention Nurse Telemedicine Study
(HINTS), clinical surveys, interviews, and medical epidemiological
surveys. These methods have their own inherent challenges and limi-
tations which led to the emergence of more technologically advanced
screening such as scans, colonoscopies and others [Olson et al., 2013;
Nuti et al., 2014].
GT is now attracting attention in the use of detecting cancer in-

formation seeking incidences and related topics. The correlation be-
tween GT data and reported data on breast mammography, cervi-
cal pap smears, prostrate (PSA) and colonoscopy screening is adding
value to the traditional surveillance systems [Schootman et al., 2015].
This is accomplished by a GT search (up to 30 words) which produces
a Relative Search Volume (RSV) scaled to the highest search propor-
tion per time scale. GT can aid and support the inherent questions in
such traditional surveys.
7.3 methodology
7.3.1 Google flu trends development
The early stages of the development of GT had its challenges where

outbreaks of the Swine Flu in 2009 were completely missed with other
examples [Nuti et al., 2014; Lotto et al., 2017; Schootman et al., 2015].
Since then the tool has been refined and the wording criteria or syn-
tactic algorithms have been further developed and enhanced for tar-
geted accuracy. One of the primary requirements for the efficacy of a
GT result is that there must be a large population utilising the web ex-
tensively and preferably within a sophisticated environment of web
search user population; only then can the trend be deemed reliable
and accurate to a large degree [Nuti et al., 2014; Schootman et al.,

2015].
GT is largely a web-based tool proving to be as accurate and re-

liable as any traditional surveillance methods, and any technology
or tool currently deployed by global organisations, such as the Cen-
tre for Disease Control (CDC), based in Atlanta, USA [Lotto et al.,
2017]. The CDC will typically rely on laboratory results and clinical
data to publish these regional outbreaks of flu but this takes time;
the retrospective application creates a time lag of approximately two
weeks [Carneiro, 2009; Olson et al., 2013].
GT is convincing the medical community that the retrospective ac-

curacy is sharpening and becoming more reliable with comparative
analysis conducted on a regular basis [Nuti et al., 2014; Young and
Zhang, 2017]. However, there are challenges from researchers who
are questioning the frameworks of criteria used and the consistency
of such criteria, as they remain unclear or undocumented purely from
a research tool point of view [Carneiro, 2009; Schootman et al., 2015].
The scientists want to see more transparency and documented ac-
counts of frameworks of these search criteria Brownstein et al. [2009].
In this study we applied specific search criteria in the GT tool

utilising the most common, or rather, the most prevalent communica-
ble and non-communicable diseases appearing in the South African
healthcare environment. These disease entities are most prevalent in
public health clinics and hospitals and currently being dealt with on
a daily basis. These diseases are rated among the highest causes of
death and morbidity such as HIV, tuberculosis, diabetes, cardiac fail-
ure and cancer.
A simple narrative was applied in drafting a comparative narra-

tive and applying the search criteria within the Google tool. The set-
tings as per time frame were standardised over five years of historical
data for all the listed scenarios concerning South Africa and its re-
gions. The results of this search criteria for South Africa are presented
in the figures for Gauteng, Western Cape, Northern Cape, North West,
Free State and Eastern Cape. From these clinical observations, certain
inferences were made and presented.
Nevertheless, respected researchers are questioning the validity,

consistency and transparency of the methods adopted and used by
GT [Nuti et al., 2014; Olson et al., 2013]. It eventually brings the trust-
worthiness of such a tool over time. The GT tool has its limitations
and restrictions in: the actual plotting of the incident or event of such
a disease. Further, this includes the disease outcome trend or inter-
ventions implemented, which may be based on a sampling bias, with
certain events and regions [Carneiro, 2009; Brownstein et al., 2009;
Schootman et al., 2015]. GT does not map any specified data, but only
7.4 results 137
focuses on the internet search communication thereof, hence, its ef-

ficacy in the early detection of a disease event in a specified region,
with sophisticated internet users, over a specified time frame will cer-
tainly supplement other monitoring and surveillance tools [Brown-
stein et al., 2009; Carneiro, 2009].
The advantage of GT is that it is consistently proving to have a

reliable lead time of nine days on any other traditional tool, in the
early detection of a disease event [Nuti et al., 2014; Olson et al., 2013].
Ultimately, the internet portals and systems are proving to be a pow-
erful communication channel and a dominant source of information
on emerging patterns of diseases but some remain cautious of the
actual methods deployed [Brownstein et al., 2009].
7.4 results
7.4.1 Google trends scenarios - South African healthcare
The following scenarios or hypotheses support the GT predictive

trend of the following events using the GT tool, within the South
African context.
7.4.2 Tobacco smoking can result in an increase in lung cancers
Tobacco smoking is a legal drug that is chronically debilitating caus-

ing potential non-communicable diseases such as lung cancer and car-
diovascular diseases resulting in death [Rotich and Onyancha, 2017].
Mainstream medical literature has extensively documented the

chronic use of tobacco smoking that will inevitably cause a cascade
of diseases, with lung cancer being the most prevalent [Young and
Zhang, 2017; Zhang et al., 2015]. The WHO is on a global drive to
reduce tobacco smoking considers that approximately five million
people globally, die from direct tobacco smoking, annually [WHO,
2013]. In Figure 12 shows the trend analysis for smoking, versus lung
cancer results, in South Africa.
7.4.2.1 Trends
In the presence of tobacco smoking there is a trend of lung cancer

awareness. The increase or peaks are directly related to the increases
of lung cancer. Lung cancer needs more awareness due to the sig-
nificant gap between the trends. Mpumalanga shows a trend of the
Figure 12: Smoking versus lung cancer within South Africa
Figure 13: Smoking versus lung cancer within regions in South Africa
highest rate of smoking with the least awareness of lung cancer, as

seen in Figure 13.
7.4.2.2 Inferences
The trend of smoking interest and the awareness of lung cancer would
need an awareness campaign or more patient education. Intervention
by these campaigns will close the gaps between the trends.
7.4.3 Diabetic neuropathy and lower limb amputations
Within mainstream diabetic clinical medicine the main cause of leg

amputations is the loss of sensation, due to chronic diabetes with:
clinical neuropathy, or diabetic neuropathy [Pickwell et al., 2015].
This neuropathic syndrome leads to complications and ulcerations

due to undetected mechanical pressure on the feet [Vaquero et al.,
2019]. Ulcerative sepsis eventually leads to chronic infection of the
7.4 results 139
foot and limb, which can necessitate a lower leg amputation, if left
untreated or undetected, in a chronically mismanaged diabetic [Pick-
well et al., 2015].
The key is preventative interventions in detecting pressure and the

changes in skin temperature. Early warning sensors, or (wearables)
are a possible mechanism in early detection of temperature changes
in the limb [Vaquero et al., 2019]. Sensors can aid in early detection of
a causative factor.
Figure 14: Trend analysis of diabetic neuropathy versus lower limb amputa-
tion in South Africa
Figure 15: Trend analysis of diabetic neuropathy versus lower limb amputa-
tion as per region in South Africa
7.4.3.1 Trends
There is a visible fluctuating trend of activity with a high level of

activity with the amputations trend. Amputation awareness is preva-
lent - above neuropathy. Neuropathy awareness is significantly low
in KZN, Western Cape, and Gauteng whereas amputation awareness

is much higher in these regions as seen in Figure 14.
7.4.3.2 Inferences
There is a national need for prompt awareness of potential limb am-

putations due to neuropathy which needs to be addressed at regional
level, starting with KZN and Western Cape and Gauteng, in that or-
der. This makes logical sense since, the diabetic population is concen-
trated in the Indian community of KZN and the Coloured communi-
ties of the Western Cape region, as presented in Figure 15 [Sou, 2016].
Medical science which are available to corroborate these figures of
concentration which are immediately indicated via GT, thus making
it a credible and viable tool for medical trend monitoring.
7.4.4 Breast cancer screening
One of the most frequently diagnosed cancers is breast cancer which

can be in excess of a million diagnoses a year [Andersen et al., 2011;
Sugrue et al., 2016]. The outcome of such a disease can be fatal, unless
an early warning mechanism is utilised constantly. Breast cancer is
the leading cancer of all cancers, amongst women [Sugrue et al., 2016].
It remains the focus of awareness and trend analysis, globally.
Figure 16: Trend analysis of breast cancer screening and breast cancer within
South Africa
.
7.4 results 141
Figure 17: Breast cancer screening versus breast cancer as per region within
South Africa
.
7.4.4.1 Trends
Breast cancer is prevalent with spikes of increases. This trend annu-

ally spikes over the October period. Breast cancer interest volume
is lowest in June, annually. Screening interest is low compared to
dealing with cancer and breast cancer is most prevalent in the West-
ern Cape region, followed by Northern Cape, as shown in Figure 16.
There is a definite trend for breast cancer screening, compared to lung
cancer screening. Breast cancer screening remains low in comparison
to the disease interest.
7.4.4.2 Inferences
More campaigns for screening directed at the greater population are

required. The South African National Department of Health (NDOH)
needs to intervene with direct campaigns of screening in the whole
of the Northern Cape region, according to regional trends where it
would have the largest impact [Harrison, 2010]. Figure 17 shows this
result within South Africa.
7.4.5 The prevalence of diabetes and cardiac arrest
The prevalence of diabetes will eventually lead to an increase in con-

gestive cardiac diseases, commonly known as heart failure or heart
attacks. It remains common knowledge in the medical literature and
well documented in most medical journals, being one of many cas-
cading pathologies of diabetes [Kannel and McGee, 1979].
Figure 18: Trend analysis of diabetic prevalence and heart failure within
South Africa
.
Figure 19: Trend analysis of diabetic prevalence and heart failure within
South Africa as per region
.
7.4.5.1 Trends
The diabetic trend is more pronounced over time, compared to car-

diac failure, which is in parallel with the diabetic trend, albeit at a far
lower level. Diabetes spikes and the volume trend are similar to that
of heart failure but on a lesser scale. There is a wide gap between the
two trends. The diabetes trend is most prevalent in the North West
Region. There tends to be far less awareness of cardiac failures/dis-
eases amongst the diabetic population in general across the SA region
shown in Figure 18.
7.4.5.2 Inferences
This trend analysis shows a great need for more awareness cam-
paigns, directed at the diabetic population considering the correla-
7.4 results 143
tion between diabetes and cardiac failures, or potential heart attacks.

These campaigns must emphasise that Cardiac awareness must start
early for all diabetic management programmes, incorporating lifestyle
change programmes. Presented in Figure 19.
7.4.6 The most prevalent diseases in South Africa are cancers, diabetes,
HIV and Tuberculosis (TB)
Currently these diseases have the greatest financial cost to both the
private and public healthcare systems [Archer, 2016; Ruxwana, 2014].
They fall within the communicable and non-communicable diseases
of South Africa [Archer, 2016; Herselman et al., 2016]. There is cur-
rently approximately 6,8 million people within SA living with HIV [Har-
rison, 2010; Theron, 2016].
A similar prevalent communicable disease is Tuberculosis (TB).

However, the most prevalent in SA are HIV and Cancer, according to
GT. TB is a global disease that realises 10 million new cases annually
and over a million deaths per year [Zhou et al., 2011]. The major line
of defence is early detection and early surveillance by the traditional
method implemented by, which is lagging behind the adoption of GT
analysis [Ortiz et al., 2011; Pervaiz et al., 2012; Zhou et al., 2011].
In the developed world, such as the United States of America,

various studies are now being conducted with the use of big data
search engine analysis for the monitoring of new HIV diagnoses per
state [Young and Zhang, 2017]. This is done in collaboration with the
Centre for Disease Control (CDC) by comparing their statistics col-
lated through their own traditional monitoring tools such as surveys,
interviews and laboratory reports. The only disadvantage of these tra-
ditional methods is that they carry a time lag and are costly, whereas
the GT tool is proving to be a feasible and reliable tool in predicting
new HIV cases, at state level in the USA [Zhou et al., 2011].
7.4.6.1 Trends
Trend is showing the most prevalent diseases in South Africa as HIV,

cancer, diabetes, TB and cardiac diseases, in that order, according to
GT in modern SA (presented in Figure 20). Cancer and HIV are the
most prevalent trends in SA which correlates with medical science
and NDOH statistics. TB is the fourth highest disease trend. Cardio
vascular diseases or heart attack trends are fifth and the last of non-
communicable diseases but is emerging in modern SA. Communica-
ble diseases supersede the non-communicable diseases. Eastern Cape
and North West region are the most prevalent for cancers.
Figure 20: Trend analysis of cancer, diabetes, HIV and tuberculosis(TB)

within South Africa
..
Figure 21: Trend analysis of cancer, diabetes, HIV and tuberculosis(TB)

within South Africa per region
..
7.4.6.2 Inferences
The current trend is that the non-communicable disease, such as dia-

betes and cancers, are the less prevalent according to the trends. This
clearly indicates to the NDOH, and government, to refocus funding
efforts towards emerging disease patterns, such as the non-communicable
diseases, with a decisive plan in place to cater for this emerging
disease trend. HIV and TB are indeed prevalent and currently at-
tracts attention, but one needs to see the same focus on the non-
communicable diseases in the near future. The regions that should
receive more focus and funding are the Eastern Cape and North West
regions according to the regional trend analysis shown in Figure 21.
7.5 discussion 145
7.5 discussion
The concept of web-based information seeking and mapping in pre-

dicting epidemics, through the use of trending analysis, is now widespread
[Telfer and Woodburn, 2015]. On-line social networks are now ac-
cepted reliable research tools that are currently being enhanced, and
making increasing improvements on search criteria. This is evident
in these trend analysis outcomes being compared with traditional
surveillance medical tools, currently in place [Ortiz et al., 2011; Rotich
and Onyancha, 2017; Schootman et al., 2015]. Trend analysis tools,
such as GT, will inherently carry a large degree of concern amongst
the general populace as they are in the public domain [Nuti et al.,
2014; Pervaiz et al., 2012].
The above examples lean towards the fact that these web-based
tools are becoming important sources of valuable information, that
may affect communities and sectors of certain population groups,
presenting as disease epidemics. This invariably has an impact on
the public health measures taken to prevent, or circumvent such dis-
ease trends. However, this requires careful analysis and verification
of such findings [Brownstein et al., 2009; Carneiro, 2009; Ricketts and
Silva, 2017].
Medical science is expanding its horizons with internet search en-

gine tools by analysing the relationship between internet user searches
on actual mortality rates and the causes of death, within the United
States. The actual data on the ground was collated from the CDC
morbidity and mortality report. GT variables proved to be a reliable
measure or mortality, proving accurately the top ten causes of mor-
bidity, but GT cannot identify race, gender or age [Ricketts and Silva,
2017; Zhou et al., 2011].
It is becoming more evident that modern software monitoring

tools such as GT analysis are playing a more critical role in the early
detection and monitoring rates of epidemiological diseases. The tra-
ditional contemporary methods used by global organisations, such as
the Centre for Disease Control (CDC), are now experiencing time lags
of two to twelve weeks on average for the detection of cancers, HIV,
TB, and flu strains [Sugrue et al., 2016; Zhou et al., 2011].
Central to the vision of the Fourth Industrial Revolution (4iR) and

the United Nations Sustainable Development Goals (SDGs) for 2030,
are the advanced data analytical capabilities of the internet with: both
being challenged to contribute to future digital economies [Jeffery,
2016].
While GT is just one of many tools to be exploited by internet

users in this regard, many challenges lie ahead to refine the mechan-
ics behind the tool in order to make valuable contributions to future

health outcomes. The limitations or drawbacks of such search tools
such as GT are manifold, with respect to other methodologies that
have been adequately documented, which hinders the reproducibil-
ity of certain trend searches in many respects [Schootman et al., 2015;
Sugrue et al., 2016; Young and Zhang, 2017].
The GT studies analysed were primarily used for the following

outputs:
1. Causal inference depicting direct relations of such trends.
2. Surveillance of epidemic trends.
3. Description of epidemics and precision of identification.
Certain scientists do argue that a certain amount of positive result

bias can exist due to the novelty of the tool where a certain amount of
reluctance exists in admitting or declaring negative feedback on the
actual tool [Olson et al., 2013; Sugrue et al., 2016]. Researchers need
to undertake more searches and compare retrospectively with suffi-
cient documentation in order to create a set of best practices going
forward [Brownstein et al., 2009]; thus to ensure the reliability and
authenticity of the tool. Researchers should strive to document those
methodologies and publish them on a wider scale with screenshots
included [Sugrue et al., 2016].
7.6 conclusion
Google Trends is a tool in the public domain where the general popu-
lation can extract and derive meaningful insights about the behaviour
of the population on certain topics or events such as disease epi-
demics. This is due to significant changes in healthcare strategy and
healthcare policy. The outcome of such an epidemiological trend has
a lead time on these insights which produces clinical value in the
field. Much of the body of knowledge in this particular field of study
of surveillance studies certainly validates GT outputs against tradi-
tional medical surveillance data analysis. The GT results presented
have similar applications in other areas of medicine and disease mon-
itoring efforts in South Africa whereby the general population can
derive meaningful insights of such disease trends.
In this chapter we have chosen to present the distinct advantages of

analysing various medical conditions with simple software tools at
our disposal. The use of medical surveillance and predictive analysis

is more prevalent than ever before, through the use of software tools.
Google Trends (GT) is an internet research tool, originally utilised for
the surveillance of flu outbreaks and trends.
The search criteria using GT within a typical healthcare scenario

or disease trend was applied to the most common diseases found
within the South African healthcare environment. The diseases anal-
ysed with the GT tool were; Cancer screening; tobacco related cancers,
diabetic amputations and neuropthy; breast cancers; diabetic preva-
lence in heart failure and the prevalence of tuberculosis within South
Africa. The trends and results were analysed and presented. The re-
sults were similar to those found using traditional medical surveil-
lance tools.
In the next chapter we have managed to present a simple tool in

decrypting clinical notes into simple language, for all to understand.
8 N AT U R A L L A N G U A G E P R O C E S S I N G ( N L P ) :
D E C RY P T I N G C L I N I C A L N O T E S
N
atural Language Processing (NLP) presents a set of tech-
niques that is finding application in modern healthcare, for
the extraction and generation of text. Clinical notes are clas-
sically originated and derived from various sources, such as reports,
referral letters, discharge notes and clinical summaries.
In this chapter, we present algorithms and a tool for enabling each

member of a medical team to read and understand each others med-
ical notes, using NLP techniques. We refer to this process as the de-
cryption, or the deciphering, of complex clinical notes into simple
readable language. We have presented a tool called the Clinical Note
Translator (CNT). Based on the replacement of technical terms in clin-
ical notes, CNT translates these notes to plain text. It is a solution
expected to assist multidisciplinary medical teams, in understanding
specialist clinical notes.
8.1 introduction
Healthcare is a profession where effective communication amongst

the whole medical team and, the patient, is of paramount importance.
Such effective communication should include clear and legible clini-
cal notes. Such effective communication, is not always the case, com-
pared to other industries and professions [Collobert et al., 2011; Liddy,
2001]. The concept of NLP is the computational approach to analysing
text, and more recently the spoken linguistic analysis, based on sets of
theories and computational technologies, for the purpose of produc-
ing simple user friendly records [Cawsey and Jones, 1997; Friedman,
1999; Liddy, 2001]. NLP can be seen as a range of computational tech-
niques applied to naturally occurring texts, or text mining, for the
sole purpose of achieving a natural language, presented in a simple
format [Liddy, 2001].
The underlying essence of any medical record is effective commu-

nication amongst medical personnel, including the patient, and it re-
mains critical with respect to clearly stated medical records [Cawsey
and Jones, 1997; Friedman, 1999]. These applications were not de-
void of challenges in the translating of clinical records into struc-
148
tured reports and clinical notes [Murdoch and Detsky, 2013; Nad-
karni et al., 2011]. The difficulty when treating a patient within a
multi-disciplinary team (MDT), in a clinical setting, is to understand
each other’s clinical notes, summaries and narratives, across the dif-
fering medical specialties within the medical team. Understanding
the differing narratives of each others clinical notes, remains a frus-
tration within any medical team and institute [Porter, 2010a].
It is evident that clear communication (verbal and written) in

healthcare is critical between professionals and patients [Collobert
et al., 2011; Friedman, 1999; Hangu, 2018]. It means legible clinical
notes, within public or private hospitals. Medical notes, hand writ-
ten or sound clipped, are generally presented in differing views, ex-
pressiveness, ambiguity, vagueness and interpretations of the clinical
situation at hand [Friedman, 1999; Neveol and Zweigenbaum, 2015].
The most common problem being faced was the complexities and
ambiguity of medical terminology, the multitudes of meanings and
differing interpretations of clinical data, in establishing a clinically ac-
curate synopsis of the condition, or management of such a medical
condition [Afzal et al., 2018; Trivedi, 2018].
NLP within healthcare originated back in the 1940s where many

technological attempts, and strides, have been made ever since to per-
fect the science of interpreting natural language into a computerised
format or simple text. Machine Translation (MT) was the precursor to
NLP, in the form of a question-and-answer basis. Moreover, the early
projects initiated in the Second World War inspired many projects,
post the war [Liddy, 2001]. This marked the birth of cryptography
and language translation as an accepted science.
The first systems took a simplistic view of word-orders, which

used dictionary reference for the appropriate words. Those words
were simply re-ordered with a few accompanying rules based on the
target language [Liddy, 2001]. It was only much later in 1957 that syn-
tactic structures, came to light [Cawsey and Jones, 1997; Collobert et
al., 2011]. This introduced the concept of generative grammar [Liddy,
2001]. It was not long before syntactic theory of parsing algorithms,
motivated further development [Liddy, 2001; Trivedi, 2018].
Through the 1980s, further development of NLP techniques and

software technologies emerged, creating further statistical NLP [Nad-
karni et al., 2011]. The concept gave rise to statistical parsing which
was based on probabilities, determined through machine learning or
annotating those large bodies of text. It further disambiguates the
language, so that many detailed rules could be condensed and re-
placed with statistical-frequency information. Statistical approaches
soon developed into mainstream NLP, as they rely on vast amounts of
data [Friedman, 1999; Nadkarni et al., 2011]. The volumes of data are
150 natural language processing(nlp): decrypting clinical notes
increasing exponentially, where analysts were predicting that health-

care data will soar to 35 Zeta bytes by 2020 [Schmarzo, 2016]; expand-
ing it approximately 43 times the volume of healthcare data found in
2009 [Gui, 2016].
The contribution of this chapter includes an applied algorithm in

order to extract plain text, from complex medical reports. It serves to
propose an approach to solving the problem, in deriving a simple so-
lution within medical teams, by deciphering and understanding each
others clinical notes. The solution is based on accepted NLP rules and
standards, being applied to clinical narratives, within a medical teams
notes. Further, it brings an approach to simplifying specific medical
domain knowledge areas, by means of applying an engineered algo-
rithm, per a medical sub-specialty or domain. This method will ma-
ture with time, including lessons being learned within the medical
domain.
The chapter defines the background work done in healthcare. The

methodology is presented including the results of such an algorithm,
applied in the healthcare landscape. The discussions of the outcomes
of such results and their implications by utilising the algorithm is pre-
sented. Finally the conclusions are outlined, regarding the extraction
of plain text from an array of complex medical reports.
8.2 background and related work
8.2.1 NLP in healthcare
The application of NLP within healthcare has been attempted over

previous decades in translating medical records. This consisted of
notes from the generation of medical diagnoses to the transforma-
tion of free text clinical notes, into legible documents and reporting
structures [Murdoch and Detsky, 2013; Trivedi, 2018].
NLP is considered as a specific discipline of Artificial Intelligence

(AI), striving for human-like performance. AI consists of three main
tools namely; Machine Language (ML) techniques - for structured
data, deep learning techniques and NLP techniques - for unstruc-
tured data [Chopra et al., 2013; Jiang et al., 2017]. NLP originated as a
combination of artificial intelligence and linguistics [Nadkarni et al.,
2011]. It was further combined with the combination of NLP and In-
formation Retrieval (IR). Currently, NLP has been further developed
within many fields, expanding those knowledge bases.
It is through this challenge that a large degree of structure needs

to be well-defined and tagged, through the effective application of
8.2 background and related work 151
NLP, within the healthcare sector [Afzal et al., 2018; Hangu, 2018].
This can further produce well-structured coherent medical notes and
records, devoid of ambiguity [Cawsey and Jones, 1997; Friedman,
1999]. These techniques have been used extensively in decision sup-
port analysis within healthcare.
The early approaches to language processing or NLP are found

within the following categories:
1. deep linguistic basis - seen in logical or rules-based systems.

These rule-based systems comprise of rules and conditions to
be followed [Nadkarni et al., 2011],
2. statistical - these methods consist primarily of mathematical

techniques depending largely on statistical models such as the
Hiddem Markov Model (HMM) which is used extensively in
speech recognition, and
3. connectionist - similar to statistical approach but it combines

with various theories of representation. Thus allowing infer-
ences and the manipulation of logic formulae [Nadkarni et al.,
2011].
A patients medical records are invariably shared amongst fellow

professionals and professions, without ever meeting each other in the
process. However, medical terminology and interpretation can dif-
fer [Cawsey and Jones, 1997; Friedman, 1999].
8.2.2 Basic structure
NLP enters the realm by applying codes and tags to certain constructs
of the language, giving it a coding index of terms and definitions. In
order to understand the medical textual language, the NLP system
must understand a few concepts:
1. syntax- is the structure of sentences such as the meanings, verbs,

nouns, [Friedman, 1999],
2. semantics- relate to the meanings of the words and their respec-

tive combinations in giving meaning to the sentence such as the
description of the patient’s pain to be severe, and
3. domain knowledge- the information of the subject matter or

concepts which can vary considerably in different systems, crit-
ical in reducing ambiguity and different meanings or interpre-
tations of the subject matter at hand [Friedman, 1999].
Further, domain knowledge is organised into a type of hierarchy

model which gives a relation to similar concepts.
In China and England, modern healthcare systems are becoming

more intelligent, as these systems have the ability to pass medical
board licensing examinations [Iroju and Olaleke, 2015]. NLP systems
are gaining momentum and recognition in the healthcare sector, but
this is not without medico-legal and technological challenges.
The uptake of NLP is hindered by variation in the language of ter-

minology across healthcare, including its many medical sub-specialties.
The respective sub-specialities of medicine, include the differing en-
coding diagnostic criteria governed by various global bodies. It in-
cludes the International Codes of Diagnoses - ICD 10 or 11 - diagnos-
tic coding versus other forms of encoding, laid down by the World
Health Organisation (WHO) [WHO, 2013]. It includes the adaptation
of the respective NLP systems to the medical sub-speciality which
requires development time in such events.
8.2.3 Parsing
Parsing is a designated NLP task that will attach syntactic structures

to certain sentences or clinical notes according to some form of gram-
mar. These current parsers are broken down into two types - con-
stituency parsers and dependency parsers. These dependency parsers
are currently attracting more attention and are further categorised
into two approaches - transition based and graph based [Zhang et al.,
2019].
The most commonly used parsing systems currently being used

include the systems developed by the Stanford parsers [Wang et al.,
2018a; Zhang et al., 2019]. The traditional Stanford parser has used
medical lexicons where deep learning dependency parsers have been
minimally used for the clinical landscape [Zhang et al., 2019]. The
most progressive of such parsers further include the Bist parser, De-
pendency tf and the jPTDP-parser [Zhang et al., 2019].
Nevertheless, NLP systems are being used extensively to bring

structure and linguistic comprehension to vast unstructured clinical
narratives in order to improve the decision making processes, to ulti-
mately improve value based outcomes, efficiencies and reduce health-
care cost [Iroju and Olaleke, 2015; Koleck et al., 2019; Porter and Teis-
berg, 2008].
8.3 the methodology
NLP within healthcare has historically produced various techniques

in generating reports, briefings, discharge letters, progress notes and
8.3 the methodology 153
various other medical narratives. In addition, expert systems sup-

porting clinical decision-making capabilities for various medical sub-
specialties have been developed. The techniques applied, have been
ratified and validated in differing areas of medical sub-specialties.
However, these were primarily rules based and applied through sim-
plistic NLP methods such as applying rules to strings of text, then
stringing them together, in a process, termed "realisation".
input : clinical_note
output : plain_text
plaintext += clinical_note.Replace(p_meaning);
return plain_text
Algorithmus 8.3.1 : Algorithm translates clinical notes
Recent work is now enriching these systems with highly sophisti-

cated developed algorithms. They are being applied across electronic
health records. They are producing improved outputs with much
higher percentage rates of accuracy. Therefore NLP systems with ap-
plied algorithms, are used to extract meaningful information from un-
structured data sets, predominantly found in electronic health records
[Iroju and Olaleke, 2015; Porter and Teisberg, 2008].
A medical condition, such as an "ischaemic limb" (diminished

blood supply to the leg), is not necessarily coded under the Interna-
tional Codes of Diagnosis version 10 or 11 (ICD 10) coding structure,
as it presents as a manifestation of numerous sign and symptoms.
Those signs and symptoms can be separate diagnoses of underlying
diseases, such as diabetes, which is encoded under the ICD 10 cod-
ing system. Therefore, in order to extract meaningful data to predict
a condition, it now becomes far more beneficial than identifying a
condition [Porter and Teisberg, 2008].
It makes logical sense to develop an effective algorithm in pre-

dicting the clinical outcome, hence preventative medicine being the
logical pathway. Moreover, in developing and applying an effective al-
gorithm for clinical narratives and notes, thereby extracting meaning-
ful value. A clinical management plan can be implemented, creating
a favourable clinical outcome, in preventing the envisaged medical
’condition’ from occurring [Porter and Teisberg, 2008].
The development of an algorithm, in extracting and translating

meaningful clinical value, from other specialists’ clinical notes, im-
proves the medical teams ability to easily understand the original
file notes. The process followed is: sentence detection, word tokeni-
sation, section identification, contextual information and concept. Fi-
nally, the text is identified and translated from a designated dictio-
nary of terms whereby the text is extracted, generated, and then trans-
lated as shown in Figure 23.
It effectively creates user friendly plain text from complex med-

ical reports, from different specialities. The related keywords, from
sequelae of signs and symptoms, were then categorised and isolated,
as per the condition established by the medical disciplinary team. In
addition, a bodily location rule, is presented as an algorithm output,
described in Figure 22 and adapted from [Iroju and Olaleke, 2015].
Figure 22: The process of deciphering clinical notes into plain text.
8.4 results
The end result of the solution included a simple NLP translator. It

makes reference to a clinical dictionary, compiled with relevant med-
ical terminology. It then processes relevant differing reports, from
other specialists.
This algorithm made use of commonly accepted NLP rules, being

applied to medical notes. A medical dictionary, relevant to the spe-
ciality or designated field of study, was compiled. The terms were
tagged and the clinical information was extracted from this dictio-
nary and presented in the Clinical Note Translator (CNT), as a simple
comprehension of the medical event or terminology involved, shown
in Figure 23.
8.5 discussion 155
Figure 23: Clinical notes translator(the deciphering of complex medical

notes notes into readable text).
8.5 discussion
NLP within the healthcare domain undergoes a staged process where

the computer system will comprehend a piece of unstructured med-
ical text written in English, or any other language. It then strives to
present the same text into a structured set of medical terms having
followed a set of levels of processes and rules, as per Figure 22. The
levels are as follows:
1. Phonological analysis - applies mostly to speech analysis,
2. Morphological analysis - refers to chunking of sentence compo-

nents,
3. Lexical analysis - includes the tokenisation of certain structures

of words,
4. Syntactic analysis - the parsing, or syntax analysis in deriving

a meaning of basic text and data structures thereby conforming
to grammar rules,
5. Semantic analysis - relating to the syntactic structures from clin-

ical phrases, clauses and sentences, in addition to paragraphs
as a whole, to their meanings, and
6. Pragmatic analysis - is where the actual meaning of a sentence

is extracted and finally established.
The presentation of clinical text or medical notes poses many chal-

lenges for NLP in general. These are often written notes, bullet points,
acronyms and telegraphic phrases, making them highly ambiguous.

The Unified Medical Language System (UMLS) Metahesaurus was
developed, as a repository of biomedical vocabularies, to resolve or
disambiguate clinical text [Townsend, 2013].
Certain studies show that the presence of abbreviated ambigu-

ity can be as high as 54.3% [Townsend, 2013]. Word disambiguation
poses a challenge to NLP, and yet can also be circumvented, by the
synthesis of clinical speech to clinical text, which further adds value
and efficiency to NLP [Townsend, 2013].
NLP is considered to be a specific discipline of Artificial Intel-

ligence (AI), striving for human-like performance. AI mainly com-
prises of three main devices or tools called Machine Language (ML)
techniques (for structured data), deep learning techniques and NLP
techniques - for unstructured data [Chopra et al., 2013; Jiang et al.,
2017].
The unstructured data comprises of clinical medical journals, med-

ical reports and even speech. The objective of the NLP procedures is
to produce machine readable structured data, which are then anal-
ysed by Machine Language (ML) techniques. The main categories of
AI research, within medical domains that utilise NLP, are cancers,
neurology and cardiovascular diseases [Chopra et al., 2013; Jiang et
al., 2017].
The primary forms of NLP systems or systems aiding NLP, are

Information Retrieval (IR) and Machine Translation (MT), with the
more recently Machine Language (ML) tools available. These are be-
coming a focal point of the Fourth Industrial Revolution (4iR) [Cawsey
and Jones, 1997; Collobert et al., 2011].
The application of NLP within the South African healthcare con-

text will attract further attention, due to the South African govern-
ments alignment with the eHealth strategy. This includes the arrival
of the National Health Insurance (NHI) plan and the Fourth Indus-
trial Revolution (4iR) [Harrison, 2010; Haywood, 2016; Nat, 2012]. The
digitalisation of the healthcare industry is ever present and will revo-
lutionise healthcare across private and public healthcare [Neveol and
Zweigenbaum, 2015; Pang et al., 2018; Reddy and Sharma, 2016].
In addition to tagging or coding, the system attaches or tags a

specific operational structured coding language, extensively used in
the healthcare landscape. The medical diagnostic coding system is
referred to as the International Coding Diagnoses (ICD 10), coding
system for diagnoses. The SNOMED Chapter 2 coding system, is a
coding system for medical procedures, performed in a clinic or hos-
pital [Benson, 2013; Hinchley, 2007; Mead, 2006].
8.5 discussion 157
Both systems offer a large degree of standardisation of terminolo-

gies and interpretations, within modern healthcare systems. They are
available in many other global languages [Kaplan and Porter, 2017].
Most NLP systems today produce standard terminologies and out-
puts, using large scale vocabularies integration. These add to the
many reference points of dictionaries and classifications of conditions
or pathologies [Friedman, 1999].
The specific tasks, within NLP, will depend on the tool and obvi-
ously the subject content, containing a few low and high-level tasks.
These are Part-Of-Speech tagging (POS), chunking, Named Entity
Recognition (NER) and Semantic Role Labelling (SRL) [Collobert et
al., 2011; Nadkarni et al., 2011]:
1. Part-Of-Speech tagging (POS) - is a syntactic function whereby

the sentence word is labelled a noun, verb or object etc., which
are commonly referred to as POS classifiers trained on win-
dows of text. These are then fed to a bi-directional decoding
algorithm which will tag text, including combinations of words,
often reaching a 97% accuracy rate [Collobert et al., 2011],
2. Chunking is called shallow parsing which determines labelling

segments of a sentence, being a verb or noun phrase (VP or NP).
Only one tag is assigned to each word which is often encoded
as a begin: chunk (B-NP) or inside- chunk (I-NP) [Collobert et
al., 2011],
3. Name Entity Recognition (NER) labels atomic elements within

a sentence into categories being a person or location. Both NER
and chunking are based on the CoNLL2003 setup, which is a
benchmark data set according and aligned to Reuters data, and
4. Semantic Role Labelling (SRL) strives to give a semantic role to

a syntactic feature of the sentence, which is essentially based on
the arguments of verbs [Collobert et al., 2011].
A recent system has managed to diagnose 55 paediatric medical

conditions, far better than junior interns. These systems are now man-
aging and handling a far broader aspect of scope and input of clinical
medicine [Carrell et al., 2017]. The earlier system was the IBM Watson
system, a medical question and answer system developed by IBM
that created a global sensation, outpacing all the competitors, on a
live show and winning a Jeopardy contest on US television [Nadkarni
et al., 2011].
NLP should be adopted widely, it will need to be adapted to a

variety of diverse clinical care environments, Electronic Healthcare
Record(eHR) systems, reporting styles and geographic regions with
differing policies [Carrell et al., 2017; Liddy, 2001]. These adaptations

are critical in achieving consistent results from the NLP system.
8.5.1 Current NLP challenges
The automated medical diagnosis domain, largely based on NLP, is

proving difficult in being accepted by the medical profession, based
on the following challenges:
1. legal liability: the healthcare service provider and not the soft-
ware application, remains responsible for the diagnosis of the
patient [Nadkarni et al., 2011]. This remains the most contentious
issues of data ownership and responsibility documenting the ac-
tual diagnosis or medical information being published,
2. reference-content reliability: how is the reliability of informa-

tion justified and objective? [Nadkarni et al., 2011]. The question
remains if the author of a reference to reliable information, is
not displaying any conflict of interest at any given time, which
is difficult to manage or police,
3. unstructured language text: the limited role of NLP when it

comes to making a diagnosis, based on unstructured text, where
it was traditionally based on structured data sets, in basic med-
ical domains [Nadkarni et al., 2011]. The new NLP system must
understand the corpora of text and terminology within a spe-
cific domain to analyse it effectively,
4. adaptation of an NLP system: the most underestimated entity

of adapting a system to a specific medical landscape or do-
main. It requires substantial time and is often overlooked, when
working a project of time and material as a cost factor [Carrell
et al., 2017]. For instance, the psychiatric language domain is
completely different to the Podiatric medical science language
domain, which will require extensive adaptation or learning
within the system,
5. multi-site adaptation of the NLP system: will require local ex-

pertise and time to understand and incorporate the local poli-
cies and systems which will impact the clinical documentation [Car-
rell et al., 2017],
6. volumes of system idiosyncrasies: the sheer volume of changes,

software modifications and testing cycles needed in the domain,
can be time consuming and overwhelming. Further, this will re-
quire software engineering best practices which will include a
modular design with intelligent algorithmic development strate-
gies [Carrell et al., 2017],
8.5 discussion 159
7. disparate data sources: vast amounts of data are coming from

disparate systems within the clinical setting which must be adapted
to manage the volume and differing data sources, feeding the
NLP system,
8. incompatible languages and abbreviations: hence multiple vo-

cabularies can and do refer to the same diagnosis or pathol-
ogy concept. It includes many varying abbreviations within the
medical language which proves difficult for reading medical
texts [Iroju and Olaleke, 2015],
9. negation of clinical texts: negation can be explicit, or not, when

stating that a condition is present or absent such as lungs that,
show no signs of fluid, which means that ’lungs that are clear
on auscultation’ ie. the word "lungs" is absent of abnormal lung
sounds. Negations can be semantically ambiguous [Iroju and
Olaleke, 2015],
10. incompatible vocabularies in medicine: the expression of diverse

names to mean the same concept. It led the Unified Medical
Language System (UMLS) to control these diverse meanings,
which is now the accepted dictionary, or ontology of biomedical
concepts [Dinov, 2016; Iroju and Olaleke, 2015]. The UMLS cur-
rently, has more than 200 English vocabularies which covers ap-
proximately three million medical terms. The medical language
is riddled with differing grammars and semantics,
11. System Nomenclature of Medicine - Clinical Terms (SNOMED-

CT): which was developed by Snomed International, is owned
and governed by the International Health Terminology Stan-
dards Development Organisation [Benson, 2013]. They cover
most disease findings, procedures, diagnoses and pharmaceuti-
cals in a set of 300 000 codes of medical terms [Iroju and Olaleke,
2015; Sou, 2016],
12. Logical Observation Identifier Names and Codes (LOINC): LOINC

is a clinical terminology system used for clinical laboratory re-
sults, which supports clinical research globally. LOINC codes
and names are used universally in Electronic Health Records [Iroju
and Olaleke, 2015],
13. the World Health Organisation (WHO): sets of classification of

codes which are based on global WHO standards, and com-
mon practice for all participating countries to develop and im-
plement, within their respective healthcare systems,
14. International Classification of Function (ICF), is a broad view of

body function, health and disability,
15. International Classification of Diagnoses (ICD 10/11): univer-

sally legislated for all healthcare systems which must adopt and
adhere to it [NDOH, 2014a; Nat, 2012],
16. International Classification of Health Interventions (ICHI)- this

is to provide a standard tool for reporting and analysing health
interventions in order to statistically report on. These are inter-
ventions performed by various members of a multidisciplinary
team such as doctors, surgeons, physiotherapists, nurses, health-
care workers, podiatrists and diabetologists, amongst others in
the medical team [NDOH, 2014a; Nat, 2012], and
17. missing metadata layer: traditional mainstream NLP systems,

such as Google and Yahoo search systems, are built on a layer
of metadata being developed over the years which is clearly
missing within the medical ecosystem of most modernised clin-
ical environments. Electronic Health Record systems have not
reached that level of maturity, nor are they as advanced, as the
Google metadata layering is inadequate [Dinov, 2016; Trivedi,
2018].
While the US government has already spent$29 billion on eHR,

the only remedy is time, and further medical ecosystem meta-
data strategy development, for the slow and arduous uptake of
eHR [Kleynhans, 2011a]. eHR currently, can only offer about 10
% of required data, for big data analysis and mining, which is
not nearly the required percentage to perform AI interventions,
such as NLP. However, it remains a sound starting point in
building the meta data layer, for any healthcare environment [Di-
nov, 2016; Trivedi, 2018].
One will need to focus on training these NLP models within the
actual domain specific medical specialty when building an NLP sys-
tem, as they differ considerably. Off the shelf NLP systems do not
work on healthcare text. Spark NLP for Healthcare, is being devel-
oped and adapted for most healthcare knowledge domains. The sys-
tem is open source and has a deep learning framework embedded in
the library, which is an extension of Spark ML. The IBM Watson is
the gold standard and the pioneer in the NLP field, which includes
both ML and NLP modules. Recently Watson joined up with Quest
Diagnostics which offers the AI genetic diagnostic analysis [Jiang et
al., 2017].
8.6 conclusion
The healthcare landscape is currently embracing the capabilities of

NLP systems. Primarily, this is to analyse large sets of unstructured
clinical narratives and records, in order to gain access to a vast bank

of clinical insights, to aid the doctors and providers, in their treat-
ment plans for their respective patients. However, not all have the ex-
pert skills in adapting and dealing with these systems. These systems
include clinical domains which present a multitude of challenges,
in transforming and preparing these unstructured data sets, for ad-
vanced NLP analysis. Clinical text, narratives, reports, scanned data,
demographic and biomedical data are voluminous and complex, orig-
inating from varied sources. The data is often stored away in medical
files and lies dormant, where many valuable insights need to be ex-
tracted, through a simple method, such as an algorithm. NLP, offers a
solution. It was presented in the form of an algorithm where complex
clinical notes can easily be analysed, tagged, sorted and decrypted
into simple user friendly language, for any medical team member to
clearly understand. It should include the patients understanding of
the decrypted medical notes by the newly designed NLP algorithm.
Finally, the ultimate objective is to produce legible, and user friendly
clinical notes. This chapter presented a possible solution in the de-
velopment of an algorithm in aiding the process of deciphering clini-
cal notes, from various medical sub specialities, into readable clinical
notes, for all to utilise.
In this chapter we have presented the Clinical Note Translator (CNT),

as a simple solution with input from varied sources of clinical notes
and narratives. They are tagged and referenced in the clinical trans-
lator, where it creates an output of a re-engineered simple natural
language, that is understood by all members of the medical team. Ob-
viously the clinical notes or reports of the output file is filled with
definitions and explanations. These translations would need further
refining, in order to achieve the fluidity of language. As the system
development matures, over time, it will further refine the text out-
put into more simple language, by the respective medical team. In
the next chapter we present the emerging technology of robotics in
healthcare.
Part IV
R O B O T I C S A N D M O B I L E A P P L I C AT I O N S
The applications of 4iR technologies are presented. These

AI technologies are applied: the anonymisation of patient
data, the decryption of clinical notes, disease trend analy-
sis, applied robotics and a medical chatbot, in support of
a medical assessment of a diabetic limb condition.
Further, we present the value of telemedicine, telesurgery

and digital healthcare as one of the cornerstones of the
Fourth Industrial Revolution (4iR) within healthcare. We
presented two solutions. The first was the use of robotic
technology in healthcare, by way of a modern application
of AI, in a medical Moonboot. It is geared for the early
detection of the devastating effects of a Diabetic Foot Syn-
drome (DFS), in preventing lower leg amputations. The
second technology we presented, was the applied set of
rules within a medical chatbot questionnaire, aimed at in-
forming the general public in establishing a cursory risk
profile, for the prevention of a possible, or imminent, Dia-
betic Foot Syndrome (DFS).
The section presents applications of digital healthcare, namely,

Robotics and Chatbots. Chapter 9 discusses the Robotics
in Healthcare, and Chapter 10 discusses The MedBot.
9 R O B O T I C S I N H E A LT H C A R E : T H E R A B O O T
I
ntroducing new technologies into mainstream medicine, surgery
and nursing care appear to be the current forerunners with the
advancing fourth industrial revolution. However, telemedicine
and telesurgery have been around for well over thirty years. Robotics
has now become a far more viable undertaking than ever before, due
to modern technologies, which include the digitalisation of health-
care [Mesko, 2017].
9.1 introduction
The convergence of certain disciplines of science such as engineering,

computer science and electronics is constantly being thrust upon us,
which calls for celebration, or concern. No industry will be further
affected than the healthcare industry. It will experience the biggest
impact, ever witnessed with the development and proliferation of
robotic engineering in healthcare. It will encompass the development
of robotic devices and sensory monitors, and keeping track of the
data flow from these sensors in a consistent and accurate manner
[Karandikar and Tayade, 2014; Reddy and Sharma, 2016].
Medical robotic technology presents an exciting phase of modern

medicine, considering the widespread shortages of medical person-
nel. The shortage of skilled supportive nursing and assisted living
care, is a global problem Figure 24. Robots can be used to aid peo-
ple with physical, sensory, and cognitive impairments [Riek, 2017].
Robots are essentially embodied systems that create big changes in
the physical world of a patient, by engaging a multitude of sensor
data in order to make a decision.
Medical robots are evolving from a myriad of autonomous ap-

plications or systems, ranging from tele-operated platforms to au-
tonomously controlled medical robots [Payne and Yang, 2014]. The
uptake of medical robots is slow and cumbersome, due to the high
capital costs involved [Gkegkes et al., 2017]. However, hand controlled
robots, directed by surgeons, are coming to the forefront of modern
surgery, due to the advancement of AI software technology [Gkegkes
et al., 2017; Payne and Yang, 2014]. Artificial Intelligence (AI) is the
common platform, closely related to robotic technology, with a strong
164
collaborative dependence on AI technologies from robotic surgery

and, nursing care to medical sensor systems monitoring a patient [Schreuder
and Verheijen, 2009; Wiederhold, 2017].
Robotic surgery (RS) is attracting interest within the field of medicine,

as it aids and supports minimal invasive surgical procedures, such as
organ transplantation [Randell et al., 2014]. The surgical profession
has witnessed some profound advances, over the last two decades,
with the advancement of key-hole surgery and minimal invasive tech-
niques that are based on new technologies. Clinicians and policymak-
ers are now addressing robotic surgery. Surgical teams processes, will
become increasingly efficient within their respective surgical units [Ran-
dell et al., 2014; Riek, 2017].
Given that many institutions still depend on traditional surgical

and nursing procedures and processes, nursing robots are gaining
special attention due to the valuable assistance, for both patient and
nurses, in a clinical or nursing environment. A big proportion of nurs-
ing is the primary point of care, for the activities of daily living (ADL),
which are the mundane tasks of nursing. It is in alignment with the ex-
panding elderly population which is placing more demand on health-
care workers, globally [Karandikar and Tayade, 2014; Kumar, 2018].
Through the application of basic AI technologies in aiding a patients
daily living activities, or informing them through sensor monitoring,
many benefits are gained which ultimately add to the quality of the
patient’s life.
Figure 24: The Robo Nurse aiding patients in Italy. Source: www.pri.org [Ku-
mar, 2018; Riek, 2017].
.
166 robotics in healthcare: theraboot
The chapter endeavours to highlight the distinct advances, bene-

fits, risks and ethical issues involved in healthcare robotics. Further,
an applied AI design solution, is proposed for a medical condition
such as a diabetic foot. The diabetic foot has the potential to be fa-
tal if ignored, resulting from sepsis originating in a simple pressure
ulcer of the foot. Applied AI solutions within a clinical situation can
monitor, detect, inform, and predict these medical risks to the patient,
such as the sensor and monitoring technology. The chapter further in-
cludes the inherent benefits and challenges with the digital advance-
ment of medicine and patient care, in general.
9.2 robotics in healthcare
9.2.1 Medical robotics
The definition and perception of a medical robot is largely a mis-

understood concept. The term robot originated, or derived from the
word ’Robota’, meaning ’compulsory behaviour’. It performs the func-
tions of a human being, without any feeling or emotion Figure 25.
It can mean many functions such as the lifting or moving a heavy
patient: bringing meals or other devices; patient interaction with fa-
cial recognition; surgical procedures by a remote surgeon through
telemedicine and the administering of medication to patients [Karandikar
and Tayade, 2014; Kumar, 2018]. The benefits are numerous, consid-
ering the ageing populations of most countries where pensioner pop-
ulations are increasing, at a rate faster than the formalised worker
population [Herselman et al., 2016; Schwab, 2017]
Robotic surgery (RS) originated in 1985. The first robotic surgery

was performed by a robotic arm, called Puma 560 that conducted
non-laparoscopic neurological biopsies. Twenty years later, in 2002,
the first transatlantic telesurgery was performed between New York
and France, over an ATM network. The procedure was a cholecystec-
tomy [Cazac and Radu, 2014]. The surgical procedure proved to be a
success, thus marking it as the first of its kind [Giulianotti et al., 2003;
Martin, 2016]. Robots can be fully autonomous or fully tele-operated
with affiliated systems, such as sensors and algorithms for the pro-
cessing of sensor data [Riek, 2017].
The advancement of robotic software applications, in clinical prac-

tice, is evident within the proliferation of medical literature [Karandikar
and Tayade, 2014; Riek, 2017]. Robotic surgery raises ethical and moral
questions across the healthcare spectrum. However, this technology is
already in full implementation and gaining wide popularity consid-
ering the global shortage of healthcare and nursing skills. Robotic
9.2 robotics in healthcare 167
Figure 25: The Robo Nurse aiding patients in Italy. Source: www.pri.org [Ku-
mar, 2018; Riek, 2017].
.
technology Figure 24 is being widely accepted and enjoyed by many

physicians, nurses, patients and global industry leaders alike [Riek,
2017].
Further, there is a groundswell of knowledge regarding the topic

of value-based outcomes medicine, being a strategic re-directive for
healthcare, called ’outcomes-based medicine’. This is centred on the
application of quality-based medicine, ultimately serving the outcome
of improved basic human function [Porter et al., 2016; Porter, 2010a].
It means attaining the normal basic human functions, as fast as pos-
sible, at minimal cost and sustainable over longer periods of time.
Medical robots are going to fill this gap in the nursing care business,
shown in Figure 25, by aiding and supporting patients in gaining nor-
mal human function, for every day living activities [Pang et al., 2018;
Riek, 2017].
Currently, 20% of the world population experience difficulty in

their daily living activities, which varies from the short term to the
long term, or temporary. The said population is reported to be in the
region of 190 million that are suffering from this impairment [Riek,
2017]. This impairment can relate to any daily activity from combing
hair, moving around, feeding themselves, cognitive functions, house-
keeping to more advanced activities such as finance or problem solv-
ing [Riek, 2017].
Medical robots have the ability to learn and adapt to their respec-
tive environments, including the mannerisms of the patient, through
their own AI sensors of patterns. Thus, providing further physical
support and reducing the cost of labour, compared to a real nurse
performing the very same menial tasks for a patients daily living
demands. The technologies essentially become the new descriptors,
within a care setting. Robotics can aid caregivers, patients and clini-
cians alike, in a similar care setting with varying degrees of disrup-
tion and impacts [Karandikar and Tayade, 2014; Pang et al., 2018; Riek,
2017].
Robots have the ability to deliver the daily tasks of care-giving

which present in the form of the three D’s (dirty, dull and danger-
ous), which are of interest to clinical staff. The use of robotics presents
in the form of Direct Robotic User(DRU), termed as primary benefi-
ciaries Figure 25. Clinicians and caregivers are the secondary bene-
ficiaries where the tertiary beneficiaries are policymakers and other
groups garnering an indirect benefit [Riek, 2017]. Moreover, there are
supportive technologies within robotics which comprise of the de-
cision support tools, management and charting. These are software
engineering technologies based in Artificial Intelligence (AI) [Kumar,
2018].
Robotic technology can bridge care-giving gaps in modern health-

care and aid healthcare workers, and patients alike which includes
tele-medicine and robotic surgery amongst others. These recent robotic
advances are testament to this progress and all relate to health and
wellness. However, there remains a lack of clinical value trials, ac-
tually showing the benefits derived from robotics compared to stan-
dard treatment protocols of mainstream conservative medicine [Riek,
2017].
Medical robots are placed in three main categories: inside the

body, outside the body and on the body. Inside the body - consists
mainly of the micro robotics and surgical intervention robotics which
can perform a range of functions, such as localised therapy and pre-
cise surgical interventions, such as biopsies and tissue removal [Ku-
mar, 2018; Riek, 2017].
On the body - are the wearable robotics that mainly comprise

wearable robotics in the form of prostheses, orthoses and exoskele-
tons, such as replacing a leg that enhances the function of such a limb,
in addition to enabling the function. These can be powered knee and
ankle prostheses [Riek, 2017].
Exoskeletons have been particularly useful for application in paral-

ysis and muscular disorders. Outside the body - are being used in
primarily clinical settings with highly infectious diseases. These are
mobile manipulators that aid in surgical procedures [Riek, 2017].
Future robotics is the emergence and fusion of clinical and me-

chanical engineering, with computer science, when designing these
future robots [Karandikar and Tayade, 2014]. Whilst medical robotics
9.3 robotic surgery 169
are AI, their applications bring exponential changes within health-

care. These vast changes are markedly present in those environments,
that minimise human risk in high risk environments [Karandikar and
Tayade, 2014; Martin, 2016].
The clinical settings for telerobotic technology can vary from toxic
areas to defusing bombs and performing tele-surgery, on high risk
infected patients. However, the uptake remains slow and arduous,
with hurdles such as ethical and moral barriers, coupled with high
risk to humans. No matter the evolution of the technology, robotic
engineering is gaining traction in the medical healthcare world.
9.3 robotic surgery
The last twenty years have produced the most radical changes within
surgery. During the 90’s the large abdominal cut in general surgery,
gave way to specific laparoscopic interventions. Those techniques then
extended to cancer surgery. The surgeon was able to undertake small
incisions and perform key-hole surgery. Cameras were inserted, thereby
eliminating major trauma to the abdomen. This kind of surgery was
limiting and challenging for any surgeon, which hindered its uptake
in mainstream surgery. Nevertheless, key-hole surgery drastically in-
creased the benefits, whilst improving the return to normal human
function [Randell et al., 2014].
Robotic surgery serves to lessen these restrictions within laparo-

scopic techniques whereby robotic arms are remotely controlled by
the surgeon. The robot has full control of the arms and camera, through
the surgeons remote guidance, who now enjoys more freedom of
movement and superior visuals. The surgery is performed through
the camera with a three dimensional view on a separate platform Fig-
ure 26 [Randell et al., 2014]. Therefore the surgeon exercises superior
skill, through precision and control, which is otherwise, difficult in
traditional laparoscopy.
9.3.1 Minimally Invasive Surgery (MIS)
MIS was introduced to circumvent laparoscopic surgery. Various small

surgical interventions are carried out through a minor cut, or inci-
sion in the abdominal area. Surgical instruments and specialist cam-
eras are inserted. Those thin tubes are referred to as trocars which
serve as gateways for instruments to pass through. MIS offers signif-
icant decrease in trauma to the body, with far less risk of infection
and improved recovery, with minimal scarring [Avgousti et al., 2016;
Levi Sandri et al., 2017].
The robotic arms involved are managed by the medical expert

through a master controller, or manipulator, a foot control and hand
movements. The robotic arm can hold surgical tools such as a forceps
or scissor. The second arm holds a scope providing a visual of the
surgical field. The third arm holds the high frequency scalpel knife
for cutting tissue or tumors in order to minimise bleeding [Avgousti
et al., 2016; Levi Sandri et al., 2017].
An important feature of telerobotic surgical systems being used

for MIS, is the use of haptic interfaces, which are implemented on
robotic systems called the Surgeons Operating Force Feedback Inter-
face Eindhoven (SOFIE). This system is based on a master-slave con-
trol system architecture Figure 28, with the slave being a robotic arm
frame consisting of three independent manipulators (two for surgical
tools and the other one for the scope). In certain cases of MIS proce-
dures, only a single incision is made, for all the instruments to pass
through and is called a Single Port Access (SPA) surgery [Avgousti et
al., 2016; Levi Sandri et al., 2017].
Specialised systems have been developed to circumvent the pres-

ence of a surgical sister, by the development of the integration of sev-
eral semi automatic telerobotic surgical systems that integrate with
each other, called the Trauma Pot. Originally developed for the bat-
tlefield which consists of a Da Vinci telesurgical robot, a scrub nurse,
a dispensing system and an automated surgical instrument rack Fig-
ure 26. The surgeon has three arms at his disposal where one arm
holds the endoscope, and the other two arms manipulate the instru-
ments accordingly. In addition there is an X-ray machine for diagnosis
and a 2D fluoroscope to support most procedures, which essentially
eliminates the need for a scrub nurse [Avgousti et al., 2016; Levi San-
dri et al., 2017].
General surgery and long range telerobotic systems include three

main surgical sites namely: the doctor’s or master slave site, the pa-
tient recipient site of surgical intervention, and a vast network or plat-
form interconnecting these sites. The patient site will have two manip-
ulators. The surgeon site has a laptop running the surgeons graphical
interface software, with two surgical devices (Phantom Omni devices)
and a USB foot pedal and a video feed [Avgousti et al., 2016; Levi San-
dri et al., 2017].
The system can operate a laparoscopic arm with approximately

four instrument arms, directly attached to the operating table. The
slave robot is remotely controlled, across the network with feedback
mechanisms to the master-slave interfaces, at the surgeon’s side Fig-
ure 28. It is the basic layout of the telerobotic surgical device system,
applied across all major surgical specialties, from neurosurgery, to
general and orthopaedic surgery [Avgousti et al., 2016].
Figure 26: The TeleRobotic daVinci surgical system. Source:

www.ResearchGate.com and daVinci surgical systems [Av-
gousti et al., 2016; Levi Sandri et al., 2017].
..
9.3.2 Telerobotic systems
The largest impact is in minimal invasive telesurgery, now displacing

laparoscopic surgery of the old order. These are teleoperated robotic
systems allowing the doctor to perform surgical procedures, diag-
noses and treatments over vast distances. Thus, utilising a wired net-
work or a wireless network, depending on the geographic regional
infrastructure. Telemedicine eliminates the need for the doctor to be
present in areas, where such specialist skills are not available, thus
eliminating many risks for the doctor whilst bringing much needed
medical skill to a region or village [Avgousti et al., 2016; Levi Sandri
et al., 2017].
Robotic systems entered the medical landscape in the mid 80’s

and have undergone an evolutionary path creating major impacts
within neurosurgery, orthopaedic surgery and general surgery. They
have not evolved without their challenges, but telerobotic systems
will continue to make significant changes in clinical practice [Av-
gousti et al., 2016].
The mechanics of such a system are where the remote manipulator

delivering the surgery is controlled by the operator or doctor. The
doctor is referred to as the master, and the manipulator system is
referred to as the slave system Figure 27. It is a typical ’master-slave’
system where a sensory feedback channel is utilised in the form of
position commands, visuals and other sensory feedback information
exchanged between the master and slave systems.
Figure 27: The Telerobotic surgical team. Courtesy reddit.com [Avgousti et

al., 2016].
.
The tele-operated robot performs precision surgical techniques in

closed-loop environments, where the surgeon(remote operator), sits
in proximity, or remotely, undertaking the surgery or clinical proce-
dure, being performed in another town or country [Avgousti et al.,
2016; Levi Sandri et al., 2017].
The local site that includes the human operator site, or the master-
slave, consists of all the components needed such as keyboards, mon-
itors, joysticks and input and output devices. The remote site called
the recipient or slave site, consists of the robotic manipulation system
with the patient in a clinical setting and clinicians present. When a
surgical intervention is applied it is commonly referred to as telesurgery
[Avgousti et al., 2016; Levi Sandri et al., 2017].
The concept of telepresence means the surgeon or ’master’ opera-

tor requires the information feedback of the remote clinical environ-
ment to be identical to his or her presence within that clinical environ-
ment. However this presence diminishes with distance variances and
communications networks supporting such systems Figure 26. Obvi-
ously local area networks with fibre connections have superior and
stable bandwidth, and are more reliable than a wireless connection.
It can further be improved with a Local Area Network (LAN) system,
such as the Da Vinci system, operated over a short distance by using
a dedicated fibre optic cable [Avgousti et al., 2016].
The controls and network support remain a big factor in the relia-
bility of any teleoperated system. Network nodes may have differing
policies and throughput, which have an impact on the buffering and
queuing of data packets on networks. Regardless, bandwidth relia-

bility and speed have improved in leaps and bounds over the last
decade, where latency periods in digital networks have improved
vastly, critical for the reliability in telerobotic and telesurgery sys-
tems [Avgousti et al., 2016; Levi Sandri et al., 2017].
The data flows or traffic is of importance as its the "blood" of the

bilateral telerobotic systems, in differing categories, namely:
1. real-time control data - bidirectional data flow of the control

system of both the master and slave system,
2. medical video stream - transmitted from slave to master and

therefore the clinical capacity of the streaming medical video is
not compromised at all within the network, and a
3. high-level management data - covers the data from slave to mas-

ter which sets the sampling frequency which requires a stable
data flow connection, such as TCP/IP [Avgousti et al., 2016].
There are various kinds of data involved in such robotic systems:
1. synchronisation flags - bidirectional byte packets transmitted

asynchronously, to synchronise the events between master oper-
ator and slave, at the clinical site of treatment,
2. robot control data - control data flow is bidirectional, which

pinpoints set points between master and slave, at the patient
clinical side,
3. video conference data - data throughput for video channels, pro-

vided the video protocols are correct and of the correct technol-
ogy, and
4. ultrasound video - data transmitted from the master to the re-

cipient robot site.
It is a multi-modal system and application where the two remote sites

are connected via a TCP connection for data and image transmission.
There is an interest and acceptance of telerobotic technology amongst

healthcare practitioners, in general, and most are planning to imple-
ment such systems in the near future. This includes robotic technol-
ogy for treatment plans to performing telesurgery within their own
clinical practices Figure 28.
The benefits of robotic surgery to the surgeon are numerous and

include: improved surgical precision with less tremor, improved ma-
noeuvrability in a surgical field, precise theatre timing regarding fund-
ing, reduced tissue damage, less risk of travelling, increased level of
health with better research and the acquisition of new knowledge and
surgical skills.
Figure 28: The TeleRobotic daVinci surgical system. Source:

www.ResearchGate.com and daVinci surgical systems [Av-
gousti et al., 2016; Iroju and Olaleke, 2015].
.
The benefits to the patient are as impressive; with national cover-

age, access to a top class surgeon, no long distance travelling, faster
recovery times and reduced costs for the patient [Cazac and Radu,
2014].
Telerobotic surgery has developed so rapidly in the last decade

that it is now becoming prominent within organ transplant surgery.
It is proving to bring many benefits to the surgeon and patient alike,
even bringing down costs with reduced post operative pain and hos-
pital recovery time [Levi Sandri et al., 2017]. Knowledge bases are
now proving that the benefits of telerobotic surgery far outweigh the
costs involved, especially when one scales it up to cover larger remote
communities [Cazac and Radu, 2014].
9.4 internet of things (iot)
Recently, the services of the internet have produced advanced biomed-

ical sensors and sensing technology, that enables the physical moni-
toring of a patient’s condition or progress [Pang et al., 2018]. IoT is
seen as an ecosystem that integrates, hardware, software, devices, ob-
jects and people over an integrated network which enables all these
objects to collate the data in the monitoring of the patient’s condition
or injury [Cabestany et al., 2018]. The objective is to enable a smart
monitoring mechanism with advanced treatment decisions and even
options for the doctors and patient.
These sensing and monitoring devices are becoming highly cus-

tomised and personalised for the patient, considering that they are
9.5 the theraboot 175
sleek, natural looking, unencumbered, long lasting, non visible, hands

free and networked in their natural environments. These sensing de-
vices are often embedded in clothing with direct contact with the
person’s body or embedded into a wearable system or device worn
by the patient, such as clothing or a shoe, in monitoring the patients
condition [Armstrong et al., 2005; Jegede et al., 2015].
The IoT monitoring devices can monitor conditions at the site of

insertion, such as pressure, humidity, and temperature. For instance
the placing of such sensing devices in the innersole of a shoe-insert,
has the ability to monitor unwarranted changes of the foot, within
the confines of a shoe worn by a diabetic patient. The changes or
sensing data is detected and is streamed from the transponder via
Bluetooth wireless technology, to a data analytical platform or appli-
cation, where it is monitored by a device with built-in alerts or simply
crude warning signals. The essence of such a mechanism, is early de-
tection of such foot conditions or pathologies, originating from that
hot spot of the patient’s foot. It is achieved by strategically placing
the sensors in the critical pressure areas of the insert for feedback
and monitoring [Armstrong et al., 2005; Jegede et al., 2015].
Early detection is critical in preventing foot infections, which could

lead to a possible surgical intervention, or even amputation [Arm-
strong et al., 2005; Jegede et al., 2015].
9.5 the theraboot
A MoonBoot is a well constructed robust motion controlled walking

plastic boot, which is a below-knee therapeutic boot, or removeable
plastic cast, indicated for the stabilisation of various foot and an-
kle injuries ranging from fractures, ankle sprains, tendon and liga-
ment tears to complex pathologies, presenting in the foot and ankle.
The MoonBoot is a well structured supportive structure to aid ambu-
lation whilst being injured as shown in Figure 29 [Armstrong et al.,
2005; Zhang et al., 2013].
The MoonBoot is primarily used for foot and ankle fractures, by

offloading weight, with no inherent technology or embedded intelli-
gence, as yet. Having researched and inquired extensively across the
industry, no sensing or monitoring devices are found in any of the
commercially available MoonBoot models, to date, besides a mechani-
cal pump device providing inflated cushioning.
The causal reasons for the specific use of a MoonBoot Figure 29

is specifically indicated for trauma, such as a ’Jones Fracture’- a fifth
metatarsal fracture. This is due to the fact that Jones fractures are
slow to heal, with high degrees of non-union. Other modalities of
Figure 29: The standard MoonBoot. Source: www.Google-scholar.com and

Amazon.com. [Armstrong et al., 2017, 2005].
.
treatments include strapping, tubigrip, plaster casting and hard shoes,

with varying degrees of recovery. Stress fractures of the fifth and the
second metatarsal, are also justified pathologies for MoonBoot man-
agement, proven to be effective as a treatment regimen [Armstrong et
al., 2005; Katz et al., 2005].
There is evidence of custom made MoonBoots indicated for badly

deformed and fractured ankles that require custom casting and de-
sign, termed a ’Rapid Prototype’, as a design for such a customised
MoonBoot [Diegel et al., 2006]. Essentially, it would involve the con-
struction of a customised newly constructed boot that takes the form
of the deformed foot or ankle joint. The traditional method of treat-
ment was the old plaster casting method, with a rocker-bottom stop-
per or sole under the foot, which had been the mainstay regimen
for decades. The MoonBoot or Removable Cast Walker (RCW), is now
widely utilised for the aforementioned traumatologies and foot con-
ditions [Armstrong et al., 2017, 2005; Keenan et al., 2013].
Moreover, the existence of a Smart Insole devised and developed

by a team in Canada, has adapted such technology into a smart shoe
only. They devised the smart shoe, by applying smart sensors in an
insole. It is then inserted into the shoe, to monitor the status of a dia-
betic foot condition. Essentially it monitors the pressure, temperature
and humidity only, in order to prevent further complications leading
to infection and amputation [Jegede et al., 2015]. The technology will
be further enhanced and adapted, producing an advanced MoonBoot,
described in this chapter.
Figure 30: The LAN layout as presented for the Smart Insole. Source [Jegede
et al., 2015].
The plantar sensors are placed in the ’hot spots’ of the foot, where
a total of eight ’hot spots’ were isolated. The designated hot spots
were one sensor in the heel area, two sensors in the midline of the
midfoot, three sensors placed across the first, third and fifth metatarsal
head where a sensor was finally placed on the first and third toe. This
is covering all the vital pressure zones of the plantar aspect of the hu-
man foot [Jegede et al., 2015].
Figure 31: The Smart Insole with adapted sensors. Source [Jegede et al.,
2015].
9.5.1 The diabetic smart boot- TheraBoot
The MoonBoot is widely used for the off-loading of plantar foot pres-
sures within a neuropathic diabetic foot, or anaesthetic foot, in or-
der to prevent foot callouses and ultimately ulceration. The reasons
for specifically addressing such a condition are manifold. Diabetes
is at pandemic level, globally, placing a major fiscal drain on health
budgets. In the USA amputations are at critically high levels, due to
diabetic foot conditions leading to ulcers and gangrenous feet. The
condition is equally rife in the UK and European countries, with the
developing countries, not far behind. The World Health Organisation
has prescribed globally accepted, evidence based interventions, for
these diabetic conditions, in addressing the prevention of diabetic re-
lated plantar foot ulcers [Armstrong et al., 2018; Zhang et al., 2013].
Medical science has widely published the causal factors in diabetic

foot amputation. It is the loss of sensation leading to a plantar callus
formation followed by a foot ulcer, if not treated. The loss of sensation
is called neuropathy, which leads to a neuropathic ulcer, described as
the precursor to any plantar foot ulceration [Razavian et al., 2015]. A
foot ulcer will inevitably lead to an infectious gangrenous foot, if left
untreated [Zhang et al., 2013]. The objective of the off-loading mech-
anism, is to disperse the pressure away from the point of the plantar
callous formation, the prevention of such a condition is key [Arm-
strong et al., 2018].
Medical, customised, off-loading footwear is indicated for the pri-

mary and secondary line of preventative footwear for plantar foot ul-
cerations. The MoonBoot is known as one of the ultimate off-loading
efficient mechanisms, currently available [Jegede et al., 2015; Katz et
al., 2005]. Moreover, it eliminates patient non adherence to a strict pro-
tocol of treatment by wearing such a boot permanently, only remov-
ing it when sleeping. Many clinics and hospitals do not always adopt
these treatment protocols, despite extensive evidence based interna-
tional guidelines, proving its efficacy [Jegede et al., 2015; Pickwell et

al., 2015].
The accepted design of the boot is ingenious and makes structural

logical sense from an engineering perspective - simple in application.
Moreover, one can enhance this structure with the simple application
of additional artificial intelligence technologies, through sensing and
monitoring technologies, within the whole contact area of the boot.
As described earlier, this was done by the Canadian team who

addressed sensor monitoring of the insole of the shoe, by applying
smart sensors, shown in Figure 32, in an insole, placed in the shoe. It
then monitors the status of the diabetic foot condition on a wireless
mechanism of biofeedback to a mobile application, in order to sense
and monitor the pressure of the prominent pressure areas of the foot.
This could ultimately lead to the prevention of further complications
that could, potentially lead to infection [Jegede et al., 2015].
In a similar fashion, additional sensors are placed in the boot cre-

ating a biofeedback mechanism in order to monitor and collate the
data in a structured manner. The monitored pressure inputs are ex-
tracted, from the patient’s foot pressure readings, through a wireless
Bluetooth to a central server for further assimilation and data analy-
sis. This was based on their original computer engineering approach,
by measuring the basic parameters or ’redflags’ for a diabetic foot
condition, namely: temperature, pressure and humidity [Armstrong
et al., 2017; Jegede et al., 2015].
These parameters were based on extensive research and findings

published by renowned specialists, in the field of diabetology. They
made extensive use of temperature readings, by self monitoring de-
vices, infrared monitoring devices and home based dermal thermom-
etry, in order to self detect injury [Armstrong et al., 2017, 2018; Jegede
et al., 2015].
9.5.2 The new therapeutic boot - TheraBoot
The advancement of the design of the innersole sensing device has

been further adapted and engineered into a medical MoonBoot. We
have put forth a design of this new sensing and monitoring design
with an added feature of therapy, through the zones of the new boot,
called the TheraBoot. The application of AI, in the normal MoonBoot,
is simple in structure, demarcated into domains, thus addressing a
specific anatomical domain of the foot and lower leg, appearing in
the figure Figure 32.
The envisaged design enhanced enhanced the MoonBoot - TheraBoot

- with the application of AI technologies, through sensors placed in
the specified zones of the boot accordingly. We would further isolate
pressure points, or hot spots, in the ankle area, namely the medial
and lateral malleoli, the anterior tibial area and posterior leg area.
Figure 32: The modified concept of a MoonBoot or TheraBoot with zones of

demarcated sensors.
The AI driving the sensors in the biofeedback system in the appli-

cation is enabled to monitor and trigger a mechanical devices in the
boot. The devices are triggered to flush any excess fluid, sequentially
from Zone 1 to Zone 2, to Zone 3, then to Zone 4. The limb fluid is
then pushed up and absorbed into the upper leg area. The patient is
relieved with the automated flushing of the pooled fluid, away from
the leg and foot. This then markedly reducing swelling in the leg,
preventing any further pooling of fluid and, thereby diverting focal

pressure from the pressure points on the lateral and medial malleoli
of the ankle space. It includes the flushing of the front of the shin
(anterior tibial area).
The sensor monitoring points in the lower limb will serve as feed-
back mechanisms in relaying critical data, regarding pressures and
temperatures, as presented in figure Figure 31. In addition to apply-
ing therapeutic applications such as mechanically induced massag-
ing or vibratory therapy - it facilitates increased blood flow and mild
flushing of oedematous fluid, from the most distal region, or Zone 1,
upwards against gravity, towards the upper leg and away from the
foot ankle area. It will follow a monitored sequential action, in the
boot by flushing and massaging the lower leg, preventing any stasis
ulcers, as shown in Figure 32.
The boot, appearing in Figure 32, has designated zones in cover-

ing the vital anatomical zones of the lower leg and foot. The zones
are depicted in the monitoring software with electronic sensors and
activators for each designated zone in the boot:
1. Zone 1. The forefoot, comprised of metatarsals and digits is

a critical component of the plantar weight-bearing area of the
foot.
2. Zone 2. The midfoot and ankle area, comprised mainly of the

talar and subtalar joints, adjoining the heel or calcaneal area.
It is a primary weight-bearing area, especially around the heel,
that commonly ulcerates in a diabetic foot.
3. Zone 3. The shin area, above the main ankle area where mostly
venous ulcerations occur, but critical to monitor for skin temper-
ature changes.
4. Zone 4. The area just below the knee and covering the upper
shin area and more importantly, the anterior and posterior tib-
ialis muscle area.
Each Zone will have an area of applied intelligence by virtue of

the inserted temperature, moisture and pressure sensors that monitor
and send messages, via a wireless channel to the master application
on a mobile device which then integrates to a central server similarly
shown in Figure 32. The applied sensors in the boot will be embedded
in the contact surface of the inner boot, and strategically placed at
prominent pressure bearing areas in order to monitor and create a
biofeedback mechanism. This feedback can be performed wirelessly
through a Bluetooth or similar communication technology.
An important feature is a physical or mechanical massaging mech-

anism, in combating swollen lower legs and ankles, common amongst
the elderly. The therapy activators, receive messages, to activate ac-

cording to the biofeedback mechanism. A specific programmed mode
of action will activate, such as a massaging therapy that would be
coordinated or harmonised upward from the forefoot Zone 1, thus
flushing and massaging the forefoot up the leg to Zone 4 shown
in Figure 32. The rate and frequency will be automated according
to the feedback, with reference to an algorithm. The feedback and ap-
plied intelligence feedback loops, would naturally incorporate a level
of therapy and be governed accordingly. A manual feature for set-
tings, will also be allowable by choice, with strict governance settings
applied.
This creates a highly effective and automated electronically en-

abled flushing of the lower leg. The indications can potentially, in-
clude conditions such as tendonitis, fasciitis, sesamoiditis and many
other acute and chronic inflammatory conditions. The care giver or
physiotherapist need not intervene or apply any therapy whatsoever,
besides managing the mechanism correctly. This is a true example of
applied AI, aiding the patient in an automated manner. It equates to
passive therapy, whilst the patient is resting or even sleeping.
The objective of sensing devices in the TheraBoot must be capa-

ble of collecting large data sets, measuring temperature, pressure and
moisture, for effective big data analysis. It is facilitated through smart
algorithms, in order to extract valuable insights, in addition to ther-
apy for the patient. These clinical insights are analysed at the point
of patient contact, on a mobile device application and at the central
server application platform, where more extensive data analysis can
be performed.
The sensing and monitoring technology architecture, is a simple

multi layered architecture consisting of four layers as shown in Fig-
ure 30:
1. First layer: sensory in the form of a patch or wearable.
2. Second layer: network layer.
3. Third layer: services layer - meeting patients requirements.
4. Fourth Layer: interface layer - interactions with other users or

applications.
9.5.3 Applying AI to other pressure injuries
Hospitalisation has its clinical merits, as well as many risks. One of

these risks is prolonged periods of bed rest to heal, from sickness
or injury. It can cause another type of chronic injury, called pressure
injuries. The old term is better known as bed sores, or pressure ulcers,
commonly referred to as PUs.
The latest terminology of such a condition is referred to as a pres-

sure injury. It commonly occurs when the pressure inhibits the cir-
culation on a prominent weight bearing area of the body, whilst the
patient is lying down or sitting for prolonged periods of time, with
a loss of pressure sensation. Another consideration as a contributing
risk factor in the development of such an injury, is moisture. It can
typically be urine or perspiration, that will increase the development
of a bed sore, and ultimately ulceration [Mulder, 2015].
It is believed that the accepted measure of capillary filling pres-

sure is 30 mm Hg reading, which is the golden standard. Any pres-
sure exceeding this capillary filling pressure will cause ischaemia,
and ultimately skin damage. The surface tissue cannot withstand the
pressure via the simple mechanism of superficial blood circulation
which then begins to degrade. The surface pressure ultimately causes
a pressure injury to the skin of the patient [Armstrong et al., 2018;
Fremmelevholm and Soegaard, 2019; Obagi et al., 2019].
The pressure injury problems arise when the patient is bedridden

for prolonged periods of time, in excess of three hours, not moving
at all. The circulation is hindered, and pressure sores develop around
the elbows, buttocks, heels and even back region. Pressure sores can
be complex to heal, and even prove to be fatal when infected. Hence,
prevention is key, through judicious nursing, with the application of
smart materials to alleviate the pressure on the body points, in con-
tact with the bed mattress or the hospital ambulatory chair. A pres-
sure injury will affect those patients that cannot move easily due to
spinal injuries, head injuries, being comatose and diabetics with se-
vere loss of pain sensitivity [Armstrong et al., 2018; Fremmelevholm
and Soegaard, 2019; Obagi et al., 2019].
A pressure injury or ulcer will develop over a period of time

through four stages:
1. Stage 1. the skin appears red and has a touch of warmth
2. Stage 2. discoloured skin with blistering and no blanching
3. Stage 3. some tissue damage to the dermis is visible and deep

set in nature
4. Stage 4. the tissue is damaged with oozing can be infected and

damage is deep.
9.5.3.1 The hospital bed pressure injury or ’bed sore’.
The method of treating a pressure injury is a whole clinical regimen

but for the purposes of this design the most important is the allevi-
ation of pressure on the affected area shown in Figure 33. Naturally,
there are other treatments, such as the debridement of the ulcer, clean-
ing the wound and finally applying dressings on a regular rotation.
This would be followed up by medication if needed [Armstrong et al.,
2018].
Figure 33: The padded hospital bed: note the pressure pads. Source:
www.Medicalexpo.com
The cost implications of the healing process are reported to be

in the region of $129 000 per patient, for a stage four-pressure ulcer
alone [Brem et al., 2010]. The costs are also high where the emphasis
is placed on the prevention of such a condition developing. It will
include, amongst others, applying splints, cushions and a pneumatic
mattress, on the patients bed which is the current modus of best prac-
tice [Fremmelevholm and Soegaard, 2019; Obagi et al., 2019].
The mainstream technology of treating a pressure injury, in a hos-

pital bed-ridden patient, is a pressure mapping system, found only
within a handful of high-tech hospitals in Europe and the USA. It is
a sophisticated mat filled with sensors and monitoring of pressure
capabilities (as shown in Figure 34). The readings are conveyed to
an interface, where the system translates the various pressure read-
ings to designated colours shown on the screen. One can clearly vi-
sualise the oranges and reds as dangerous pressure zones [Hanson et
al., 2009]. The mechanism is an exceptionally good graphical display,
but only of the pressure readings. It lacks the applied intelligence to
read the temperature and moisture content. Nevertheless, it does give
the car-giver an indication of the area of pressure, on the patient.
Figure 34: The MAP sensor mat for a hospital bed: A-Remote monitor de-
vice; B-Sensor mat under sheet; C-Monitoring of pressures; D-
Sensor Data. Source:https://www.XSensor.com hospital systems
.
The application of a smart mattress with built-in sensors or recep-

tors to create a clinical feedback mechanism, will prove its worth in
medical science. The current mattresses are of the most modern ma-
terial and, some are equipped with a pneumatic pump mechanism
to increase or decrease the pressure of the mattress. Unfortunately,

this has to be manually monitored and managed by the care-giver, as
shown in Figure 36.
The technologically advanced hospital beds shown in Figure 36

tend to only have an electronic pump which is manually or electron-
ically operated to a certain pressure, on assessment of the patient. It
is reliant on a responsible care-giver to exercise good judgement in a
reliable manner. Applied AI can effectively automate this monitoring
with precision.
If one could add sensors in monitoring the temperature, pressure

and moisture content, it could prove to be life changing in prevent-
ing the development of a stage three or four pressure ulcers. It can
be achieved by collecting the data over a period of time and alert-
ing the care-giver to any change in the status of the static patient as
shown in Figure 34. The simple application of AI through a sensory
biofeedback mechanism, to another monitor or coupled/linked to an
overhead ECG monitor with compatible interfaces, to integrate sim-
plistically.
The same concept of applying AI to a mattress on a hospital bed,

can be applied to the padded seat of a wheel chair as shown in Fig-
ure 35. It covers the buttocks and coccyx area of the wheel chair
bound patient. The seat would have sensors monitoring the pressure,
temperature and moisture content within the target area.
Figure 35: A wheelchairs padded seat - note the pressure cushion.

Source: https://www.drivemedical.com
.
The system would be interfaced via wireless technology to a mo-

bile application or warning device, held by the care-giver, giving a
9.6 medical ethics and medical robots 187
pre-emptive warning to alert to the duration and moisture content of

the patient seating area. It serves to give the necessary preemptive
preventative warning signs, for the prevention of a pressure injury, or
ulcer forming. The suggested design of this technology is similar to
that of the TheraBoot sensor monitoring technology. The common pa-
rameters of temperature, pressure and humidity are constantly moni-
tored and, the data collated, for pattern recognition of large data sets,
over long periods of time.
Figure 36: The smart hospital bed - note the smart pressure mat.
Source: https://www.Boditrak.com
9.6 medical ethics and medical robots
The moral dilemma of using robotic technology for telesurgery or

rehabilitative therapy in aiding the patient, will exist for years. There
is no one single defining moral theory per se’ [Pagallo, 2017]. The one
argument is that robots hold the key to medical personnel shortages.
Others argue that robots will replace healthcare workers. They see it
as immoral to allow human care to decline with robots [Levi Sandri
et al., 2017]. It will lead to a decline in responsibility of the robot
delivering the care. It can be selectively personal and subjective, to
a degree. It reflects back to the old debate of placing robots and AI
systems on a battlefield, or in our houses to cater for our daily tasks.
It is further applied to meta-ethics, applied ethics and moral theories.
Without philosophising further, it gravitates to the concepts of right
and wrong.
The robotic behaviour portrays an outcome of behaviour, which

should be based on morals and principles. Based on this premise, the
designer, being a human, must develop such principles in the medi-
cal robotics. The goal is to control the behavioural outcomes of such
a robot divinely and ethically, basically - robo-ethics. It essentially

combines the principles of law, ethics and technology, in meeting the
required behavioural outcomes of the medical robot. The rule of law
must and should always be applied in designing the robot, without
diving into the philosophy of criminal and civil law [Pagallo, 2017;
Weber, 2018].
In simplistic terms, the design stance will supersede any behavioural

"intentions" of the robot. The robot can never be charged accordingly,
as it does not know the law and cannot exhibit any crime of intent
or negligence for that matter. However, AI technology that drives the
robot, can be said to possess the ability to exhibit negligence, and/or,
the intent of a crime. The debate surrounding the ethics of medical
robots continues [Pagallo, 2017; Weber, 2018].
The only smart way is to create deregulated zones, or special Re-

search and Development (R and D) zones to circumvent these ’fuzzy-
zones’ within the legal system. It further gives more latitude towards
improving the behavioural outcomes of robotic development, based
on ethically moral behaviour at all times [Pagallo, 2017; Weber, 2018].
9.7 challenges
Considering all the benefits, with distinct and proven advances that
telerobotics and general robotics bring into the clinical healthcare sec-
tor, it certainly has its challenges and future opportunities, which
must be highlighted. A decade ago, one of the major challenges, was
the cost factor considering the capital costs of setting up a telerobotic
station, with all its encompassing technologies and network systems,
in building such a platform.
The cost has shrunk substantially due to the improved advances of

technology, including Moore’s Law of sliding scales. However, most
of the challenges remain, where medical science is unforgiving re-
garding ethical issues, risks, governance and compliance. However,
there is an ongoing struggle in gaining acceptance of telerobotic pro-
cedures versus conventional surgical procedures being performed by
a whole surgical team [Cazac and Radu, 2014; Weber, 2018].
Listed below are some of the major challenges lying ahead for the
full acceptance and adoption of medical robotics:
1. acceptability and usability - the challenge of usability and accep-

tance, remains the largest challenge, whereby the rate of aban-
donment of surgical robots remains high. A surgeon is a highly
educated person with little technological knowledge and skill,
who will abandon the telerobotic system if it fails to meet his
9.8 conclusion 189
or her expectations [Cazac and Radu, 2014]. The designs are

evolving in becoming more user-friendly and adaptable to any
environment where surgeons demand for less robotic complex-
ity, is gaining momentum [Kumar, 2018; Riek, 2017],
2. safety and reliability - this is always a sensitive topic aligned to

the moral acceptance, as a back drop, to how sustainably safe
and reliable a medical robot can actually be,
3. regulatory approvals - the approval process is tedious and eats

into the development cycle as there are no global standards of
approval, which makes it costly,
4. cost of outlay - the development costs are astronomical and re-

lates to strict rules of safety and reliability factors which have
to be considered,
5. interdisciplinary development approach - the development cy-

cle will require a high degree of collaboration and consultation
between doctors, engineers, and bio-ethicist’s,
6. human factors - the gathering of inputs, from all stakeholders,

from the human touch to care delivery, is a first for the whole
design team,
7. software tools - the pre-operative software tools to enable the

surgeon or operator to make informed decisions,
8. telecommunication networks - bandwidth and the reliability of

such networks remains a challenge, especially for long distance
tele-operations, and
9. physicians training - medical schools will need to re design pro-

grammes and courses to up-skill surgeons and doctors, in all
relevant technologies, with applied training and development
simulations [Kumar, 2018; Riek, 2017].
9.8 conclusion
Healthcare and medical robotics is one of the most ambitious forays

into medical science. The technology is breaking all the conventional
ethical rules, beliefs, fears, and resistance to adopt, yet it is bound
to bring the most benefit, to all the stakeholders in the clinical cycle.
These robots have been evolving and gaining momentum in accep-
tance and trust, from the patient, over the last few years. Many chal-
lenges lie ahead and are proving to be part of a difficult undertaking.
Robotics in healthcare is in the process of being integrated into mod-
ern healthcare, and society at large.
The application of healthcare monitoring and sensory AI and IoT

technologies, in healthcare devices is also developing with great strides.
The small applications of sensory monitoring devices and applica-
tions in an ambulatory MoonBoot, or a hospital mattress, can produce
significant changes in clinical outcomes. Early detection and preven-
tion of such pending pathologies can cut hospital costs, and shorten
hospital stays by months, giving a more positive clinical outcome of
such conditions.
Modern medical science is based on scientific principles such as

’evidence-based medicine’ and more evidence and research is needed
in order to bridge the gaps. Research and industrial communities
need to collaborate and share data thus laying down some firm evidence-
based cases for medical robotic interventions going forward. It goes
hand in hand with other emerging technologies such as medical big
data analytics and electronic Health Records (eHR’s), enriching the
case for more advanced artificial intelligence (AI), that will drive med-
ical robotic adoption, in the near future.
The designers of such robotic technology must closely consult

with clinicians, care-givers, medical ethicist’s, surgeons and doctors in
order to establish the finer detailed technical requirements that will al-
ways feed into the next level of robotic complexity and sophistication
of robotic technology. Robotic technology is evolving and developing
daily in tandem with a narrowing gap in robotic literacy, amongst
all healthcare professionals. Medical robotic technology is going to
prove to be a major game-changer in the near future.
In this chapter we have presented the evolution and the application

of medical robotics within healthcare, including its clinical and surgi-
cal applications and benefits. The medical robot has been described
as telerobotic surgery. It primarily took place for the first time, in
remote New York, with the surgeons sitting in France. Telerobotic
surgical systems are becoming common place, in modern surgical
units, despite the large capital costs. The daVinci robotic surgical sys-
tem, with its advancing technologies, is the most advanced surgical
robotic system in the world. The technology is gaining momentum
as bandwidth develops and broadens, as with the modern daVinci
system. However, public healthcare is finding it difficult to justify the
capital outlays.
Nursing robots will be playing a more prominent and important

role with care in assisted therapies due to the shortage of trained
skills in healthcare.
The application of AI technologies is also taking place in med-

ical devices, such as the MoonBoot, indicated for fractures and the
off-loading of body weight from the chronic diabetic foot, through
the wearing of such a device. This modified sensory and monitoring
mechanism, with the Internet of Things (IOT), was applied to the com-
mon MoonBoot. The design of the modified and enhanced MoonBoot is
called the TheraBoot.
The same technology has the potential to be adapted and ap-

plied to other medical devices such as hospital beds, mattresses and
wheelchairs, in the prevention of pressure ulcers or bed sores. These
injuries have debilitating effects on any bed ridden patient. A pro-
posed solution in monitoring and treating pressure sores, through the
applied sensors and AI, in these devices was described. Finally, the
challenges of robotic therapy and healthcare were detailed with the
future ethical challenges of the role of robots in healthcare discussed.
In the next chapter we present the application of mobile technology,
or mobile applications in the form of a chatbot.
10 M O B I L E A P P L I C AT I O N : H E A LT H C A R E C H AT B O T
S
ocial media is one of the most efficient ways to reach a large
group of targeted customers or patients, through various mo-
bile devices and applications. The mobile application has given
rise to a myriad of avenues where the patient and the medical care
worker actively interact, in finding a solution or treatment regimen, to
inform or educate the user regarding a medical inquiry [Nacinovich,
2011]. Moreover, it supports the feedback mechanism from a patient
or user, regarding a targeted healthcare campaign, or a selected pop-
ulation or group in gathering medical research data.
10.1 introduction
During recent years, the world has experienced an unprecedented

growth in the mobile cellular market [Botha et al., 2016; Chib, 2013].
Almost two thirds of the world’s population own a mobile phone.
Currently, the trend is set to expand from the 4.8 billion mobile sub-
scribers to approximately 6 billion in the year 2023. The sub-Saharan
region of Africa has a penetration of 50%. It has been speculated
that the average person in Africa will own a mobile phone before
they own a toilet. The current trend is undoubtedly proving this
fact [Botha et al., 2016; Chib, 2013].
Mobile health is commonly referred to as mHealth, termed by the

World Health Organisation (WHO), which is an umbrella term within
the healthcare sector being aided and supported by mobile devices,
monitoring, personal digital assistance’s (PDA) and any other wire-
less mobile devices [Botha et al., 2016]. It is a rapidly growing sec-
tor, as a subset of the SA National Department of Health (NDOH).
The eHealth strategy was outlined by the WHO and Gazetted by the
South African government [Coleman et al., 2011; Kotzé and Alberts,
2017; Nat, 2012].
The South African National Department of Health (NDOH) has

embarked on a strategic imperative to adopt and apply mobile tech-
nologies to the public healthcare sector in earnest. The strategic drive
is based in the national mHealth strategy as a sub-strategy of the
eHealth strategy [Herselman et al., 2016; Seebregts et al., 2018].
192
To clarify the definitions of mHealth, in the context of the eHealth

strategy, eHealth is the execution of healthcare services, through in-
formation and communication technology (ICT), where mHealth re-
mains a subset of eHealth, executing healthcare services, where the
solutions are designed based on mobile wireless technologies. The
two entities are complimentary and do not operate independently,
but collaboratively [Seebregts et al., 2018; HST, 2016].
The eHealth strategy, applied to all WHO approved governments,

has an mHealth strategic imperative and, South Africa is no different.
SA has always taken the lead in the region, regarding national health
systems (DHIS), whereby it boasts a national framework of strategic
outputs or foundations, leading the eHealth strategy for public health.
The flagship project of MomConnect has been successfully integrated
into maternal and child health services. The project’s, SMS-platform
facility, is used by more than 60% of pregnant women in SA, who
have easy access to antenatal services [Nat, 2012; Wolmarans et al.,
2014; WHO, 2013].
The penetration of mobile phones in Africa, and the world, has

created a shift in consumer demand for applications in order to en-
gage various services and technologies such as banking services and
social media to basic communication through social media. To this
end, mobile coverage is now offering extended services and engaging
consumers on a socioeconomic basis, in providing solutions through
effective communication alone, amongst others. Mobile access in the
healthcare sector has proven itself within the developed countries
with the proliferation of hundreds of mobile applications, across the
healthcare sector [Botha et al., 2016; Ouma et al., 2011; Nacinovich,
2011].
mHealth has many applications in the public and private health-

care domains which are: data collection, patient remote monitoring or
sensing, educational drives, disease tracking or monitoring of trends
in outbreaks, and treatment solutions support. Equally, there are many
applications of uses such as: professional development for service
providers, emergency response systems, telemedicine, and human re-
source management, amongst others [Botha et al., 2016; Ouma et al.,
2011]. It includes campaigns aimed at patients, supply chains, fun-
ders, management, and providers.
South Africa has a more advanced mobile communications plat-

form than other African countries. However, there remains remote
areas with little to no mobile infrastructure, which is needed in sup-
porting the mHealth strategy, currently being implemented. Structures
and frameworks are in place to breach this gap with policies and
frameworks for the privacy of patient data, including the urgency for
interoperability standards [Botha et al., 2016; Herselman et al., 2016].
194 mobile application: healthcare chatbot
While there are numerous challenges such as lack of management,

governance, standards, policies and communication, the breadth of
mobile penetration remains widespread [Herselman et al., 2016]. Therein
lies the opportunities for further development of mobile healthcare
services, such as the communications interactive platform, in the form
of a medical healthcare Chatbot. It is to serve the patient consumer
category, or the healthcare provider category [Brinkel et al., 2014;
Chib, 2013].
A healthcare Chatbot is understood within the healthcare indus-

try to be, a first line knowledge base, for any patient who is prepared
to converse with a mobile application. The chatbot reduces the con-
gestion and complexity of medical questions that healthcare organi-
sations and call centres must field, on a daily basis.
This chapter endeavours to highlight the distinct advances and

the derived benefits of the application of mobile and wireless tech-
nologies. It includes the risks and ethical issues involved that will be
discussed in the literature, on this advancing field of medical tech-
nology, with its inherent benefits for global medicine. It includes a
simple design of a medical Chatbot or MedBot, in addressing every-
day health queries from a concerned patient.
10.2 background to mobile applications
10.2.1 mHealth
According to numerous analysts, the saturation point for mobile phones

is approaching the global population, of 8 billion in 2021, with Africa
proving to be the most aggressive in reaching this target. Africa has
immense penetration of mobile phones, across the whole continent.
In addition, Africa owns a quarter of the world’s pathologies, which
makes it a potential market for any mobile health application [Brinkel
et al., 2014; Mesko, 2017].
Mobile interventions take on vast proportion, where the interven-

tion of technology can directly offer improved access, delivery, evi-
dence, and quality based healthcare outcomes. It can span from the
use of mobile phones in the collection of data, to communication with
the patient. It includes the delivery of care through wireless devices,
for real time monitoring of medication and compliance [Tomlinson et
al., 2013]. Despite the slow uptake in such interventions, they require
a robust platform of evidence based outcomes, needed in order to
create a wider uptake.
10.2 background to mobile applications 195
The magical appeal that mobile technology holds for the access to
medicine is immense. It overcomes land and geographical limitations
such as bad roads and forests, in low income areas. The promise of
bringing healthcare to the masses in the remotest of places is by far
the most appealing to any government, as it eradicates the dangers
of travelling. The access to specialist doctors, for remote areas, can be
immediately realised through mobile health technologies [Betjeman
et al., 2013; Brinkel et al., 2014; Tomlinson et al., 2013].
Technology must be used to its full potential within the public

domain or rural healthcare. The pressing issues in treating malaria,
HIV and Tuberculosis (TB) is now a global and local issue with the
population in remote areas. The World Health Organisation (WHO)
propagates the use of ICT in healthcare, to be utilised to great effect.
It includes eHealth strategies and technologies, where mobile health
is one facet of such an intervention [Wolmarans et al., 2014].
The use of mobile health or telehealth in developing countries is

now being propagated by the WHO, on a large scale. These solutions
can include mobile telemedicine, patient monitoring, disease trend
monitoring, clinic performance monitoring, drug dispensing tracking
technologies and many other applications of such services and tech-
nologies. This even reaches as far as drone technology, currently be-
ing introduced into many rural low income areas, with sparse access
to transportation and decent roads [Ouma et al., 2011; WHO, 2013].
Over the last forty years, great strides have been made to improve
healthcare amongst the poor and rural populations, globally. This is
propagated by the international bodies such as the United Nations
(UN) and the WHO. The concept of primary care was introduced
by the ’Alma Ata Agreement’ in 1978. It has been followed through
by the Millennium Development Goals (MDG) in providing macro
level output objectives, primarily focusing on the reduction of child
mortality, improving maternal health, combatting communicable dis-
eases and improving access to healthcare, amongst others. In addition
to these are the strategic objectives of eHealth and mHealth inter-
ventions, in achieving such goals [Coovadia et al., 2009; WHO, 2012,
2013].
One such classical example conducted by the South African health-

care department (NDOH), of an applied mhealth initiative, was the
MomConnect project. The objective of this initiative, was to connect
pregnant mothers with maternal and child health services. All com-
munication was conducted via SMS regarding clinic appointments
and information, related to the development of her baby, or its condi-
tion. Closely linked to this initiative was another project called Ncediso.
It is a clinical decision support mobile application, providing commu-
nity healthcare workers with relevant information [WHO, 2012]. Both
initiatives were based in rural areas, providing vital health informa-

tion at the point of care [Barron et al., 2016].
10.2.2 mHealth platforms
The benefits of mHealth will require a large degree of standards, frame-

works, technologies, networks with bandwidth, and platforms of de-
livery, under the guidance of effective governance and management
structures within a clinical environment, to simply facilitate and sup-
port the whole structure [Ouma et al., 2011; Ouma, 2013]. Technical
and medical skills will undoubtedly impose strain on labour require-
ments, as these areas are remote with scarce skills, and about to mi-
grate to the large cities [Barron et al., 2016].
The major challenges in adopting such mHealth interventions are

many, but the more common ones include scaling up and sustain-
ability. Many others include demand-side barriers like ignorance of
such knowledge and health services which, for example, constrains
a mother in seeking help. This can be overcome by short messaging
services (SMS) [Barron et al., 2016]. Another major challenge is the
awareness of such campaigns [Hampshire et al., 2015]. The SA Na-
tional Department of Health is aware of these constraints and chal-
lenges and is putting structures in place to address this further.
Mobile services have many challenges and require the infrastruc-

ture is to be stable and robust, with a high degree of reliability [Ouma
et al., 2011]. It makes no sense for sophisticated mHealth services, to
be, based on an unstable platform with little reliability. The appli-
cations can be directed at various actors, from the service provider
to the doctor, the patient, the hospital or nurse for educational cam-
paigns or messaging functions. The applications are manifold and rel-
atively isolated and underdeveloped within the South African public
healthcare sector [Brinkel et al., 2014; Chib, 2013; Gurman et al., 2012].
A simple solution for an applied mobile health application, to es-

tablish a value-add to the patient, is called the healthcare Chatbot,
or MedBot.
10.3 healthcare chatbots
A Chatbot is a relatively new concept, widely utilised in the private

insurance, financial and banking sector, to name a few. It is a novel
human-machine interaction underpinned by eHealth strategies and
applications. In particular, this human-machine interaction mecha-
nisms utilising a program, are designed and engineered to interact
10.3 healthcare chatbots 197
with a patients request or concern regarding a certain medical condi-

tion or ailment [Amato et al., 2017; Schwab, 2017].
The healthcare Chatbot is currently gaining focus in the industry,

although still under developed. According to many analysts, technol-
ogists, and medical specialists, the opportunities within healthcare
are endless. Medical and healthcare provider’s most meticulous de-
liverable, to any patient, are assurances and communication of their
treatment plans on a regular basis. Chatbot’s can be immensely help-
ful by virtue of relieving the pressure on medical staff for mundane
communications, medical management campaigns, vital information,
programmes and treatment plans, for the chronic patient. Chatbots
become virtual assistants or algorithm driven dialogues with patients,
driven and geared by AI rules [Mesko, 2017; Mechael, 2009].
Chatbots are becoming more sophisticated with time to the point

of detecting medical conditions through the conversational interface.
These are termed as ’smartbots’, that can assist with laboratory re-
sults presentations through other advanced means such as Natural
Language Programming (NLP) and the advancement of digital health,
critical for the development of eHealth. They have the ability to mon-
itor and sensor large medical data sets, from smart devices collecting
such big data [Bates, 2019; Mesko, 2017].
Technology corporates are developing large-scale platforms such

as; Babylon health, Ada health and Your.MD, which make them ac-
cessible to patients and healthcare. Chatbots are now designed and
built into Electronic Healthcare Records (eHR) systems, in assisting
practitioners with treatment plans, repeat medications and general
updates for the patient [Amato et al., 2017; Bates, 2019; Mesko, 2017].
In addition, the Health On-line (HOlMES) chatbot module, was

designed for patients to alleviate any machine bias in the system,
by understanding the natural language of the user. It makes use of
the IBM Watson Application Program Interfaces (API) purely by un-
derstanding the natural language which draws on more than four
terabytes of data. The decision making cluster produces the logical
decision making outcome on the ’Apache Spark Cluster’, executed
over the Databricks infrastructure [Amato et al., 2017].
We have proposed an interaction program in a real-world clini-

cal concept. The medical ChatBot, or MedBot, has been applied in a
medical decision support system having the goal of providing useful
recommendations and possible diagnoses, concerning several signs
and symptoms, all pointing to a specific medical condition, from a
myriad of signs and symptoms.
The MedBot will interact with the patient in asking a few simple
questions, on the state of their health. This will establish a medical
profile or status of the patient which will follow a few conversational

questions, in order to inform the patient of a suggested medical status
or treatment plan. The MedBot may not make medical diagnoses based
on their answers received, as this is in breach of the medical ethical
laws.
10.4 methodology
The method approach to designing the MedBot was simplistic in na-

ture. We followed the Design Science Research method (DSR). It con-
cerns the application of the design of something unique. It implies the
application of meaningful structure and order, through a logical pro-
cess. It is most relevant in the scientific and engineering fields which
can include building architecture, engineering sciences, town plan-
ning, management sciences and computer science. It involves solving
problems through creative innovations and practical solutions within
the sciences, through an applied research methodology, viewed as
a set of activities creating something completely novel, commonly
known as an "artefact" [Gacenga et al., 2012; Hevner et al., 2004].
It is a research methodology based upon an outcomes-based pro-

cess. The process includes the iterative and evaluation process within
the project, grounded on systematic, testable, and logical methods. It
forms the platform for solving defined problems through creative in-
novation. The process of DSR defines the ideas, rules, technicalities,
and the end product, wherein the analysis, design, implementation
and data system are most efficiently achieved [Gacenga et al., 2012;
Hevner et al., 2004].
10.5 the design of the medbot ver 3.0 - html
A common ChatBot is nothing more than a computer program, or

an artificial intelligence (AI) object, which conducts a conversation
via auditory or textual methods. They convincingly simulate how
a human would behave as a conversational partner. ChatBots are
used in dialogue systems for practical purposes, including the seek-
ing of information. They use complex natural language processing
systems, but many of them scan for keywords in the input, then they
build replies with closely matching keywords, from a database. AI
enabled ChatBots are about learning and improving from every inter-
action. They basically learn from every conversation your business is
engaged in. Your AI driven chatBot will improve over time, indepen-
dent of human interference [Chang, 2020; Denecke et al., 2019].
10.6 the implementation and results of the medbot 199
The design of the MedBot is based on a working algorithm which

follows basic principles of the platform. The editor aids in developing
a conversation through NLP constructs. The interactions are carefully
coordinated and designed in order to help patients in choosing the
most logical pathway, by answering a few simple questions, starting
from a general level of inquiry of the patients status, signs and symp-
toms. It means asking for a positive yes, or negative no, regarding the
specific questions, from a general level, up to a possible final diagno-
sis. The design of the MedBot has the inherent logic of the diagnostic
signs and symptoms, in the conversational algorithm in reaching an
outcome, for the patient.
10.6 the implementation and results of the medbot
The MedBot, Figure 37, is based on the true NLP concept of a MedBot,
which follows the course of a typical conversation in seeking the in-
sights of a patient’s signs and symptoms.
The process follows the flow of a conversation. It is as an exam-

ple of a patient inquiring on the symptoms about a typical migraine,
or chronic headache that the patient is experiencing. The MedBot is
following the flow of questions gaining insights, or appearing to be
humanlike in understanding the flow of symptom information.
The MedBot analysis is basic in illustrating the logic, and the con-
cept of the value it can add to a patient’s understanding of the asso-
ciated risks of such a condition, in the comfort of a remote setting on
a smartphone, without having to book formal medical appointments,
initially Figure 37. When the patient has such knowledge, it gives
him/her an informed understanding and the confidence to approach
a professional, in order to follow up on the management of a high
risk condition. This presents exercising preventative healthcare at its
highest level, through applied artificial intelligence with the MedBot.
Figure 37: The MedBot conversational screening desktop. (1/4)

10.6 the implementation and results of the medbot 201
This practically applies, drilling down into a specific medical con-

dition, or diagnosis such as diabetes, as an example. Diabetes Mellitus
presents and manifests further medical conditions, such as Diabetic
Foot Syndrome (DFS), amongst others. The aforementioned condition
will possess an array of signs and symptoms whereby the MedBot can
be further enhanced or streamlined, in order to field enquiries on the
condition and deliver educated advice back to the patient via an in-
terface.
Figure 41: The MedBot conversational screening on a smartphone. (1/6)
However, the diagnosis remains an indication only, and must not

be misconstrued as a final diagnosis. A final consultation to estab-
lish a definitive diagnosis of the condition, will be strongly recom-
mended [Amato et al., 2017]. We then further analysed the said exam-
ple as an extension to the basic MedBot depicted in Figure 41.
10.7 the logic of a medbot for a diabetic foot
The medical community has widely published the causal factors in

diabetic foot amputation, being loss of sensation, bad circulation and
chronic trauma causing corns and calluses 2. Leading to a potential
plantar foot ulcer, this can eventually degenerate further with sepsis,
possibly leading to amputation, unless it is treated aggressively in the
10.7 the logic of a medbot for a diabetic foot 203
early stages. The loss of sensation is called neuropathy, leading to a

neuropathic ulcer, described as the precursor to plantar foot ulcera-
tions [Armstrong et al., 2018]. A plantar foot ulcer can potentially lead
to a gangrenous foot, if left untreated [Pickwell et al., 2015; Zhang et
al., 2013]. The most critical complication of such a medical event is
septicaemia or death, if neglected.
Medically customised off-loading footwear and debriding the cal-

lus formation, are indicated for the primary and secondary line of
preventative medical care. It can be in the form of foot devices, such
as a MoonBoot, or a foot insert, which are off-loading efficient mecha-
nism for any abnormal forces in the foot and limb. The diabetic foot
condition is critical, but also, a financial drain on the patient and the
economic system, for any government’s public health. The crux of the
matter in treating a chronic foot condition is early detection, through
a few simple screening questions and monitoring devices. The clini-

cal answers from the patient can be of critical value in taking action
to prevent this condition, or only monitoring it in the early stages
of detection [Cavanagh and Bus, 2010; Jegede et al., 2015; Katz et al.,
2005].
The following chatbot questions are basic in nature directed at

the diabetic patient, in establishing the presentation of a diabetic foot
condition. This is a patient, who will show a level of risk, for such
a condition, through the interaction with a Chatbot, via a messaging
mechanism answering questions.
For the purposes of demonstrating the design and logic of the

conversational flow, the MedBot will detect the presence of a Diabetic
Foot Syndrome (DFS), or condition, through the use of a simple flow
of trained questions. The patient would have to follow this up by

seeking a medical professional opinion, regarding the risks associated
with a DFS, displayed by the Medbot, on a general level. The flow of
the conversational questions is the following:
1. Are you a Diabetic patient and receiving treatment for this con-
dition?
2. How long have you suffered from diabetes?
3. Do you suffer from a loss of sensation in your feet?
4. Have you ever been treated or tested by a medical professional

for insensate/loss of feeling in both feet?
5. Are you aware of any structural foot problem i.e. bunions, flat
feet, claw toes or skew toes?
6. Do you have any corns or calluses on your feet?
7. Have you ever received professional treatment for the corns/-

calluses?
8. Have you ever received corrective therapy i.e. foot inserts or

wedges?
9. Have you ever suffered from a previous foot ulcer?
10. Have you ever been hospitalised, or could not walk for a period
of time, due to this condition?
The Medbot’s interaction can only suggest the possible risks of

such condition developing. It can never establish a definitive diag-
nosis. This can only be done by a medical professional, through a
physical examination.
Obviously the MedBot is basic in nature, and would require further

refining in another study. However, it gives credence to an important
applied AI interface, to the patient in a basic understanding of his or
her foot condition, being a confirmed diabetic, type 1 or type 2.
10.8 evaluation and application of the medbot
10.8.1 The MedBot for diabetic foot syndrome
The design of the desktop version of MedBot was implemented with C-

Sharp (C#) programming language. MedBot was programmed to ask
the above questions. Scores were assigned to each question. The basic
ten questions will give an indication and inform the user of such a
condition. However, it is not a definitive diagnosis and therefore the
user or patient is strongly urged to seek further medical examination
and advice.
The application is a demonstration of the interaction between a

patient and an AI object, albeit a MedBot. It is as responsive as a real
person probing the answers, from the patient.
10.8.2 The evaluation of the MedBot
An evaluation was conducted in South Africa where the audience in-

cluded a focus group of experts, consisting of specialists and other
experts, in the field of treating diabetic feet. This demographic was
selected in considering them to be experts and futuristic thought lead-
ers in foreseeing novel ways towards a positive outcome. This is mak-
ing a valuable contribution to medical science in the management of
the diabetic patient, specifically related to the foot.
Most of the the experts strongly agreed on the design concept of

the MedBot, and its design definitely has the ability to guide and in-
form a patient, on the risks of having Diabetic Foot Syndrome (DFS).
In addition, they agreed that the concept of its design in the treatment
and management of a diabetic foot, would make a valuable contribu-
tion to medical science.
10.8.3 The application of the MedBot
The MedBot was further developed to serve as an interactive digital

assistant offering medical guidance, advice, referrals and, awareness
of risks, but not a diagnosis. The appeal of a MedBot is that it, tends
to attract the attention of the patient far more, from an interest per-
spective. MedBot was used to assess the risk profile, on the patient’s
foot condition. It follows the same question and answer application.
The initial screening question will immediately establish, if the

respondent is indeed within the correct parameters or category of a
10.9 conclusion 209
chronic diabetic patient, potentially at risk of a DFS. From here on

in, the questions will follow the same diagnostic logic. If the patient
shows a high risk, it suggests a possible high risk of developing into
DFS. The patient is instructed in the MedBot to seek a professional
medical opinion.
10.9 conclusion
The applications of mHealth, encompassing mobile technologies, are

so profound that they encourage further investment and research.
This chapter has explained the background to eHealth providing the
environment and platform to execute such mobile Health campaigns.
They enable the most valuable technologies, such as a cell phone, in
providing the basis to value-driven clinical outcomes, in remote ru-
ral areas. These interventions are no doubt the precursors to much
larger projects, as the technological platforms evolve, transforming
rural healthcare.
An example of a simple decision support mechanism, or inter-

active aid for a diabetic patient at a high risk, is a medical chatbot.
The MedBot, was designed and developed in order to illustrate the
practical application of such a simple AI tool, that can be utilised
by the most remote rural, or urban region. The MedBot, provides the
means to enriching and enlightening the patient of risks, specifically
relating to Diabetic Foot Syndrome (DFS), by means of a smartphone
application Figure 41.
A simple JAVA application and an HTML MedBot were designed

and presented to illustrate the value of such AI technologies. The MedBot
is ultimately a preventative tool, designed from AI software. The sim-
ple AI tool has the potential to save limbs, lives and unnecessary
cost for the patient, in addressing a common Diabetic Foot Syndrome
(DFS).
Many challenges lie ahead with interoperability of systems, stan-

dardisation of technologies and skills development of such technolo-
gies. These facets form the basis of the ultimate development of the
coming ‘Digital Healthcare 4.0’ era, aligned with the fourth industrial
revolution of healthcare, in South Africa.
In this chapter we presented specific benefits of mobile technology

and its impacts on health, especially the extended reach to certain
communities of patients in the most remote rural regions. The ben-
efits of mobile AI tools, are to be found in targeting specific medi-

cal conditions, or communities with designated campaigns. Notwith-
standing rural areas, many local suburban communities can be en-
gaged, such as chronic diabetic patients, showing signs of high-risk
factors. These risk factors can be easily detected through the simple
design of AI tools such as the MedBot interface.
The design of a simple MedBot interaction was analysed, designed

and presented, utilising basic program languages. A further design,
was applied with a specific set of conversational questions, regarding
a medical condition such as the risk profiling of a chronic diabetic
patient, showing signs of Diabetic Foot Syndrome (DFS). The tool
contains specific high level questions, targeting a designated category
of patients. It further highlights the derived risk factors, then provides
information to the patient through a clinical status assessment, before
it is too late.
The emphasis in the design is patient education by simply inform-

ing the patient, through an applied conversational interface. It discov-
ers an early risk detection of Diabetic Foot Syndrome (DFS). If this
condition is neglected, it can potentially result in foot ulcerations and
possibly amputation. The MedBot is ultimately a preventative design
of an AI software tool, saving limbs, lives and unnecessary cost for
the state and patient, in addressing Diabetic Foot Syndrome (DFS).
mHealth remains largely under-developed where the potential for

such development of interventions requires further research and de-
velopment. In addition, there remain a large deal of skills develop-
ment and technical expertise in executing these mobile health inter-
ventions. mHealth is largely driven by the eHealth policies and strate-
gies of all the WHO member states, globally. South Africa has en-
acted and promulgated its domestic eHealth strategies, according to
the WHO protocol being rolled out globally. It forms the platform
for mHealth services in the greater context of public healthcare. How-
ever, many challenges lie ahead with interoperability of systems, stan-
dardisation of technologies, and skills development of such technolo-
gies which remain the largest challenge for national healthcare in
South Africa.
In the next chapter we present the final evaluation of the presented

AI tools designs, being reviewed by specialists in the medical field.
.
Part V
E VA L U AT I O N A N D C O N C L U S I O N
This part contains Chapter 11 that presents evaluations

of the designs of the 4IR contributions of the thesis, dis-
cussed in Section 3.4 of Chapter 3, Chapter 6, Chapter 7,
Chapter 8, Chapter 9 and Chapter 10.
A survey questionnaire covering all five of the above con-

tributions, was reviewed by a focal group of healthcare
experts and specialist participants. They were resident at
South African universities, medical institutions, and health-
care organisations. The analysis showed that most of the
participants have a special interest in these subjects in
their professional capacities. They all emphasised the added
value from the contributions listed, through an electronic
survey. The detailed results are depicted in the graphs and
diagrams as referenced.
An evaluation of approximately 20 participants was con-

cluded, from a targeted pool of more than 40 specialist re-
spondents. The outcome, showed that the majority of the
participants was positively in agreement, and convinced
of the value being introduced by the proposed designs
of the digital healthcare tools. The 4IR tools are expected
to relieve healthcare workers of mundane tasks which ul-
timately creates improved value-based health outcomes.
Thus, the feasibility and acceptance of such designs of the
aforementioned healthcare AI tools, within the ambit of
4IR, were well received within a pool of healthcare spe-
cialists. The designs of tools were well received where they
felt that such designs would improve value-based health-
care outcomes.
Chapter 12 concludes the dissertation and presents future

work
E VA L U AT I O N
11
I
n this thesis, we presented tools for the following: the decryp-
tion of medical notes, medical chat-bots, disease trend analysis,
anonymisation of personal medical details, and the MoonBoot de-
sign.
The developed tools are design proposals, with no third-party in-

volvement of persons or subjects, besides the expert respondents in
the validation of these developed AI tools:
1. The Moonboot: https://tinyurl.com/y56txf4l
2. The MedBot: https://tinyurl.com/y5bmoxf3
3. Medical note decryption: https://tinyurl.com/y567f7v9
4. Anonymisation algorithm: https://forms.gle/Y5rezSw5N2C3dtzc8
5. Trend analysis: https://tinyurl.com/y23jezf9
This study combines all five designs of the AI healthcare tools,

as defined above. The evaluation includes both direct and indirect
evaluation of these designs through a final survey questionnaire. It is
directed at healthcare practitioners, specialists in their respective ar-
eas of healthcare, to share their insights and opinions on these digital
tool designs for healthcare.
11.1 moonboot
Practitioner perceptions
The chapter presents the results from a survey evaluation on a group

of practitioner’s understanding and their acceptance of the AI tech-
nology, as a monitoring and therapeutic tool, in a diabetic MoonBoot.
In addition, the new design of the digital MoonBoot, and the benefits
in aiding the prevention and management of a chronic diabetic foot
syndrome were also evaluated.
The online survey was conducted with a Google survey tool through
the University of Johannesburg in South Africa. Most of the respon-
dents are specialists in their fields of healthcare. The practitioner was
213
214 evaluation
a specialist in his/her field of expertise and registered with the SA

Health Professions Council (HPC). The inclusion criteria are further
detailed in the methodology section. A total of 20 responses were re-
ceived from a wide spectrum of healthcare specialists - all experts in
their respective fields of medicine.
From the survey, 45% of respondents were Podiatrists, 20% doc-

tors or surgeons, while the remaining 30% were therapists and nurses
of a specialist nature.
Considering that of the respondents, 95% have treated diabetics

in all facets of healthcare where the majority of 85% understood the
concept of the fourth industrial revolution, including its impending
impacts on healthcare. Surprisingly, only 15% admitted to not under-
standing the concept.
All the respondents agreed and believed in the benefits regarding

the concept of the AI MoonBoot design. It showed that 60% believed
that such a design would have benefit at the primary healthcare level
in the public healthcare sector. This aligns to the current analysis and
situation of the public healthcare sector, being incapacitated and un-
derfunded with little to no digital healthcare in rural areas.
From the survey, all the practitioners believed that the design of
the MoonBoot would aid a practitioner in the management of a chronic
diabetic foot and decreasing the risk profile of a diabetic patient, with
a chronic diabetic foot, requiring medical care and specialist treat-
ment.
Figure 47: Specialty groupings.
11.1.1 Practitioners suggestions
The responses received from these specialist practitioners, suggested

that the devices would have to be robust enough and with replaceable
components for Third World conditions. The massage mechanism
would have to be finely tuned in detecting infection as this mecha-
nism could spread it.
11.1 moonboot 215
(a) Previously treated diabetics.
(b) Understanding of 4iR concepts.
Figure 48: The potential benefits of such a design on the improved off-
loading of pressure points in the modified AI Moonboot.
(a) The Improved prevention of a dia- (b) The aiding of a diabetic patient.
betic foot syndrome.
(c) Understanding of 4iR concepts. (d) The potential benefits of this design
at the primary healthcare level.
Figure 49: Survey: opinions on the benefits of the MoonBoot/ TheraBoot

design.
It was not clear how often a person needs to wear the boot to be
beneficial. Depending on the everyday demands of the person and
216 evaluation
(a) The potential lessening of the risk

profile for a diabetic foot.
(b) Assisting a practitioner in treating a (c) Decreasing the swelling in a diabetic

diabetic foot. foot.
Figure 50: Survey: Expert opinions on the potential benefits of the Moon-
Boot/ TheraBoot design.
the environment in which he/she works, the moonboot can either be

a facilitator to perform all activities required of the patient, or it can
be a barrier. For example, in sedentary activities the moonboot will
be a facilitator of good health but construction type of activities will
be a barrier, the moonboot will be destroyed.
If the person only wears the moonboot at night, none of these

concerns are valid. The actual ’Gold Standard’ for off-loading is the
total contact cast, based on ’the rule of thirds’. The current Moonboot
is very useful in acute management of Charcot foot. An AI Moonboot
could be even better and possibly an added benefit or indication.
Evidence shows that the Moonboot is removed by the patient

when it should not. Availability and cost will always remain factors
in public and private health. One can apply a total contact cast in a
few minutes from cheap materials which are readily available. The
device shows good modality for off-loading areas where there are
neuropathic ulcerations.
In summary the above responses were positive, interesting, and

encouraging. The majority agreed on the derived value and benefits
the design will bring to the management of a diabetic foot. The sug-
gestions of added risks and benefits were noted. They need to be
considered in the sequence of events in the structure of the algorithm,
through the zones of the TheraBoot. The design was well received by
11.2 medbot 217
all practitioners and a positive response is clearly noted, with the ben-
efits proving to be of immense benefit in the prevention and treatment
of diabetic foot syndrome, currently prevalent amongst diabetics.
11.2 medbot
11.2.1 Practitioner perceptions
This section presents the results from a survey evaluation on practi-

tioner’s acceptance of the AI technology of a medical chatbot, referred
to as MedBot.
The online survey was conducted at the University of Johannes-

burg in South Africa. Most of the respondents are specialists in their
field of healthcare, and a total of 21 responses were received from a
wide spectrum of healthcare specialists.
Reflecting on the survey, 38.1% of the respondents were podia-

trists and, 19% were doctors, while the remaining 38.1% were thera-
pists and nurses of a specialist nature.
It was evident that 90.5% understood the concept of the medical

chatbot in all facets of healthcare, whilst 9.5% did not fully under-
stand the concept of the MedBot. It was interesting that 61.9% admit-
ted to interacting with a chatbot in other subject domains.
The MedBot concept was agreed to by 52.4% who believed in the

benefits in the concept of the MedBot design. A mere 23.8% disagreed
on its value as a valid digital tool within their practice, where 19%
remained neutral on the design concept of the MedBot.
The survey highlighted that 38.1% strongly believed that such a

design of the MedBot would aid and support a patient, albeit only by
means of information sharing and making basic bookings in general.
The survey showed that 57.1% agreed that the MedBot would be of
benefit at the primary healthcare level in the public sector, whereas
23.8% strongly agreed on the MedBot concept. The remainder were
neutral.
In addition 47.6% agreed, with 23.8% strongly agreeing, that a MedBot

definitely has the ability to guide and inform a patient on the risks of
having a Diabetic Foot Syndrome (DFS). The concept of the MedBot, in
the treatment and management of a diabetic foot was largely agreed
upon and well received.
218 evaluation
Most of the practitioners agreed that the design of the AI MedBot

will aid a practitioner and lessen the risk profile of a diabetic patient,
regarding the management of a diabetic foot condition.
(a) Specialty groupings. (b) Understands the concept of

the MedBot.
(c) The interaction with a MedBot. (d) The relevance and value of a MedBot
Design.
Figure 51: Survey: The potential benefits - design of the MedBot.
(a) The MedBot in aiding a patient. (b) The MedBot informing a diabetic pa-
tient.
(c) Understanding of 4iR concepts. (d) The potential benefits of such a de-
sign at the primary Healthcare level.
Figure 52: Survey: The potential benefits - design of the MedBot.

11.3 clinical note decryption algorithm 219
The outcome of the study survey further revealed that a few com-
ments from various practitioners regarding the MedBot, were varied.
The following comments were gleaned from the survey:
"From feeling utter frustration when the chatbot misinterprets an-

swers and sends one into a loop of missing the point. ChatBots are
only as good as the questions they are given to ask."
"The reliability of the software and hardware in general can pose

a challenge, certainly in the case of a life threatening scenario. In spite
of the complexity of many medical chatbots, they remain invaluable
to interact with a human interface. I temporarily live in the UK and
getting into the health system took two months through these "help-
ful" systems. No need to speak to people that is a basic need, even
more so when you have a health concern."
"I remain concerned about the lack of the human touch that plays
such an important role in patient management - could never replace
a medical foot examination."
11.3 clinical note decryption algorithm
The online survey presents the results from a survey evaluation on the
practitioner’s acceptance of the AI technology of a clinical notes, de-
cryption algorithm. It is simply converting medical practitioner notes,
clinical notes and reports into user friendly narratives, to be under-
stood by the whole medical team. Most of the respondents are spe-
cialists in healthcare and, a total of 21 responses were received from
a wide spectrum of healthcare specialists.
From the survey, 38.1% of the respondents were podiatrists and,

19% were doctors, while the remaining 38.1% were therapists and
nurses of a specialist nature. The remainder of the respondents were
chiropractors. The majority (95%), understood the merits and the con-
cept of a clinical note decryption algorithm, in order to decipher med-
ical reports.
Most of the practitioners - 57.1% - strongly agreed, while 33.3%

agreed in general that patients should be treated in a multi-clinical
environment with differing specialties, treating the patient holistically.
A total of 66.7% of practitioners are familiar with the concept of the
4iR, or digital health.
220 evaluation
A majority of, 99% agreed that other specialist’s medical notes

are not always easily understood. The practical use and application
of a clinical note algorithm design was agreed by 100%, who also
believed in the benefits of the design. All agreed on its value as a
valid potential digital tool, within their practice.
Collectively, 38.1% strongly agreed, in addition to 42.9% who agreed,

on the added value of this specific design of the clinical note al-
gorithm translator, which produced more clarity of patient medical
notes than the original specialist report. Again, over 90% agreed that
the clinical note algorithm would be of benefit, within their respective
clinical environments.
Most of practitioners did agree, in concept, that the design of this

digital tool (clinical note decryption algorithm) will aid a practitioner.
It will assist in their clinical practices in treating a patient, in collabo-
ration with other specialists.
(a) Specialty groupings. (b) Understands the concept of the NLP

decryption tool.
(c) The importance of internal collabora- (d) The application of the tool in a clinic.
tion.
Figure 53: The potential benefits - design of the medical notes decryption
tool.
The following comments were noted from the practitioners:
"Consensus between different professions may take time. I can see

this working well with routine tasks such as screening, assessment,
measuring outcomes but not so much for person-centred intervention
11.4 anonymisation of personal medical details algorithm 221
(a) The medical report is easier to under- (b) In considering the tool within their
stand with this tool. practice.
(c) Understanding of 4iR concepts. (d) In aiding a practitioner with the clar-
ity of others medical notes.
Figure 54: Survey: The potential benefits - design of the medical notes de-
cryption tool.
where there are too many variables. The personal touch and wanting
to be more human and less machine clinician might be lost."
"I think we often misunderstood somebody else’s notes because

of either their poor writing (just can’t read what it says) or unknown
abbreviations not commonly used by all disciplines. Using decryption
software may help but it may also misunderstand what is meant and
for safety in healthcare I think it is always better to rather go directly
to the other discipline and talk about what they meant. I think with
this program mistakes can come in which may be detrimental to the
patient."
"If the program will allow for the differences in the use of En-
glish by practitioners who are not English first language speakers.
Language is so much more than words. Could be a useful tool in
understating other medical specialists’ notes."
11.4 anonymisation of personal medical details algo-

rithm
The online survey presents the results from a survey evaluation on

the practitioners’ acceptance of the anonymisation algorithm, of the
222 evaluation
medical details of a patient, whilst applying analytical tools and in-

terventions. It is in alignment with compliancy rules regarding the
privacy of personal identifiable data of a patient file, or data sets, in a
clinical or research scenario. The algorithm endeavours to de-identify,
any personal medical detail related to a specific patient.
The medical algorithm simply de-identifies an individual patient,

within a data set of a patient’s records. Most of the respondents are
healthcare specialists and analysts. A total of 20 responses were re-
ceived from a wide spectrum of healthcare specialists.
The respondents consisted of the following: 40% of the respon-

dents were podiatrists, 20% were doctors, while 20% were therapists
and nurses of a specialist nature. The remainder of 20% were data spe-
cialists. The majority (95%) understood the merits and the concept of
an NLP, medical note algorithm, to decipher medical reports.
From the survey of practitioners, 57.1% strongly agreed with 33.3%

agreeing in general that patients should be treated in a multi-clinical
environment, with differing specialties by treating the patient holisti-
cally.
Further, the survey revealed that 99% collectively agreed that other
specialists medical notes, are not always easily understood. In addi-
tion 66.7% of practitioners are familiar with the concept of the 4iR or
Digital Health in itself.
The practical use and application of a clinical note algorithm de-

sign, were agreed on by 100%, who believed in the benefits of the
concept of the clinical note decryption tool design. All agreed on its
value as a valid potential digital tool within their practice.
Further, an observation of 38.1% strongly agreed, in addition to

42.9% that agreed, in the value of this particular design of the medi-
cal note algorithm. The consensus was that it would improve the clar-
ity of clinical medical notes. The survey further revealed that, 90%
agreed that the clinical note algorithm would be of benefit within
their respective clinical environments.
Most of the practitioners did agree on the concept of the design

of the clinical note decryption algorithm. This was in addition, to
aiding a practitioner in treating a patient, in collaboration with other
specialists, within their clinical practices.
11.4.2 Practitioners Suggestions
The following comments were noted from two respondents:

11.5 disease trend analysis 223
(a) Specialty groupings. (b) Understands the concept of the

anonymisation tool.
(c) The regular managing of patient de- (d) The relevance of this AI Tool in a
tails. clinic.
Figure 55: Survey: the potential benefits - design of the anonymisation of

patient details tool.
"Anonymisation of personal medical details is an important part

of medical data analysis. Clinical health systems are invariably ex-
posed to potential hacking, security remains a priority. The anonymi-
sation algorithm, will therefore add to the security and certainty that
personal data, remains confidential."
"An effectively designed store architecture, for the data layers

would be critical to ensure that de-anonymisation is contained, pre-
venting the anonymous data from being cross-referenced, with vari-
ant data sources, for re-identifying the anonymous data source. If not
mitigated, this risk is otherwise achievable through smart machine
learning and AI."
11.5 disease trend analysis
The online survey presents the results from a survey evaluation on

the practitioners understanding and the acceptance of the of a dis-
ease trend analysis tool, such as the Google Trends (GT) analysis tool.
It is primarily utilised in preventing and predicting a disease trend, or
pandemic. The GT tool, has been used extensively by the Centre for
Disease Control (CDC). It has proven to be more effective than con-
224 evaluation
(a) The aiding of this tool in predictive (b) In promoting the tool within their
analysis. practice.
(c) Understanding of 4iR concepts and (d) Aiding a practitioner by unidentify-

digital healthcare ing personal medical details.
(e) The benefits of the tool in predictive analysis.
Figure 56: Survey: the potential benefits - design of the anonymisation of

patient details tool.
ventional monitoring and prediction tools, in mainstream epidemi-

ological analysis. By monitoring and surveying internet traffic of a
category or specific disease trend, it can trace and monitor a disease
trend, with a lead time of approximately eight days.
the survey consisted of 20 respondents, from a wide spectrum of

healthcare specialists. Collectively the majority understood and ap-
proved of GT, as an effective monitoring tool, within their respective
clinical practices.
The survey, 36.8% of the respondents were podiatrists, 15.8% were

doctors, while the remaining 42.1% were therapists and nurses, of a
specialist nature. The remainder were chiropractors and homeopaths.
The majority (95%) understood the merits and the concept of disease
trend analysis, on the current Google platform.
11.5 disease trend analysis 225
Collectively, 80% agreed, on the relevance of such a GT tool in

gaining critical disease insights into medical conditions or disease
trends per region within South Africa. The remaining 20% disagreed,
or remained neutral in this observation.
Over 80% of the practitioners agreed that the GT tool would help
in the prediction of a medical outcome, while 85% agreed that such a
tool would be of benefit in their practices.
Most of the practitioners were in favour of the practical applica-

tion and use of a Google trend tool, as a highly effective surveillance
and monitoring tool. All agreed on its value as a potential valid tool,
within their practices.
11.5.2 Practitioners Suggestions
The respondents delivered negative and positive observations, on the

actual GT tool in a clinical context:
"Dr Google often gives incorrect information."
"Google "feed" you what they think you want to see, according
to your searches. I have googled many things related to patients and
then such a programme might think that it is related to me. In fact I
had an unpleasant experience due to such an assumption."
"I specialise and research hypnotherapy that is still in its early sci-
entific development stages. At this stage such a tool might rather add
to disinformation than to true scientific knowledge. It might change
in future, but people tend to think that they can provide therapy after
reading some information on Google."
"Trends can help different professions in understanding the trajec-

tories of treatment and gain in health as a result of treatment. It will
help with countering isolated decision making concerning treatment.
It will also prevent certain professions from believing they have the
only solution to the treatment plan. Other professions will become
more empowered to make equally valid treatment recommendations,
promoting the multi-disciplinary approach to treatment."
"In a real-time world, the availability of analyzed trends both lo-

cally and globally, can only be of insightful value, and a precursor for
preventative medicine practice."
"We have been doing analysis of data and trying to establish pre-
dictors of function but the data captured to date, is not tidy and often
too unreliable. AI can help us to get clean data, identify the data cap-
turers who are not reliable or need more training. The data can help
us to validate our tools across different populations and in different
226 evaluation
(a) Specialty groupings. (b) Understands the concept of trend

analysis on the Google platform
(GT).
(c) The assistance of the GT tool in pre- (d) The relevance of the GT Tool in gain-
ventative healthcare. ing disease insights per region in SA.
(e) The GT tool in predicting medical outcomes.
(f) Foreseeing further benefits to clinical

practice.
Figure 57: Survey: the potential medical benefits - disease trend analysis
tool.
settings. The benefits will be dependent on broad acceptance of the

tool."
Finally, in Table 2, the high level summary depicts a general ac-

ceptance and derived benefit of the aforementioned AI tools design,
within modern healthcare and digital medicine.
Table 2 reflects in the 451.2 positive (agreeable) responses against

the 48.8 neutral or disagreeing responses. It supports the concept of
the advent of the fourth industrial revolution in modern healthcare,
or rather, Healthcare 4.0. Thus, proving that these tools will bring
added clinical value with improved medical outcomes in the predic-
tion, monitoring and ultimately the prevention of unnecessary health-
care and expenditure.
Table 2: Practitioner perception of AI and digital tools in the 4iR

Practitioner Perception (%)
No AI Tool
Agree/Strongly Agree Disagree/Neutral
1 MoonBoot 100.0 0
2 MedBot Design 76.2 23.8
3 Medical Note Decryption 100 0
4 Anonymisation of Data 95.0 5.0
5 Disease Trend Analysis 80.0 20.0
Sum 451.2 48.8
In this chapter, we have analysed and evaluated the design, with its ef-
fects and impacts of the AI tools under discussion, within healthcare
settings, from 20 individual specialist respondents that participated.
The evaluation overwhelmingly showed that all five AI tool designs,
would be useful tools for healthcare. The designs of these tools have
proven to be well positioned, for the eHealth strategy, and ultimately
within the fourth industrial revolution, for South African healthcare.
In Chapter 12, we conclude this thesis and present future work.

12 CONCLUSION AND FUTURE WORK
12.1 conclusion
T
hroughout this thesis, we have given a detailed analysis of the
architectural landscape of the South African healthcare sector
with its deficiencies, gaps and areas for improvement, from
a technological and economic perspective. The Fourth Industrial Rev-
olution (4IR), or ’Digital Healthcare 4.0’, brings with it a sweeping
change in the operational technology of healthcare. It includes many
digital innovations, in the healthcare sector.
We have addressed these contributions through the application of

Artificial Intelligence (AI) tool designs, on the South African public
and private healthcare care sectors, within the context of the Fourth
Industrial Revolution (4IR), by presenting the following:
1. Architectural analysis of the South African healthcare landscape: we

have presented an in-depth view of the economics and dynam-
ics of the South African healthcare sector, with an understand-
ing of its current challenges, facing the strategic intent of eHealth
and the impacts of 4IR technologies.
2. Big data anonymisation algorithm: enabling the analysis of big

healthcare data sets, through an innovative algorithm anomymis-
ing personal medical data.
3. NLP decryption algorithm: the usefulness and effectiveness of de-

crypting, or translating, other specialist’s clinical notes into user
friendly medical notes.
4. Disease trend analysis: descriptions of how a search tool such

as Google Trends, can be applied in the healthcare field for
medical condition trends other, than flu trends.
5. Medical chatbot, MedBot: medical chatbots have now become

mainstream AI tools, to aid and assist practitioners and pa-
tient’s, in informing them of possible conditions, such as a high
risk foot condition, for a diabetic patient, as an example.
6. Robotics technology - TheraBoot: the application of sensors and

AI tools in a uniquely designed Moonboot, called the TheraBoot.
228
12.2 future work 229
All five designs were well accepted by a focus group of experts,

in support of 4IR technologies, in the South African healthcare
context. It signifies a positive uptake of AI technologies, in Dig-
ital Healthcare 4.0, for private and public healthcare. Moreover,
this supports and parallels the eHealth strategy, put forth by
the South African government, in adopting the digitalisation
of healthcare. The evaluation, overwhelmingly indicated that
the designs would be of benefit for healthcare and in particu-
lar their own clinical environments. The designs of these tools
have proven to be well positioned, for the eHealth strategy, and
ultimately in the 4IR, for South African healthcare.
However, there remains many gaps, and interventions are, needed

in the short and long-term, to overcome these obstacles in the
South African public and private healthcare context. Based on
the foundations needed for 4IR and eHealth infrastructure and
strategies, we have suggested some solutions to overcome these
obstacles. We have attempted to lay those critical foundations,
for such a platform, in the SA healthcare landscape.
12.2 future work
12.2.1 South African public health (NDOH)
The world of evolving global technology has reached unprecedented

levels across all industries. These could be exploited in the healthcare
sector, thus unlocking further value for the sector:
1. The SA government needs to utilise high end technologies such

as Cloud technology, for standardisation of technologies that
could be utilised as an outsourced option, with minimal infras-
tructure and capital needed,
2. broadband connectivity should be a basic human right, but it

remains too expensive and generally out of reach - especially in
the rural areas- it could be applied on a large scale for rapid
development,
3. internet platform 3.0 – new current modernised internet plat-

form consisting collectively of the SMAC technologies (Social
media, mobile technologies, analytics and Cloud), desperately
needs enhanced development for healthcare,
4. the effective use of ICT in improving data analytics and the

operational use of Governments healthcare supply chain,
230 conclusion and future work
5. the inter-connectivity of public health hospitals with centralised

databases, with good governance principles in place i.e., Cloud
technologies,
6. the need for fair peer-review systems in the public and private
healthcare sectors, collaborating on effective clinical outcomes,
7. the need to up-skill – the development of professional and ad-

ministrative staff on the latest technologies facing 4IR, in the
public healthcare setting,
8. big data analytics and modelling - Business Intelligence/Data

mining (BI), and
9. knowledge management strategies and frameworks for Enter-

prise Information Management systems (EIM).
Biomedical informatics can identify early indicators of chronic dis-

eases, enabling health workers and medical aid companies to begin
addressing healthcare needs, at early onset. Data analytics can ex-
plore healthcare transactional data, thereby identifying which mem-
bers are at risk, for a range of illnesses - including diabetes, asthma,
congestive heart failure and coronary artery disease.
These categories of diseases are all falling within the four main
burdens of disease, as an example. Data analysis identifies the most
relevant variables, including those high-risk patients, developing chronic
conditions. It can construct patterns, enabling early recognition of
such disease trends [Kotzé and Alberts, 2017; Mead, 2006; Perry, 1998].
Furthermore, within the SA public healthcare sector, the need for

effectively managing patient data, remains critical. It will further un-
lock value and knowledge, for the local populace, provided that the
following high level macro-initiatives within the public sector, takes
place:
1. Primary Health Care (PHC) must be implemented along ro-

bust and sustainable business architectures with measurable
outcomes - universal access to healthcare,
2. the need for public health to embrace technology to its full ad-
vantage, such as the standardisation of technologies across the
landscape. It includes data analytics and supply chain technolo-
gies, incorporating good operational governance over current
process [Gentry, 1996],
3. to seek guidance and consult with the private sector healthcare

leaders in developing standardised technologies to collaborate
on joint integrations with Public Private Partnerships (PPPs).
The idea is to leverage off theses sophisticated environments.
The focal areas of collaboration be in: pharmaceutical coding,
diagnostic related groups, diagnostic coding schema, and pro-

cedural coding schema,
4. standards of clinical content,
5. critical need in upgrading hospitals/colleges to acceptable stan-

dards - focused need on large projects or mega projects; across
SA for over 4200 hospitals and clinics at provincial, district and
primary level [Abbott and Ade-Ibijola, 2018],
6. increased use of technology to improve administration and op-

erational aspects of procurement in the national supply chain,
education and development of the human resource issue in pub-
lic health [Mars and Seebregts, 2008],
7. interconnectivity of public hospitals with encompassing networks,

ensuring the centralisation of the Relational Data Base Manage-
ment Systems (RDBMs) [Blumenfeld, 1997; Fayad, 1996],
8. establish ethical boundaries for risk-sharing with better analyt-

ics through established interconnectivity and big data analyt-
ics [Hastie et al., 2001; Frigo and Interviewee, 2017],
9. fair medical peer review systems to monitor standards in pri-

vate and public healthcare, through managing and monitoring
effective treatment outcomes,
10. the most critical need, is the effective development and training
of nursing and medical personnel across the board, in the public
healthcare sector,
11. nursing colleges to be upgraded and integrated,
12. doctor, intern, nursing shortages to be addressed, and
13. the review of the whole remuneration or grading system for

health [Archer, 2016; Harrison, 2010; Jeffery, 2016; Sonnier, 2016].
These macro-initiatives, remain critical factors for the South African

healthcare landscape, in addressing many ICT challenges. However,
these initiatives are partly being attempted, for the arrival of NHI and
the eHealth strategy. In addition, it encompasses the 4IR approach.
We have detailed the initiatives down to a micro-level basis, in de-

scribing the approaches needed and possible technological solutions
to address these critical issues on hand.
12.2.2 Big data in healthcare
The clinical value of big data, biomedical data and transformative

medical knowledge, lies in the feedback loops of such knowledge.
The implementation of these trained lessons, back into the design of
healthcare systems and protocols is key, to be of any valuable out-
come, in the modern healthcare landscape. The use of data anonymi-
sation in aiding healthcare analysis and analysts, will become main-
stream.
In future, when the healthcare industry matures, new skills devel-

opment programs, must be instituted which will address the medico
legal ethical issues regarding the privacy preservation of personal
health data. Anonymisation, or de-identification of personal medical
data, will become centre stage with big data analysis in healthcare. It
has yet to reach that stage.
12.2.3 Robotic technology and the TheraBoot
Healthcare and medical robotics are part of a giant leap for health-
care, and these robots have been evolving and gaining immense trust
and acceptance by the medical community and patient alike. It is the
ultimate in modernised digital healthcare, poised at the cutting edge
of technology. Nursing robots will be more prominent and play an
important role with nursing care, addressing the shortage of trained
skills in healthcare.
We designed and presented one small example of this modernised

technology in healthcare monitoring and sensory AI technologies, in
healthcare devices, such as the medical TheraBoot.
The applications of sensory monitoring devices and applications

in an ambulatory MoonBoot, have the ability to be adapted and ap-
plied across other medical devices, such as medical mattresses, wheel-
chair seats and, hospital bed mattresses in the prevention and treat-
ment of pressure ulcers. Early detection and prevention of such pend-
ing pathologies can reduce hospital costs and, shorten hospital stays
by months giving a more positive clinical outcome, of such condi-
tions.
Medical science is generally based on scientific principles such

as ’evidence-based medicine’. However, more evidence and research
are needed to bridge these gaps. Research and industrial communi-
ties, need to share data in laying down firm evidence-based cases for
medical robotic interventions. The architectures of such robotic tech-
nology must be closely aligned with clinicians, engineers and medi-
cal ethicists, in order to establish the detailed technical requirements,

feeding into the next level of robotic development. Robotic technology
is evolving and going to be a major player in the future of healthcare,
globally.
12.2.4 Trend analysis in healthcare
Google Trend is a tool in the public domain. The general population

can extract and derive meaningful insights and patterns, about the be-
haviour of the population, on certain topics or events, such as disease
epidemics.
However, further rigorous studies are needed to assess and evalu-

ate these causal inferences and the validity of the GT tool for surveil-
lance of disease trends and other medical conditions of a population.
It is, however, dependent on the densification of mobile technology
and ICT modernisation of such a population. Hence the coming of
4IR and digitalisation of communities, needs to rapidly develop and
expand accordingly.
12.2.5 Natural language processing - decryption of clinical notes
Translating clinical notes, or reports, into readable user friendly notes,

for the whole team of doctors and nurses, to simply understand has
always been a hindrance to good healthcare in the form of value-
based health outcomes. However, it poses a new set of technicalities
and challenges.
Clinical notes remain crucial to a positive medical outcome, or

prevention of a medical event. One needs to derive clinical insights,
suggest a solution, make a clinical decision, or offer a treatment plan
from the efficient decryption of clinical notes. It can include the pro-
cess being fully automated, with the aid of technology, such as the
reading of lung X-Rays. It then teaches the application to make a
definitive diagnosis. It is currently under development at Stanford
University, where software engineers have developed a diagnostic
tool covering over 200 diagnoses for lung conditions, through Ma-
chine Language (ML), as an example.
The opportunities for improving the technologies in support of

clinical note decryption, are endless. In future, we will explore the
concept of information extraction, from medical notes and reports,
through a mechanism of cleverly harmonising a set of AI and ML
algorithms, in order to facilitate such an effort.
Considering all the benefits, with distinct and proven advances

that telerobotics and general robotics bring to the clinical healthcare
sector, it certainly has its challenges and future opportunities, which
must be highlighted. One of the major challenges, a decade ago, was
the cost factor considering the capital costs of setting up a telerobotic
station, with all its associated technologies and network systems, in
building such a platform.
The cost has decreased considerably due to the improved advances

of technology, including Moore’s law of sliding scales. However, most
of the challenges remain, where medical science is unforgiving re-
garding ethical issues, risks, governance, and compliance. However,
the ongoing struggle remains in gaining acceptance of telerobotic pro-
cedures versus the conventional surgical procedure being performed
by a whole surgical team [Cazac and Radu, 2014; Weber, 2018].
From the premise of this work, it would be interesting to investi-

gate similar aspects on other platforms. Furthermore, we will investi-
gate, how one can use similar techniques in other areas.
The following technologies are experiencing explosive growth:
12.2.6 New AI health tools
One idea is to integrate computer science and information system

topics more directly into a messaging “game”. For example, by using
tweets and their (re-)distribution, students can study network effects,
or simulate distributed computing/algorithms via the messages sent.
12.2.7 Healthcare and data science
Other than social media the grammar rules designed in this work can
be applied in the synthesis of SMS’s and emails. This would be an
improvement on Sell-Bot’s advertising ability, by adding both SMS
and email marketing capabilities to reach a wider audience.
12.2.8 Machine-learning tools
For a digital tool like MedBot, generating interactive images about the
conditions and treatment approaches, would improve the effective-
ness of the tool. These visuals add to the patients understanding and
awareness of the condition.
12.2.9 Policy and governance frameworks
In South Africa, the successful implementation of universal health

coverage via NHI, necessitates a sound foundation of quality care
across all healthcare systems. To date, policies and interventions have
achieved moderate success, but a holistic approach is required to re-
store trust and confidence in health services, across the public and
private sectors. The proposed framework presented in this chapter
builds on and aids current policies and initiatives, provides stake-
holders with a common language of quality, and serves as a tool to
facilitate policy coherence and locate initiatives in the quality cycle.
In conclusion, the WHO handbook for developing a National Qual-

ity Policy and Strategy (NQPS), can support efforts to institution-
alise a culture of quality across the healthcare system. The WHO
reminder regarding challenges to overcome when implementing a na-
tional quality strategic framework is extremely valuable. Developing
an integrated, comprehensive quality strategic framework, focusing
on the healthcare sector of South Africa, will eventually produce the
fruits of the fourth industrial revolution in South Africa.
BIBLIOGRAPHY
[Abbott and Ade-Ibijola 2018] S. Abbott and A. Ade-Ibijola. Architec-

tural analysis of the South African Public Healthcare industry,.
In ICICIS - Proceedings of the 3rd International- Conference on The
Internet, Cyber Security and Information Systems, Gaborone, 2018.
[Abbott and Ade-Ibijola 2019a] S. Abbott and A. Ade-Ibijola. Algo-

rithms and a tool for automatic decryption of clinical notes. In
2019 6th International Conference on Soft Computing and Machine
Intelligence (ISCMI), pages 137–143. IEEE, 2019.
[Abbott and Ade-Ibijola 2019b] S. Abbott and A. Ade-Ibijola. Trend

analysis: A decision tool in SA healthcare. In 2019 International
Multidisciplinary Information Technology and Engineering Confer-
ence (IMITEC), pages 1–7. IEEE, 2019.
[Abinaya 2015] K. Abinaya. Data miningwith big data ehealth service

using map reduce,. International Journal of Advanced Research in
Computer and Communication Engineering, 4(2):123–127, 2015.
[Ackoff 1989] L. Ackoff. From data to wisdom. Journal of Applied

Systems Analysis, 16:3–9, 1989.
[Adebesin et al. ] F. Adebesin, P. Kotze, D. van Greunen, and R. Fos-

ter. Barriers and challenges to the adoption of e-health stan-
dards in africa,. CSIR.
[Adenuga et al. 2015] Olugbenga A Adenuga, Ray M Kekwaletswe,

and Alfred Coleman. eHealth integration and interoperabil-
ity issues: towards a solution through enterprise architecture.
Health information science and systems, 3(1):1–8, 2015.
[Africa 2012] Brand South Africa. Healthcare in South Africa.

https://www.brandsouthafrica.com/south-africa-fast-
facts/health-facts/health, 2012. Accessed: 2019-07-24.
[Afzal et al. 2018] N. Afzal, V. Mallipeddi, S. Sohn, H. Liu,

R. Chaudry, and C. Scott. Natural Language Processing of clin-
ical notes for identification of critical limb ischaemia. Interna-
tional Journal of Medical Informatics, 111:83–100, 2018.
[Agbo et al. 2019] Cornelius C Agbo, Qusay H Mahmoud, and

J Mikael Eklund. Blockchain technology in healthcare: a system-
atic review. In Healthcare, volume 7, page 56. Multidisciplinary
Digital Publishing Institute, 2019.
237
238 Bibliography
[Ahmed et al. 2017] Mohamed Nooman Ahmed, Andeep S Toor,

Kelsey O’Neil, and Dawson Friedland. Cognitive computing
and the future of health care cognitive computing and the fu-
ture of healthcare: the cognitive power of ibm watson has the
potential to transform global personalized medicine. IEEE pulse,
8(3):4–9, 2017.
[Ajayi et al. 2019] Olasupo O Ajayi, Antoine B Bagula, and Kun

Ma. Fourth industrial revolution for development: The rele-
vance of cloud federation in healthcare support. arXiv preprint
arXiv:1911.01708, 2019.
[Alvesson and Deetz 2000] Mats Alvesson and Stanley Deetz. A

framework for critical research. Doing critical management re-
search, pages 135–165, 2000.
[Amato et al. 2017] Flora Amato, Stefano Marrone, Vincenzo

Moscato, Gabriele Piantadosi, Antonio Picariello, and Carlo
Sansone. Automatizing healthcare. In Chatbots meet eHealth,
pages 40–49, 2017.
[Amiriam et al. 2017] P. Amiriam, P. Lang, and F.Van Loggerenberg.

Big Data in Healthcare: Extracting Knowledge from Point-of-Care Ma-
chines. Springer Briefs, Switzerland, 2017.
[Andersen et al. 2011] W. Andersen, S. Rosenberg, and K. Hormudz.

The incidence of Breast Cancer in the United States: Current
and future trends,. JNCI, 12(4):1397–1402, 2011.
[Archer 2011] N. Archer. Personal health records: a scoping review,.

Journal of Am Med Info Assoc, 18(10):515–522, 2011.
[Archer 2016] C. Archer. NHI: Paying more and getting less. Johan-
nesburg South Africa, 2016.
[Ardebesin 2013] F. Ardebesin. A review of interoperability stan-

dards in ehealth and imperatives for their adoption in africa,.
HISJ, 50, 2013.
[Armstrong et al. 2004] David G Armstrong, Lawrence A Lavery,

Brent P Nixon, and Andrew JM Boulton. It’s not what you
put on, but what you take off: techniques for debriding and
off-loading the diabetic foot wound. Clinical infectious diseases,
39(Supplement_2):S92–S99, 2004.
[Armstrong et al. 2005] David G Armstrong, Lawrence A Lavery,

Stephanie Wu, and Andrew JM Boulton. Evaluation of remov-
able and irremovable cast walkers in the healing of diabetic foot
wounds: a randomized controlled trial. Diabetes Care, 28(3):551–
554, 2005.
Bibliography 239
[Armstrong et al. 2017] David G. Armstrong, Andrew J. M. Boulton,

and Sicco A Bus. Diabetic foot ulcers and their recurrence. The
New England journal of medicine, 376 24:2367–2375, 2017.
[Armstrong et al. 2018] David G Armstrong, AJ Meyr, H Sanfey,

JF Eidt, JL Mills, and JA Billings. Basic principles of wound
management. UpToDate, Waltham, MA.(Accessed on December 14,
2016), 2018.
[Arndt and Bigelow 1998] M. Arndt and B. Bigelow. Reengineer-

ing – déjà vu all over again,. Health Care Management Review,
23(3):58–66, 1998.
[Avgousti et al. 2016] Sotiris Avgousti, Eftychios G Christoforou, An-

dreas S Panayides, Sotos Voskarides, Cyril Novales, Laurence
Nouaille, Constantinos S Pattichis, and Pierre Vieyres. Medical
telerobotic systems: current status and future trends. Biomedical
engineering online, 15(1):96, 2016.
[Awad et al. 2018] Atheer Awad, Sarah J Trenfield, Simon Gaisford,

and Abdul W Basit. 3d printed medicines: A new branch of dig-
ital healthcare. International journal of pharmaceutics, 548(1):586–
596, 2018.
[Ayentimi and Burgess 2019] Desmond Tutu Ayentimi and John

Burgess. Is the fourth industrial revolution relevant to sub-
sahara africa? Technology Analysis and Strategic Management,
31(6):641–652, 2019.
[Badimo 2018] Kgabo H Badimo. The Impact of the Fourth Industrial

Revolution on public service delivery. 2018.
[Baines et al. 2019] Darrin Baines, Lotte Stig Nørgaard, Charlotte

Rossing, et al. The fourth industrial revolution: Will it change
pharmacy practice? Research in Social and Administrative Phar-
macy, 2019.
[Bali 2013] R. Bali. Pervasive Health Knowledge Management. Springer,

NY, 2013.
[Barron et al. 2016] Peter Barron, Yogan Pillay, Antonio Fernandes,

Jane Sebidi, and Rob Allen. The MomConnect mHealth initia-
tive in south africa: Early impact on the supply side of MCH
services. Journal of public health policy, 37(2):201–212, 2016.
[Bateman 2013] Chris Bateman. Drug stock-outs: Inept supply-chain

management and corruption. South African Medical Journal,
103(9):600–602, 2013.
[Bates 2019] M. Bates. Health care chatbots are here to help. IEEE
pulse, 10(3):12–14, 2019.
240 Bibliography
[Bauer 2018] G. Bauer. Delivering value-based care with ehealth

serivices,. In Foundation of the American College of Healthcare Ex-
ecutives, Connecticut USA. 2018.
[BCX 2016] BCX. Health and healthcare in a Digitalised Future, 2016.
[Beck et al. 2012] E. Beck, A. Craig, J. Beeson, S. Bourn, J. Goodloe,

H. Moy, B. Myers, E. Racht, and L. White. A healthcare de-
livery strategy to improve access, outcomes and value,. Mobile
Integrated Health Practice, 8(1):1–9, 2012.
[Begg et al. 2018] Kerrin Begg, Punithasvaree Mamdoo, Lilian Dud-

ley, Justin Engelbrecht, Gail Andrews, and Lebogang Lebese.
Development of a national strategic framework for a high-
quality health system in south africa. South African health review,
2018(1):77–85, 2018.
[Belle et al. 2015] A. Belle, R. Thiagarajm, S.M. Soroushunder, F. Na-

vidi, D. Beard, and K. Najarian. Big data analystics in health-
care,. Biomedical Research International, 2015:1–16, 2015.
[Benson and Jatoi 2012] R. Benson and I. Jatoi. The global breast can-
cer burden,. Future Medicine, 8(6):1–12, 2012.
[Benson 2013] T. Benson. Principles of Health Interoperability: HL-7 and

SNOMED and FHIR method. Springer Publications, London, 3rd
edition, 2013.
[Berghout et al. 2015] M. Berghout, J.Van Exel, L. Leensvaart, and

J. Cramm. Healthcare professionals views on patient centered
care in hospitals,. BMC Health Services Research, 15:385, 2015.
[Berlinger 2006] N Berlinger. Robotic surgery — squeezing into tight

places. New England Journal of Medicine, 354(20):2099–2101, 2006.
PMID: 16707746.
[Betjeman et al. 2013] Thomas J Betjeman, Samara E Soghoian, and

Mark P Foran. mhealth in sub-saharan africa. International jour-
nal of telemedicine and applications, 2013:6, 2013.
[Bhattacherjee 2012] Anol Bhattacherjee. Social science research:

Principles, methods, and practices. 2012.
[Bisandu 2016] Desmond Bala Bisandu. Design science research

methodology in computer science and information systems. In-
ternational Journal of Information Technology, 5(4):55–60, 2016.
[Bishop 2015] Felicity L Bishop. Using mixed methods research

designs in health psychology: An illustrated discussion from
a pragmatist perspective. British journal of health psychology,
20(1):5–20, 2015.
Bibliography 241
[Blumenfeld 1997] C. Blumenfeld. Integrating knowledge bases at

the point of care,. Health Management Technology, 18(7):44–46,
1997.
[Botha et al. 2016] A Botha, Marlien Herselman, Ronell Alberts,

Thomas Fogwill, Matthew Chetty, and Paul Geldenhuys. Phase
1a: Literature overview of health in south africa. CSIR, 2016.
[Boulton et al. 2018] Andrew JM Boulton, David G Armstrong,

Robert S Kirsner, Christopher E Attinger, Lawrence A Lavery,
Benjamin A Lipsky, Joseph L Mills Sr, and John S Steinberg. Di-
agnosis and management of diabetic foot complications. 2018.
[Brem et al. 2010] Harold Brem, Jason Maggi, David Nierman, Linda
Rolnitzky, David Bell, Robert Rennert, Michael Golinko, Alan
Yan, Courtney Lyder, and Bruce Vladeck. High cost of stage iv
pressure ulcers. The American Journal of Surgery, 200(4):473–477,
2010.
[Brinkel et al. 2014] Johanna Brinkel, Alexander Krämer, Ralf

Krumkamp, Jürgen May, and Julius Fobil. Mobile phone-
based mhealth approaches for public health surveillance in
sub-saharan africa: a systematic review. International journal
of environmental research and public health, 11(11):11559–11582,
2014.
[Brooke 2002] Carole Brooke. What does it mean to be ‘critical’in is

research? Journal of Information Technology, 17(2):49–57, 2002.
[Brown and Dueñas 2020] Megan EL Brown and Angelique N

Dueñas. A medical science educator’s guide to selecting a re-
search paradigm: building a basis for better research. Medical
Science Educator, 30(1):545–553, 2020.
[Brownstein et al. 2009] J. Brownstein, C. Freifeld, and C. Madoff.

Digital disease detection-harnessing the web for Public Health
surveillance,. New England Journal of Medicine, pages 2153–2157,
2009.
[Broyles et al. 1998] R. Broyles, E. Brandt, and D. Biard-Holmes. A

practical method of adjusting for risk in prospective costs of
capitated systems,. Health Care Management Review, 23(2):63–75,
1998.
[Busain 2017] R. Busain. IBM Watson Health, 2017.
[Cabestany et al. 2018] Joan Cabestany, Daniel Rodriguez-Martín,

Carlos Pérez, and Albert Sama. Artificial intelligence contribu-
tion to ehealth application. In 2018 25th International Conference"
Mixed Design of Integrated Circuits and System"(MIXDES), pages
15–21. IEEE, 2018.
242 Bibliography
[Calhoun 1997] I. Calhoun. Migrating from a transaction based to an

information based managed care system,. Health Management
Technology, 18(5):72–75, 1997.
[Carestream 2013] Carestream. How to achieve Interoperability in health-

care,, 2013.
[Carneiro 2009] P. Carneiro. Mylonakis. A web based tool for real

time surveillance of disease outbreaks,. Clinical Infectious Dis-
eases, pages 1557–1564, 2009.
[Carr and Moore 2003] C Carr and S Moore. Ihe: a model for driving
adoption of standards. Computerized Medical Imaging and Graph-
ics, 27(2-3):137–146, 2003.
[Carrell et al. 2017] D. Carrell, R. Schoen, D. Leffler, M. Morris,

S. Rose, A. Baer, S. Crockett, and K. Dean. Challenges in adapt-
ing existing clinical natural language processing systems to mul-
tiple diverse healthcare settings,. JAMIA, 24(5):986, 2017.
[Cavanagh and Bus 2010] Peter R Cavanagh and Sicco A Bus. Off-
loading the diabetic foot for ulcer prevention and healing. Jour-
nal of the American Podiatric Medical Association, 100(5):360–368,
2010.
[Cawsey and Jones 1997] A. Cawsey and R. Jones. Natural language

generation in health care,. Journal of American Medical Informa-
tion Assoc, pages 473–482, 1997.
[Cawsey 1997] L. Cawsey. The application of big data - healthcare

trends,. JAMA, 1997.
[Cazac and Radu 2014] C Cazac and G Radu. Telesurgery–an effi-

cient interdisciplinary approach used to improve the health care
system. Journal of medicine and life, 7(Spec Iss 3):137, 2014.
[Chang 2020] Anthony Chang. The role of artificial intelligence in

digital health. In Digital Health Entrepreneurship, pages 71–81.
Springer, 2020.
[Chapman et al. 2000] C. Chapman, J. Clinton, R. Kerber, and

T. Khabaza. Crisp-dm 1.0 step by step data mining guide. Jour-
nal of Healthcare Informatica, 2000.
[Chawla and Davis 2013] N. Chawla and D. Davis. Bringing big data
to personalized healthcare: A patient-centered framework,. Jour-
nal of General Internal Medicine, 28(3):660–665, 2013.
[Chen and Asch 2017] J. Chen and M. Asch. Machine learning and
prediction in medicine:beyond the peak of inflated expectations,.
NEJM, 376(26):2496–2507, 2017.
Bibliography 243
[Chen et al. 2012] Hsinchun Chen, Roger HL Chiang, and Veda C

Storey. Business intelligence and analytics: From big data to
big impact. MIS quarterly, pages 1165–1188, 2012.
[Chen et al. 2014] Min Chen, Shiwen Mao, and Yunhao Liu. Big data:
A survey. Mobile networks and applications, 19(2):171–209, 2014.
[Cheruto 2021] Sowon Karen Cheruto. How the healthcare-seeking

socio-cultural context shapes maternal health clients’ mhealth
utilisation in a kenyan context. 2021.
[Chetty and Yamin 2015] Girija Chetty and Mohammad Yamin. In-
telligent human activity recognition scheme for ehealth applica-
tions. Malaysian Journal of Computer Science, 28(1):59–69, 2015.
[Chib 2013] Arul Chib. The promise and peril of mhealth in develop-
ing countries. Mobile Media and Communication, 1(1):69–75, 2013.
[Cho et al. 2018] Ji-Eun Cho, Jun Sang Yoo, Kyoung Eun Kim,
Sung Tae Cho, Woo Seok Jang, Ki Hun Cho, and Wan-Hee Lee.
Systematic review of appropriate robotic intervention for gait
function in subacute stroke patients. BioMed research interna-
tional, 2018, 2018.
[Chopra et al. 2013] A. Chopra, A. Prashar, and C. Sain. Natural Lan-

guage Processing. Open Access Journal of Scientific, Technology and
Engineering Research, 1(4):131–146, 2013.
[Chou 2018] Shuo-Yan Chou. The fourth industrial revolution: Digi-

tal fusion with internet of things. Journal of International Affairs,
72(1), 2018.
[Chowles 2014] T. Chowles. The time is right for electronic health

records (ehr). Natal, 2014.
[Chute and French 2019] Chaloner Chute and Tara French. Introduc-
ing care 4.0: an integrated care paradigm built on industry 4.0
capabilities. International journal of environmental research and pub-
lic health, 16(12):2247, 2019.
[Clark 1993] M. Clark. Creating customer value: Information-chain-

based management.,. The Executive’s Journal, no. Fall, page 13–18,
1993.
[Coleman et al. 2011] Alfred Coleman, Marlien E Herselman, and Da-

lenca Potass. E-health readiness assessment for e-health frame-
work for africa: a case study of hospitals in south africa. In
International Conference on Electronic Healthcare, pages 162–169.
Springer, 2011.
[Coleman 2012] A. Coleman. E Health Readiness Assessment for E

Health Framework for Africa: A Case Study of Hospitals in South
244 Bibliography
Africa,, 2012.
[Collden and Hellostrom 2018] C. Collden and A. Hellostrom. Value-

based healthcare translated: a complimentary view of imple-
mentation,. BMC Health Services Research, 18(681):11, 2018.
[Collobert et al. 2011] R. Collobert, J. Weston, L. Bottou, M. Karlen,

and P. Kuksa. Natural language processing from scratch,. Jour-
nal of Machine Learning, 10(3):2493–2537, 2011.
[Commission and others 2013a] National Planning Commission et al.

National development plan vision 2030. 2013.
[Commission and others 2013b] NDP Commission et al. National de-

velopment plan vision 2030. Gazette, 2013.
[Coovadia et al. 2009] H. Coovadia, R. Jewkes, P. Barron, D. Sanders,

and D. McIntyre. The health and health system of south africa:
historical roots of current public health challenges. The Lancet,
374(9692):817–834, 2009.
[Corp 2013] McKinsey Corp. The big data revolution in healthcare:

Accelerating value and innovation. 2013.
[Cupoli et al. 2014] Patricia Cupoli, S Earley, and D Henderson.

Dama-dmbok2 framework. DAMA International, 2014.
[Dahl and Boulos 2014] T Dahl and M Boulos. Robots in health and
social care: A complementary technology to home care and tele-
healthcare. Robotics, 3(1):1–21, 2014.
[Davenport 1993] T. Davenport. Process Innovation: Reengineering

Work through Information Technology. Harvard Business School
Press, Boston MA, 1993.
[Davies 2000] B. Davies. A review of robotics in surgery. Journal of

Engineering in Medicine, 214(1):129–140, 2000.
[Denecke et al. 2019] Kerstin Denecke, Mauro Tschanz, Tim Lucas

Dorner, and Richard May. Intelligent conversational agents in
healthcare: Hype or hope? Studies in health technology and infor-
matics, 259:77–84, 2019.
[Dhingra and Dabas 2020] Dhulika Dhingra and Aashima Dabas.

Global strategy on digital health. Indian pediatrics, 57(4):356–358,
2020.
[DHMIS 2011] DHMIS. District Health Management Information Sys-

tem, South Africa, (DHMIS) Policy. National Department of
Health, SA, 2011.
[Diegel et al. 2006] O Diegel, WL Xu, and J Potgieter. A case study of

rapid prototype as design in educational engineering projects.
Bibliography 245
International Journal of Engineering Education, 22(2):350, 2006.
[Dinov 2016] G. Dinov. Methodologica challenges and analytical op-

portunities for modeling and interpreting Big Healthcare Data,.
Open Access: Giga Science, 5(12):1–15, 2016.
[Duwe 2004] H. Duwe. Relationships between healthcare and re-

search records,. Institute of Aerospace Medicine, German Aerospace
Center (DLR), Journal of Aerospace Medicine, 2004.
[Ellis 2014] B. Ellis. Real-time analytics: Techniques to analyze and visual-

ize streaming data. John Wiley and Sons, 2014.
[Elton and Riordan 2016] J. Elton and A. Riordan. Next Generation

Business Models and Strategies, in Healthcare Disrupted. Wiley and
Sons, 2016.
[Erb 2012] B. Erb. Concurrent programming for scalable web archi-

tectures. 2012.
[Essack 2011] S. Essack. Part III Antibiotic supply chain and manage-
ment, 2011.
[Fatehi et al. 2018] Farhad Fatehi, Anish Menon, and Dominique Bird.
Diabetes care in the digital era: a synoptic overview. Current
diabetes reports, 18(7):38, 2018.
[Fayad 1996] U. Fayad. Advances in Knowledge Discovery and Data Min-

ing. MIT Press, MA, 1996.
[Feeley et al. 2020] Thomas W Feeley, Zachary Landman, and

Michael E Porter. The agenda for the next generation of health
care information technology. NEJM Catalyst Innovations in Care
Delivery, 1(3), 2020.
[Fis 2016] Predictors of 30-day readmission following in-patient re-

habilitation for patients at high risk for hospital for hospital re
admissions. Physical Therapy, 96(1):62–70, 2016.
[Flyvberg 2014] B. Flyvberg. What you should know about mag-

aprojects and why: An overview. Project Management Journal,
45(2):6–19, 2014.
[Foster 2013] Rosemary Foster. Assessing the development process

of the ehealth strategy for south africa against the recommen-
dations of the who/itu national ehealth strategy toolkit. Journal
of the International Society for Telemedicine and eHealth, 1(2):62–72,
2013.
[Fremmelevholm and Soegaard 2019] Aase Fremmelevholm and

Knaerke Soegaard. Pressure ulcer prevention in hospitals: a
successful nurse-led clinical quality improvement intervention.
246 Bibliography
British Journal of Nursing, 28(6):S6–S11, 2019.
[Friedman 1999] C. Friedman. Natural language processing and its

future in medicine,. Academic Medicine, 74(8), 1999.
[Frigo and Interviewee 2017] M. Frigo and Interviewee. Managing

Healthcare Costs and Value, 2017.
[Gacenga et al. 2012] Francis Gacenga, Aileen Cater-Steel, Mark Tole-

man, and Wui-Gee Tan. A proposal and evaluation of a design
method in design science research. Electronic Journal of Business
Research Methods, 10(2):89–100, 2012.
[Gartner 2015] Gartner. Big Data in Healthcare: Analytics, 2015.
[Gartner 2018] Gartner. Gartner Insights for Industry for 2018 to 2019,
2018.
[Gastrow 2020] MICHAEL Gastrow. Policy options for the fourth in-
dustrial revolution in south africa. 2020.
[Geldenhuys and Botha 2015] P. Geldenhuys and A. Botha. “on the

design of an electronic health patient registration system,” in
proceedings of the 9th idia conference, idia2015, beyond devel-
opment. time for a new ict4d paradigm? Zanzibar, 2015.
[Gentry 1996] J. Gentry. Process Over Function: Preparing for Reengi-

neering in Health Care, 1996.
[Ghobakhloo 2020] Morteza Ghobakhloo. Industry 4.0, digitization,

and opportunities for sustainability. Journal of cleaner production,
252:119869, 2020.
[Giulianotti et al. 2003] Pier Cristoforo Giulianotti, Andrea Coratti,

Marta Angelini, Fabio Sbrana, Simone Cecconi, Tommaso
Balestracci, and Giuseppe Caravaglios. Robotics in general
surgery: personal experience in a large community hospital.
Archives of surgery, 138(7):777–784, 2003.
[Gkegkes et al. 2017] Ioannis D Gkegkes, Ioannis A Mamais, and

Christos Iavazzo. Robotics in general surgery: A systematic cost
assessment. Journal of minimal access surgery, 13(4):243, 2017.
[Goh et al. 2020] Teik Chiang Goh, Mohd Yazid Bajuri, Sivapathasun-
daram C Nadarajah, Abdul Halim Abdul Rashid, Suhaila Ba-
haruddin, and Kamarul Syariza Zamri. Clinical and bacteriolog-
ical profile of diabetic foot infections in a tertiary care. Journal
of foot and ankle research, 13(1):1–8, 2020.
[Google 2018] Google. Google Trends, 2018.
[Government 2012a] SA Government. National development plan

2030: Our future-make it work. 2012.
Bibliography 247
[Government 2012b] South African Government. National Develop-

ment Plan: Strategic Plan,, 2012.
[Government 2016] SA. Government. Public Service and Administra-

tion -DPSA, 2016.
[Goy et al. 2019] A Goy, S Nishtar, V Dzau, C Balatbat, and R Diabo.

Health and healthcare in the fourth industrial revolution: Global
future council on the future of health and healthcare 2016-2018.
In Proceedings of the World Economic Forum, 2019.
[Gregor and Hevner 2013] Shirley Gregor and Alan R Hevner. Po-
sitioning and presenting design science research for maximum
impact. MIS quarterly, pages 337–355, 2013.
[Gröger 2018] Christoph Gröger. Building an industry 4.0 analytics

platform. Datenbank-Spektrum, 18(1):5–14, 2018.
[Grossman and Mazzucco 2002] R. Grossman and C. Mazzucco.

Dataspace: a data web for the exploratory analysis and mining
of data. Computing in Science and Engineering, 4(4):44–51, 2002.
[Gui et al. 2016] Hao Gui, Rong Zheng, Chao Ma, Hao Fan, and Liya
Xu. An architecture for healthcare big data management and
analysis. In International conference on health information science,
pages 154–160. Springer, 2016.
[Gui 2016] H. Gui. An architecture for healthcare Big Data manage-

ment and analysis,. Springer Int Publishing, 10(7):154 –161, 2016.
[Gurman et al. 2012] Tilly A Gurman, Sara E Rubin, and Amira A

Roess. Effectiveness of mhealth behavior change communica-
tion interventions in developing countries: a systematic review
of the literature. Journal of Health Communication, 17(1):82–104,
2012.
[Habermas 1976] Jürgen Habermas. Some distinctions in universal

pragmatics. Theory and Society, 3(2):155–167, 1976.
[Halamka et al. 2008] John D Halamka, Kenneth D Mandl, and

Paul C Tang. Early experiences with personal health records.
Journal of the American Medical Informatics Association, 15(1):1–7,
2008.
[Hampshire et al. 2015] Kate Hampshire, Gina Porter, Samuel Asiedu

Owusu, Simon Mariwah, Albert Abane, Elsbeth Robson, Alister
Munthali, Ariane DeLannoy, Andisiwe Bango, Nwabisa Gungu-
luza, et al. Informal m-health: How are young people using
mobile phones to bridge healthcare gaps in sub-saharan africa?
Social Science and Medicine, 142:90–99, 2015.
248 Bibliography
[Han and Kamber 2000] J. Han and M. Kamber. Data Mining: Con-
cepts and Techniques. Morgan, NY, 2000.
[Hangu 2018] Y. Hangu. A machine learning based natural language

question and answering system for healthcare data search using
complex queries. In IEEE International Conference on Big Data. Big
Data, NY, 2018.
[Hanson et al. 2009] Darlene S Hanson, Diane Langemo, Julie Ander-

son, Ptrt Thompson, and Susan Hunter. Can pressure mapping
prevent ulcers. Nursing, 39(6):50–51, 2009.
[Harrison 2010] D. Harrison. An Overview of Healthcare in South

Africa 1994-2010: Priorities, progress and prospects for new
gains,. Henry J Kaiser Family Foundation, Muldersdrift, Johannes-
burg SA, 2010.
[Haslinda and Sarinah 2009] A. Haslinda and A. Sarinah. A review

of knowledge management models.,. The Journal of International
Social Research, 2(9), 2009.
[Hastie et al. 2001] R. Hastie, R. Tibshirani, and J. Friedman. The el-

ements of statistical learning: Data Mining inference and prediction.
Springer, NY, 2001.
[Haywood 2012] M. Haywood. Monitoring our Health, 2012.
[Haywood 2016] M. Haywood. The NHI and SA Public Healthcare.

Section 27 Newsletter, 2016.
[Herselman and Botha 2016a] M. Herselman and A Botha. Ap-

proaches and Experiences: Towards building a South African Digi-
tal health Innovation Ecosystem. Meraka Institute, CSIR, Pretoria,
2016.
[Herselman and Botha 2016b] ME Herselman and A Botha. Strate-

gies, Approaches and Experiences: Towards building a South African
Digital Health Innovation Ecosystem, 2016.
[Herselman et al. 2016] M. Herselman, A. Botha, R. Alberts, and

T. Fogwell. Conceptualisaiton of a digital health innovation ecosys-
tem. Meraka Institute South Africa, Pretoria, 2016.
[Hevner et al. 2004] Alan R Hevner, Salvatore T March, Jinsoo Park,

and Sudha Ram. Design science in information systems re-
search. MIS quarterly, pages 75–105, 2004.
[Hevner 2007] Alan R Hevner. A three cycle view of design sci-

ence research. Scandinavian journal of information systems, 19(2):4,
2007.
Bibliography 249
[Higman et al. 2019] Susan Higman, Vikas Dwivedi, Alpha

Nsaghurwe, Moses Busiga, Hermes Sotter Rulagirwa, Dasha
Smith, Chris Wright, Ssanyu Nyinondi, and Edwin Nyella.
Designing interoperable health information systems using
enterprise architecture approach in resource-limited countries:
A literature review. The International journal of health planning
and management, 34(1):85–99, 2019.
[Hilferty 2018] D. Hilferty. Getting healthcare right,. The Journal of

Law, 46:829–830, 2018.
[HIMSS 2013] HIMSS. Interoperability Standards, 2013.
[Hinchley 2007] Andrew Hinchley. Understanding version 3. A

primer on the HL7 Version, 3, 2007.
[Hirschheim 1985] Rudy Hirschheim. Information systems episte-

mology: An historical perspective. Research methods in informa-
tion systems, 9:13–35, 1985.
[HiSP 2012] HiSP. HiSP: Projects - Health Information System Pro-

gramme, 2012.
[HMT 2017] HMT. Health Management Technology Archives, 2017.
[Horkheimer 1972] Max Horkheimer. The social function of philoso-

phy. Critical theory: Selected essays, pages 253–72, 1972.
[Howcroft and Trauth 2004] Debra Howcroft and Eileen M Trauth.

The choice of critical information systems research. In Infor-
mation systems research, pages 195–211. Springer, 2004.
[Howie et al. 2019] Lynn J Howie, Nancy S Scher, Laleh Amiri-

Kordestani, Lijun Zhang, Bellinda L King-Kallimanis, Yasmin
Choudhry, Jason Schroeder, Kirsten B Goldberg, Paul G Kluetz,
Amna Ibrahim, et al. Fda approval summary: pertuzumab for
adjuvant treatment of her2-positive early breast cancer. Clinical
Cancer Research, 25(10):2949–2955, 2019.
[HST 2012] HST. Hospital Transformation Plan Concerns, 2012.
[HST 2016] HST. Health Systems Trust-Publications,, 2016.
[Huang et al. 2017] X. Huang, P. Baade, D. Youlden, P. Youl, and

W. Hu. Google as a control tool in queensland. BMC Cancer,
17:1–9, 2017.
[IHE 2016] IHE. The Integration of Health Domains, 2016.
[Iroju and Olaleke 2015] O. Iroju and J. Olaleke. A Systematic review

of Natural Language Processing in healthcare,. International
Journal of Information Technology and Computer Sciences, 8(7):44–
50, 2015.
250 Bibliography
[Issom et al. 2015] David-Zacharie Issom, Ashenafi Zebene Woldare-

gay, Taridzo Chomutare, Meghan Bradway, Eirik Årsand, and
Gunnar Hartvigsen. Mobile applications for people with dia-
betes published between 2010 and 2015. Diabetes Management,
5(6):539–550, 2015.
[Javaid and Haleem 2019] Mohd Javaid and Abid Haleem. Industry
4.0 applications in medical field: a brief review. Current Medicine
Research and Practice, 9(3):102–109, 2019.
[Jeffery 2016] A. Jeffery. Healthcare and the NHI. Johannesburg South

Africa, 2016.
[Jegede et al. 2015] O Jegede, K Ferens, B Griffith, and B Podaima. A

smart shoe to prevent and manage diabetic foot diseases. In Int’l
Conf. Health informatics and medical systems, pages 47–54, 2015.
[Jiang et al. 2017] F. Jiang, H. Zhui, H. Li, present Wang, Y. Intelli-

gence in healthcare past, and future. British Medical Journal,
2(1):230–240, 2017.
[Jones et al. 2014] Daniel L Jones, Robert C Brewster, and Rob Phillips.
Promoter architecture dictates cell-to-cell variability in gene ex-
pression. Science, 346(6216):1533–1536, 2014.
[Kaboub 2008] Fadhel Kaboub. Positivist paradigm. Encyclopaedia of

counselling, 2(2):343, 2008.
[Kahil et al. 2019] Moustafa Sadek Kahil, Abdelkrim Bouramoul, and

Makhlouf Derdour. Big data and interactive visualization:
Overview on challenges, techniques and tools. In International
Conference on Advanced Intelligent Systems for Sustainable Develop-
ment, pages 157–167. Springer, 2019.
[Kankanhalli et al. 2005] Atreyi Kankanhalli, Bernard CY Tan, and

Kwok-Kee Wei. Contributing knowledge to electronic knowl-
edge repositories: an empirical investigation. MIS quarterly,
pages 113–143, 2005.
[Kannel and McGee 1979] W. Kannel and D. McGee. Diabetes and

Cardiovascular Disease,. JAMA, 241:2035–2038, 1979.
[Kaplan and Porter ] R. Kaplan and M. Porter. How to solve the the
cost crisis in health care. HBR, 11:47.
[Kaplan and Porter 2017] R. Kaplan and M. Porter. Managing health-

care costs and value. Strategic Finance, 34(10):25–35, 2017.
[Kapoor et al. 2020] Aditya Kapoor, Santanu Guha, Mrinal Kanti Das,
Kewal C Goswami, and Rakesh Yadav. Digital healthcare: The
only solution for better healthcare during covid-19 pandemic.
Indian Heart Journal, 2020.
Bibliography 251
[Karandikar and Tayade 2014] Pratibha M Karandikar and Motilal C

Tayade. Application of robotics technology in clinical practice
in india. Asian Journal of Medical Sciences, 5(1):29–33, 2014.
[Katuu 2016a] S. Katuu. Transforming south africa’s health sector.

Journal of Science and Technology Policy Management, 7:330–345,
2016.
[Katuu 2016b] Shadrack Katuu. Transforming south africa’s health

sector: The ehealth strategy, the implementation of electronic
document and records management systems (edrms) and the
utility of maturity models. Journal of Science and Technology Policy
Management, 7(3):330–345, 2016.
[Katuu 2018] Shadrack Katuu. Healthcare systems: typologies,

framework models, and south africa’s health sector. International
Journal of Health Governance, 23(2):134–148, 2018.
[Katuu 2019] Shadrack Katuu. Health information systems, ehealth

strategy, and the management of health records: The quest to
transform south africa’s public health sector. In Healthcare Policy
and Reform: Concepts, Methodologies, Tools, and Applications, pages
493–517. IGI Global, 2019.
[Katz et al. 2005] Ira A Katz, Anthony Harlan, Bresta Miranda-Palma,

Luz Prieto-Sanchez, David G Armstrong, John H Bowker,
Mark S Mizel, and Andrew JM Boulton. A randomized trial of
two irremovable off-loading devices in the management of plan-
tar neuropathic diabetic foot ulcers. Diabetes Care, 28(3):555–559,
2005.
[Kaur et al. 2019] P. Kaur, M. Sharma, and M. Mittal. Big data and ma-
chine learning based secure healthcare framework. In ICCIDS,
New Dehli. 2019.
[Kawamoto et al. 2014] K. Kawamoto, C. Martin, K. Williams,

C.Hunter Ming-Chieu, C. Staes, and B. Bray. Value driven out-
comes (vdo): a pragmatic, modular, and extensible software
framework for understanding and improving healthcare costs
and outcomes. AMIA, 2014.
[Keen 1994] J. Keen. Information Management in Health. Services, UK-

Bristol: Open University Press, 1994.
[Keenan et al. 2013] ACM Keenan, ACM Wood, N Beattie, RM Boyle,

FC Doogan, and C Court-Brown. The treatment and out-
comes of “jones” fractures. In Orthopaedic Proceedings, vol-
ume 95, pages 0–4. The British Editorial Society of Bone and
Joint Surgery, 2013.
252 Bibliography
[Kempen 2019] Annalise Kempen. The 4th industrial revolution-

hidden threats to human and cybersecurity? Servamus
Community-based Safety and Security Magazine, 112(10):10–12,
2019.
[Kerr et al. 2018] D Kerr, C Axelrod, C Hoppe, and DC Klonoff. Di-

abetes and technology in 2030: a utopian or dystopian future?
Diabetic Medicine, 35(4):498–503, 2018.
[Khalifa 2013] M. Khalifa. Barriers to health information systems

and electronic medical record implementation: A field study of
saudi arabian hospitals. In The 3rd International Conference on
Current and Future Trends of Information and, King Faisal Special-
ist. Hospital and Research Center, Jeddah 21499, Saudi Arabia,
2013.
[Khumalo et al. 2012] N. Khumalo, M. Willie, and E. “what are the

key determinants of hospital admissions, re-admission rate and
day case rate within the south african medical schemes popula-
tion,”. Farmeconomia, 13(2):83–100, 2012.
[Kleynhans 2011a] A. Kleynhans. Is South Africa ready for a National

Electronic Health Record (EHR)? SBL, Pretoria, Unisa, 2011.
[Kleynhans 2011b] Adele-mari Kleynhans. Is South Africa ready for a

national Electronic Health Record (EHR)? PhD thesis, 2011.
[Klimko 2001] G. Klimko. Knowledge management and maturity

models: building common understanding. In Second Conference
on Knowledge Management. Slovenia, 2001.
[Koen 2019] Hildegarde Koen. An introduction to artificial intelli-

gence. 2019.
[Koleck et al. 2019] T. Koleck, C. Dreisbach, P. Bourne, and S. Bakken.

Natural language processing of symptoms documented in free
text narratives of eHealth records: A systematic review,. JMAIA,
26(4):364–376, 2019.
[Kotzé and Alberts 2017] Paula Kotzé and Ronell Alberts. Towards
a conceptual model for an e-government interoperability frame-
work for south africa. In ICEIS (3), pages 493–506, 2017.
[Kotze 2014] P. Kotze. A conceptual data model for a primary health-

care patient-centric electronic medcial record system. In Proceed-
ings of the IASTED International Conference, Gaborone Botswana,
2014.
[Kraftová et al. 2018] Ivana Kraftová, Iveta Doudová, and Radim

Miláček. At the threshold of the fourth industrial revolution:
who gets who loses. 2018.
Bibliography 253
[Kumar et al. 2020] Ravinder Kumar, Rajesh Kr Singh, and Yogesh Kr

Dwivedi. Application of industry 4.0 technologies in smes for
ethical and sustainable operations: Analysis of challenges. Jour-
nal of cleaner production, 275:124063, 2020.
[Kumar 2018] R SreeRaja Kumar. Robotic nursing in health care de-

livery. International Journal of Nursing Education, 10(3), 2018.
[Kvasny and Richardson 2006] Lynette Kvasny and Helen Richard-

son. Critical research in information systems: looking forward,
looking back. Information Technology & People, 2006.
[Lavery et al. 1998] Lawrence A Lavery, David G Armstrong,

Steven A Vela, Terri L Quebedeaux, and John G Fleischli. Prac-
tical criteria for screening patients at high risk for diabetic foot
ulceration. Archives of internal medicine, 158(2):157–162, 1998.
[Lee and Lim 2017] Jai Yon Lee and Jae Young Lim. The prospect of
the fourth industrial revolution and home healthcare in super-
aged society. Annals of Geriatric Medicine and Research, 21(3):95–
100, 2017.
[Lee et al. 2016] S. Lee, K. Kawamuto, R. Hess, C. Park, J. Young,

C. Hunter, S. Johnson, S. Gulbrandsen, D. Horton, and
K. Graves. Implementation of a value-driven outcomes program
to identify high variability in clinical costs and outcomes and
association with reduced cost and improved quality,. Journal of
American Medical Association, 316(10):1061–1074, 2016.
[Levi Sandri et al. 2017] Giovanni B Levi Sandri, Edoardo de Werra,

Gianluca Mascianà, Francesco Guerra, Gabriele Spoletini, and
Quirino Lai. The use of robotic surgery in abdominal or-
gan transplantation: A literature review. Clinical transplantation,
31(1):e12856, 2017.
[Liddy 2001] E. Liddy. Natural Language Processing. Encyclopaedia

of Library Science and Information Science, 2001.
[Limaye et al. 2017] Rupali J Limaye, Tara M Sullivan, Scott Dalessan-

dro, and Ann Hendrix Jenkins. Looking through a social lens:
conceptualising social aspects of knowledge management for
global health practitioners. Journal of public health research, 6(1),
2017.
[Lin et al. 2018] Rongheng Lin, Zezhou Ye, Hao Wang, and Budan
Wu. Chronic diseases and health monitoring big data: A survey.
IEEE reviews in biomedical engineering, 11:275–288, 2018.
[Lotto et al. 2017] M. Lotto, P. Aiguira, D. Rios, and M. Machado.

Analysis of the interests of Google users on toothache informa-
tion. PloS One Journal, 72:1–17, 2017.
254 Bibliography
[Lozano-Rubí et al. 2016] Raimundo Lozano-Rubí, Adolfo Muñoz

Carrero, Pablo Serrano Balazote, and Xavier Pastor. Ontocr: A
cen/iso-13606 clinical repository based on ontologies. Journal of
biomedical informatics, 60:224–233, 2016.
[Machado et al. 2012] L. Machado, V. Bafna, A. Boxwala, W. Chap-

man, and K. Chaudhuri. idash: integrating data for analysis,
anonymization and sharing,. Journal of American Medical Infor-
mation Association, 19:196–200, 2012.
[Magazines 2012] Health Magazines. Healthcare review: Healthcare

in africa - rising to the challenge. Healthcare Review, 23, 2012.
[Mahomed 2018] S Mahomed. Healthcare, artificial intelligence and

the fourth industrial revolution: Ethical, social and legal consid-
erations. South African Journal of Bioethics and Law, 11(2):93–95,
2018.
[Maina and Singh 2020] Anthony M Maina and Upasana G Singh.

Why national ehealth strategies matter-an exploratory study of
ehealth strategies of african countries. In 2020 International Con-
ference on Electrical and Electronics Engineering (ICE3), pages 670–
675. IEEE, 2020.
[Makovhololo 2018] Modikwa Luwi Makovhololo. Effects of gwea

implementation on ict standardisation across sa government de-
partments. In 2018 Open Innovations Conference (OI), pages 339–
345. IEEE, 2018.
[Mars and Seebregts 2008] Maurice Mars and Chris Seebregts.

Country case study for e-health south africa. Online)
URL: http://ehealthconnection. org/files/resources/County% 20Case%
20Study% 20for% 20eHealth% 20South% 20Africa. pdf.[Accessed:
12 March 2019], 2008.
[Martin 2016] POHL Martin. Robotic systems in healthcare with par-

ticular reference to innovation in the ‘fourth industrial revolu-
tion’. Journal of International and Advanced Japanese Studies, 8:17–
33, 2016.
[Masilela et al. 2013] Thulani Clifford Masilela, Rosemary Foster, and

Matthew Chetty. The ehealth strategy for south africa 2012-2016:
how far are we? South African Health Review, 2013(1):15–24, 2013.
[Masood and Sonntag 2020] Tariq Masood and Paul Sonntag. Indus-
try 4.0: Adoption challenges and benefits for smes. Computers
in Industry, 121:103261, 2020.
[Matsoso and Fryatt 2013] Malebona Precious Matsoso and R Fryatt.

National health insurance: the first 18 months. SAMJ: South
African Medical Journal, 103(3):154–155, 2013.
Bibliography 255
[Mayosi 2012] B. Mayosi. Health in south africa: changes and chal-

lenges since 2009. The Lancet, 20, 2012.
[McArthy 2001] J. McArthy. Phenomenal data mining from data to

phenomenah. Computer Science Dpt, 2001.
[Mead 2006] C. Mead. Data interchange standards in healthcare it-

computable semantic interoperability: Now possible but diffi-
cult, do we need a better mousetrap. Journal of Health Information
Management, 20(1), 2006.
[Mechael 2009] P. Mechael. The case for mhealth in developing coun-

tries. Innovations: Technology, Governance, Globalization, 4(1):103–
118, 2009.
[Mesko 2017] B. Mesko. A guide to artificial intelligence in health-

care. The Medical Futurist, 2017.
[Metagroup 2005] Metagroup. Research and Development in Systems,

2005.
[Mettler 2016] Matthias Mettler. Blockchain technology in healthcare:

The revolution starts here. In 2016 IEEE 18th international confer-
ence on e-health networking, applications and services (Healthcom),
pages 1–3. IEEE, 2016.
[Meyers et al. 2012] D. Meyers, J. Durlak, and A. Wandersman. The

quality implementation framework: A synthesis of critical steps
in the implementation process. AJPsych, 50:462–480, 2012.
[Mgudiwa and Iyamu 2018 04] S. Mgudiwa and T. Iyamu. Integration

of social media with healthcare big data for improved service delivery,
2018-04.
[Middleton 2016] Middleton. Motor and cognitive functional status

are associated with 30 day unplanned rehospitalisation follow-
ing postacute care in medicare fee-for-service beneficiaries. Gen-
eral Internal Medicine, 31(12):1427–34, 2016.
[Milne et al. 2019] Barry J Milne, June Atkinson, Tony Blakely, Hilary
Day, Jeroen Douwes, Sheree Gibb, Meisha Nicolson, Nichola
Shackleton, Andrew Sporle, and Andrea Teng. Data resource
profile: The new zealand integrated data infrastructure (idi). In-
ternational journal of epidemiology, 2019.
[Mittalstadt and Floridi 2015] B. Mittalstadt and M. Floridi. The

Ethics of Big Data: Current and Foreseeable Issues in Biomedical Con-
texts,, 2015.
[Morrison 2016] Ciarán Morrison. Defining digital health. 2016.

256 Bibliography
[Moyo 2012] B.M. Moyo. Health in south africa: changes and chal-
lenges since 2009,. The Lancet, 380(9858), 2012.
[Mudaly et al. 2013] T. Mudaly, A. Pillay, and C. Seebregts. Architec-

tural Frameworks for Developing National Health Information Sys-
tems in Low and Middle Income Countries,, 2013.
[Mulder 2015] G Mulder. Principles of wound management and en-

hanced tissue repair for the podiatric physician. JAPMA, 2015.
[Murdoch and Detsky 2013] T.B. Murdoch and A.S. Detsky. The in-
evitable application of Big Data to healthcare,. Journal of Ameri-
can Medical Association, 309(13):1351–60, 2013.
[Mwachofi 2008] A. Mwachofi. Who killed healthcare?: America’s 2

trillion dollar medical problem and the consumer driven cure,.
JCI, 118(1):5–5, 2008.
[Myers and Klein 2011] Michael D Myers and Heinz K Klein. A set
of principles for conducting critical research in information sys-
tems. MIS quarterly, pages 17–36, 2011.
[Nacinovich 2011] M. Nacinovich. Defining mHealth, 2011.
[Nadkarni et al. 2011] P. Nadkarni, L. Machado, and W. Chapman.

Natural Language Processing: An Introduction. Journal of Amer-
ican Information Association, 18:54–551, 2011.
[Nat 2012] National eHealth Strategy 2012 to 2017, 2012.
[Nations 2015] United Nations. 17 Sustainable Development Goals -

SDG, 2015.
[NDOH 2003] NDOH. South African National Health Act No. 61 of 2003,
2003.
[NDOH 2005] NDOH. South African National Health Act No. 61 of 2003,
2005.
[NDOH 2014a] NDOH. HNSF Document., 2014.
[NDOH 2014b] NDOH. National Health Normative Standards Frame-

work for Interoperability in eHealth in South Africa, 2014.
[NDOH 2015] NDOH. White Paper on National Health Insurance,, 2015.
[NDOH 2016] NDOH. South African Health News-NHI, 2016.
[NDOH 2017] NDOH. South African National Health Insurance (NHI)

White Paper, 2017.
[NEEDHAM 2010] W NEEDHAM. Government-wide enterprise ar-

chitecture (gwea) framework implementation guide. South
Africa: Government Information Technology Officer’s Council (GI-
Bibliography 257
TOC) of South Africa, 2010.
[Neveol and Zweigenbaum 2015] A. Neveol and P. Zweigenbaum.

Clinical Natural Language Processing in 2014: Foundational
Methods Supporting Efficient Healthcare. IMIA Year Book of
Medical Informatics, 10(1):1200–1234, 2015.
[Nguyen 2015] Nguyen. Functional status at discharge and 30 day

readmission risk in copd,. Respiratory Medicine, 109(2):238–246,
2015.
[Ngwenya 2018] Mandlenkosi Ngwenya. Health systems data interop-

erability and implementation. PhD thesis, 2018.
[Niranjanamurthy et al. 2019] M. Niranjanamurthy, B. Nithya, and

S. Jagannatha. Analysis of blockchain technology: pros, cons
and swot. Cluster Computing, 22:18, 11 2019.
[NIST 2015] NIST. National Institute of Standards and Technology: Big

Data, 2015.
[Nonaka 1994] I. Nonaka. A dynamic theory of organisational knowl-

edge creation.,. Organizational Science, 5:14–37, 1994.
[Nonaki 1991] I. Nonaki. The knowledge creating company,. Harvard

Business Review, page 96–104, 1991.
[Novak and Bridwell 2019] Karina Novak and Larry Bridwell. The
future of healthcare in africa. In Competition Forum, volume 17,
pages 431–438. American Society for Competitiveness, 2019.
[Nuti et al. 2014] S. Nuti, B. Wayda, I. Ranasighe, S. Wang, R. Dreyer,

and S. Chen. The use of Google Trends in Healthcare research:
A Systematic Review,. PLoS One Journal, 9:1–49, 2014.
[Obagi et al. 2019] Z. Obagi, O. Toledo, G. Damiani, A. Grada, and

V. Falanga Langa. Principles of wound dressings: A review. Sur-
gical technology international, 35, 2019.
[Ochian et al. 2014] A. Ochian, G. Sucio, O. Fratu, and V. Sucilo. An

overview of cloud middleware services for the interconnection
of healthcare platforms,. IEEE, 978(1):4, 2014.
[Olson et al. 2013] D. Olson, K. Konty, M. Paladini, C. Vibound, and

L. Simonsen. Reassessing the Google Flu Trends data for De-
tection of Seasonal and Pandemic Influenza: A comparative epi-
demiological study at three geographical scales,. PLoS One Jour-
nal, pages 1–11, 2013.
[Oracle 2014] Oracle. Mega Project Management: Reducing Risk and

Complexity Across the Value Chain, 2014.
258 Bibliography
[Ortiz et al. 2011] J. Ortiz, H. Zhou, and K. Newzil. Monitoring in-

fluenza activity in the United States: A comparison of tradi-
tional surveillance systems. Plos One Journal, 6:1–12, 2011.
[Ouma et al. 2011] Stella Ouma, Marlien Herselman, and D Van Gre-
unen. Factors that influence m-health implementations in re-
source constrained areas in the developing world. 2011.
[Ouma 2013] Stella Ouma. An m-health user experience framework

for the public health care sector. 2013.
[Pagallo 2017] Ugo Pagallo. When morals ain’t enough: Robots,

ethics, and the rules of the law. Minds and Machines, 27(4):625–
638, 2017.
[Pang et al. 2018] Zhibo Pang, Geng Yang, Ridha Khedri, and Yuan-
Ting Zhang. Introduction to the special section: convergence
of automation technology, biomedical engineering, and health
informatics toward the healthcare 4.0. IEEE Reviews in Biomedical
Engineering, 11:249–259, 2018.
[Partington 2002] David Partington. Essential skills for management re-

search. Sage, 2002.
[Passchier 2017] RV Passchier. Exploring the barriers to implement-

ing national health insurance in south africa: the people’s per-
spective. South African Medical Journal, 107(10), 2017.
[Payne and Yang 2014] Christopher J Payne and Guang-Zhong Yang.

Hand-held medical robots. Annals of biomedical engineering,
42(8):1594–1605, 2014.
[Pearson et al. 2012] Jake Pearson, David Rowlands, and Ruth Highet.
Autologous blood injection to treat achilles tendinopathy? a ran-
domized controlled trial. Journal of sport rehabilitation, 21(3):218–
224, 2012.
[Peffers et al. 2007] Ken Peffers, Tuure Tuunanen, Marcus A Rothen-

berger, and Samir Chatterjee. A design science research method-
ology for information systems research. Journal of management
information systems, 24(3):45–77, 2007.
[Perry 1998] A. Perry. Healthcare Financial Management. MIT, Spring

ed., NY, 1998.
[Pervaiz et al. 2012] F. Pervaiz, M. Pervaiz, and S. Umar. Flubreaks:

Early epidemic detection from Google Flu Trends,. Journal of
Medical Information Research, 14(5):1–9, 2012.
[Peters and Lipsky 2013] Edgar JG Peters and Benjamin A Lipsky. Di-
agnosis and management of infection in the diabetic foot. Medi-
cal Clinics, 97(5):911–946, 2013.
Bibliography 259
[Peters et al. 2018] Brian S Peters, Priscila R Armijo, Crystal Krause,

Songita A Choudhury, and Dmitry Oleynikov. Review of emerg-
ing surgical robotic technology. Surgical endoscopy, 32(4):1636–
1655, 2018.
[Peters 2014] D. Peters. Republished research: Implementation re-

search: What it is and how to do it,. British Journal of Sports
Medicine, 48(8):7, 2014.
[Petersen et al. 2015] Virginia Petersen, Caesar Vundule, and Carin

Koster. South africa’s socpen: A legacy system which still
delivers. Jakarta. http://www. opml. co. uk/sites/default/files/Poster
20South 20Africa_FINAL, 2015.
[Piateski and Frawley 1991] G. Piateski and W. Frawley. Knowledge

Discovery in Databases (KDD). MIT Press, NY, 1991.
[Pickwell et al. 2015] K. Pickwell, V. Siersma, M. Kars, M. Edmonds,

and P. Holstein. Predictors of lower-extremity amputation in pa-
tients with an infected diabetic foot ulcer,. Diabetes Care, 38:852–
857, 2015.
[Polanyi 1997] M. Polanyi. “Tacit Knowledge,” in Knowledge in Organi-

sations, 1997. Chapter 7.
[Politzer 2015] E. Politzer. Creating a better future: Four scenarios

for how digital technologies could change the world,. Journal of
International Affairs, 72(1), 2015.
[Porter and Teisberg 2008] M. Porter and E. Teisberg. How physi-

cians can change the future of health care,. Journal of American
Medical Association, 297(10):1103–11012, 2008.
[Porter et al. 2016] Michael E Porter, Stefan Larsson, and Thomas H

Lee. Standardizing patient outcomes measurement. New Eng-
land Journal of Medicine, 374(6):504–506, 2016.
[Porter 2009] M. Porter. A Strategy for Healthcare Reform-Toward a

value-based System,, 2009.
[Porter 2010a] M. Porter. Value in Healthcare. New England Journal of

Medicine, 363(2):2477–81, 2010.
[Porter 2010b] Michael E Porter. Measuring health outcomes: the out-

comes hierarchy. N Engl J Med, 363:2477–81, 2010.
[Pramanik et al. 2019] P. Pramanik, B. Upadhyaya, S. Pal, and T. Pal.

Internet of things, smart sensors, and pervasive systems: Enabling
connected and pervasive healthcare in Healthcare Data Analysis,
2019.
260 Bibliography
[Prather et al. 1997] J.C. Prather, L. Lobach, L.K. Goodwin, and M.L.
Hage. Medical data mining: knowledge discovery in a clinical
data warehouse,. Health Management Technology, page 101–105,
1997.
[Pregibon 1997] N. Pregibon. Data mining. Statistical Computing and

Graphics, US, 1997.
[Press 2000] Microsoft Press. “Microsoft Briefing,” in Microsoft Product

Launch, 2000.
[Prisecaru 2016] Petre Prisecaru. Challenges of the fourth industrial

revolution. Knowledge Horizons. Economics, 8(1):57, 2016.
[Provost and Fawcett 2013] Foster Provost and Tom Fawcett. Data
science and its relationship to big data and data-driven decision
making. Big data, 1(1):51–59, 2013.
[Radebe 2021] Fani Radebe. Towards a mobile application to aid law

enforcement in diagnosing and preventing mobile bully-victim
behaviour in eastern free state high schools of south africa. 2021.
[Ragupathi and Ragupathi 2014] W. Ragupathi and V. Ragupathi.

Big data analytics in healthcare: promise and potential,. Health
Information Science and Systems, no.:2–3, 2014.
[Randell et al. 2014] Rebecca Randell, Joanne Greenhalgh, Jon Hind-

marsh, Dawn Dowding, David Jayne, Alan Pearman, Peter
Gardner, Julie Croft, and Alwyn Kotze. Integration of robotic
surgery into routine practice and impacts on communication,
collaboration, and decision making: a realist process evaluation
protocol. Implementation Science, 9(1):52, 2014.
[Razavian et al. 2015] N. Razavian, S. Blecker, A. Maclallan, S. Nigam,

and D. Sontag. Population level prediction of type 2 dia-
betes from claims data to and analysis risk factors,. Big Data,
3(4):277–282, 2015.
[Reddy and Sharma 2016] Priyananadan Reddy and Brahm Sharma.

Digitalisation: The future of health care. Journal of Business Man-
agement, (11), 2016.
[Ricketts and Silva 2017] C. Ricketts and C. Silva. An analysis of mor-

bidity and mortality using Google Trends,. Journal of Human
Behaviour in the Social Environment, 27:559–570, 2017.
[Riek 2017] Laurel D Riek. Healthcare robotics. arXiv preprint

arXiv:1704.03931, 2017.
[Roche 2006] Ltd Roche. FDA Approves Herceptin for the Adjuvant
Treatment of HER2-Positive Node-Positive Breast Cancer, 2006.
Bibliography 261
[Romero et al. 2016] David Romero, Johan Stahre, Thorsten Wuest,

Ovidiu Noran, Peter Bernus, Åsa Fast-Berglund, and Dominic
Gorecky. Towards an operator 4.0 typology: a human-centric
perspective on the fourth industrial revolution technologies.
In International conference on computers and industrial engineering
(CIE46) proceedings, 2016.
[Rotich and Onyancha 2017] D. Rotich and L. Onyancha. Trends and

patterns of medical and health research at the University of
Kenya between 2002 and 2014: an Informatics study,. Journal
of Library Science and Information, 82(2):20–33, 2017.
[Rupali 2017] J. Rupali. Looking through a social lens: conceptualis-

ing social aspects of knowledge management for global health
practitioners,. Journal of Public Health Research, 6:761–768, 2017.
[Ruxwana 2010] N. Ruxwana. Ict applications as e-health solutions in

rural healthcare in the Eastern Cape province of South Africa,.
Health Information Management Journal, 39(1), 2010.
[Ruxwana 2014] N. Ruxwana. A generic quality assurance model

(GQAM) for successful e-health implementations in rural hos-
pitals in South Africa,. Health Information Management Journal,
43(1):26–36, 2014.
[Ryan 2006] Anne B Ryan. Post-positivist approaches to research. Re-

searching and Writing your Thesis: a guide for postgraduate students,
pages 12–26, 2006.
[Santoni de Sio and van Wynsberghe 2016] Filippo Santoni de Sio

and Aimee van Wynsberghe. When should we use care robots?
the nature-of-activities approach. Science and Engineering Ethics,
22(6), 2016.
[Savova et al. 2010] Guergana K Savova, James J Masanz, Philip V

Ogren, Jiaping Zheng, Sunghwan Sohn, Karin C Kipper-Schuler,
and Christopher G Chute. Mayo clinical text analysis and
knowledge extraction system (ctakes): architecture, component
evaluation and applications. Journal of the American Medical In-
formatics Association, 17(5):507–513, 2010.
[Schabetsberger et al. 2010] Thomas Schabetsberger, Florian Wozak,

Basel Katt, Richard Mair, Bernhard Hirsch, et al. Implemen-
tation of a secure and interoperable generic e-health infrastruc-
ture for shared electronic health records based on ihe integra-
tion profiles. In Medinfo, pages 889–893, 2010.
[Schmarzo 2016] B. Schmarzo. Big Data MBA: Driving Business Strate-

gies with Data Science. Wiley, USA, 2016.
262 Bibliography
[Schonfeldt et al. 2011] Anzel Schonfeldt, Thulani Masilela, Peter Bar-

ron, and Rene English. Health information systems in south
africa. South African health review, 2011(1):81–89, 2011.
[Schootman et al. 2015] M. Schootman, A. Toor, P. Rehg, D.B. Jeffe,

and A. McQeen. The utility of Google Trends data to examine
the interest in cancer screening,. British Medical Journal, pages
1–8, 2015.
[Schreuder and Verheijen 2009] HWR Schreuder and RHM Verheijen.

Robotic surgery. BJOG: An International Journal of Obstetrics and
Gynaecology, 116(2):198–213, 2009.
[Schwab and Davis 2018] Klaus Schwab and Nicholas Davis. Shaping
the future of the fourth industrial revolution. Currency, 2018.
[Schwab 2017] Klaus Schwab. The fourth industrial revolution. Cur-

rency, 2017.
[Scott et al. 2018] J. Scott, R. Dunscombe, D. Evans, M. Mukherjee,

and J. Wyatt. Learning health systems need to bridge the two
cultures of clinical informatics and data science,. Innovation in
Health Informatics, 25(2):126, 2018.
[Seara et al. 2016] G. Seara, A. Paya, and J. Mayol. Value-based health-

care delivery in the digital era,. Journal of Euro Psychology,
1(860):1, 2016.
[Sedano et al. 2011] FJ Farfán Sedano, M Terrón Cuadrado, Y Castel-

lanos Clemente, P Serrano Balazote, D Moner Cano, and M Rob-
les Viejo. Patient summary and medicines reconciliation: appli-
cation of the iso/cen en 13606 standard in clinical practice. Stud
Health Technol Inform, 166:189–196, 2011.
[Seebregts et al. 2006] C Seebregts, M Mars, C Fourie, Y Singh, and

K Weyer. Inexpensive open source tb and hiv electronic medical
record system (openmrs) in south africa collaborating toward
an emr for developing countries. In Proceedings of the AMIA
Symposium, pages 11–15, 2006.
[Seebregts et al. 2018] Christopher Seebregts, Pierre Dane, Annie Neo

Parsons, Thomas Fogwill, Debbie Rogers, Marcha Bekker, Vin-
cent Shaw, and Peter Barron. Designing for scale: optimising
the health information system architecture for mobile mater-
nal health messaging in south africa (momconnect). BmJ global
health, 3(Suppl 2):e000563, 2018.
[Serfontein 2016] J. Serfontein. The unanticipated operational com-

plexities in the nhi,. Johannesburg SA, 2016.
Bibliography 263
[Sharma and Kshetri 2020] Ravi Sharma and Nir Kshetri. Digital
healthcare: Historical development, applications, and future research
directions, 2020.
[Sharma et al. 2018] N. Sharma, A. Panwar, and U. Sugandh. Big data

in healthcare: A literature survey,. International Journal of Recent
Research Aspects, 5:127–133, 2018.
[Shid 2014] Shid. Functional status ouptperforms comorbidities in

predicting an acute care readmission in medically complex pa-
tients,. Gen Intern Medicine, 30(110):1427–34, 2014.
[Shinde 2016] K. Shinde. Real time monitoring system in healthcare

with hadoop,. Research Journal, 1(1):15–19, 2016.
[Siegfried et al. 2017] Nandi Siegfried, Thomas Wilkinson, and Karen

Hofman. Where from and where to for health technology as-
sessment in south africa? a legal and policy landscape analysis.
South African Health Review, 2017(1):42–48, 2017.
[Singh et al. 2012] Nagendra K Singh, Deepak K Gupta, Pawan K

Jayaswal, Ajay K Mahato, Sutapa Dutta, Sangeeta Singh, She-
fali Bhutani, Vivek Dogra, Bikram P Singh, Giriraj Kumawat,
et al. The first draft of the pigeonpea genome sequence. Journal
of plant biochemistry and biotechnology, 21(1):98–112, 2012.
[Singh et al. 2017] Deepika Singh, Johannes Kropf, Sten Hanke, and
Andreas Holzinger. Ambient assisted living technologies from
the perspectives of older people and professionals. In Interna-
tional Cross-Domain Conference for Machine Learning and Knowl-
edge Extraction, pages 255–266. Springer, 2017.
[Sirkin et al. 2005] Harold L Sirkin, Perry Keenan, and Alan Jackson.
The hard side of change management. HBR’s 10 Must Reads on
Change, 99, 2005.
[Skilton and Hovsepian 2017] Mark Skilton and Felix Hovsepian. The
4th Industrial Revolution: Responding to the Impact of Artificial In-
telligence on Business. Springer, 2017.
[Ślusarczyk 2018] Beata Ślusarczyk. Industry 4.0: Are we ready? Pol-

ish Journal of Management Studies, 17, 2018.
[Smith 2013] N. Smith. Econex-Trade competition and Applied Economics,

2013.
[Soley-Bori 2014] Soley-Bori. Functional status and hospital readmis-

sion using the medical expenditure panel review,. General Inter-
nal Medicine, 30(7):965–72, 2014.
[Sonnier 2016] Paul Sonnier. Definition digital health. Paul Sonnier-

Story of Digital Health, 2016.
264 Bibliography
[Sou 2016] NDoH: National Strategic Plan/National Health Insurance.

www.health.gov.za./., 2016. [Accessed 10 09 2019].
[Spidlen 2004] J. Spidlen. MUDRLite – Health Record Tailored to your

needs,, 2004.
[Stahl 2008] Bernd Carsten Stahl. The ethical nature of critical re-
search in information systems. Information systems journal,
18(2):137–163, 2008.
[Stevens and Chen 2017] M. Stevens and J.H. Chen. Machine learn-
ing and prediction in medicine: Beyond the peak inflated expec-
tations,. NEJM, page 2507–2508, 2017.
[Sugrue et al. 2016] R. Sugrue, E. Carthy, M. Kelly, and K. Sweeney.

Science or popular media: What drives breast cancer online
activity? The Breast Journal: Short Communication, 24:189 – 193,
2016.
[Talby 2019] D. Talby. Lessons learned building natural language process-

ing systems in Healthcare, 2019.
[Tan et al. 2016] Alan Tan, Hutan Ashrafian, Alasdair J Scott, Sam E
Mason, Leanne Harling, Thanos Athanasiou, and Ara Darzi.
Robotic surgery: disruptive innovation or unfulfilled promise?
a systematic review and meta-analysis of the first 30 years. Sur-
gical endoscopy, 30(10):4330–4352, 2016.
[Tan 2005] J. Tan. E-Health Care Information Systems: An Introduction

for Students and Professionals. Jossey-Bass, California, 2005.
[Tedesco 2018] Tedesco. Impact of rehabilitation on mortalitiy and

readmissions after surgery for hip fracture,. BMC Health Sciences
Research, 18:701, 2018.
[Telfer and Woodburn 2015] S. Telfer and J. Woodburn. Let me

Google that for you: a time series analysis of seasonality in in-
ternet search trends for terms related to foot and ankle pain,.
Journal of Foot and Ankle, 27(8):1–10, 2015.
[Telkom 2015] Telkom. Tomorrow Starts Today: Healthcare and Technol-

ogy: ICT Challenges in the SA Healthcare Industry,, 2015.
[Theron 2016] N. Theron. Comments on select aspects of the NHI White

Paper,, 2016.
[Thomas 2013] Nontuthuzelo Thomas. Electronic health record: Op-

portunities to harness this technology locally. 2013.
[Tiffany et al. 2016] C. Tiffany, B. Liu, K. Bozic, and K. Teisberg. Value-

base healthcare: Person centered measurement: Focusing on the
three c’s,. BMC Health Services Research, 475(10):315–317, 2016.
Bibliography 265
[Toffler and Alvin 1980] Alvin Toffler and Toffler Alvin. The third
wave, volume 484. Bantam books New York, 1980.
[Tomlinson et al. 2013] Mark Tomlinson, Mary Jane Rotheram-Borus,

Leslie Swartz, and Alexander C Tsai. Scaling up mhealth: where
is the evidence? PLoS medicine, 10(2), 2013.
[Townsend 2013] H. Townsend. The promise and progress of NLP

for improved care,. Journal of American Health Information Associ-
ation, 84(2):44–45, 2013.
[Trivedi 2018] G. Trivedi. Towards interactive Natural Language Pro-

cessing. In Clinical Care in IEEE International Conference on Health-
care Informatics. USA, 2018.
[Valentijn and Vrijhoef 2017] P. Valentijn and B. Vrijhoef. Value-based

integrated care: Exploring strategies to enhance the uptake of
integrated care,. In 17th International Conference on Integrated.
Care, Dublin, 2017.
[Valentijn et al. 2016] P. Valentijn, C. Biermann, and M. Bruijnzeels.

Value based integrated(renal) care: setting a development
agenda for research and implementation strategies,. BMC Health
Service Research, 16:330, 2016.
[Vaquero et al. 2019] J. Vaquero, A. Encinas, A. Dios, J. Bullon,

A. Nova, J. Gonzalez, and C. Carbajo. Review of wearables to
Monitor Foot Temperature in Diabetic patients,. MDPI, 17:776–
783, 2019.
[Vayena and Blasimme 2018] E. Vayena and A. Blasimme. The Ethics

of Big Data: Current and Foreseeable Issues in Biomedical Contexts,,
2018.
[Venable et al. 2017] John R Venable, Jan Pries-Heje, and Richard L

Baskerville. Choosing a design science research methodology.
2017.
[Venkantraman and Loh 1994] N. Venkantraman and N. Loh. The

shifting logic of the it organisation: From technical portfolio to
relationship portfolio,. The Executives Journal, 10(2):5–11, 1994.
[Vigilante et al. 2019] Kevin Vigilante, Steve Escaravage, and Mike

McConnell. Big data and the intelligence community—lessons
for health care. New England Journal of Medicine, 380(20):1888–
1890, 2019.
[Walsham 2005] Geoff Walsham. Learning about being critical. Infor-

mation Systems Journal, 15(2):111–117, 2005.
[Wamba et al. 2015] S. Wamba, S. Akter, G. Chopin, and D. Gnanzou.

How ‘big data’ can make big impact: Findings from a systematic
266 Bibliography
review and a longitudinal case,. International Journal of Produc-

tion Economics, 12(31):34, 2015.
[Wang and Alexander 2016] L. Wang and C. Alexander. Big data an-
alytics as applied to diabetes management. European Journal of
Clinical and Biomedical Sciences, 2(5):29–38, 2016.
[Wang et al. 2018a] Y. Wang, L. Kung, and T.A. Byrd. Big data analyt-
ics: Understanding its capabilities and potential benefits,. Tech-
nological forecasting and Social Change, 126:3–13, 2018.
[Wang et al. 2018b] Yanshan Wang, Liwei Wang, Majid Rastegar-

Mojarad, Sungrim Moon, Feichen Shen, Naveed Afzal, Sijia Liu,
Yuqun Zeng, Saeed Mehrabi, Sunghwan Sohn, et al. Clinical
information extraction applications: a literature review. Journal
of Biomedical Informatics, 77:34–49, 2018.
[Weber and Stein 2018] Lynne M Weber and Joel Stein. The use of
robots in stroke rehabilitation: a narrative review. NeuroRehabil-
itation, 43(1):99–110, 2018.
[Weber 2018] AS Weber. Emerging medical ethical issues in health-

care and medical robotics. International Journal of Mechanical En-
gineering and Robotics Research, 7(6), 2018.
[Weeks 2013] R. Weeks. The convergence of systemic threads shap-

ing a future south african healthcare dispensation: A technology
management perspective,. Acta Commercii, 13(1), 2013.
[Weisinger 2016] D. Weisinger. Formtek-Cloud Computing: Business

Move to the ’Third Platform’ — the Nexus of Forces,, 2016.
[Weiss and Indutkya 1997] S.M. Weiss and N. Indutkya. Predictive

Data Mining: A practical guide. Morgan, NY, 1997.
[Westphal and Biaxten 1998] C. Westphal and T. Biaxten. Data Min-

ing solutions: Methods and Tools for Solving RealWorld Problems.
Wiley, NY, 1998.
[WHO 2006] WHO. Building foundations for eHealth: progress of Mem-

ber States: report of the WHO Global Observatory for eHealth. World
Health Organization, 2006.
[WHO 2012] WHO. National eHealth strategy toolkit. International

Telecommunication Union, 2012.
[WHO 2013] WHO. WHO country cooperation strategy 2013-2014:

South africa. 2013.
[WHO 2016a] WHO. Millenium Development Goals, 2016.
[WHO 2016b] WHO. Monitoring and evaluating digital health inter-

ventions: a practical guide to conducting research and assess-
Bibliography 267
ment. 2016.
[WHO 2016c] WHO. World Health Organisation SDG’s Millenium De-

velopment Goals: MDG’s to SDG’s, 2016.
[Wiederhold 2017] Brenda K Wiederhold. Robotic Technology Remains

a Necessary Part of Healthcare’s Future Editorial, 2017.
[Wiig 2004] K. Wiig. Knowledge management: An introduction and

perspective,. Journal of Knowldge Management, 1, 2004.
[Wiley and Mathews 2017] L. Wiley and G. Mathews. Health care

system transformation and integration: A call to action for pub-
lic health,. Journal of Law, Medicine and Ethics, 45(1):94–97, 2017.
[Witten 2000] I. Witten. Data Mining. Morgan Kaufman, NY, 2000.
[Wolmarans et al. 2014] M. Wolmarans, W. Solomon, G. Tanna, J. Ven-

ter, A. Parsons, M. Chetty, and M. Dombo. ehealth programme
reference implementation in primary health care facilities. South
African Health Review, 2014(1):35–43, 2014.
[Yin 1989] R. Yin. Case Study Research. Sage Pub, California, 1989.
[Young and Zhang 2017] S. Young and O. Zhang. Using search en-
gine big data for predicting new hiv diagnoses. Plos One Journal,
13(7):1–8, 2017.
[Young 2016] Montgomery Young. Private vs. public healthcare in

south africa. 2016.
[Zhang et al. 2013] Angela J Zhang, Michele Garret, and Steven

Miller. Bullosis diabeticorum: case report and review. The New
Zealand Medical Journal (Online), 126(1371), 2013.
[Zhang et al. 2015] Z. Zhang, X. Zheng, D. Zeng, and S. Leischow. In-

formation seeking regarding Tobacco and Lung Cancer: Effects
of Seasonality. PloS One, 10:1–11, 2015.
[Zhang et al. 2018] Peng Zhang, Douglas C Schmidt, Jules White, and
Gunther Lenz. Blockchain technology use cases in healthcare. In
Advances in computers, volume 111, pages 1–41. Elsevier, 2018.
[Zhang et al. 2019] Yaoyun Zhang, Firat Tiryaki, Min Jiang, and Hua
Xu. Parsing clinical text using the state-of-the-art deep learning
based parsers: a systematic comparison. BMC medical informatics
and decision making, 19(3):53–114, 2019.
[Zheng et al. 2021] Ting Zheng, Marco Ardolino, Andrea Bacchetti,

and Marco Perona. The applications of industry 4.0 technolo-
gies in manufacturing context: a systematic literature review.
International Journal of Production Research, 59(6):1922–1954, 2021.
268 Bibliography
[Zhou et al. 2011] X. Zhou, J. Ye, and Y. Feng. Tuberculosis surveil-

lance by analysing Google Trends,. IEEE, 58(8):1–14, 2011.
colophon
T
he work is based on research supported by the University of
Johannesburg of South Africa. Any opinion, findings and con-
clusions or recommendations expressed in this material are
those of the author and therefore the University of Johannesburg does
not accept liability in regard thereto.
This document was typeset using the typographical look-and-feel

classicthesis developed by André Miede. The style was inspired by
Robert Bringhurst’s seminal book on typography “The Elements of Ty-
pographic Style”. classicthesis is available for both LATEX and LYXat:
https://bitbucket.org/amiede/classicthesis/. The author of this
thesis has made many modifications to the classicthesis template
to arrive at this final version
South African Healthcare and the Fourth Industrial Revolution: New

Applications of Technology
Simon Abbott | 2022

abbott821@gmail.com
Copyright © University of Johannesburg, South Africa

Abbott SG Wtms

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Abbott SG Wtms

Uploaded by

Copyright:

Available Formats

COPYRIGHT AND CITATION CONSIDERATIONS FOR THIS THESIS/ DISSERTATION

How to cite this thesis

Simon George Abbott

A thesis submitted for the degree of

Information Technology Management

Department of Applied Information Systems

Supervisor: Professor Abejide Ade-Ibijola

Copyright © University of Johannesburg, Johannesburg, South Africa

All rights reserved. No part of this publication may be stored in a

This work is dedicated to my sister for her continuous support.

I, Simon G Abbott, hereby declare the contents of this doctoral the-

[1] Abbott, S.G. and Ade-Ibijola, A. (2018). Architectural analysis of

[2] Abbott, S.G. and Ade-Ibijola, A. (2019). Trend analysis: a deci-

[3] Abbott, S.G. and Ade-Ibijola, A. (2019). Algorithms and a tool

[4] Abbott, S.G., Ade-Ibijola, A. and Roestenburg, W. (2021). Big

[5] Abbott, S.G. and Ade-Ibijola, A. (2022). An Architectural Anal-

[6] Abbott, S.G. and Ade-Ibijola, A. (2022). Big Data in Healthcare.

For their support during this Doctoral programme, I give thanks to

• Formal Structures Lab — Nikita Patel, Lloyd Radebe, Safwaan Pa-

Sponsors. This research was supported by the following grants and

• University of Johannesburg Supervisor-Linked Bursary (2018 to

• Prof. Abejide Ade-Ibijola’s research funds (for other expenses

1. Architectural analysis of the South African healthcare landscape: we

2. Healthcare robotic designs: we have presented the design of a mod-

3. Trend analysis: we have presented a case scenario of searching

4. Clinical note decryption tool: we have presented an algorithm for

5. Clinical chatbot: we have presented a clinical chatbot application

1. Medical healthcare sciences, an in-depth insight into the eco-

2. Computing methodologies, artificial intelligence, natural lan-

1. Presented The Impacts of 4iR on the South African Healthcare

2. Presented An Architectural Analysis of the South African

3. Presented An Architectural Analysis of the SA Healthcare

4. Presented a public lecture on the analysis of the South African

i introduction, literature and related work 1

3.1.14 Summary of SA Healthcare Landscape . . . . . 39

ii theoretical background and research method-

5.2.2 DSR in Information Systems or Technology . . 103

iiicontributions of 4ir technologies to south african

Google trends scenarios - South African health-

iv robotics and mobile applications 163

10.2.1 mHealth . . . . . . . . . . . . . . . . . . . . . . . 194

v evaluation and conclusion 212

Figure 1 The economic overview of the SA healthcare

Figure 21 Trend analysis of cancer, diabetes, HIV and tu-

Figure 40 The MedBot conversational screening desktop.

Table 2 Practitioner perception of AI and digital tools

LIC Low Income Countries

SRL Semantic Role Labelling

The South African healthcare industry is currently under-

In addition, the WHO is prescribing eHealth policies, glob-

The construct of the vision of the impact of the Fourth Industrial

The United Nations (UN) Millennium Development Goals (MDG’s)

The South African National Department of Health (NDoH) has

Further to this initiative, is the World Health Organisation (WHO)

South Africa is in dire need to upgrade and modernise its pub-

1.1 problem definition

In this section, we define the problem addressed in this thesis.

By presenting the challenges faced by the South African healthcare

The South African public healthcare sector is lacking in imple-

the application of artificial intelligence (ai): To ex-

the resultant impacts of such ai applications: To imple-

the application of such ai applications: To evaluate addi-

Further, this includes the convergence of health and technology

1.1.2 New applications of technologies in South African healthcare