Professional Documents
Culture Documents
Ttat Hit
Ttat Hit
By
BAHAE SAMHAN
DOCTOR OF PHILOSOPHY
MAY 2016
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
ProQuest 10139705
Published by ProQuest LLC (2016). Copyright of the Dissertation is held by the Author.
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
© Copyright by BAHAE SAMHAN, 2016
All Rights Reserved
To the Faculty of Washington State University:
___________________________________
K.D. Joshi, Ph.D., Chair
___________________________________
Kenneth Butterfield, Ph.D.
___________________________________
Terence Saldanha, Ph.D.
ii
ACKNOWLEDGMENTS
committee, Prof. K.D. Joshi for her continuous support of my Ph.D study, for her patience,
motivation, and immense knowledge. I am very fortunate to have had you as my advisor and as
the chair of my dissertation committee. I could not have imagined having a better mentor for my
Ph.D study. I am confident that your guidance will have a lasting positive impact on my future
career. There are no words great enough to express my sincerest appreciation and gratitude to all
I would also like to thank the rest of my dissertation committee members: Dr. Kenneth
Butterfield, and Dr. Terence Saldanha, for their insightful comments and encouragement. Dr.
Saldanha, you have been of tremendous help for me. You agreed to step in and join my
dissertation committee when it was most-needed. I could not imagine completing my dissertation
without your presence, perspective, and guidance. Dr. Butterfield, I really appreciate you for
always being there for me. I will be forever grateful for your continuous support for me as a
member on my dissertation committee and as well as my department chair. You are a real role
model to me. I sincerely hope that our relationship continues throughout my career.
My sincere thanks also goes to Dr. Mauricio Featherman, I honestly cherished each
moment I worked with you teaching students how to program. I have learned a lot from you and
it has been my honor to have you as a great mentor and a close friend. I trust that our relationship
Special thanks goes to Prof. Deborah Compeau. Although we shared a relatively short
period of time at WSU, but I am forever grateful for all of your constructive feedback that you
iii
never hesitated to provide on my work. Thank you for your wisdom and continuous
encouragement.
Next, I would like to thank my family for all of their support, love, and encouragement
throughout this process. To my great father Dr. Marouf Samhan, you have always been the
source of my inspiration. If I can be half as good a scholar as you have always been, then that
will be my greatest achievement. Thank you for all of your teachings, wise advise, endless
support, and encouragement. To my dear mom Aida, you are everything to me. Thank you for
always being there for me. You always encouraged me to pursue my academic career because
you believed in me. I am very happy that I was able to make your dream come true. To my
lovely wife Sara, I am truly blessed to have you in my life. I deeply acknowledge all of what you
had to go through to make this journey possible. Thank you for believing in me and always
teaching me how to make the impossible reachable. I would have not been able to accomplish
this without you in my life. To our beautiful sons Adam and Alex, you have gave true meaning
to my life and a reason to keep going no matter what. My brother Waseem, thank you for always
being my “go-to” person when I needed help. My sweet sister Katherine, thank you for always
reminding me that when the going gets tough, the tough gets going. I am truly blessed to have all
of you in my life.
Finally, I must mention the students in my cohort who have been an important part of this
process. Majid Dadgar, Joseph Taylor, and Benyawareth (Yaa) Nithithanatchinnapat. I am truly
grateful to have had you all along. Majid, you have been more than a brother to me. I would also
like to thank my senior cohort: Tanya Beaulieu, Chris Callif, Xiaolin Lin, and Todd Martin. I am
grateful that I have had you as colleagues and friends. Special thanks goes to my dear friend
from the junior cohort, Mina Jafarijoo. Talking with you has always reenergized me and kept me
sane. I am sure you will be finishing your dissertation sooner than you can imagine. To everyone
else that I may have unintentionally forgot, please accept my apologies and I truly thank you.
iv
“One child, one teacher, one book, and one pen can change the world. Education is the only
v
WHY DO PEOPLE RESIST HEALTHCARE IT? LITERATURE ANALYSIS, MODEL
Abstract
Health Information Technology (HIT) has the potential of improving the overall
resistance and avoidance behaviors. IS researchers have provided valuable insights about these
behaviors towards IT in general. Very limited work has aimed to explain these behaviors in
healthcare settings and especially towards Electronic Health Record (EHR) systems. In this
dissertation, we investigate the resistance and avoidance behaviors in the context of a hospital,
specifically having the EHR system as the focal technology for my study. Overall, the
dissertation is comprised of five chapters and subscribes to a mixed-methods approach. The first
chapter will provide an extensive review the literature to highlight the current status of research
on technology resistance and the main research gaps. In the second chapter, we empirically
validate the User Resistance Model (Kim and Kankanhalli, 2009). We aim to explain the
healthcare providers’ resistance to change from the paper-based recording system to the new
EHR system. Based on an analysis of survey data from healthcare providers, we investigate the
effects of the key determinants from URM on EHR resistance. The third chapter is a qualitative
vi
research in which we collect responses from healthcare providers using open-ended questions.
We use the Reasoned Causal Mapping (RCM) methodology to uncover the main predictors of
EHR resistance and the key concepts shaping those predictors. In the fourth chapter, we aim to
explain the EHR avoidance behaviors after the implementation of the system is complete and has
been enforced to all users. We adopt the Technology Threat Avoidance Theory (TTAT) (Liang
and Xue, 2009) and empirically test the complete proposed conceptual model. Based on an
analysis of survey data from healthcare providers, we investigate the effects of perceived threats
and perceived avoidability on avoidance motivations, and the direct effects of avoidance
motivation on avoidance behavior. In the fifth chapter, we use RCM to reveal the main
constructs impacting EHR avoidance as well as the key concepts forming these constructs. We
analyze qualitative data collected from healthcare providers using open-ended questions.
vii
TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS............................................................................................. iii
ABSTRACT ................................................................................................................... vi
1. ABSTRACT ...................................................................................................... 1
2. INTRODUCTION ............................................................................................. 2
3. METHODOLOGY ............................................................................................ 4
5. IMPLICATIONS ..............................................................................................14
viii
6. FUTURE WORK .............................................................................................15
7. REFERENCES .................................................................................................17
1. ABSTRACT .....................................................................................................24
2. INTRODUCTION ............................................................................................25
User Resistance...........................................................................................43
7. METHODOLOGY ...........................................................................................46
8. RESULTS ........................................................................................................49
ix
9. DISCUSSION OF FINDINGS..........................................................................53
Contribution to Research.............................................................................58
1. ABSTRACT .....................................................................................................73
2. INTRODUCTION ............................................................................................74
Data Analysis..............................................................................................83
6. RESULTS ........................................................................................................86
x
EHRS Perceived Value ...............................................................................92
7. DISCUSSION ....................................................................................................97
1. ABSTRACT ...................................................................................................108
xi
Sample and Data Collection ......................................................................121
1. ABSTRACT ...................................................................................................145
5. RESULTS ......................................................................................................152
xii
Emotion Focused Coping .......................................................................... 157
APPENDIX
xiii
LIST OF TABLES
3.3 An Illustration of the 5-Step Procedure for Constructing Causal Map ...............84
xiv
3.9 EHRS Perceived Value Revealed Key Concepts ...............................................93
4.5 Items Loadings and CFA Model Fit Statistics ................................................. 143
4.9 Suggested Model Fit Indices TTAT Model Fit Indices ....................................131
5.1 An Illustration of the 4-Step Procedure for Constructing Causal Maps ............ 150
5.2 Revealed Constructs of EHR Avoidance and Its Concepts .............................. 153
xv
LIST OF FIGURES
5.3 Aggregated Revealed Causal Map of EHR Threat Avoidance ......................... 159
xvi
Dedication
For my dad, who has been there for me with every step.
xvii
CHAPTER ONE:
ABSTRACT
argued that HIT is being resisted by the same people who are expected to use it the most. IS
research in this area is still considered as under-researched. Prior Research has mainly focused
on the adoption aspect of the technology with limited effort towards explaining resistance
behaviors especially in healthcare organizations. HIT investments are high and expectations of
getting better performance, after its implementation are thus high. HIT, if meaningfully used, has
the potential of changing the way of applying healthcare services and highly benefit the
technology was resisted. This paper provides an extensive review of the literature to uncover the
current status of the research in this area, to highlight the main research gaps that need to be
addressed, and to reveal how relatively little attention has been paid to understand HIT
resistance. Aiming to provide a research agenda that would encourage scholars to conduct
1
CHAPTER ONE:
INTRODUCTION
(Bhattacherjee and Hikmet, 2007; and Poon et al, 2004), and despite the effort of IS researchers
in this area, it still remains understudied (Karahanna et al, 1999). This paper provides an
extensive review of the literature to uncover the current status of the research in this area, to
highlight the main research gaps that need to be addressed, and to reveal how relatively little
attention has been paid to understand HIT resistance. User resistance to information systems has
been identified as one of the salient reasons for the failure of new systems (Kim and
Kankanhalli, 2009). This is evident in the healthcare setting, (Spil et al, 2004) developed a ‘USE
IT’ model describing four dimensions of HIT use based on users’ needs: resistance, relevance,
requirements, and the availability of resources. The authors found that resistance was the
strongest determinant of not-using the electronic prescription systems (EPS). The technology
was not used in 72% of the cases, and that resistance was the cumulative consequence of the
remaining three determinants. Another example was reported on the Cedars-Sinai Medical
Center at Los Angeles where doctors resisted to use the newly installed Computerized Physician
Order Entry (CPOE) system, which caused the system to fail and result in a complete withdrawal
from implementation after it was already implemented in two-thirds of the 870-bed hospital
(Bhattacherjee et al, 2013). System failure caused by resistance has a potential of negative
impacts on users. This is very crucial in the healthcare setting where impacts are directly aimed
on patients. Result could be life-threatening cases, for example, the resistance of using a blood
2
sugar monitor by a type 2 diabetic patient may result in many health complications, especially if
the patient is in rural area and must rely on a mobile monitoring device to alert the doctor and the
HITs are expected to increase patient safety, reduce cost of healthcare delivery, and
improve efficiency in the healthcare industry (Bhattacherjee and Hikmet, 2007), yet these
technologies, in many cases, are facing wide resistance (Lapointe and Rivard, 2005). User
resistance to new HITs needs to be understood and managed in order to mitigate HIT failures
There is an extensive body of research that has focused on technology adoption. It started
in 1989 when the Technology Acceptance Model (TAM) was first introduced (Davis, 1989).
Today, More than 345 articles focusing on technology acceptance have been published in the top
19 journals of the IS community in the last 20 years (Williams et al, 2009). However, it is argued
that technology in healthcare context has unique characteristics that makes it different from the
more general environments in which technology acceptance models have been successfully
tested in prior studies, therefor these models have less explanatory power when applied in
healthcare settings (Smith et al., 2014). Also, Practitioners usually use complex and difficult
medical equipment, so “hard to use” fails to explain the non-acceptance for them (Hare et al,
2006). Difficulty of use has a potential effect on patients’ perspectives towards using or resisting
the HIT. However, no studies, that includes patients in the examination of HIT resistance, were
found. Moreover, these theories are designed to explain voluntary IT usage and have limited
(Bhattacherjee et al, 2013). Thus, it is believed that applying the acceptance models in healthcare
3
settings is not sufficient to completely explain the individuals’ behaviors towards new HIT and a
different perspective is still required to provide more insights regarding user resistance to HIT.
Some early IS research was conducted to uncover the concept of resistance, for example,
(Jha et al, 2008; Bhattacherjee and Hikmet, 2007; and Markus, 1983). However, the review
conducted by (Lapointe and Rivard, 2005), looked into 20 IS/IT journals over the past 25 years
and found 43 articles that considered resistance as an implementation issue. They reported that
only 9 of the 43 articles had defined resistance. Resistance is not only under researched but also
shapes a practice phenomenon in the healthcare sector, healthcare lags behind other industries in
disciplines to describe the current status of this literature and to analyze the literature to provide
researchers with a clear understanding of the emerging research gaps that need to be addressed in
future studies.
METHODOLOGY
To uncover the current status of the research and find the main research gaps, we
examined the extant literature. As shown in Figure 1.1, we started with conducting an extensive
search for technology resistance in different IS and healthcare journals and key IS conferences.
The keywords used were: IS, IT, adoption barriers, resistance, rejection, and healthcare.
4
Figure 1.1: Search, review, and analysis process
This search resulted in 121 articles that covered the four areas of research as depicted in
Figure 1.2.
5
Figure 1.2. Literature Review Space
Section 1 included studies about HIT with no resistance behavior, so these papers were
considered irrelevant to our research. Section 2 included papers about resistance to healthcare but
not to the technology used. These papers were also considered irrelevant to our research. Section
3 included papers that looked into resistance to IT, in general, not specifically in healthcare
settings. These papers are considered relevant to our research because HIT is merely a subset of
understanding of resistance in the healthcare settings. Section 4 included the most relevant
papers of our research, these papers looked into the resistance behavior towards technology in
healthcare settings.
Only 47 studies were considered as relevant after eliminating all irrelevant articles. The
papers were then coded, analyzed and synthesized in order to highlight the emerging gaps in this
area.
6
EMERGING RESEARCH GAPS
This study aims to uncover the research gaps in the IT resistance literature with a focus
on HIT resistance. In order to reveal the research gaps of HIT resistance it was essential to
uncover those of IT resistance in general, and then format it in the context of healthcare. To find
those gaps, It was first necessary to characterize the current status of the IS research by
summarizing the literature that studied IT resistance in both the general context and the
healthcare context. It was clear that despite researchers’ attempts to understand IT resistance, this
area is still considered under researched (Laumer and Eckhardt, 2012). There are critical gaps in
the IT resistance literature, especially in healthcare settings, that need to be addressed (Poon et
al, 2004).
In the next section we provide a list of the research gaps revealed from the literature
review and analysis. Table 1.1 provides a visual summary of these gaps.
Prior research that examined IT resistance, explained the resistance to IT using a sub-set
of factors that shape resistance behavior, such as loss of power (Markas and Hornik, 1996),
political concerns (Lin and Ashcraft, 1990), evaluation of the net gain and the change in the
status quo (Joshi, 1991), fear of losing jobs because of the introduction to the new technology
(Yoon et al, 1995), bad experiences perceived from the change to a newer technology and lack of
user involvement (Gill, 1996). However, (Karahanna et al, 1999) argued that prior perspectives
on IT resistance are still valuable, but the antecedents of IT resistance evolve and expanded over
time as individuals start to use IT more regularly and on different scales in different areas. The
study conducted by (Bhattacherjee and Hikmet, 2007) found that the prior research on IT
resistance is still limited and fragmented with no unified theory to explain the resistance behavior
7
towards technologies. In a recent study, (Laumer and Eckhardt, 2012) reviewed 9 IS models and
theories of resistance and concluded that IS literature still lacks a unified theory of user
resistance. The review conducted in this study indicates that the IS literature still needs to find
ways to define and measure IT resistance beyond the prior perspectives and demands a unified
framework explaining the IT resistance phenomenon and how similar or different is it from the
Patients are major stakeholders in the healthcare process Campbell et al, 2009). Today
the technology is advancing rapidly and is facilitating healthcare options to patients through a
variety of methods such as medical mobile apps that are widely spread and easily accessed by
patients (Ahmad et al, 2002). Also there is a wide spread of diagnostic tools that are available
and easily accessed over the web (Yager et al, 2006). However in most of the HIT studies such
as (Bhattacherjee and Hikmet, 2007; Joshi, 1991; and Lapointe and Rivard, 2005) the focus was
only on the practitioners, physicians, nurses, managers or Clinical laboratory personnel but not
the patients. It is important to understand who is resisting the technology and why are they
resisting it. This includes not only practitioners but also patients. Section 2 depicted in figure1 is
the area of articles that looked into patients’ resistance to healthcare not to the technology. These
patients could be resisting the HIT as a result of resisting the care itself. The indirect resistance
of patients towards healthcare occurred via the technology. While examining patients’ resistance,
it is crucial to explain what patients are actually resisting, and uncouple the resistance behavior
to understand whether the patients are rejecting the technology or the care itself.
The literature showed limited work on patients’ resistance to HIT. This is a major
research gap and researchers are encouraged to include patients in their future research designs.
8
Patients might have different reasons for resisting the technologies and uncovering these reasons
HIT resistance is very critical as it has a direct effect on human health, the resistance of
HIT may result in many medical errors that could lead to serious injuries and deaths (Poon et al,
2004). However, HIT resistance is clearly present in the U.S. according to (Jha et al., 2008) who
conducted a study comparing the level of HIT adoption in seven developed nations, and revealed
that U.S.A ranked last along with Canada with only 10 to 30 percent adoption rate. Also, within
the U.S. organizations, health technologies are being resisted more than technologies in different
HIT resistance is still not well understood (Laumer and Eckhardt, 2012), and a better
understanding of user resistance to HIT may help design better systems that are functional and
also adoptable by their potential users (Shah et al, 2006), it will also help in conducting
appropriate interventions to minimize the resistance behavior (Poon et al, 2004). However, the
review of the literature showed that very little attention has been given to the resistance of HIT.
In 2004 Poon et al., (Poon et al, 2004) found that this critical area is was considered under
researched in the IS literature. Ten years later, still very few studies are found on HIT resistance,
only 17 articles were conducted in the healthcare setting. Researchers are encouraged to give
more attention to this critical area and conduct research that will benefit academic research and
provide practitioners with a better understanding of the phenomenon and potentially suggest
The Nature of Technology Support- System Functionalities and Supported Work Processes
9
The literature review uncovered a number of different technologies used when examining
performed task has a potential impact on individuals’ behavior towards the technology (Haymes,
2008), and could possibly change the behavior of potential users towards resistance. HIT used in
prior research confined to electronic health records (EHR), computerized patient orders entry
systems (CPOES) and other systems used for front desk clerk tasks such as admission, discharge,
and transfer tasks. Exploring the different technologies used for different work processes, such as
diagnostics, care giving, monitoring, etc., and examining resistance towards those technologies
The prior literature defined and conceptualized resistance in a variety of ways. Resistance
behaviors were characterized by low levels of use, by lack of use, or by dysfunctional use
(Martinko et al, 1996). One study conducted by Keen (Keen, 1981) defined resistance as “social
inertia”. Some other definitions were based on the concept of resistance to change, e.g., (Zaltman
and Duncan, 1977), others defined it based on behavioral expressions and opposition to use, e.g.,
(Markus, 1983). Not only defined differently, resistance was also explained in a variety of ways;
Markus (Markus, 1983) explained the resistance behavior as a result of individuals perception on
power gain or loss associated with the change. This perspective is similar to the net gain concept
that Joshi (Joshi, 1991) proposed in his findings, he posits that users react to change by
evaluating their net gain based on change in their inputs and outcomes and comparing their
relative outcomes with that of other user groups and the employer. Bahattacherjee and Hikmet
10
(Bhattacherjee and Hikmet, 2007) is one of the few papers that explains resistance to change as
an important aspect of organizational change. They studied resistance in the healthcare setting
and also explained resistance differently building on the dual factor structure of IT usage
proposed by (Cenfetelli, 2004) that posits inhibitors as the main reason of resistance, defining
inhibitors as the negative factors that discourage IT usage when present, but have no effect when
absent.
This summary of IT resistance conceptualized in the extant literature illustrates that this
construct is multifaceted and needs to be developed further. The IS scholars should strive for a
conceptualization of the resistance behavior will impede scholarship in the research area which is
in its incipient stages. A rich and deep conceptualization of this multifaceted construct will allow
IS research has provided limited insights on the role of behavioral and social beliefs in
shaping IT resistance. Prior research mainly focused on perceived threat as the belief that shifts
individuals’ behavior towards resisting the technology (Laumer, 2011). This is evident in studies
that focused on IT resistance as well as on HIT resistance. For example, (Markas and Hornik,
Hikmet (Bhattacherjee and Hikmet, 2007) used it to explain HIT resistant. However, perceived
threat is a broad concepts that may include a number of antecedents and causes. Prior research
has characterized threat differently, Bahattacherjee and Hikmet (Bhattacherjee and Hikmet,
2007) explained threat as physician’s loss of control over their work, which is similar to that
observed by (Lapointe and Rivard, 2005). Another attempt to explain resistance using the
11
concept of threat is the work of (Kim and Kankanhalli, 2009) who defined threat as the perceived
costs verses the perceived benefits that potential users consider before rejecting the technology.
This is similar to the findings of Joshi (Joshi, 1991). Perceived threat still has no one unified
definition or explanation. Other beliefs were considered when explaining resistance; (Cenfetelli,
2004) categorized potential users’ beliefs into inhibitors and enablers that would discourage or
encourage the usage behavior. However, a general explanation was given to show how inhibitors
discourage usage with no further investigation on what these inhibitors could be. (Markas and
Hornik, 1996) proposed that both fear and stress generated from the introduction of the
technology into the world of the potential user has impact on the resistance behavior. They
explained how we must be mindful that people “do not resist new technology, rather they resist
the effect that the new technology may have on their own lives.” This goes in hand with the
study of (Martinko et al, 1996) that also posited that resistance is affected by individuals’ beliefs.
They explained that negative outcomes result from individuals’ beliefs towards technology
usage. (Enns et al., 2003) studied how uncertainty and the loss of status can change the balance
Not only behavioral believes require further theorization, also social beliefs impact must
be considered when researching resistance. Opinions and behavior of individuals are influenced
by their social networks (Laumer and Eckhardt, 2012). This is explained by the social impact
theory which shows how social influence has an important impact on individuals’ behaviors.
influence of other users (Keen, 1981). However, limited attention has been devoted to the study
of social influence on resisting technologies within organizations (Laumer, 2011). Also limited
attention has been given to the social influence stemmed from private sources outside of the
12
organization such as family and friends (Laumer and Eckhardt, 2012). Our review found that
only (Eckhardt et al, 2009) studied social influence on individual resistance. They expanded the
work of UTAUT (Venkatesh et al, 2003) to explain the impact of social influence on non-
adoption intentions, and these influences were limited to work place colleagues and superiors not
including the influence of other individuals outside of the organizations’ boundaries, such as
friends and family. This is very important to examine, especially when measuring its impact on
patients resistance to HIT. Patients, in most cases are not employees of the healthcare facility,
and in some cases might be resisting these technologies while being remote from the healthcare
facility, therefor the influence of employees at the facility has limited impact on patients’
decisions. Articles in section 2 of figure1 revealed that patients resist receiving care from the
health providers for reasons related to their personal beliefs such as religion. When examining
resistance to HIT, it is important to uncouple patients’ resistance to the technology and their
resistance to the actual care based on their beliefs and previous experiences. Moreover, patients
may be resisting care as a result of resisting the technology implemented. Both behavioral and
social beliefs impact on resistance are interesting areas that still needs more investigating.
The IS scholars should strive for a more comprehensive understanding of the role of
individuals’ believes in IT resistance. Understanding the impact of beliefs on resistance and how
these beliefs are related helps in developing a better understanding of resistance (Laumer and
Eckhardt, 2012).
The review of prior literature indicates that not only demographic variables of the
potential users affect IT resistance, but also personality factors such as levels of flexibility, desire
for change and growth, and tolerance for ambiguity are closely related to resistance (Majchurzak
13
et al, 1988). The case study of IT resistance at the Hmong American Partnership (Vang, 2008)
revealed that “the one antecedent that withstood time in relation to user acceptance or rejection
of information technology was individual indifferences, such as a user’s age, education, and
gender”. IS research has considered individual differences in a variety of studies. However, few
2011) only one study (Klaus et al, 2010) considered individual differences when examined IT
resistance.
The literature review confirms that few number of articles studied the impact of
individual differences on IT resistance and none were conducted in the healthcare setting.
Studying the impact of individual differences on IT resistance could result in interesting findings
IMPLICATIONS
This study has a number of implications for research. First, our analysis of the current
status of resistance in the IS literature was not only limited to HIT resistance, rather it provided a
14
general analysis on IT resistance including healthcare settings. Second, it provides research gaps
analysis and highlights a research agenda for future studies that are encouraged to build on the
findings and address the emerging research gaps that were looked over in the past. Third, it
summarized the comprehensive description of the core concepts of resistance to IT and HIT.
Finally, the study encourages researchers to start aiming their work towards finding a unified
theory of resistance that considers individual believes and differences, as well as to include
user resistance to IT, system designers will be able to build better systems that are functional and
also adoptable by their potential users (Poon et al, 2004). Furthermore, understanding how
resistance could be treated will have a positive impact on system success (Kim and Kankanhalli,
2009). This paper seeks to find answers to IT resistance in general and focuses on the HIT
resistance behavior. Reducing the resistance to HIT by patients and practitioners will help in
avoiding catastrophic medical errors (Kohn et al, 2000) and will help in improving the overall
quality, efficiency and performance of the healthcare provider (Bhattacherjee et al, 2006). This
study aims to encourage researchers to build on the emerging gaps to provide clear guidelines
that potentially could help in building and implementing technologies with lower rates of
resistance.
FUTURE WORK
We hope that this study will spark interest in the IT resistance phenomena in general and
unexplored yet potentially fertile area of research. In particular, future work needs to design
15
research studies that are patient centric. Future research needs to decouple the effects of the
factors that contribute to patients’ resistance to healthcare from their resistance to IT used to
support the care they are receiving. We encourage scholars to address the research gaps
uncovered in this study by conducting empirical research in this area in an aim to discover the
significant antecedents shaping the HIT resistance behavior. We also hope that this study will
stimulate more theoretical contributions in this area to form a better understanding of the
technology resistance and designers in building more adoptable systems. The main contribution
of this paper is in shaping the future research agenda by revealing the research gaps in this
critical area. By revealing the research status this work encourages researchers to focus on the
key research gaps while conducting studies related to IT and HIT resistance.
16
REFERENCES
Agarwal, R 2000. "Individual acceptance of information technologies," Framing the domains of
IT management: Projecting the future through the past, pp. 85-104.
Agarwal, R, 2010. "Research commentary-The digital transformation of healthcare: current
status and the road ahead," Information Systems Research, 21, 4, pp. 796-809.
Ahmad, A., Teater P., Bentley T.D., Kuhen L., Kumar R.R., Thomas A., and Mekhijan H.S.,
2002. “Key attributes of a successful physician order entry system implementation in a
multi-hospital environment,” Journal of American Medical Informatics Association, 9, 1,
pp. 16–24.
Ajzen, I. and Fishbein, M., 1980. “Understanding attitudes and predicting social behavior,”
Englewood Cliffs, NJ: Prentice-Hall.
Ajzen, I., 1991. “The theory of planned behavior.” Organizational Behavior and Human
Decision Processes”, 50, pp. 179–211.
Anderson, Catherine L., and Ritu Agarwal, 2011. "The digitization of healthcare: boundary risks,
emotion, and consumer willingness to disclose personal health information," Information
Systems Research, 22, 3, pp. 469-490.
Ang, J., and Pavri, F., 1994. "A Survey and Critique of the Impacts of Information Technology,"
International Journal of Information Management, 14, 2, pp. 122-133.
Angst, C., R. Agarwal, V. Sambamurthy, and K. Kelley, 2010. “Social contagion and
information technology diffusion: The adoption of electronic medical records in U.S.
hospitals,” Management Science, 56, 8, pp. 1219–1241.
Baker, L., T. H. Wagner, S. Singer, M. K. Bundorf, 2003. “Use of the Internet and e-mail for
health care information: Results from a national survey,” Journal of American Medical
Association, 289, 18, pp. 2400–2406.
Bates, W., Leape, L.L., Cullen, D.J., Laird, N., Peterson, L.A., and Teich J.M., 1998. “Effect of
computerized physician order entry and team intervention on prevention of serious
medication errors,” Journal of American Medical Association, 280, 15, pp. 1311–1316.
Bhattacherjee, A., and Hikmet, N., 2007. "Physicians' resistance toward healthcare information
technology: a theoretical model and empirical test," European Journal of Information
Systems, 16, 6, pp. 725-737.
Bhattacherjee, A., Hikmet, N., Menachemi, N., Kayhan, V. O., & Brooks, R. G., 2006. “The
differential performance effects of healthcare information technology adoption,”
Information Systems Management, 24, 1, pp. 5-14.
17
Bhattacherjee, A., and Hikmet, N., 2007. "Physicians' resistance toward healthcare information
technologies: a dual-factor model," 40th Annual Hawaii International Conference on
System Science, IEEE.
Bhattacherjee, A, Davis C., and Hikmet N., 2013. "Physician Reactions to Healthcare IT: An
Activity-Theoretic Analysis," 46th Hawaii International Conference on system science,
IEEE.
Bhattacherjee, A., Davis, C., Hikmet, N., and Kayhan, V., 2008. "User reactions to information
technology: Evidence from the healthcare sector," Association for Information Systems.
Borzekowski, R, 2002. “Health care finance and the early adoption of hospital information
systems,” Divisions of Research & Statistics and Monetary Affairs, Federal Reserve
Board. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=337960
Ram, B., Mala, Grocott, P., and Weir, H., 2008. "Issues and challenges of involving users in
medical device development," Health Expectations, 11, 1, pp. 63-71.
Campbell, S.M., Shield, T., Rogers, A., and Gask, L., 2004. "How do stakeholder groups vary in
a Delphi technique about primary mental health care and what factors influence their
ratings?" Quality and Safety in Health Care 13.6 (2004), pp. 428-434.
Campbell E., K. Guappone, D. Sittig, R. Dykstra, J. S. Ash., 2009. “Computerized provider order
entry adoption: Implications for clinical workflow,” Journal of General Internal
Medicine 24, 1, pp. 21–26.
Cenfetelli R.T., 2004. “Inhibitors and enablers and dual factor concepts in technology usage,”
Journal of the Association of Information Systems, 5, 11, (2004), pp. 472–492.
Chen, J.M., Suen, M.W., Chiu, S.C., and Lin, M.J., 2004. “Social dilemma studies in Taiwan,”
International Journal of Psychology, Vol. 39, No. 5-6, pp. 577-577.
Chen, P., and Hitt, L. M., 2002. “Measuring Switching Costs and the Determinants of Customer
Retention in Internet-Enabled Online Brokerage Industry,” Information Systems
Research, 13, 2, pp. 255-274.
Chiasson, M., E. Davidson, 2002. “Getting the two to dance: Examining barriers to health
information systems research in mainstream IS journals”, Proceedings of the 35th Annual
Hawaii International Conference on System Science, Big Island, HI, pp. 1966-1974.
Chismar, William G., and Wiley-Patton, S., 2003. "Does the extended technology acceptance
model apply to physicians," System Sciences Proceedings of the 36th Annual Hawaii
International Conference on System Science. IEEE.
Coetsee, L. D., 1993. "A Practical Model for the Management of Resistance to Change: An
Analysis of Political Resistance in South Africa," International Journal of Public
Administration, 16, 11, pp. 1815-1858.
18
Coetsee, L. D., 1999. "From Resistance to Commitment," Public Administration Quarterly, 23,
2, pp. 204-222.
Cutler, D. M., N. E. Feldman, J. R. Horwitz, 2005.“U.S. adoption of computerized physician
order entry systems,” Health Affairs, 24, 6, (2005), pp. 1654–1663.
Davis, F. D., 1989. “Perceived Usefulness, Perceived Ease of Use, and User Acceptance of
Information Technology,” MIS Quarterly, 13, 3, pp. 319-340.
DeSanctis, G., and Courtney, J.F., 1983. "Toward Friendly User MIS Implementation,"
Communications of the ACM, 26, 10, pp. 732-738.
Doolin, B., 2004. "Power and resistance in the implementation of a medical management
information system," Information Systems Journal, 14.4, pp. 343-362.
Eckhardt, A., Laumer S., and Weitzel, T., 2009. "Who influences whom and quest; analyzing
workplace referents' social influence on IT adoption and non-adoption," Journal of
Information Technology, 24, 1, pp. 11-24.
England, I., D. Stewart, S. Walker, 2000. “Information technology adoption in health care:
When organizations and technology collide,” Australian Health Review, 23, 3, pp. 176–
185.
Enns, H.G., Huff, S.L., and Higgins, C.A., 2003. "CIO Lateral Influence Behaviors: Gaining
Peers' Commitment to Strategic Information Systems," MIS Quarterly, 27, 1, pp. 155-
175.
Fishbein, Martin, and Ajzen, I., 1975. “Belief, attitude, intention and behavior: An introduction
to theory and research.”
Fisher, S., 2008. “Improving Physician Adoption of CPOE Systems,”
http://www.strategiestoperform.com/volume3_issue2/docs/ImprovingPhysicianAdoption.
pdf
Ford, Jeffrey D., Laurie W. Ford, and Angelo D'Amelio, 2008. "Resistance to change: The rest
of the story,” Academy of Management Review, 33, 2, pp. 362-377.
Freudenhiem, M., 2004. “Many hospitals resist computerized patient care,” New York Times.
http://www.nytimes.com/2004/04/06/technology/06errors.html.
Gill, T. Grandon, 1996. "Expert systems usage: task change and intrinsic motivation," MIS
quarterly, pp.301-329.
Goh, Jie Mein, Guodong Gao, and Ritu Agarwal, 2011. "Evolving work routines: adaptive
routinization of information technology in healthcare," Information Systems Research,
22,3, pp. 565-585.
Goldschmidt, Peter G., 2005. "HIT and MIS: implications of health information technology and
medical information systems," Communications of the ACM, 48, 10, pp. 68-74.
19
Hare, K., Whitworth, B., Deek, F., and Norris, T., 2006. “Identifying Barriers to IT Adoption in
Clinical Health Care: A Pilot Study,” Health Care and Informatics Review Online, ISSN
1176-4201, http://hcro.enigma.co.nz/website/index.cfm
Haymes, T., 2008. "The three-e strategy for overcoming resistance to technological change,"
Educause Quarterly, 31, 4, pp. 67-69.
Hirschheim, R., and Newman, M., 1988. “Information Systems and User Resistance: Theory and
Practice,” Computer Journal, 31, 5, pp. 398-408.
Holmes, Thomas J., and James Schmitz, 1995. "Resistance to new technology and trade between
areas." Federal Reserve Bank of Minneapolis Quarterly Review, 19, 1, pp. 2-17.
Jha, A. K., Doolan, D., Grandt, D., Scott, T., and Bates, D. W., 2008. “The use of health
information technology in seven nations,” International journal of medical informatics,
77, 12, pp. 848-854.
Jiang, James J., Waleed A. Muhanna, and Gary Klein, 2000. "User resistance and strategies for
promoting acceptance across system types," Information & Management, 37, 1, PP. 25-
36.
Joshi, K., 1991. “A Model of User Perspective on Change: The Case of Information Systems
Technology Implementation,” MIS Quarterly, 15, 2, pp. 229-242.
Karahanna, Elena, Detmar W. Straub, and Norman L. Chervany, 1999. "Information technology
adoption across time: A cross-sectional comparison of pre-adoption and post-adoption
beliefs,” MIS quarterly, 23, 2.
Kazley, A., Y. Ozcan, 2007. “Organizational and environmental determinants of hospital EMR
adoption: A national study,” Journal of Medical Systems, 31, 5, pp. 375–384.
Keen, P.G., 1981. "Information Systems and Organizational Change," Communications of the
ACM, 24, 1, pp. 24-33.
Kim, H.W., and Kankanhalli, A., 2009. "Investigating user resistance to information systems
implementation: a status quo bias perspective,” MIS quarterly, 33, 3.
Klaus, T., S. C. Wingreen and J. E. Blanton, 2010. "Resistant Groups in Enterprise System
Implementations: A Q-Methodology Examination," Jouransl of Information Technology,
25, 1, pp. 91-106.
Kohn, Linda T., Janet M. Corrigan, and Molla S. Donaldson, 2000. "To err is human: building a
safer health system: A report of the Committee on Quality of Health Care in America,”
Institute of Medicine.
Lapointe, L., and Rivard, S., 2005. “A Multiple Model of Resistance to Information Technology
Implementation,” MIS Quarterly, 29, 3, pp. 461-491.
Lapointe, L., and Rivard, S., 2007. “A Triple Take on Information System Implementation”,
Organization Science, 18, 1, pp.89-107.
20
Latané, B., 1981. “The psychology of social impact,” American Psychologist, 36, pp. 343-356.
Laumer, S., and Eckhardt, A., 2012. "Why do people reject technologies: a review of user
resistance theories," Information Systems Theory. Springer New York, pp. 63-86.
Laumer, S., 2011. “Why Do People Reject Technologies–A Literature-Based Discussion of the
Phenomena Resistance to Change In Information Systems and Managerial Psychology
Research”
Lee, H.G., and Clark, T.H., 1997. "Market Process Reengineering through Electronic Market
Systems: Opportunities and Challenges," Journal of Management Information Systems,
13, 3, Winter, pp. 113-137.
Lehoux, P., 2008. “Why examining the desirability of health technology matters,” Politiques de
Santé.
Leonardi, Paul M., 2009. "Why do people reject new technologies and stymie organizational
changes of which they are in favor? Exploring misalignments between social interactions
and materiality,” Human Communication Research, 35, 3, 407-441.
Lin, E.M., and Phillip Ashcraft, 1990. "A case of system-development in a hostile environment,”
Journal of Systems Management 41, 4, pp. 11-14.
Lin, C., Lin I., and Roan J., 2012. "Barriers to physicians’ adoption of healthcare information
technology: an empirical Study on multiple hospitals," Journal of medical systems, 36, 3,
pp. 1965-1977.
Lorence, D.P., and Richards, M.C., 2003. “Adoption of regulatory compliance programs across
United States healthcare organizations: a view of institutional disobedience,” Health
Services Management Research, 16, 3, pp. 67–178.
Majchrzak, A., and Cotton, J., 1988. "A longitudinal study of adjustment to technological
change: From mass to computer‐automated batch production," Journal of Occupational
Psychology, 61, 1, pp. 43-66.
Marakas, G. M., and Hornik, S., 1996. “Passive Resistance Misuse: Overt Support and Covert
Recalcitrance in IS Implementation,” European Journal of Information Systems, 5, 3, pp.
208-219.
Markus M. L., 1983. “Power, Politics, and MIS Implementation”, Communications of the ACM,
26, 6, pp. 430-444.
Martinko, M. J., Henry, J. W., and Zmud, R. W., 1996. “An Attributional Explanation of
Individual Resistance to the Introduction of Information Technologies in the Workplace,”
Behavior and Information Technology, 15, 5, pp. 313-330.
Massaro T.A., 1993. “Introducing physician order entry at a major academic medical center:
Impact on organizational culture and behavior,” Academic Medicine 68, 1, pp. 20–25.
21
Meissonier, R., and Emmanuel H., 2009. "Attitudes to information technology in health care
professions," European Conference on Information Systems.
Mekhjian, S., Kumar R., Kuehn, Bentley, Teater, P., Thomas, A., Payne, B., and Ahmad A.,
2002. "Immediate benefits realized following implementation of physician order entry at
an academic medical center," Journal of the American Medical Informatics Association,
9, 5, pp. 529-539.
Gonzalez, P. A., and Chan, Y.E., 2013. “Exploring Healthcare Professionals’ Resistance to
Using Wireless Communication Technology,” Proceedings of the 19th Americas
Conference on Information Systems, Chicago, Illinois.
Parthasarathy, M., and Bhattacherjee A., 1998. “Understanding post adoption behavior in the
context of online services,” Information Systems Research 9, 4, 362–379.
Piderit, S.K., 2000. “Rethinking resistance and recognizing ambivalence: a multidimensional
view of attitudes toward an organizational change,” Academy of Management Review 25,
4, pp. 783-794.
Poon, E., Blumenthal, D., Jaggi, T., Honour, M., Bates, D.W., and Kaushal, R., 2004.
“Overcoming barriers to adopting and implementing computerized physician order entry
systems in U.S. hospitals,” Health Affairs 23, 4, pp. 184–190.
President’s Information Technology Advisory Committee (2004), Revolutionizing Health Care
through Information Technology, National Coordination Office for Information
Technology Research and Development, Arlington, Virginia. http://www.nitrd.gov
Shah, S., and Robinson I., 2006. "User involvement in healthcare technology development and
assessment: structured literature review,” International Journal of Health Care Quality
Assurance, 19, 6, pp. 500-515.
Smith, Tim, Gerald G., and Ramirez, A., 2014. "Investigating the influence of psychological
ownership and resistance on usage intention among physicians,” 47th Hawaii
International Conference on System Science, IEEE.
Spil, T. A., Schuring, R. W., and Michel-Verkerke, M. B., 2004. “Electronic prescription system:
do the professionals use it?” International Journal of Healthcare Technology and
Management, 6, 1, pp. 32-55.
Taylor, R., Manzo, J., and Sinnett, M., 2002. "Quantifying value for physician order-entry
systems: a balance of cost and quality." Healthcare financial management: journal of the
Healthcare Financial Management Association, 56, 7, pp. 44-48.
Kate, T., Haverkamp, S., Mahmood, F., and Feldberg, F., 2010. “Social network influences on
technology acceptance: A matter of tie strength, centrality and density,” BLED
Proceedings, Paper, 40.
22
Tulu, Bengisu, Richard J. Burkhard, and Thomas A. Horan, 2006. "Continuing use of medical
information systems by medical professionals: empirical evaluation of a work system
model,” Communications of the AIS, 18, 31, pp. 641–656.
Vang, T. S., 2008. “User resistance to information technology: A case study of the Hmong
American partnership,” ProQuest.
Venkatesh, V., M.G. Morris, G.B. Davis, and F.D. Davis, 2003. “User Acceptance of
Information Technology: Toward A Unifying View,” MIS Quarterly, 27, 3, pp. 425-478.
Venkatesh, V., and Susan A. B., 2001. "A longitudinal investigation of personal computers in
homes: adoption determinants and emerging challenges." MIS quarterly, pp. 71-102.
Williams, Michael D., Yogesh K. Dwivedi, Banita Lal, and Schwarz, A., 2009. "Contemporary
trends and issues in IT adoption and diffusion research," Journal of Information
Technology, 24, no. 1, pp. 1-10.
Yager, P., Edwards, T., Fu, E., Helton, K., Nelson, K., Tam, M. R., & Weigl, B. H., 2006.
“Microfluidic diagnostic technologies for global public health,” Nature, 442, pp. 412-
418.
Yoon, Y., and Guimaraes, T., 1995. "Assessing expert systems impact on users' jobs," Journal of
Management Information Systems, 12, 1, pp. 225-249.
Zaltman, G., and Duncan R., 1977. “Strategies for Planned Change,” John Wiley & Sons, New
York, NY.
23
CHAPTER TWO:
ABSTRACT
Despite the positive effects of electronic health records within healthcare practices, the
adoption rate of such systems is still low and faces resistance from healthcare providers. In this
paper we aim to understand why healthcare providers resist to change from the traditional
paper-based recording systems to the newly implemented electronic health record systems. Our
study tests the user resistance model using a questionnaire to collect data from a random sample
and enforced a new EHR system. Data was collected in 2 stages: pilot study and main study. 226
valid respondents participated in the main study survey. Results were then analyzed to describe
and evaluate resistance to change to the new EHR system. Our finding contributes to IS research
help EHR system developers in the process of updating, designing, and developing these systems
24
CHAPTER TWO:
INTRODUCTION
Electronic Health Records (EHR) are expected to bring a number of benefits to the health
organization. EHR can increase patient safety, reduce cost of healthcare delivery, and improve
efficiency in the healthcare industry (NITRD, 2004; Hillestad et al, 2005). However, switching
to EHR is challenging as it requires hospitals and its staff to change the way work is conducted
(Sheikh et al, 2011). For instance, in the U.S. the Meaningful-Use Mandate, enforced by the
Department of Health and Human Services, requires that all healthcare organizations must adopt
and use EHR meaningfully (Blumenthal and Tavenner, 2010). However, it has now been over
six years since this initiative was launched, and early reports show that it is still a challenge for
physicians and other healthcare providers to accomplish tasks using the new EHR systems
(Heath and Appan, 2014). In 2009 it was estimated that 73% of EHR implementations are, “not
using the system as intended 12 months after implementation” (Terry, 2012). This is evident on a
global scale. International case studies from 2013 showed that the U.S., the U.K., France, and
One of the primary challenges of EHR usage is the labor intensive data entry that
healthcare providers have little tolerance for because they believe it impedes with their primary
goal of taking care of their patients (Cantrill, 2010). Additionally, issues such as technical
problems and design difficulties (Cantrill, 2010), and security problems and data privacy issues
(Fu, 2009) also pose usage obstacles. These challenges are contributing in the shaping of the
25
of the phenomenon may help design systems that are functional, adoptable by their potential
users, and have higher success rates (Bhattacherjee and Hikmet, 2007; Shea and Hripcsak, 2010;
Despite that resisting to change to new EHR systems is a global phenomenon, this topic
is still considered underexplored (Samhan and Joshi, 2015). There is a lack of empirically tested
models aiming to explain “why” and “how” healthcare providers resist EHR systems (Boonstra
and Broekhuis, 2010; Weeger et al., 2011; Olaniran, 2015). Therefore, in an effort to explain
healthcare providers’ motivations to resist the switch to EHR systems, in this study we examine
the following research question - Why do healthcare providers resist to switch to EHR systems?
Because we aim to capture resistance at early stages of the EHR implementation it was
appropriate to investigate the resistance to change behaviors where healthcare providers are
switching from a Paper Based Recording System (PBRS) to a new EHR system.
To our knowledge, this work is the first attempt in the IS literature to apply a
implementation stage of a new EHR system. Our study conducts an empirical test of the User
Resistance Model (URM) (Kim and Kankanhalli, 2009) to assess resistance to change from
PBRS to EHR. We contribute to literature by extending and enhancing this model by, first,
applying the model in a healthcare context. Second, this study captures resistance to change
right after the implementation of the new EHR system, whereas this model in the past has been
posited and tested during the pre IT system implementation stage. This could serve as a
comparative between resistance at different stages (i.e. pre and post implementation). Third,
while URM provided a conceptual argument only on the effect of loss aversion on the status quo
bias, we empirically validate the principle of loss aversion in our model. Finally, while URM
26
limited social influence to that from peers and coworkers only, we include peers, colleagues, and
LITERATURE REVIEW
Prior research defined and conceptualized resistance to change differently. The work of
(Keen, 1981) viewed resistance to change from the costs-benefits perspective, he argued that
resistance to change occurs if the costs of change are perceived as greater than the benefits. This
is similar to the conceptualization used in the Equity Implementation Model (EIM) (Joshi, 1991)
which suggests that the greater the perceived inequity the greater the resulting distress would be,
and people with greater distress are more likely to resist it by attempting to minimize their inputs
and others’ outcomes as well as attempting to increase others’ inputs. In a more recent study
(Bahattacherjee and Hikmet, 2007) incorporated resistance to change with the Technology
Acceptance Model (TAM) (Davis, 1989). They explained user resistance to change by
explaining physician resistance to HIT that was a result of the perceived threats. This is similar
to the work of (Walter and Lopez, 2008) who posited that perceived threat influences intentions
to use which results in resistance to change behaviors. The User Resistance Theory (Kim and
Kankanhalli, 2009) posited the user resistance model which explained resistance to change from
associated with a new IS implementation. However, it is found that there is still a need for a
unified theory of resistance in IS (Samhan and Joshi, 2015). Further, our literature review
revealed that prior IS research has captured resistance to EHR systems mainly by capturing
lower levels of adoption. (Table 2.1) summarizes prior studies on EHR resistance.
27
Boonstra and Broekhuis, 2010 Lower Levels of EHR adoption Review of both Pre and Conceptual
Post Implementation Literature
Review
Burns et al, 2015 Work-Around Behaviors Post Implementation Survey
28
Ngafeeson and Midha, 2014 A covert or overt intention Post Implementation Survey
that opposes change towards
the use of an information
system
Nöltner and Krönung, 2015 Lower Levels of EHR Post Implementation Literature
Adoption Review
Meade et al, 2009 Lower Levels of EHR adoption Post Implementation Survey
Menachemi et al, 2007 Lower Levels of EHR adoption Post Implementation Survey
Miller et al, 2004 Lower Levels of EHR adoption Post Implementation Interviews
Olaniran, 2015 Lower Levels of EHR Post Implementation Conceptual
Adoption
Pizziferri et al, 2005 Lower Levels of EHR Post Implementation Qualitative
Adoption Observations
Poon et al, 2006 Opposing adoption of HIT Post Implementation Interviews
29
One of the most insightful reviews of EHR adoption (Najaftorkaman et al, 2015)
categorized prior findings into 8 main factors and argued that these factors has the potential to
impact EHR adoption levels. And that lower levels of adoptions would be interpreted as
resistance. These factors are: Individual factors such as age and gender, physiological factors
such as satisfaction and attitudes, behavioral factors such as behavioral change and automatic
factors such as size and age of the organization, financial factors such as start-up costs and on-
going costs, legal factors such as privacy and security issues, and finally technical factors such as
technical training and technical support. The work of (Desroches et al., 2008) investigated EHR
resistance in the same direction and found that the most common hurdles for EHR adoption are
financial cost related issues, not finding a system that meets healthcare providers’ needs, and
concerns that the EHR system would become “obsolete”. Other research emphasized on the
concept of low self-efficacy as the main contributor towards having lower EHR adoption rates
(Jha et al, 2008; Olaniran, 2015). The work of (Hung et al, 2013) also posited that computer self-
efficacy influences EHR adoption but also included perceived service levels (reliability and
responsiveness of the EHR system) along with perceived risks as contributors that impact
healthcare providers’ decision of EHR adoption. Additionally, the work of (Alghamdi, 2015)
posited that the main hurdles of EHR adoption are self-efficacy, privacy and security concerns
cost of implementation, and EHR maintenance “down time”. The task technology fit theory was
used to explain EHR resistance by lower levels of adoption (Gan and Cao, 2014). They argued
that adoption of EHR may be explained by the intention of the organization to improve
organizational performance through the fit between task and technology. Furthermore, the
30
concept of strength of attitude was used to explain EHR resistance (Nöltner and Krönung, 2015).
Prior empirical research on EHR resistance focused on finding solutions to the resistance
behavior rather than testing the resistance variable and conceptualize it based on well-defined
antecedents. For example, (Hewitt and McLeod, 2011; Dinev et al, 2016; Burns et al, 2015)
argued that security and privacy concerns may be a main hurdle in the process of EHR diffusion.
(Angst and Agarwal, 2009) investigated whether individuals can be persuaded to change their
attitudes and “opt-in” behavioral intentions toward EHRs, and allow their medical information to
be digitized even in the presence of significant privacy concerns. They found that an individual’s
concerns for information privacy (CFIP) interact with argument framing and issue involvement
to affect attitudes toward the use of EHR, and that attitude toward EHR use and CFIP directly
influence opt-in behavioral intentions. Their work shows that even when people have high
concerns for privacy, their attitudes can still be positively altered with appropriate message
framing. Another important study was conducted by (Weeger et al, 2011) who focused on the
perceived risks that could lead to EHR resistance. They explained how physicians associate
several risks with adopting EHR, these risks are: performance risks, social risks, psychological
risks, and privacy risks. They argue that these risks will need to be mitigated to enable proper
user acceptance of EHR. However, the study had no exploration to other antecedents to EHR
resistance and mainly focused on the risks associated with the introduction to EHR systems.
Other studies explored the phenomenon using qualitative methods. The study conducted
by (Poon et al, 2006) argued that the limited incentives given for EHR adoption to care providers
is effecting the diffusion of the technology. Furthermore, they argue that the high cost associated
with the implementation of EHR is shaping a burden on small physician practices and in turn is
31
limiting its diffusion. (Lapointe and Rivard, 2005) conducted case studies on EHR resistance.
They studied the phenomenon on group levels over time. They found that group resistance
behaviors vary during implementation. When a system is introduced, users in a group will first
assess it in terms of the interplay between its features and individual and/or organizational-level
initial conditions. They then make projections about the consequences of its use. If expected
In summary, our review uncovered interesting research gaps that need to be explored
through future research. There are limited IS research studies on EHR resistance that considered
theoretical lenses other than adoption theories. Additionally, we find that prior studies aimed to
In this study we aim not only to understand resistance towards EHR systems but also to
address the research gaps we found in the literature. This study is the first to apply the
comprehensive theoretically grounded model of the URM with a focus on the early
INVESTIGATIVE CONTEXT
The target hospital for this study is a public governmental hospital located in the capital
city of the Jordanian kingdom, Amman. The hospital consists of multiple health departments and
33 specialists’ clinics. The total inpatient capacity is 450 beds, with a plan of adding 100 more
beds by early 2016. The hospital has recently implemented a new EHR system that is mandated
to all employees. It is one of two health organizations in the country that have completely
implemented the system. The project started in 2009, implemented in 2014, and was enforced
early 2015. No more paper records are being used at the hospital. The new EHR system is named
“Hakeem”, the Arabic word for “Doctor”. Hakeem was implemented through a governmental
32
program which is the first national e-health initiative in Jordan. The system was designed and
technology-driven company established in early 2009. EHS is a partnership between the main
Technology, Royal Medical Health Awareness Society and Private Hospitals Association.
According to EHS, 4.5 million JDs (around $6.4 million) were invested in specialized
portal as well as a transactional system. It serves several key functions, such as storage, retrieval
and updating of the EHR of patients cared for by all of the participating healthcare facilities in
how employees perform their tasks. Employees had to switch from using the previous paper-
based system to the new EHR system Hakeem. When entering data into the system, they are
asked to complete the data entry of each form before being able to move to the next page. This is
taking some of the freedom they had with the paper based system. Additionally, the system is
only accessible from specific computers that are located within the hospital which challenges
Understanding resistance to change to the new way of working with the EHR system
requires special attention to the unique characteristics of the healthcare context. Healthcare
organizations differ from other environments in many ways. First, healthcare providers leverage
a different level of power which makes them capable of resisting the change to a newly
implemented technology even if it was mandated by the organization. Second, physicians are
seldom employed by the hospital (Halamka, 2011) which allows them more freedom of choice
and less managerial control over their decision of using or resisting a specific technology. Third,
33
healthcare providers usually perceive the use of some technologies as unnecessary tasks that
hurdles them from performing their original job of providing care to patients, and some consider
these technologies inappropriate for the physicians to use because it requires certain tasks that
represents threat to their professional status (Lapointe and Rivard, 2005). Some healthcare
providers have also expressed concerns about EHR systems affecting their communication with
patients such as loss of eye contact, which is against the social norm for physicians to avoid
using the computer while with the patient (Cotea, 2010). Additionally it was found that
healthcare providers fear that the EHR systems will take time away from actually caring for the
patient and make the art of their job impersonal (Sassen, 2009). Finally, healthcare providers do
not believe EHR systems are useful if compared to the PBRS (Stream, 2009).
Given the uniqueness of the healthcare context, it becomes essential not to investigate
resistance to change as simple as less EHR adoption behaviors. The healthcare context makes it
different from the more general “corporate settings” in which adoption theories have been
successfully tested in prior studies, therefor when applying these theories in the healthcare
context they will have less explanatory power (Smith et al, 2014). Also, Practitioners usually use
complex and difficult medical equipment, so “hard to use” fails to explain the non-acceptance for
them (Hare et al, 2006). Moreover, these theories are designed to explain voluntary IT usage and
have limited capabilities in explaining mandated IT usage which is present in the healthcare
settings (Bhattacherjee and Hikmet, 2013). It is believed that only applying the acceptance
models, with no extensions, in the healthcare settings is not sufficient to completely explain the
individuals’ resistance to change behaviors towards new EHR systems. Thus a different
perspective is required to provide more insights regarding resistance to change to EHR systems
34
while still leveraging the technology acceptance literature in examining resistance to change, and
not define resistance as the opposite of usage or adoption (Bhattacherjee and Hikmet, 2007).
THEORETICAL BACKGROUND
The URM was for this study for many reasons. First, it aims to explain resistance to
change associated with the implementation of a new IS system. Our theoretical development
focuses on the early implementation stage of the EHR system. In this stage the hospital is
keeping both the old paper-based record system (PBRS) and the new EHR system in place, and
employees are gradually changing from the PBRS to the EHR. In a later stage the EHR system
will be mandated to all employees and the PBRS will no longer be available for them to use. In
the context of our study, we aim to capture the resistance behaviors reported by employees when
changing from the PBRS to the EHR system. Second, it integrates different perspectives of user
resistance including the Status Quo Bias Theory (SQBT) (Samuelson and Zekhauser, 1988)
which will help explaining employees’ resistance to change due to the preference to stay with the
current status of working with the PBRS. Additionally, the URM includes the Equity
Implementation Model (EIM) (Joshi, 1991) perspective which will help in explaining
employees’ resistance to change to the EHR system based on cost-benefit analysis of the change
or the “net equity” associated with the change to the new EHR system. Third, the URM is a
comprehensive model which draws from previous literature various antecedents for technology
acceptance and resistance. Which allows us to integrate concepts from the technology acceptance
literature including Attitude, Subjective Norms, and Behavioral Control from the Theory of
Planned Behavior (TPB) (Ajzen, 1991). Fourth, prior research that tested the URM on EHR
systems were not found in the IS literature. Finally, replicating the URM in a different context
35
for the first time will help in making findings more generalizable and will contribute to the
progress of research in this area. (Nosek et al., 2015) reported that research progress relies on
both innovations which points out possible paths and replication which points out likely paths.
Thus, this work will increase certainty when findings are reproduced and promote innovation
explain user resistance to change. Namely it integrates the Equity Implementation Model (EIM)
(Joshi, 1991), the Status Quo Bias Theory (SQBT) (Samuelson and Zekhauser, 1988), and the
Theory of Planned Behavior (TPB) (Ajzen, 1991). In this section we briefly explain each of
these theories and how do these concepts correspond to the constructs of the URM.
The Equity Implementation Model posits that individuals evaluate the change associated
with the new IS implementation based on the “net equity”. The net equity is perceived as the
with the new IS. The outcomes are measured as (increase in outcomes – decrease in outcomes)
and inputs are measured as (increase in inputs – decrease in inputs). Additionally, net equity is
determined after individuals compare the change in their relative outcomes with that of the
employer and that of other individuals in the reference group. If Inequity (decline in net gain) is
perceived, then individuals would be resistant to the change; the greater the inequity the greater
The SQBT provides a theoretically driven explanations of resistance associated with the
new IS implementation based on the evaluation of the current status of the individual and the
perceived future status of the individual after accepting the change. It posits that resistance can
be due to the preference to stay with the current situation. The SQBT is explained using three
36
categories: (1) Rational Decision Making; this includes the evaluation of costs and benefits
associated with the change. If costs were perceived to be larger than benefits then this would lead
to the status quo bias. SQBT identifies two types of costs in this category: First, Transition Costs,
and these are the costs incurred in the process of adapting to the new situation, it includes costs
that happen during the change (transient costs) and costs that result from the change (permanent
costs). Second, Uncertainty Costs, and these are explained as the perception of risk associated
with the change to the new situation. Switching to a new IS can trigger uncertainty costs on
users because they may be unsure and anxious about the resulting changes. (2) Cognitive
Misperception; this is mainly explained by the psychological concept of loss aversion. Loss
aversion is manifested when individuals weigh losses more than gains while making decisions on
value, so relatively small losses associated with the change may be perceived as larger as they
actually are (Kahneman and Tversky 1979). (3) Psychological Commitment; Samuelson and
Zeckhauser explained this category using three factors: First, sunk cost, and these refer to
previous commitments, which may cause resistance to change. For example, if the current status
requires certain skills that a person had put time and effort into mastering those skills, and the
future alternative would require complete different set of skills, then the individual would
perceive his skills as sunk costs that may affect negatively on his decision to switch to the new
status. Second, Social norms, and these refer to the perceptions of people at work about the
change. These social norms may influence the individual’s status quo bias. Third, Efforts to Feel
in Control, and these are individuals’ desires to determine and control their own situation. If an
individual feels that s/he will lose control by changing to an unknown or unfamiliar way of
37
The Technology Acceptance (TA) literature was used in building the URM. The TA
theories have enriched the URM by including concepts such as social norms, control beliefs and
attitudes from the TPB. According to TPB, behavioral beliefs produce a favorable or unfavorable
attitude toward the behavior, normative beliefs result in subjective norms, and control beliefs
increases perceived behavioral controls. Because colleagues are usually the important referents
for individuals in work-related issues (Kim and Kankanhalli, 2009), the URM has incorporated
the concepts of colleague opinion which corresponds to the social norms from the SQBT.
Additionally, the perceived behavioral controls are accounted for by both external and internal
controls in URM. Self-efficacy for change and organizational support for change respectively
represent the internal and external means of achieving control of the changed situation (Ajzen
2002). Moreover, attitude is represented by perceived value in URM, which refers to the overall
The integrative framework of URM consists 7 main constructs. In this section we provide
a definition for each of these constructs in terms of the context of our study, and how does the
EIM, SQBT and TA theories correspond to each of the model constructs (see Table 2.2).
38
Net Benefits SQBT
Perceived The net benefits of switching from the PBRS to the
Loss Aversion SQBT
Value new EHR system.
Attitude TA
Control SQBT Individual’s confidence in own ability to change and
Self-Efficacy
adapt with the new EHR system.
for Change Behavioral Control TA
Control SQBT
Organizational The perceived facilitation provided by the hospital to
Support for make individual’s adaptation to new EHR system
Change easier.
Behavioral Control TA
Colleague Social Norms SQBT The perception that colleagues favor the change to the
Opinion Subjective Norms TA new EHR system.
User Any act of opposition to change from the PBRS to the
User Resistance URM
Resistance EHR system.
Switching Benefits
Switching Benefits in our study refers to the benefits perceived after changing from the
PBRS to the new way of working with the EHR system. Switching benefits correspond to the
increase in outcomes (e.g., improved job productivity) and the decrease in inputs (e.g.,
performing job faster) from the EIM. Based on URM, we theorize that the switch to the new
EHR would result in benefits as enhancements in the job productivity and performances of
employees, as a result these enhancements would be associated with certain rewards, therefor it
would increase the perceived value of the change. Thus, we adapt the URM hypothesis that
higher switching benefits would increase the perceived value of change to the new system.
Switching Costs in our study refers to the costs perceived after changing from the PBRS
to the new way of working with the EHR system. Switching costs correspond to the decrease in
outcomes (e.g., less job productivity) and the increase in inputs (e.g., performing a task slower)
from the EIM. Additionally, switching costs correspond to the transition costs, uncertainty costs,
Switching Costs
39
First, Transition costs. In our study these are the costs associated with the change from
the PBRS to the new EHR system. Any switch from the current status quo to a new status incurs
transition costs (Kim, 2011). When undertaking the transition to the new system, users must
acquire new skills and knowledge in order to use the new system and to perform their tasks
(Burnham et al, 2003). Users would strongly oppose the change when it requires transition costs
and have a tendency to keep their current status quo (Samuelson and Zekhauser, 1988).
Second, Uncertainty costs. These are the perceived levels of risk on users’ performances
associated with the change (Burnham et al, 2003). If users have limited knowledge about the
new system, they will be uncertain about their job performance with this system, and would
perceive this as a risk (Whitten and Wakefield, 2006), this would produce unpleasant
psychological reactions such as anxiety, which in turn would influence their status quo bias
(Inder and O’Brien, 2003).Thus, users will prefer to maintain the status quo (i.e. PBRS) and
Third, Sunk costs. These are the investments and costs already put into mastering the
current way of working (Kim, 2011). In our study, these are the costs in time and effort that have
already been invested by users in mastering the PBRS. The larger is the past investment, the
Based on URM, we theorized that switching costs would negatively affect the net
benefits of the switch because net benefits are evaluated by weighing benefits against costs.
Additionally, the corresponding costs from the SQB (i.e. transition costs, uncertainty costs, and
sunk costs) would strongly influence the individual’s decision on change. Thus, we adapt the
URM hypothesis that switching costs have negative effects on perceived value and positive
effects on resistance.
40
Perceived Value
Perceived value in our study refers to the perceived net benefits (benefits vs costs) of
switching from the PBRS to the new way of working with the EHR system. Perceived value
corresponds to the net equity and net benefits concepts from EIM and SQBT respectively.
Additionally, perceived value corresponds to the concept of loss aversion from SQBT because it
contributes explanations of how the perceived value of change is assessed (i.e. losses are viewed
larger than they actually are). Following Kim and Kankanhalli’s definition of the construct, we
define perceived value in our study as the net benefits of switching from the PBRS to the new
EHR system. Based on URM, we theorize that people tend to like maximizing value when
making a decision (Sirdeshmukh et al, 2002), if the perceived value of the change is low, then
users are likely to have greater resistance to change (i.e. switching to the new EHR system). On
the other hand, if the perceived value is high, then users are likely to have lower resistance to
switching to the new EHR system. Thus, we adapt the URM hypothesis that perceived value has
Self-Efficacy for Change in or study refers to the internal means that can enhance
achieving control of the changed situation (i.e. switching to the new EHR system) (Ajzen 2002).
Self-efficacy corresponds to the control concept from the SQBT. If individuals feels that there
will be a loss in control associated with the change to the new unfamiliar way of working, then
status quo bias would be observed. This loss of control may be manifested as low self-efficacy.
Following Kim and Kankanhalli’s definition of the construct, we define self-efficacy for change
as an individual’s confidence in his or her own ability to change and adapt with the new EHR
system. Based on URM, we theorized that difficulties employees face during the change to the
41
EHR system will be viewed as threats to be avoided or skills to be mastered depending on the
level of self-efficacy the employee has (Bandura 1995). So individuals with high levels of self-
efficacy would face the change confidently and, on the other hand, individuals with low levels of
self-efficacy would be more inclined to resist the change. Additionally, individuals with high
levels of self-efficacy for change would be less likely to experience anxiety and uncertainty
regarding the change (Bandura 1995; Compeau et al. 1999), which implies less switching costs
(i.e. uncertainty and transition costs). Thus, we adapt both of URM hypotheses: First, that self-
efficacy for change has negative effect on user resistance, and second, that self-efficacy for
Organizational Support for Change, like the Self-Efficacy for Change, it also corresponds
to the control concept from the SQBT. Following Kim and Kankanhalli’s definition of the
construct, we define organizational support for change as the perceived facilitation provided by
the hospital to make individual’s change to the new EHR system easier. Based on URM we
theorize that organizational support mechanisms such as training and providing resources could
affect individual’s reactions towards change to the new EHR system (Hirschheim and Newman
1988). Additionally, these mechanisms would minimized the perceived difficulty of adapting to
the new EHR system (Lewis et al. 2003), and in turn would reduce costs of time and effort
required to learn the new way of working. Thus, we adapt both of URM hypotheses: First, that
organizational support for change has a negative effect on user resistance, and second
Colleague Opinion
42
This construct corresponds to social norms from SQBT. Kim and Kankanhalli’s
definition of the construct limited social influence to only that coming from colleagues and
peers. In our study we included peers, colleagues and superiors as important others. Thus, we
used named this construct as “Social Influence” rather than “Colleague Opinion”. We define
Social Influence as the degree to which an individual perceives that important others believe he
or she should use the new EHR system (Venkatesh et al, 2003). We are including both types of
influence: (1) informational influence which comes mostly from colleagues, and (2) normative
influence which comes mostly from superiors (Burnkrant and Cousineau 1975). Building on
URM, we theorize that individuals have the tendency to conform to their important others’
opinions because of the need for social companionship and the fear of sanction for
noncompliance (Ajzen 2002; Lewis et al. 2003). So individual’s resistance behaviors are directly
affected by what other people at work think about the change to the new EHR system.
Additionally important others’ opinion about the new EHR system may alter the original
perceptions an individual has about switching costs and benefits (e.g. informational influence -
Bunkrant and Cousineau, 1975). Positive opinions toward the new change can serve to reduce
users’ uncertainty and therefore lower their perceptions of switching costs. Also, positive
opinions about the change to the new EHR would lead to a greater perception of switching
benefits among users. Thus, we adapt both of URM hypotheses that positive opinions (positive
social influence) about the change to the EHR has a negative effect on both user resistance and
User Resistance
The outcome variable of the URM, User Resistance, was constructed based on the frame
work of resistance behaviors (Bovey and Hede, 2001) which distinguishes between overt
43
resistance (physical behavior) and covert resistance (psychological mid-set), and between active
and passive resistance. Resistance can occur on a spectrum from covert passive (e.g., ignoring) to
overt active (e.g., sabotage) behaviors (see Figure 2.1). Based on Kim and Kankanhalli’s
definition of the construct, we define user resistance as any act of opposition to change from the
Control Variables
variables. First, because it is argued that employees with different positions would perceive the
change differently. While physicians consider data entry into EHR is a “clerical task” (Lapointe
and Rivard, 2005), other employees may enjoy and appreciate the change. Thus, we included
position at hospital as a control variable. Second, because social influence may have effect on
user resistance it was appropriate to control for the number of employees in a unit. Third, we
controlled for all of age, gender, and level of education to assess how individual differences may
44
Table 2.3 summarizes the research model hypothesis and Figure 2.2 illustrates the posited
structural model of URM (Kim and Kankanhalli, 2009) that will be tested in the context of this
No. Hypothesis
45
Figure 2.2: URM (Kim and Kankanhalli, 2009)
METHODOLOGY
Instrument Development
Existing validated scales were adopted for this study. We mainly adopted the instrument
items used for the URM’s original study. However all items were modified to fit the context of
our study. Social Influence included an additional item (SOI4) to capture superiors’ influence.
Also unlike URM, Perceived Value included Loss Aversion items in the survey (LAV1-6) these
were adopted from (Gächter et al, 2007). Measurement items were anchored on five-point Likert
scales (1 = strongly disagree, 5 = strongly agree). The instrument was reviewed by IS researchers
before the study. The final version of the questionnaire was developed for the main study as
46
The total estimated number of employees who work at the hospital is 5,500. Majority of
these employees are potential users of the new system. However administrative staff are less
likely to use the system as they do not participate in the care giving process, however they still
have access to the system and would be able to print out reports as needed. The sample is a
combination of all stakeholders of the system (i.e. physicians, nurses, lab specialists, radiologists,
pharmacists, admin staff, and resident medical students). Data was collected in 2 phases. In the
first phase, we collected data for the pilot study from employees who works only for the
Dermatology unit at the hospital, we collected a total of (126) responses. After conducting
Explanatory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) of the data
collected for the pilot test, it was necessary to alter some of the items to address a number of
issues related to the psychometric properties of factors. Additionally, some new items were
added to better capture the concepts of some constructs. First, SWC1 had a low loading of
(0.451). The item is developed to capture sunk costs which refers to the investments and costs
already put into mastering the current way of working (Kim, 2011). Thus, it was necessary to
reword the item to reflect the costs healthcare providers have invested in the previous PBRS, so
we added “I already have” at the beginning of the survey question. The loading of the item went
up to (0.702) after collecting data for the main study. Second, SFC1 also had a low loading of
(0.595). The item was reworded to match the same fashion as other factors of the same construct.
This resulted in an increase in its loading in the main study to become (0.761). Additionally, we
added item SFC4 into the main study survey to capture the levels of confidence healthcare
providers have in their ability to change to the new EHR system. The new item had a loading of
(0.740). Third, we included the fourth item (OGS4) to capture healthcare providers’ opinions on
the levels of assistants provided by the hospital. The new item had a loading of (0.833). Fourth,
47
because our study aims to capture social influence sourcing from all possible employees at the
hospital, it was necessary to include item SOI4 which captures social influence of superiors on
the healthcare providers. This new item had a loading of (0.846). Finally, we included the
concepts of loss aversion within the perceived value construct. We also added a fourth item
PVL4 which captured healthcare providers’ perceptions of value after considering the limitations
of Hakeem. Loss aversion had a negative loading, in the main study, and no statistical
significance (p = 0.11 > 0.05), so Loss Aversion item was dropped from the scale. However the
new item PVL4 had a loading of (0.976). After refining the survey items based on the pilot
data’s psychometric properties, we collected data for the main study from the entire hospital, but
employees who filled the first survey (i.e. pilot study) were specifically asked not to fill in the
survey for the second time. We collected a total of (226) responses, descriptive statistics are
Instrument Validation
To validate the survey instrument, the psychometric properties of the survey were
assessed by conducting Explanatory Factor Analysis (EFA) and Confirmatory Factor Analysis
48
(CFA) using MPlus 7.1 (Muthén and Muthén, 1998-2006). Cronbach’s α was performed using
SPSS 20 (IBM Corp. 2013). After the pilot data was collected, we refined the items to keep those
with statistical significant loading larger of (0.7). Cronbach’s α reliability tests for all constructs
exceeded (0.8). After the main data was collected, we again conducted CFA analysis and again
all items had a significant loading greater than (0.7) except for the dichotomous factor indicator
Loss Aversion within Perceived Value. It had a negative loading with no statistical significance
(p = 0.11 > 0.05), so Loss Aversion items was dropped from the scale. All constructs had
Cronbach’s α values exceeding (0.8). The CFA analysis provided strong support for our
measurement model, which suggested that the items under each of the constructs were
adequately measuring the constructs. Table 2.6 in the Appendix lists the items’ loadings and the
The survey responses came in (2) waves: (W1: N=164; W2: N=62). We checked for
nonresponse bias by testing the difference in means between waves (see Table 2.7). We found no
significant difference between the two respondent groups based on the sample attributes (gender,
RESULTS
49
The original research model (URM) was tested by applying Structural Equation
Modeling (SEM) using Mplus version 7.1 (Muthén and Muthén, 1998-2006). We applied the
maximum likelihood estimator with robust standard errors (MLR). Because the URM is not
saturated (i.e., not all possible regression paths were included) we evaluated the model fit
Normed X2 (NC=3.91). It is suggested that NC value between (2.0 and 6.0) indicates reasonable
fit. However, we notice that (CFI = 0.882), this indicates illness in the model fit. The rule of
thumb for the CFI and other incremental indices is that values greater than roughly (.90) may
indicate reasonably good fit of the researcher’s model (Hu and Bentler, 1999). Additionally, the
Root Mean Square Error of Approximation (RMSEA) is (0.114). According to the rules of
thumb by (Browne and Cudeck, 1993) RMSEA value between (0.05) and (0.08) suggest
reasonable error of approximation and values greater than (0.1) suggests poor fit. In order to
improve the goodness-of-fit of the model we conducted modification indices test using MPlus.
This suggests possible paths to be added to the model to improve its goodness-of-fit.
50
The suggested model and its new improved model fit indices are included in the
discussion section of this paper. The standardized path coefficients, its level of significance, and
the standard errors (S.E) are provided in (Table 2.9), and results are depicted in (Figure 2.3).
51
(*p<0.05; **p<0.01; ***p<0.001)
After comparing our findings with the hypotheses of the original theory we posit mixed
results (i.e. some hypothesis were confirmed and others were not). Our study found that the
positive effect switching benefits has on perceived value is confirmed. On the other hand
switching costs effect on perceived value was not confirmed but switching cost direct effect on
resistance was confirmed. The positive effect of Social Influence on switching benefits was
confirmed, as well as its negative affect on switching costs. However social influence had no
significant direct effect on resistance, so the original hypothesis was not confirmed. Similarly,
52
self-efficacy for change and organizational support had no direct effect on resistance, and both
hypotheses were not confirmed. The negative effect of Self-efficacy for change on switching cost
was confirmed. Finally, organization support had a positive direct effect on switching costs. This
contradicts the original hypothesis that suggested a negative relation between both constructs.
None of the control variables had significant effect on resistance. Perceived value and switching
costs effects on user resistance explained (48.6%) of its variance. Switching benefits explained
(75.1%) of the variance in perceived value. Social influence explained (19.2%) of the variance in
switching benefits. Social influence, organizational support, and self-efficacy for change effects
DISCUSSION OF FINDINGS
Our findings show that most of posited hypotheses were supported. This indicates that the
URM has the potential of being generalized and could successfully explain resistance to change
in different contexts. However, we still found that few hypotheses were not supported in addition
to a number of interesting issues surrounding the resistance to change to a new EHR system. In
this secession we provide a discussion of these findings. First, organizational support had a
positive direct effect on switching costs which in turn has a positive direct effect on resistance.
This means the more healthcare providers perceive the hospital being supportive to the change to
Hakeem the more they perceive the switch to Hakeem as being costly in terms of their transition,
uncertainty, and sunk costs. This opposes the original theory hypothesis that suggested
organizations are aware of a number of compelling issues surrounding EHR systems that keeps
them motivated to provide support to their EHR users: that EHR systems consist of complex
software and hardware that requires certain skills in order to be applied in the workplace
53
(Boonstra and Broekhuis, 2010), that one of the most important barriers to EMR adoption is a
lack of computer proficiency (Najaftorkaman et al, 2015), and that some healthcare providers are
still interested in documenting their work using PBRS instead of using EHR systems. If these
providers cannot embrace EHR in their workplaces, the resistance of EHR systems can rapidly
offer training sessions to their healthcare providers who are expected to interact with the EHR
system, as well as provide technical support to users as needed. However, healthcare providers at
the hospital may be perceiving increased support from the hospital as an indication that the
transition is difficult and requires special support and this may result in an increase in the
perceived transition costs. Furthermore, hospital support may be in the form of training sessions
which means there will be new skills to be learned and new knowledge to be mastered, and this
may increase levels of sunk costs. Moreover, hospital support for the change may be unclear on
how would this effect healthcare providers’ productivities and performances, which may lead to
an increase in their uncertainty costs. Thus, we found that hospital support had a positive effect
on switching costs. This positive effect of hospital support on switching costs is an interesting
finding especially that switching costs is a composite construct and the effect may be directed to
one or more of the types of costs. We encourage future research to further explore this finding.
Second, switching costs had no effect on perceived value. However, our findings confirm
the direct effect of switching costs on resistance. In other words, healthcare providers are more
influenced to resist the change to Hakeem when higher switching costs are perceived, however
their perception of costs has no influence on their perception of the value of the switch. So one
with high perceptions of switching costs may be influenced to resist the switch without affecting
his perceptions of the switching value. This may be different if loss aversion items were included
54
in the perceived value scale items. Loss aversion suggests that when evaluating a situation
people will value losses more than gains. This is especially true in the healthcare context where
losses could be directly aimed towards the health of patients. When healthcare providers are
delivering care with a suggested success rate of the treatment they are expected to value the
failure of the procedure heavily and direct their focus on avoiding it. We believe that, as a matter
of automatic behavior, healthcare providers will weigh costs much more than they weigh benefits
in any given evaluation case. Therefore, the loss aversion concept can be used to help explain
this finding. Additionally, our finding could be explained by the concept of perceived
consequences which refers to the cognitive evaluation of the probable consequences of the
behavior (Gagnon et al., 2010). The work of (Najaftorkaman et al, 2015) suggested that
healthcare providers’ perceived consequences of using EHR systems have a significant impact
on their adoption behaviors of the EHR. Thus, when healthcare providers know about the various
costs and consequences of using EHR systems they begin to resist EHR systems. This perception
of consequences is found to be stronger and independent from their beliefs of the value
Third, self-efficacy for change has no direct impact on user resistance. Rather, there is an
indirect effect mediated through switching costs. This result indicates that self-efficacy for
change decreases resistance by reducing the perception of switching costs. We believe that this
finding could be explained partly by the fact that we have captured self-efficacy to change to the
new way of working with Hakeem rather than self-efficacy of actually using Hakeem. Healthcare
providers may be confident in their ability to adapt to the transition process of the hospitals in
terms of redirecting the use of resources, work processes, budget allocations, or any other modes
of operation that will reshape the hospital, but not confident in their ability to use the new
55
technology on their own. It has been shown in the literature that computer literacy has a negative
impact on EHR resistance and healthcare providers with less computer skills are less likely to
adopt the systems (Stream, 2009; Terry et al, 2009). Using EHR systems could be a completely
new experience for some healthcare providers in their workplace. Some EHR users are not
confident about their ability of using it (Simon et al, 2007a; Terry et al, 2009). Therefore, we call
on future research to investigate the impact of self-efficacy of using the EHR technology on
resistance behaviors.
Fourth, social influence has no direct effect on resistance, but there is an indirect effect
mediated through switching costs. This indicates that social influence decreases resistance by
reducing the perception of switching costs. This could be partly explained by the fact that we
included items to capture influence from superiors. Some healthcare providers may be influenced
by the fear of punishment if they resist the enforced change while their superiors are supportive
of it. Further, different opinions may be perceived from different sources. Thus, the superiors
influence, when included with other type of influences (i.e. from colleagues and peers), may
have contributed to the result of having no direct relation between social influence and
resistance.
Fifth, organizational support has no direct effects on resistance. This is related to our
needs and must choose the proper method of support carefully. Our findings indicates that the
Hakeem’s costs and is not impacting on their resistance behaviors. We encourage future research
to further explore what type of support was used by the hospital in this study and why did
healthcare providers respond negatively to hospital support. Additionally, future research may
56
aim to provide a comparative analysis between different types of support provided by healthcare
Finally, our findings indicated problems with the goodness-of-fit of the tested model. We
conducted a model indices test in MPlus to find alternative models that are aligned with the
URM theory and has an improved fit. We found the best model fit achieved by adding the
following three paths: (1) direct effect of switching benefits on resistance, (2) self-efficacy on
switching benefits, and (3) organizational support on switching benefits. Additionally, the model
dropped all 3 direct effects of social influence, organizational support, and self-efficacy for
change on user resistance. The new suggested model is depicted in (Figure 2.4) and the new
57
Calculated NC 2.378
CFI 0.944
RMSEA 0.078
90 CI 0.051 – 0.106
SRMR 0.046
The suggested model had no notable change in the amount of variance explained, except
for switching benefits (28.9%) which increased from (19.2%). However the suggested model had
a better fit. NC was at (2.3), CFI increased to (0.944) which is in the recommended range (>0.9),
and RMSEA became lower (0.078) also in the reasonable range (0.05 – 0.10). The only new
significant path is the positive effect of self-efficacy for change on switching benefits. The more
healthcare providers feel confident in their skills and abilities to master the new way of working
with the EHR system the more they will perceive the change as beneficial. This falls along the
URM theory which hypothesized the negative effect self-efficacy for change has on switching
costs.
CONTRIBUTIONS
Contribution to Research
change to the new way of working using EHR systems based on net-benefit evaluation which
influences how healthcare providers would perceive the value of the switch to the new EHR
system. Resistance to change to the new EHR system is effected directly by healthcare provider’
perceived value of the EHR system and their perceived costs of switching to the new EHR
system. These findings contributes to the IS research which still requires advancements in
understanding resistance behaviors in healthcare settings. Second, we found that the URM, if
contexts. For this study we tested the model in the healthcare context for the first time and found
58
that URM’s basic solid foundations are also applicable in such stetting. Third, we found a direct
effect, which was not reported in the original theory, between self-efficacy for change and
switching benefits. Fourth, our study reveals a counterintuitive finding; the positive direct effect
organizational support has on switching costs. This interesting finding should motivate future
work to better investigate different possibilities surrounding this phenomenon. Fifth, we included
superiors’ influence in our scale when capturing social influence. This addition confirms the
evaluating the net-befit of switching to the new EHR system. Sixth, our study was conducted at
Finally, our findings show that there is no difference in the way healthcare providers are
influenced to resist the change to the new EHR system based on different demographics such as
Contributions to Practice
solutions to hospitals and other health organizations when trying to reduce resistance behaviors
towards the change to new EHR systems. Switching cost has both a direct and indirect effect on
resistance. Thus, reducing it is critical when aiming to eliminate resistance behaviors. This could
be achieved by enhancing positive opinions about the change both from colleagues and
superiors. Also, it is very crucial to understand what type of support is more suitable for the
healthcare providers. Our findings show that hospital support has a positive effect on switching
costs. This may be due to inappropriate delivery of the support (e.g. long training sessions for
increase their self-efficacy would have direct impacts on their perception of the switch in terms
59
of its costs vs benefits. Also, it is important to emphasize the advantages of the EHR system and
to explain the core benefits of the switch at early stages before the implementation of the EHR
system, this would help in increasing switching benefits and perceived value which in turn would
reduce resistance behaviors. Second, because this study is conducted in a developing country,
which is still at the first stages of digitizing healthcare organizations, it becomes very important
to understand what predicts resistance to change to new EHR systems and what do new users
consider as benefits or costs to the switch. Answers to these questions would help system
developers and system designers build improved EHR systems that overcomes sources of costs.
Third, understanding EHR resistance would help healthcare providers get the appropriate support
to reduce resistance motivations. Governments in many countries have largely invested in the
transition to EHR systems (Khalifa, 2013). Thus, being able to explain causes of HIT resistance
would help in finding solutions to make these systems more adopted and with higher success
rates. Finally, our work shows that URM could be applied successfully in different contexts.
Thus, the theory could be replicated using different technologies across industries. Findings
would help practitioners better understand resistance in the context of which it was tested in, and
that would serve as the first step towards solving problems associated with resistance behaviors.
This study has its limitations. First, in order to fully understand the positive effect
between organizational support and switching costs, further investigations are required. It is
interesting to study how different type of support provided by healthcare organizations may
influence cost perceptions of EHR systems differently. Second, because our survey data is cross-
sectional and collected from a single source, common method bias may be a concern (Xue et al,
2014). Future research can apply a longitudinal study or collect data from multiple sources to
60
validate the findings. Finally, although we have collected loss aversion in perceived value’s
scale, the items had poor loadings and thus were dropped from the survey. Future studies could
61
REFERENCES
62
Burns, A. J., Young, J., Roberts, T. L., Courtney, J. F., & Ellis, T. S., 2015. “Exploring the Role
of Contextual Integrity in Electronic Medical Record (EMR) System Workaround
Decisions: An Information Security and Privacy Perspective,” AIS Transactions on
Human-Computer Interaction, 7(3), 142-165.
Burt C.W., Sisk J.E., 2005. “Which Physicians and Practices Are Using Electronic Medical
Records?” Health Affairs, 24(5):1334-1343.
Cantrill, M., and Stephen, V., 2010. “Computers in Patient Care: The Promise and the
Challenge,” ACM Queue, (8:8).
Compeau, D., Higgins, C. A., and Huff, S. 1999. “Social Cognitive Theory and Individual
Reactions to Computing Technology: A Longitudinal Study,” MIS Quarterly (23:2), pp.
145-158.
Cotea, C., 2010. “Electronic health record adoption: Perceived barriers and facilitators,” Centre
for Military and Veterans' Health, The University of Queensland.
Davis, F. D., 1989. “Perceived Usefulness, Perceived Ease of Use, and User Acceptance of
Information Technology,” MIS Quarterly, 13, 3, pp. 319-340.
DesRoches, C. M., Campbell, E. G., Rao, S. R., Donelan, K., Ferris, T. G., Jha, A., &
Blumenthal, D., 2008. “Electronic health records in ambulatory care—a national survey
of physicians,” New England Journal of Medicine, 359(1), 50-60.
Dinev, T., Albano, V., Xu, H., D’Atri, A., & Hart, P., 2016. “Individuals’ Attitudes Towards
Electronic Health Records: A Privacy Calculus Perspective,” Advances in Healthcare
Informatics and Analytics, (pp. 19-50), Springer International Publishing.
Earnest M.A., Ross S.E., Wittevrongel L., Moore L.A., Lin C.T., 2004. “Use of a Patient-
Accessible Electronic Medical Record in a Practice for Congestive Heart Failure: Patient
and Physician Experiences,” Journal of the American Medical Informatics Association,
11(5):410 417.
Electronic Health Solutions (EHS) http://www.ehs.com.jo/en
Fu, K. 2009. “Viewpoints: Inside Risks Reducing Risks of Implantable Medical Devices,”
Communications of the ACM (52:6), pp. 25-27.
Gächter, S., Johnson, E. J., & Herrmann, A., 2007. “Individual-level loss aversion in riskless and
risky choices.”
Gagnon, M.-P., Ouimet, M., Godin, G., Rousseau, M., Labrecque, M., Leduc, Y., 2010. “Study
protocol multi-level analysis of electronic health record adoption by health care
professionals: A study protocol,” Implementation Science, 30(5), 1–10.
Gan, Q., & Cao, Q., 2014. “Adoption of Electronic Health Record System: Multiple Theoretical
Perspectives,” System Sciences (HICSS), 47th Hawaii International Conference on
System Science, (pp. 2716-2724), IEEE.
63
Halamka, J., 2011 “Healthcare is different”, http://geekdoctor.blogspot.com/2011/08/healthcare-
is-different.html
Hare, K., Whitworth, B., Deek, F., and Norris, T., 2006. “Identifying Barriers to IT Adoption in
Clinical Health Care: A Pilot Study”, Health Care and Informatics, Review Online, ISSN
1176-4201, http://hcro.enigma.co.nz/website/index.cfm
Heath, M., & Appan, R., 2014. “Impact of Information Technology Culture Conflict on
Meaningful Use of Electronic Health Records System: A Conceptual Framework.”
Hewitt, B., & McLeod, A., 2011. “Modeling Security in Acceptance of Electronic Health Record
Systems,” Journal of Information Privacy and Security, 7(3), 23-45.
Hillestad, R., Bigelow, J., Bower, A., Girosi, F., Meili, R., Scoville, R., & Taylor, R. 2005. “Can
electronic medical record systems transform health care? Potential health benefits,
savings, and costs,” Health Affairs, 24(5), 1103-1117.
Hirschheim, R., and Newman, M., 1988. “Information Systems and User Resistance: Theory and
Practice,” Computer Journal, 31, 5, pp. 398-408.
Hu, L.T. and Bentler, P.M., 1999. "Cutoff Criteria for Fit Indexes in Covariance Structure
Analysis: Conventional Criteria versus New Alternatives," Structural Equation
Modeling, 6 (1), 1-55.
Hung, S. Y., Yu, W. J., Tsai, M. F., & Yen, D. C., 2013. “What Drives Physicians' Intention to
Use Electronic Medical Record System: The Roles of Perceived Service Level, Computer
Self-Efficacy, and Perceived Risk,” PACIS, (p. 194).
IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM
Corp.
Inder, B., and O’Brien, T., 2003. “The endowment effect and the role of uncertainty,” Bull.
Econ. Res., vol. 55, no. 3, pp. 289–301.
IT-toolbox ERP Implementation Survey, 2004. www.vvsgidas.lt/get.php?id=403
Jha, A. K., DesRoches, C. M., Campbell, E. G., Donelan, K., Rao, S. R., Ferris, T. G., &
Blumenthal, D., 2009. “Use of electronic health records in US hospitals,” New England
Journal of Medicine, 360(16), 1628-1638.
Joshi, K., 1991. “A Model of User Perspective on Change: The Case of Information Systems
Technology Implementation,” MIS Quarterly, 15, 2, (1991), pp. 229-242.
Kahneman, D., & Tversky, A., 1979. “Prospect theory: An analysis of decision under risk,”
Journal of the Econometric Society, 263-291.
Keen, P.G., 1981. "Information Systems and Organizational Change," Communications of the
ACM, 24, 1, pp. 24-33.
64
Kemper A.R., Uren R.L., Clark S.J., 2006. “Adoption of Electronic Health Records in Primary
Care Pediatric Practices,” Pediatrics, 118(1):20-24.
Khalifa, M., 2013. “Barriers to health information systems and electronic medical records
implementation. A field study of Saudi Arabian hospitals,” Procedia Computer Science,
21, 335-342.
Kim, H. W., 2011. “The effects of switching costs on user resistance to enterprise systems
implementation,” Engineering Management, IEEE Transactions on, 58(3), 471-482.
Kim, Hee-Woong, and Atreyi Kankanhalli, 2009. "Investigating user resistance to information
systems implementation: a status quo bias perspective,” MIS quarterly, 33, 3.
Laerum H, Ellingsen G, Faxvaag A., 2001. “Doctors’ Use of Electronic Medical Records
Systems in Hospitals: Cross Sectional Survey,” British Medical Journal,
323(7325):1344-1348.
Lapointe, L., and Rivard, S. 2005. .“A Multiple Model of Resistance to Information Technology
Implementation,” MIS Quarterly, (29:3), pp. 461-491.
Lewis, W., Agarwal, R., and Sambamurthy, V., 2003. “Sources of Influence on Beliefs about
Information Technology Use: An Empirical Study of Knowledge Workers,” MIS
Quarterly, (27:4), pp. 657-678.
Loomis GA, Ries S, Saywell RM, Thakker NR., 2002. “If Electronic Medical Records Are So
Great, Why Aren’t Family Physicians Using them?” Journal of Family Practice,
51(7):636-641.
Ludwick DA, Doucette J., 2009. “Primary Care Physicians’ Experience with Electronic Medical
Records: Barriers to Implementation in a Fee-for-Service Environment,” International
Journal of Telemedicine and Applications, 853524.
Meade B, Buckley D, Boland M., 2009. “What Factors Affect the Use of Electronic Patient
Records by Irish GPs?” International Journal of Medical Informatics, 78(8):551-558.
Menachemi N, Langley A, Brooks RG., 2007. “The Use of Information Technologies Among
Rural and Urban Physicians in Florida,” Journal of Medical Systems, 31(6):483–488.
Miller RH, Sim I., 2004. “Physicians’ Use of Electronic Medical Records: Barriers and
Solutions,” Health Affairs, 23(2):116-126.
MPLUS (Version 6.11). [Computer Software]. Los Angeles, CA: Muthén & Muthén
Najaftorkaman, M., Ghapanchi, A. H., Talaei‐Khoei, A., & Ray, P., 2015. “A taxonomy of
antecedents to user adoption of health information systems: A synthesis of thirty years of
research,” Journal of the Association for Information Science and Technology, 66(3),
576-598.
Ngafeeson, M. N., and Midha, V., 2014. "An exploratory study of user resistance in healthcare
IT," International Journal of Electronic Finance, 8, no. 1, 74-94.
65
Nöltner, M., & Krönung, J. 2015, “Bolstering Attitudes: An Introduction to the Foundations of
Attitude Strength and Their Consideration in IS Adoption Research.”
Nosek, B. A., Alter, G., Banks, G. C., Borsboom, D., Bowman, S. D., Breckler, S. J., ... &
Contestabile, M., 2015. “Promoting an open research culture: Author guidelines for
journals could help to promote transparency, openness, and reproducibility,” Science
(New York, NY), 348(6242), 1422.
Olaniran, B. A., 2015. “Contextual factors and challenges to e-health literacy,” Knowledge
Management & E-Learning: An International Journal (KM&EL), 7(4), 616-628.
Pinaire, K. (2009). “An overview of barriers to the adoption of electronic medical records,”
MWAIS Proceedings, 14.
Pizziferri L, Kittler AF, Volk LA, Honour MM, Gupta S, Wang S, Wang T, Lippincott M, Li Q,
Bates DW., 2005. “Primary Care Physician Time Utilization Before and After
Implementation of an Electronic Health Record: A time motion Study,” Journal of
Biomedical Informatics, 38(3):176-188.
Poon, E., Blumenthal, D., Jaggi, T., Honour, M., Bates, D.W., and Kaushal, R., 2004.
“Overcoming barriers to adopting and implementing computerized physician order entry
systems in U.S. hospitals,” Health Affairs, 23, 4, pp. 184–190.
Poon, Eric G., Jennifer L. Cina, William Churchill, Nirali Patel, Erica Featherstone, Jeffrey M.
Rothschild, Carol A. Keohane, Anthony D. Whittemore, David W. Bates, and Tejal K.
Gandhi, 2006. "Medication dispensing errors and potential adverse drug events before
and after implementing bar code technology in the pharmacy," Annals of Internal
Medicine, 145, no. 6, 426-434.
President’s Information Technology Advisory Committee, 2004. Revolutionizing Health Care
through Information Technology, National Coordination Office for Information
Technology Research and Development, Arlington, Virginia, http://www.nitrd.gov
Randeree E., 2007. “Exploring Physician Adoption of EMRs: A Multi-Case Analysis,” Journal
of Medical System, 31(6):489-496.
Reardon JL, & Davidson E., 2007. “An Organizational Learning Perspective on the Assimilation
of Electronic Medical Records among Small Physician Practices,” European Journal of
Information Systems, 16:681-694.
Samhan, B., & Joshi, K. D., 2015. “Resistance of Healthcare Information Technologies;
Literature Review, Analysis, and Gaps,” System Sciences (HICSS), 2015 48th Hawaii
International Conference on (pp. 2992-3001). IEEE.
Samuelson, W., & Zeckhauser, R., 1988. “Status quo bias in decision making,” Journal of risk
and uncertainty, 1(1), 7-59. Chicago.
Sassen, E.J., 2009. “Love, hate, or indifference: How nurses really feel about the electronic
health record system,” Computers, Informatics, Nursing, 27(5), 281–287.
66
Shea, S., & Hripcsak, G., 2010. “Accelerating the use of electronic health records in physician
practices,” New England Journal of Medicine, 362(3), 192-195.
Sheikh, A., Cornford, T., Barber, N., Avery, A., Takian, A., Lichtner, V. & Morrison, Z., 2011.
“Implementation and adoption of nationwide electronic health records in secondary care
in England: final qualitative results from prospective national evaluation in “early
adopter” hospitals,” BMJ, 343, d6054.
Simon SR, Kalshal R, Cleary PD, Jenter CA, Volk LA, Oray EJ, Burdick E, Poon EG, Bates
DW., 2007b. “Physicians and Electronic Health Records: A Statewide Survey,” Archives
of Internal Medicine, 167(5):507-512.
Sirdeshmukh, D., Singh, J., and Sabol, B., 2002. “Consumer Trust, Value and Loyalty in
Relational Exchange,” Journal of Marketing, (66:1), pp. 15-37.
Smith, Tim, Gerald Grant, and Alejandro Ramirez, 2014. "Investigating the influence of
psychological ownership and resistance on usage intention among physicians,” System
Sciences (HICSS), 47th Hawaii International Conference on. IEEE.
Stone, C., 2014. “A Glimpse at EHR Implementation Around the World: The Lessons the US
Can Learn,” The Health Institute for E-Health Policy, http://www.e-
healthpolicy.org/docs/A_Glimpse_at_EHR_Implementation_Around_the_World1_Chris
Stone.pdf
Stream, G.R, 2009. “Trends in adoption of electronic health records by family physicians in
Washington State,” Informatics in Primary Care, 17(3), 145–152.
Terry AL, Thorpe CF, Giles GG, Brown JB, Harris SB, Reid GJ, Thind A, Stewart M, 2008.
“Implementing Electronic Health Records,” Canadian Family Physician, 54:730-736.
Terry, A.L., Giles, G., Brown, J.B., Thind, A., & Stewart, M., 2009. “Adoption of electronic
medical records in family practice: The providers’ perspective,” Family Medicine, 41(7),
508.
Terry, K., 2012. “EHR Adoption: U.S. Remains The Slow Poke,” Information Week. Retrieved
from http://www.informationweek.com/healthcare/electronic-healthrecords/ehr-adoption-
us-remains-the-slow-poke/d/d-id/1107410
Valdes I, Kibbe DC, Tolleson G, Kunik ME, Petersen LA., 2004. “Barriers to Proliferation of
Electronic Medical Records,” Informatics in Primary Care, 12(1):3-9.
Venkatesh, V., M.G. Morris, G.B. Davis, and F.D. Davis, 2003. “User Acceptance of
Information Technology: Toward A Unifying View,” MIS Quarterly, 27, 3, pp. 425-478.
Vishwanath A, Scamurra SD., 2007. “Barriers to the Adoption of Electronic Health Records:
Using Concept Mapping to Develop a Comprehensive Empirical Model,” Health
Informatics Journal, 13(2):119-134.
67
Walter, Z., and Lopez, M. S., 2008. “Physician Acceptance of Information Technologies: Role of
Perceived Threat to Professional Autonomy,” Decision Support Systems (46:1), pp. 206-
215.
Weeger, A., Gewald, H., & Vriesman, L. J., 2011. “Do Risk Perceptions Influence Physician's
Resistance to Use Electronic Medical Records? An Exploratory Research in German
Hospitals,” AMCIS.
Whitten, D., and Wakefield R., 2006. “Measuring switching costs in IT outsourcing services,”
Journal of Strategic Information Systems, vol. 15, pp. 219–248.
Xue, Z., Gebremichael, M., Ahmad, R., Weldu, M. L., & Bagtzoglou, A. C., 2011. “Impact of
temperature and precipitation on propagation of intestinal schistosomiasis in an irrigated
region in Ethiopia: suitability of satellite datasets,” Tropical Medicine & International
Health, 16(9), 1104-1111.
68
APPENDIX
Measurement Items
69
My peers are supportive of the change to the new way of working with
SOI2
Hakeem.
Most people whom I deal with in my job encourage my change to the new
SOI3
way of working with Hakeem.
Most of my superiors think that the change to the new way of working with
SOI4
Hakeem is a good idea.
I was able to change to the new way of working with Hakeem easily based on
SFC1
my own knowledge, skills and abilities.
I was able to change to the new way of working with Hakeem without the help
SFC2
of others.
Self-Efficacy
I was able to change to the new way of working with Hakeem reasonably well
for Change SFC3 on my own.
I had confidence in my ability to change to the new way of working with
SFC4 Hakeem without any difficulties.
My hospital provided me with guidance on how to change to the new way of
OGS1
working with Hakeem.
The management at my hospital provided the necessary help and resources to
OGS2
Organizational enable me to change to the new way of working with Hakeem.
Support for I was given the necessary support and assistance to change to the new way of
Change OGS3 working with Hakeem by my hospital.
The assistance provided by my hospital made the change to the new way of
OGS4 working with Hakeem easier.
I would rather not comply with the change to the new way of working with
RES1
Hakeem.
User I would rather not cooperate with the change to the new way of working with
RES2
Resistance Hakeem.
RES3 I oppose the change to the new way of working with Hakeem.
RES4 I do not agree with the change to the new way of working with Hakeem.
70
Instrument Validation
71
RES1 0.968 X2: 8.420
df: 2
RES2 0.992 P < 0.05
RES 2.938 1.129 0.982 CFI: 1.00
RES3 0.985 RMSEA: 0.000
CI: 0.000 - 0.102
RES4 0.919 SRMR: 0.001
72
CHAPTER THREE:
ABSTRACT
healthcare organizations. However, evidence show that EHR systems are being resisted by
healthcare providers. Little attention in IS research has been paid to understand this
phenomenon. In this study we aim to reveal the main antecedents of EHR resistance and the
concepts shaping these constructs. We use Revealed Causal Mapping (RCM) methodology to
analyze data gathered from interviewing healthcare providers who are introduced to a newly
implemented EHR system in a large international hospital. We argue that better understanding
of EHR resistance will contribute to both practice and research. Therefore, in this study, we
explore the key determinants that drive users’ resistance to EHR system.
73
CHAPTER THREE:
INTRODUCTION
Healthcare organizations are in the process of changing the way healthcare providers
document, monitor, and share information about healthcare delivery from using Paper-Based
Recording Systems (PBRS) to using Electronic Health Records (EHR) (DesRoches et al, 2013).
Using EHR is expected to reduce medical errors, improve continuity and quality of care, and
facilitate availability of patients’ medical history (Hillestad et al, 2005). However, despite the
potential benefits EHR could provide, there are evidence that suggests EHR is being resisted by
the same people who would benefit from its use. In 2013, only 44 percent of hospitals in the U.S.
reported having and using EHR systems (DesRoches et al, 2013). This phenomenon is also
taking place around the globe. Case studies conducted in all of the U.S., U.K., France, and India
show that healthcare organizations in all of those countries shared similar EHR implementation
challenges (Stone, 2014). Resistance behaviors towards technology has been identified as one of
the salient reasons for the failure of new systems (Kim and Kankanhalli, 2009). This includes
EHR and other types of Health Information Technology (HIT). One example comes from the
Cedars-Sinai Medical Center at Los Angeles, where doctors resisted to use the newly
implemented Computerized Physician Order Entry (CPOE) system, which caused the system to
fail and result in a complete withdrawal after it was already implemented in two-thirds of the
solutions, IS research in this area is found to be limited (Samhan and Joshi, 2015; Lapointe and
74
Rivard, 2005). In this paper, motivated by the urgency of understanding HER resistance coupled
by the limited prior efforts at explaining EHR resistance, we explore the following research
question: What are the key determinants that drive users’ resistance to EHR system. Our goal is
to go beyond the typical findings available in the literature and provide more insights by
conducting a qualitative analysis. We believe that while quantitative analysis could explain and
predicts EHR resistance behaviors, a qualitative analysis could provide deeper and richer insights
about individuals’ experiences with the new EHR system that help shape their resistance
ended questions to capture their perceptions of a newly implemented EHR system. To analyze
commonly used to identify constructs and linkages revealed from respondents’ statements
(Nelson et al., 2000). RCM has proven to be useful in studying emerging phenomena that require
rich and contextualized understanding (Narayanan and Armstrong, 2005). Given that our goal is
to understand the perceptions of individuals interacting directly with the newly implemented
EHR system, we believe that the RCM method is appropriate for our investigation of these
perceptions.
Some early IS research was conducted to uncover the concept of technology resistance
(e.g., Hirschheim and Newman, 1988; Keen, 1981; Markus, 1983). However, recent reviews of
the literature (Samhan, and Joshi, 2015; Lapointe and Rivard, 2005) show that limited work has
been conducted to explain resistance behaviors especially within the healthcare context. Prior
work in IS aimed to explain the resistance phenomenon based on the concept of perceived
threats. Perceived threats are identified by expressions such as "overwhelming emotional pain"
75
(Freud 1919) or "the perception of a dangerous situation" (Marakas and Hornik 1996). IS
research suggested that new technology will be resisted when people believe it will result in a
threat situation such as loss in individual’s status, revenue, or power. Perceived threat was
explained in (Lapointe and Rivard, 2005) case study of physician resistance. They posited that
when a system is introduced, users will first assess it and then make predictions about the
consequences of its use. If results were threatening, resistance behaviors will occur. Similarly,
(Bhattacherjee and Hikmet, 2007) tested healthcare providers’ behaviors towards a new
Computerized Physician Order Entry system (COPE) and considered perceived threats to be an
important agent that helps determine individuals’ attitude towards the new technology. They
suggested that higher levels of perceived threat would result in greater resistance towards the
new system. The work of (Joshi, 1991) viewed perceived threats as the inequity which leads
people to resist the implementation of a new system. The work of (Dent and Goldberg, 1999)
suggests that when people resist change they are in fact reacting to the threats associated with
that change. Another stream of research explained technology resistance by applying resistance
to change theories (e.g. Kim and Kankanhalli, 2009; Joshi, 1991; Samuelson and Zekhauser,
1988). It is suggested that resistance behaviors could be predicted by the evaluation of the net
benefits of using the new technology. Antecedents of this theoretical perspective deal with the
outcomes of changing from the current way of working to the new way after implementing the
new technology (e.g. gaining more benefits), while others focus on the costs associated with the
change. Some other work explained EHR resistance by lower levels of adoption (e.g. Boonstra
and Broekhuis, 2010; Gan and Cao, 2014; Najaftorkaman et al, 2015; Jha et al, 2008). However,
in our work we aim to conceptualize resistance distinctively from the use construct for a number
of reasons. First, the unique characteristics of the healthcare context makes it different from the
76
general corporate environment in which the techchnology use theories were applied (Smith et al.,
2014). Thus, we suggest that applied those theories will have limited explanation power when
aiming to explain EHR resistance. Second, prior technology use theories explained voluntarily
use behaviors, whereas in our study the EHR system is mandated to all healthcare providers
and Hikmet, 2013). Finally, it is found that resistance and use behaviors may coexist (Lapointe
and Rivard, 2005). For example, resistance may be manifested in the form of indirect usage or
delayed usage. This suggests that both use and resistance constructs are distinctive. In summary,
the review of prior IS literature revealed that prior findings on technology resistance are mixed
and that research on EHR resistance are limited. Additionally, we found that prior studies had
limited explanation of the concepts shaping the antecedents of resistance. For example, prior
findings suggested that perceived threat is a main antecedent to technology resistance, however
limited explanation was given to what users may have considered to be a threat. In this paper we
aim to provide a qualitative exploration to the concepts and constructs serving as determinants
THEORY IN USE
Our study is broadly informed by the theoretical foundations of the User Resistance
Model (URM) (Kim and Kankanhalli, 2009). The URM explains user resistance to the
combing three main theories: (1) the Status Quo Bias Theory (SQBT) (Samuelson and
Zeckhauser, 1988), (2) the Equity Implementation Model (EIM) (Joshi, 1991), and (3) the
Theory of Planned Behavior (TPB) (Ajzen, 1991, 2002). (Figure 3.1) shows how constructs from
the 3 theories correspond to each of the URM constructs. We draw upon the semantics of the key
77
constructs introduced in URM to help us reveal the main concepts that influences resistance
behaviors towards the EHR system. The URM suggests that user resistance is predicted by
perceived value which is determined by evaluating the costs and benefits associated with the
change to the newly implemented technology. It also includes concepts from the technology
acceptance literature including Self-Efficacy for change, Organizational Support, and Colleague
Opinion.
Figure 3.1: The Integrative Framework of URM (Kim and Kankanhalli, 2009)
According to URM, individuals begin to evaluate the new technology based on the
concept of net-benefits (Kim and Kankanhalli, 2009; Joshi, 1991; Samuelson and Zeckhauser,
1988). This evaluation would be mainly based on the concepts of switching benefits and
switching costs. Switching benefits correspond to the increase in outcomes (e.g., improved job
productivity) and the decrease in inputs (e.g., performing job faster) from the EIM. Based on
78
URM, it is suggested that the switch to the new EHR may result in benefits as enhancements in
the job productivity and performances of employees, as a result these enhancements would be
associated with certain rewards, therefor it would increase the perceived value of the change and
decrease the perceived threat of the change. The second agent in the evaluation phase is
switching costs. Switching costs correspond to the decrease in outcomes (e.g., less job
productivity) and the increase in inputs (e.g., performing a task slower) from the EIM.
Additionally, switching costs correspond to the transition costs, uncertainty costs, and sunk costs
from the SQBT. Transition costs in our study are the costs associated with the change from the
PBRS to the new EHR system. Any switch from the current status quo to a new status incurs
transition costs (Kim, 2011). When undertaking the transition to the new system, users must
acquire new skills and knowledge in order to use the new system and to perform their tasks
(Burnham et al, 2003). Users would strongly oppose the change when it requires transition costs
and have a tendency to keep their current status quo (Samuelson and Zekhauser, 1988).
Uncertainty costs are the perceived levels of risk on users’ performances associated with the
change (Burnham et al, 2003). If users have limited knowledge about the new system, they will
be uncertain about their job performance with this system, and would perceive this as a risk
(Whitten and Wakefield, 2006), this would produce unpleasant psychological reactions such as
anxiety, which in turn would influence their status quo bias (Inder and O’Brien, 2003). Sunk
costs are the investments and costs already put into mastering the current way of working (Kim,
2011). In our study, these are the costs in time and effort that have already been invested by users
in mastering the PBRS. URM suggests that switching costs would negatively affect the net
benefits of the switch because net benefits are evaluated by weighing benefits against costs.
Additionally, the corresponding costs from the SQBT (i.e. transition costs, uncertainty costs, and
79
sunk costs) would strongly influence the individual’s decision on change. Thus, switching costs
have negative effects on perceived value and positive effects on perceived threats. Which implies
Perceived value corresponds to the net equity and net benefits concepts from EIM and
SQBT respectively. Additionally, perceived value corresponds to the concept of loss aversion
from SQBT because it contributes explanations of how the perceived value of change is assessed
(i.e. losses are viewed larger than they actually are). URM suggests that people tend to like
maximizing value when making a decision (Sirdeshmukh et al, 2002), if the perceived value of
the change is low, then users are likely to have greater resistance to change (i.e. switching to the
new EHR system). On the other hand, if the perceived value is high, then users are likely to have
lower resistance to switching to the new EHR system. Thus, perceived value would have
Self-efficacy for change refers to the internal means that can enhance achieving control
of the changed situation (i.e. switching to the new EHR system) (Ajzen 2002). Self-efficacy
corresponds to the control concept from the psychological commitment category of the SQBT. If
individuals feels that there will be a loss in control associated with the change to the new
unfamiliar way of working, then status quo bias would be observed. This loss of control may be
manifested as low self-efficacy. URM suggests that different levels of self-efficacy for change
would result in different levels of anxiety and uncertainty regarding the change (Bandura 1995;
Compeau et al. 1999), which implies differences in their evaluation of the change. Individuals
with greater self-efficacy would evaluate the change as beneficial, on the other hand, individuals
80
The organizational support for change also corresponds to the control concept of the
psychological commitment category from the SQBT. URM suggests that organizational support
mechanisms such as training and providing resources would minimized the perceived difficulty
of adapting to the new EHR system (Lewis et al. 2003), and in turn would reduce costs of time
and effort required to learn the new way of working. Thus, organizational support for change
would influence individuals’ evaluations of the change. The more support provided by the
hospital, the more healthcare providers would perceive the change as beneficial, and vice-versa.
Social influence corresponds to the social norms concept from the psychological
commitment category of the SQBT. It is the degree to which an individual perceives that
important others believe he or she should use the new EHR system (Venkatesh et al, 2003).
URM has limited social influence sources to colleagues and peers but not superiors. URM,
suggests that individuals have the tendency to conform to their important others’ opinions
because of the need for social companionship and the fear of sanction for noncompliance (Ajzen
2002; Lewis et al. 2003). Important others’ opinion about the new EHR system may alter the
original perceptions an individual has about switching costs and benefits. Thus, social influence
RESEARCH METHODOLOGY
For this study we used the Revealed Causal Mapping (RCM) method, a sub-category of
cognitive mapping that involves an inductive process of revealing constructs and linkages
between constructs, all from the respondents’ statements (Narayanan and Armstrong, 2005).
RCM is used commonly in the management and organization science field, as reflected in the
special issues on the RCM method and its application in the Journal of Management Studies
(1987) and Organization Science (1996). In IS research, (Nelson et al., 2000) applied RCM to
81
identify the skills and expertise requirements in IS maintenance, while (Deng and Chi, 2012)
used this method to reveal system usage problems and causes in organizational use of business
intelligence systems. We intend to investigate what makes people oppose the change from the
PBRS and resist the use of the new EHR system, rather than to hypothesize or test cause-and-
effect relationships. Thus, we believe that the RCM method is appropriate for identifying key
constructs and linkages in this complex and emerging phenomenon. Informed by prior
The sample used for this study is a combination of all stakeholders of the system (i.e.
physicians, nurses, lab specialists, radiologists, pharmacists, admin staff, and resident medical
students). We collected responses through open ended questions. Participants received a link to
the questions on their email, and their answers were retrieved automatically after submission via
a research software. This was repeated twice, one for a pilot study and one for the main study.
For the pilot study we targeted only employees who work at the dermatology unit at the hospital.
We collected data for the main study from the entire hospital, but employees who participated in
the pilot study were specifically asked not to answer the questions for the second time (i.e. the
main study). We collected a total of (59) responses. Descriptive statistics of all (59) Interviewees
82
>60 0
Position Physician 20
Nurse 18
Radiologist 4
Lab Specialist 5
Pharmacist 4
Admin Staff 7
Medical Student 1
Total 59 (100%)
Our sample size is similar to that in prior IS studies using the RCM method (Deng and
Chi, 2012; Nelson et al., 2000). Studies based on the RCM method use the point-of-redundancy
concept to calculate the adequacy of sampling (Narayanan and Armstrong, 2005). As the point of
redundancy is not known until RCMs are constructed, the number of required respondents is a
judgment call. We adopted a sample size consistent with (Nelson et al., 2000), in which 50
experts in the IS maintenance context were interviewed. The point of redundancy in our data
analysis (as explained below) suggested that our sample was sufficient for exhausting the
Data Analysis
respondent’s explicit statements. We used a four-step process: (1) data elicitation, (2)
construction of revealed causal maps, (3) validation of the maps, and (4) interpretation of the
maps.
Step 1: Data Elicitation; we conducted the open-ended questions using Qualtrics. Table
3.2 in the Appendix includes the questions used for this study. Fifty-nine respondents typed their
responses online to the open-ended questions. In applying the RCM method, researchers often
rely on two major types of data: text-based data and interview data (Narayanan and Armstrong,
2005). As our research objective was to reveal the cognitive structure pertaining to the resistance
83
behavior domain, it was appropriate to use open-ended questions, similar to an interview
approach (Nelson et al., 2000), and to analyze the narratives of individual participants for their
experiences, perceptions, and beliefs with regard to the new EHR system.
derived revealed causal maps of perceptions step by step. First, we identified causal statements
from an informant’s narrative by looking for key words such as “because,” “therefore,” “so,”
“while,” and “if–then.” Then, we coded each causal statement into cause and effect and replaced
the “cause” and “effect” in a statement with the concepts (the key words or phrases) and linked
“cause” and “effect” with an arrowed line to obtain a causal map at the concept level. We
highlighted repeated words and phrases and grouped them into constructs. Finally, we
aggregated all concept-level maps across all informants to obtain a causal map at the construct
level, Table 3.3 shows an example of the described process. At the end a total of 23 concepts and
Table 3.3: An Illustration of the Five–Step Procedure for Constructing Causal Maps
Step 1: Identifying causal statement Example of causal statement:
“I find Hakeem threating to my job because it takes too much
time to fill data so I cannot serve more patients…. We need more
training and practice on how to use Hakeem faster.”
Step 2: Constructing raw causal maps Cause:
1. It takes too much time to fill in data.
2. We need more training on how to use it faster.
Effect:
I find Hakeem threating to my job.
Step 3: Coding Raw Phase (Coded Concept)
1. Takes much time to fill in data (Wastage of Time)
2. Needs more training and practice (lack of training)
3. Threating to my job (Perceived Threat)
84
Step 4: Recasting raw causal maps
into concept-level revealed causal
maps
Step 3: Validation; we reviewed the resistance literature to validate the concepts and
constructs identified in the RCMs. Similar to (Nelson et al., 2000), we used the point of
redundancy to evaluate the convergence of concepts elicited from the responses (Axelrod, 1976).
In our analysis, the 28 concepts converged at the 26th respondent--meaning that no new concept
was revealed in the map of the 27th respondent or beyond-- indicating the sufficiency of our
sample size. The additional relations yielded by adding the revealed causal map of each
85
Figure 3.2: Point of Redundancy
Step 4: Interpretation of the Maps; we examined the key constructs and linkages
uncovered in the map (see Figure 3.3). The constructs and their associations demonstrate what
RESULTS
Our data analysis revealed nine constructs that are essential to understanding resistance
towards the EHR System (EHRS). Four are causes: EHRS Benefits, Organizational Support,
EHRS Self-Efficacy, and Facilitating Conditions; and four are major outcomes: EHRS Costs,
EHRS Perceived Value, EHRS Perceived Threats, and EHRS Resistance. Our data analysis
revealed new constructs that were not considered in prior research on technology resistance (e.g.,
facilitating conditions). It also revealed the main concepts associated with each revealed
construct. (Table 3.4) lists the constructs and concepts identified in the study with the percentage
of each of these concepts occurrence over all of the other concepts within the same construct.
86
accomplish their tasks. Familiarity The feeling of comfort 0.39
(Compeau and with knowledge of the
Higgins, 1995) EHRS.
Required The need for resources 0.39
Training and services dedicated
to helping educate
healthcare providers on
EHRS-related topics.
Organizational The perceived Technical Services provided by 0.11
Support facilitation provided Support the hospital to help
by the hospital to make healthcare providers
individuals’ adaptation with and advice about
to EHRS easier. (Kim the EHRS.
and Kankanhalli, Training Any course of 0.89
2009) Sessions instruction on EHRS.
This may be delivered
in class, via a software
product installed on a
single computer,
through hospital’s
intranet, or over the
Internet as Web-based
training.
EHRS Costs The perceived Time Perceived waste in time 0.25
disutility an individual Consuming when using the EHRS.
would incur with the Technical Any technical problem 0.21
use of EHRS. Difficulties healthcare providers
(Derived from the encounter when using
study) the EHRS
Design Perceived costs 0.31
Difficulties associated with the
interface, design, and/or
the enforced process of
using the EHRS forms.
Access The limited possibility 0.23
Limitations of accessing the EHRS
remotely, from any
device, or at any time.
EHRS Benefits The perceived utility Design Includes the look and 0.34
an individual would Advantages feel of the interface, the
enjoy with the use of level of data
EHRS. (Derived from organization, and ease
the study) of use.
Data Data continues to be 0.39
Availability available at a required
level of performance.
Fast The response rate of the 0.20
Processing EHRS in terms of
loading forms and
storing data.
Accuracy Retrieving medical 0.07
information that is free
of mistakes and errors.
87
EHRS The overall evaluation Valuable The EHRS is of great (+)
Perceived of EHRS based on the (Worthy) worth to healthcare 0.49
Value comparison between providers. (-)
benefits and costs 0.51
(Kahneman and
Tversky 1979).
EHRS The extent to which an Lower Low effectiveness of 0.38
Perceived individual perceives Productivity productive effort.
Threat EHRS as dangerous. Evaluated in terms of
(Liang and Xue, the rate of
2009) accomplishing tasks per
unit of input (i.e.
patients).
Lower Low accomplishments 0.41
Performances of tasks or functions.
Potential Loss The probability of 0.21
of Data losing medical data
because of technical
errors, cyber-attacks, or
any other difficulty
when interacting with
the EHRS.
Facilitating The objective factors Alternative The ability to
Conditions that individuals User accomplish tasks by 0.17
perceive as facilitating having someone else
their ability to avoid interact with the EHRS.
using EHRS. Alternative The ability to 0.83
(Thompson et al., System accomplish tasks by
1991) using a method or
system other than the
EHRS.
EHRS Any act of opposition No Usage The absence of any 0.12
Resistance intended to prevent the EHRS usage behavior.
usage of EHRS Partial Usage Using part of the EHRS 0.53
(Lapointe and functionality and
Rivard, 2005; resisting other parts.
Markus, 1983) For example: using the
EHRS to retrieve data
and resist to enter the
data in the EHRS
forms.
Delayed Using the EHRS after a 0.18
Usage period of time by which
tasks become late or
postponed.
Indirect Usage Using the EHRS via an 0.17
alternative system or
person.
EHRS Self-Efficacy
88
When employees are introduced to a newly implemented system they normally manifest
different levels of confidence in their own ability to utilize the system in doing their jobs
(Venkatesh et al., 2003). Our study revealed that employees at the hospital first create judgments
of their abilities to interact with Hakeem to accomplish their tasks. Then use these judgements as
key factors that influences their perception of Hakeem’s value. Three main dimensions of self-
efficacy were revealed: Computer Literacy, Familiarity, and Required Training. Table 3.5
Computer Literacy “…what I really do not like about Hakeem (Female, 36, Nurse, Dermatology)
Familiarity “…at the beginning I thought it [Hakeem] (Male, 43, Doctor, Anesthesiology)
Required Training “The system needs a lot of training to know (Male, 28, Lab Specialist, Human
Organizational Support
employees as an exchange of their positive actions (Eisenberger et al., 1986). Thus, employees
tend to perform better to reciprocate received rewards and favorable treatment (Eisenberger et
89
al., 1986). However, our data analysis shows that employees refer to organizational support as
the facilitations provided by the hospital to make employees utilize Hakeem in an easy and more
comfortable fashion, regardless of employees’ performance. Our data analysis also revealed that
employees’ perception on Hakeem’s value is influenced by the levels of support provided by the
hospital. Two main dimensions of organizational support were revealed: Technical Support and
Training Sessions. Table 3.6 provides examples from respondents on each concept.
Technical Support “We have an on-call facility that provides (Male, 41, Office Manager,
Training Sessions “At the beginning we were taught by (Female, 33, Nurse, General
EHRS Costs
Prior research has conceptualized costs as those being associated with the switch or the
change to the new way of working with the newly implemented technology (Kim and
Kankanhalli, 2009; and Kim, 2011). However, in this study, the responces revealed that costs are
associated with the technology itself rather than the costs associated with the change. Four types
of costs were revealed from responses: Time Consuming, Technical Difficulties, Design
90
Difficulties, and Access Limitations. Table 3.7 provides examples from respondents on each
concept.
Time Consuming “…it takes a lot of time to fill-in all data (Female, 27, Nurse, Dermatology)
Technical “I do not like it when it is not working (Male, 33, Lab Specialist, Lab)
Design Difficulties “The system’s design is not very (Male, 38, Nurse, Dermatology)
me within 24 hours”
Access Limitation “…I cannot access the system from my (Male, 33, Lab Specialist, Lab)
my computer”
EHRS Benefits
91
The URM (Kim and Kankanhalli, 2009) was the first to utilize the concept of net-benefits
when theorizing technology resistance behaviors. This was adapted from the EIM (Joshi, 1991)
and the SQBT (Samuelson and Zekhauser, 1988). URM suggested that users determine how
valuable the new system would be based on evaluating the switching benefits against the
switching costs. However in this study we revealed benefits being associated with the technology
itself rather than the changing benefits. These benefits are: Design Advantages, Data
Availability, Processing Speed, and Accuracy of completing the task. Table 3.8 provides an
Design Advantages “What I really like about Hakeem is that (Female, 43, Material Management
any task”
Data Availability “I really like it because it gives (Male, 32, Nurse, ER)
Processing Speed “Hakeem changed the way of (Male, 42, Doctor, ENT)
be accomplished faster”
Accuracy “...And even patients’ labs test results now (Female, 34, Nurse, Dermatology)
92
needed accurately with no manual errors
employees’ positive feelings about Hakeem as well as their negative feelings. Positive perceived
value were coded from respondents who used words such as: “valuable”, “worthy”, and “like”
whereas negative perceived value were coded from respondents who used words such as: “not
valuable” and “dislike”. We found costs being associated directly with negative perceived
values. On the other hand benefits were associated directly with positive perceived value. This is
closely related to the conceptualization of perceived value used in the URM which suggests that
perceived value is a result of evaluating benefits against costs. Table 3.9 provides examples from
Negative Perception “I do not like it and see that it did not add (Male, 46, Nurse, Dermatology)
I can leave”
Positive Perception “If a medicine contradicts patient’s (Male, 27, Pharmacist, Internal
93
For example if the patient is pregnant
efficient”
(e.g, Bhattacherjee and Hikmet, 2007; Bhattacherjee et al, 2008; Lapointe and Rivard, 2005).
However very little was revealed on what is considered to be a threat. Users’ perceptions of
threat may vary on basis of technology, task, or context. The study conducted by (Liang and
Xue, 2009) suggested that threat would be determined by perceived susceptibility and perceived
severity. Perceived susceptibility is defined as the employees’ subjective probability that the new
system will negatively affect them. Perceived severity is defined as the extent to which
employees perceive the negative consequences caused by the system are severe. When users
believe that system makes them vulnerable to dangerous situations and that the consequence of
that is serious, only then they will perceive a threat (Liang and Xue, 2009). In this study we
revealed three dimensions of threat: Lower Productivity, Lower Performances, and Potential
Loss of Data. Table 3.10 provides examples from respondents on each of these concepts.
94
Concept Quote Respondent
Lower Productivity “…the threat of Hakeem is that it (Female, 30, Nurse, Dermatology)
patients”
Lower “…we are not working as fast as (Female, 24, Nurse, Dermatology)
Facilitating Conditions
agree that it is making an act easy to do (Thompson et al, 1991). Venkatesh et al., (2003) gave an
example of how returning items purchased online would be facilitated if no fee is charged to
return the item. For this study we define facilitating conditions as the factors that individuals
95
perceive as facilitating their ability to avoid using the new system. This is similar to the concept
of perceived avoidability by (Liang and Xue, 2009) which suggests that people will assess their
likelihood of avoiding a system by using specific safeguarding measures. We revealed two main
factors: Alternative User and Alternative System. Table 3.11 provides examples from
Alternative User “…I ask my nurse during the (Male, 51, Doctor, Primary Care Unit)
down”
EHRS Resistance
In this study we argue that resistance may coexist with or after the usage of the system,
and that resistance does not necessarily mean there is no user acceptance (Nah et al, 2004). Thus
there will be a variety of manifestations of the resistance behavior rather than only the non-usage
behavior. Our study revealed four main types of EHRS resistance behaviors: No Usage, Partial
Usage, Delayed Usage, and Indirect Usage. Table 3.12 provides examples from respondents on
96
No Usage “…to cope with the threats on (Male, 39, Doctor, Emergency Room)
needed”
Delayed Usage “…Digitizing the data could be (Female, 32, Nurse, ER)
visit”
Indirect Usage “…I understand the importance (Male, 39, Doctor, ENT)
job”
DISCUSSION
Our analysis of responses using the RCM methodology revealed the aggregated map
presented in Figure 3.3. It shows the key constructs and the associations between constructs. In
the map, an influencing factor (cause) is depicted by an oval shape and an outcome (effect) by a
square shape. The arrow indicates the link between a cause and an effect. Resistance behavior
towards EHRS was shaped by the confluence of constructs revealed in this study. The map is
based on 375 linkages identified from 59 respondents. The number associated with a linkage
reflects the percentage of that linkage over all the linkages identified.
97
Figure 3.3: Aggregated Revealed Causal Map of EHRS Resistance
Our findings show that EHRS Perceived Value is affected by EHRS Benefits, EHRS
Costs, Organizational Support, and EHRS Self-Efficacy. EHRS Benefits has a positive effect on
how employees perceive the value of Hakeem. As more benefits of Hakeem are considered, the
“…what makes it valuable is being able to view all patient information on the computer without requesting
info from different departments.” (Female, 30, Nurse, Dermatology).
On the other hand, EHRS Costs has a negative effect on how employees perceive the
value of Hakeem. The more costs of Hakeem are considered the less employees perceive
“I do not think of Hakeem as an added value…We can only access the system from certain machines inside
the hospital, we cannot use the system from outside the hospital…so sometimes I stay late to complete my
data entry into the system.” (Male, 35, Office Manager, Dermatology).
Organizational support was revealed in terms of technical support and training sessions.
It was found that the more support provided by the hospital the more the employees perceive
98
“Hakeem added value to my work…at the beginning I did not know how to use all of its features but after
we were offered some technical courses at an IT academy I really enjoyed all of options of Hakeem.”
(Male, 43, Nurse, Primary Care Unit).
However, it was reported by only one that she did not find Hakeem of high value and
later she complained about how the training sessions were provided outside of normal working
hours and it was unpaid. Thus, she found the training session as an impediment rather than a
support.
“…the main challenge is for me is to go attend Hakeem academy sessions after long days of work…these
sessions are outside of my working hours and they do not pay us extra to attend those classes.” (Female,
31, Nurse, ER).
This may result in a negative relation between organizational support and perceived
value. However, we do not have enough evidence in this study to support this argument, and we
encourage further research examine all possible attitudes towards organizational support.
EHRS Self-Efficacy had a positive effect on EHRS Perceived Value. The more
employees believe in their ability to use Hakeem, more they perceive it as valuable to them.
However, all of the responses were associated with the negative aspect of self-efficacy (i.e. only
low self-efficacy was reported) but this is posited as a positive effect because it was associated
“I do not see the difference [in value] when doctors use Hakeem…some doctors are not used to work with
computers so they are not gaining any value when asked to use a computer to do their work.” (Male, 49,
Doctor, Dermatology).
We also found that EHRS Perceived Threat was affected by ERHS costs. The more costs
were considered by employees the more the perceived Hakeem as a threat to their job.
“…when it is slow or it is hanging [not responding] some data could be lost and we will have to enter it
again, this causes a lot of time to be wasted and patients’ appointments will be delayed.” (Female, 32,
Nurse, Dermatology).
Finally, EHRS Resistance was found to be affected by EHRS Perceived Value, EHRS
Perceived Threat, and Facilitating Conditions. EHRS Perceived Value had a negative effect on
EHRS resistance. This means the less Hakeem was perceived as valuable to the employees the
99
more the employees would resist Hakeem. We did not find responses on how higher value
perceptions may result in less resistance. However, we consider this implicitly explained within
this linkage.
“…I think of Hakeem as a second choice with the old system. Today we cannot finish the patient visit
without finishing all forms and reports, this takes time and most of this information at my clinic is not
applicable because I care for kids. So I write all inputs on the patient’s folder and by the end of the day I
ask someone [nurse] to enter the data…this makes the flow of patients faster [Indirect Usage].” (Female,
41, Doctor, Pediatric Unit).
EHRS Perceived Threats was found to have a positive effect on EHRS Resistance. The
more employees consider Hakeem as a threat the more they will manifest shapes of resistance.
“Hakeem is a threat to my job because I spend too much time with the system and less with patients…this
effects my productivity…as a safeguard I like to keep the paper records in place to be used when needed.
Digitizing the data could be done after we finish the patient visit [Delayed Usage]” (Male, 32, Nurse, ER)
Facilitating Conditions was also found to have a positive effect on EHRS Resistance. The
higher the assessment results of employees’ likelihood of avoiding Hakeem by using specific
alternatives, the more likely employees would use the alternate rather than using Hakeem.
“…to avoid the threats I use the paper filing system when needed [Alternative System].”
(Male, 49, Doctor, ENT).
“…and have someone with us during the check up to input data during evaluations [Alternative User].”
CONTRIBUTIONS
For Research
Our study has a number of contributions to the IS research. The study provides
qualitative insights on the EHR resistance phenomenon which is not well understood in IS
(Samhan and Joshi, 2015). We revealed a new construct, facilitating conditions, which emerged
from this study and was not tested in any of the prior IS resistance theories. Furthermore, we
provided a broader explanation of the resistance phenomenon by combining concepts from the
URM theory, perceived threats, and the new construct: facilitating conditions. We revealed how
100
perceived threat and perceived value may both effect resistance simultaneously. This has not
been tested in any of the prior IS theories. Moreover, our results uncovered the underlying
concepts of each construct which provides a richer understanding of these constructs. Prior IS
theorizations use these constructs abstractly with very limited insights on the dimensions shaping
each construct. Our findings is the first in IS to account for all possible shapes of resistance.
Prior theories relied merely on the conceptualization of resistance as “no usage” (e.g. Kim and
Kankanhalli, 2009; Bhattacherjee and Hikmet, 2007). Finally, our findings provided a revealed
causal map which can be the basis for empirical testing in future research.
For Practice
revealed that perceived threats, perceived value, and facilitating conditions affect resistance
directly. Thus, we argue eliminating threats and facilitating conditions as well as increasing the
perceived value, which could be achieved by enhancing benefits and reducing costs, is critical
when aiming to eliminate resistance behaviors. Also, it is very crucial to understand how the
provided support may influence healthcare providers’ decisions. Our findings shows that
organizational support has a positive effect on perceived value. However, there was one response
which considered the support as an overhead which leads to negative feelings about the EHR
system. This may be due to inappropriate delivery of the support (e.g. long training sessions after
efficacy would have direct impacts on their perception of the value of the EHR system. It is also
important to explain and emphasize the advantages of the system, this would help in increasing
perceived value and reducing costs which in turn would reduce resistance behaviors.
101
Our findings could be tested on different technologies and in different contexts. This would help
better understand resistance in the context of which it was tested in, and that would serve as the
first step towards solving problems associated with resistance behaviors. Our study provides
insights about where EHRS designer efforts and resources should be directed in order to develop
Our study has a number of limitations. First, the sample size may be a concern. We
recommend conducting empirical studies on a larger scale to validate and modify the RCM
developed in this study. Second, because our data was collected from a single source, validation
concerns may arise. Future research can apply a longitudinal study or collect data from multiple
sources to validate the findings. Finally, we did not find responses on the possible effects of
social influence on the revealed constructs in this study. We encourage future research to
consider different types of social influence when testing the revealed causal map.
102
REFERENCES
Ajzen, I, 2002. Constructing a TPB questionnaire: Conceptual and methodological
considerations.
Armenakis, A. A., and Harris, S. G., 2002. Crafting a change message to create transformational
readiness. Journal of organizational change management, 15(2), 169-183.
Axelrod, R. 1976 Structure of Decision: The Cognitive Maps of Political Elites, Princeton
University Press, Princeton, NJ.
Bandura, A. (1995). Self-efficacy in changing societies. Cambridge university press.
Bates W, Leape LL, Cullen DJ, Laird N, Peterson LA and Teich JM (1998) Effect of
computerized physician order entry and team intervention on prevention of serious
medication errors. Journal of American Medical Association 280 (15), 1311–1316.
Bhattacherjee, A., and Hikmet, N. 2007. "Physicians' resistance toward healthcare information
technologies: a dual-factor model." System Sciences, 2007. HICSS 2007. 40th Annual
Hawaii International Conference on, IEEE, 2007.
Bhattacherjee, Anol, Christopher J. Davis, Neset Hikmet, and Varol Kayhan, 2008. "User
reactions to information technology: Evidence from the healthcare sector." Association
for Information Systems.
Bovey, W. H., and Hede, A. (2001). Resistance to organisational change: the role of defence
mechanisms. Journal of Managerial Psychology, 16(7), 534-548.
Burnham, T., Frels J., and Mahajan V., 2003 “Consumer switching costs: A typology,
antecedents, and consequences,” J. Acad. Marketing Sci., vol. 31, no. 2, pp. 109–203.
Burnkrant, R. E., and Cousineau, A. (1975). Informational and normative social influence in
buyer behavior. Journal of Consumer research, 206-215.
Compeau, D., Higgins, C. A., and Huff, S. (1999). Social cognitive theory and individual
reactions to computing technology: A longitudinal study. MIS quarterly, 145-158.
Davis, F. D., 1989 “Perceived Usefulness, Perceived Ease of Use, and User Acceptance of
Information Technology”, MIS Quarterly, 13, 3, pp. 319-340.
Deng, X., and Chi, L. (2012). Understanding Post-adoptive Behaviors in IS Use: A Longitudinal
Analysis of System Use Problems in the Business Intelligence Context. Journal of
Management Information System, 29(3), 305-340.
Dent, E. B., and Goldberg, S. G. (1999). Challenging “resistance to change”. The Journal of
Applied Behavioral Science, 35(1), 25-41.
DesRoches, C. M., Charles, D., Furukawa, M. F., Joshi, M. S., Kralovec, P., Mostashari, F. &
Jha, A. K. (2013). Adoption of electronic health records grows rapidly, but fewer than
half of US hospitals had at least a basic system in 2012. Health Affairs, 10-1377.
103
Deutsch, M., and Gerard, H. B. (1955). A study of normative and informational social influences
upon individual judgment. The journal of abnormal and social psychology, 51(3), 629.
Eisenberger, R., Jones, J. R., Aselage, J., and Sucharski, I. L. (2004). Perceived organizational
support. The employment relationship: Examining psychological and contextual
perspectives, 206-225.
Electronic Health Solutions (EHS) http://www.ehs.com.jo/en
Fine, M., Weis, L., Weseen, S., and Wong, L. (2000). For whom. Qualitative research,
representations, and social responsibilities. In NK Denzin and YS Lincoln (Eds.),
Handbook of qualitative research, 2, 107-131.
Freud, S. (1919). Turnings in the ways of psychoanalytic therapy. Collected papers, 2, 392-402.
FREUDENHEIM M (2004). Many hospitals resist computerized patient care. New York Times,
http://www.nytimes.com/2004/04/06/technology/06errors.html.
Halamka, John. “Healthcare is different”, (2011)
http://geekdoctor.blogspot.com/2011/08/healthcare-is-different.html
Hare, K., Whitworth, B., Deek, F., and Norris, T., 2006 “Identifying Barriers to IT Adoption in
Clinical Health Care: A Pilot Study”, Health Care and Informatics
Review Online, (2006), ISSN 1176-4201, http://hcro.enigma.co.nz/website/index.cfm
Hillestad, R., Bigelow, J., Bower, A., Girosi, F., Meili, R. and Scoville, R. 2005. Can Electronic
Medical Record Systems Transform Health Care? Potential Health Benefits, Savings, and
Costs. Health Affairs, 24, 1103-1117.
Hirschheim, R., and Newman, M. (1988). Information systems and user resistance: theory and
practice. The Computer Journal, 31(5), 398-408.
Inder, B., and O’Brien, T., “The endowment effect and the role of uncertainty,” Bull. Econ. Res.,
vol. 55, no. 3, pp. 289–301, 2003.
Joshi, K. 1991. “A Model of User Perspective on Change: The Case of Information Systems
Technology Implementation”, MIS Quarterly, 15, 2, pp. 229-242.
Keen, P.G., 1981. "Information Systems and Organizational Change," Communications of the
ACM, 24, 1, pp. 24-33.
Kelman, H. C. (1958). Compliance, identification, and internalization: Three processes of
attitude change. Journal of conflict resolution, 51-60.
Kim, H. W. (2011). The effects of switching costs on user resistance to enterprise systems
implementation. Engineering Management, IEEE Transactions on, 58(3), 471-482.
Kim, Hee-Woong, and Atreyi Kankanhalli. 2009. "Investigating user resistance to information
systems implementation: a status quo bias perspective”, MIS quarterly 33, 3.
104
Kotter, J. P., and Schlesinger, L. A. (2008). Choosing strategies for change. Harvard business
review, 86(7/8), 130.
Lapointe, L., and Rivard, S. 2005. .“A Multiple Model of Resistance to Information Technology
Implementation.” MIS Quarterly (29:3), pp. 461-491.
Lapointe, L., and Rivard, S., 2007. “A Triple Take on Information System Implementation”,
Organization Science, 18, 1, (2007), pp.89-107
Lewis, W., Agarwal, R., and Sambamurthy, V. 2003. .“Sources of Influence on Beliefs about
Information Technology Use: An Empirical Study of Knowledge Workers.” MIS
Quarterly (27:4), pp. 657-678.
Liang, H., and Xue, Y. (2009). Avoidance of information technology threats: a theoretical
perspective. MIS quarterly, 71-90.
Marakas, G. M., and Hornik, S. 1996. .“Passive Resistance Misuse: Overt Support and Covert
Recalcitrance in IS Implementation” European Journal of Information Systems (5:3), pp.
208-219.
Markus M. L. 1983. .“Power, Politics, and MIS Implementation” Communications of the ACM
(26:6), pp. 430-444.
Martinko, M. J., Henry, J. W., and Zmud, R. W., “An Attributional Explanation of Individual
Resistance to the Introduction of Information Technologies in the Workplace”, Behavior
and Information Technology, 15, 5, (1996), pp. 313-330.
Mekhijan HS, Kumar RR, Kuehn L, Bentley TD, Teater P, Thomas A, Payne B, and Ahmad A
(2002). Immediate benefits realized following implementation of physician order
entry at an Academic Medical Center. Journal of American Medical Informatics
Association 9(5), 529–539.
Miller, V. D., Johnson, J. R., and Grau, J. (1994). Antecedents to willingness to participate in a
planned organizational change.
Nah, F. F., Tan, X., and Teh, S. H. 2004. .“An Empirical Investigation on End-Users.’
Acceptance of Enterprise Systems.” Information Resources Management Journal (17:3),
pp. 32-53.
Narayanan, V.K.., and Armstrong, D. J. (2005). Causal Mapping for Research in Information
Technology. Hershey, PA: Idea Group Publishing.
Nelson, K.M., Nadkarni, S., Narayanan, V.K., and Ghods, M. (2000). Understanding Software
Operations Support Expertise: A Revealed Causal Mapping Approach. MIS Quarterly,
24(3), 475-507.
Samhan, B., and Joshi, K. D. (2015, January). Resistance of Healthcare Information
Technologies; Literature Review, Analysis, and Gaps. In System Sciences (HICSS), 2015
48th Hawaii International Conference on (pp. 2992-3001). IEEE.
105
Samuelson, W., and Zeckhauser, R. (1988). Status quo bias in decision making. Journal of risk
and uncertainty, 1(1), 7-59. Chicago
Sirdeshmukh, D., Singh, J., and Sabol, B. (2002). Consumer trust, value, and loyalty in relational
exchanges. Journal of marketing, 66(1), 15-37.
Smith, Tim, Gerald Grant, and Alejandro Ramirez. "Investigating the influence of psychological
ownership and resistance on usage intention among physicians”, System Sciences
(HICSS), 47th Hawaii International Conference on. IEEE, (2014).
SPIL TAM, SCHURING RW and MICHEL-VERKERKE MB (2004) Electronic prescription
system: do the professional USE IT. International Journal of Healthcare Technology
Management 6(1), 32–55.
Stone, C. 2014 “A Glimpse at EHR Implementation around the World: The Lessons the US Can
Lern,” The Health Institute for E-Health Policy (available at http://www.e-
healthpolicy.org/docs/A_Glimpse_at_EHR_Implementation_Around_the_World1_Chris
Stone.pdf).
TAYLOR R, MANZO J and SINNET M (2002) Quantifying value for physician order-entry
systems: a balance of cost and quality. Healthcare Financial Management 7, 44–48.
Thompson, R. L., Higgins, C. A., and Howell, J. M. “Personal Computing: Toward a Conceptual
Model of Utilization,” MIS Quarterly (15:1), 1991, pp. 124-143.
Tichy, N. (1983). The essentials of strategic change management. Journal of Business Strategy,
3(4), 55-67.
Venkatesh, V., and Davis, F. D. (2000). A theoretical extension of the technology acceptance
model: Four longitudinal field studies. Management science, 46(2), 186-204.
Venkatesh, V., Davis, F. D., and Morris, M. G. (2007). Dead or alive? The development,
trajectory and future of technology adoption research. Journal of the association for
information systems, 8(4), 267.
Venkatesh, V., M.G. Morris, G.B. Davis, and F.D. Davis, 2003. “User Acceptance of
Information Technology: Toward A Unifying View,” MIS Quarterly, 27, 3, pp. 425-478.
Wanberg, C. R., and Banas, J. T. (2000). Predictors and outcomes of openness to changes in a
reorganizing workplace. Journal of Applied Psychology, 85(1), 132.
Watson, G. (1971). Resistance to change. The American Behavioral Scientist, 14(5), 745.
Whitten, D., and Wakefield R., 2006. “Measuring switching costs in IT outsourcing services,” J.
Strategic Inf. Sys., vol. 15, pp. 219–248.
Williams, Michael D., Yogesh K. Dwivedi, Banita Lal, and Andrew Schwarz. 2009.
"Contemporary trends and issues in IT adoption and diffusion research", Journal of
Information Technology, 24, no. 1, (2009), pp. 1-10.
106
APPENDIX
Interview Questions
107
CHAPTER FOUR:
PERSPECTIVE
ABSTRACT
This study empirically tests the conceptual model of the Technology Threats Avoidance
Theory (TTAT) (Liang and Xue, 2009) in an international healthcare context to investigate
electronic healthcare avoidance behaviors. We tested the complete conceptual model and based
on the modification indices test we suggested a theoretical model with an aim to explain
avoidance behaviors towards electronic health records. We found that the electronic health
perceived effectiveness of the safeguarding measure, and the social influence of using the
safeguarding measure. We also controlled for a number of variables and found that levels of
education and number of healthcare providers in a medical unit has significant influence on
108
CHAPTER FOUR:
PERSPECTIVE
INTRODUCTION
Electronic Health Records (EHR) systems have the potential to provide numerous
benefits (Callen et al, 2013). It provides healthcare providers with important information about
patients’ medical history including previous diagnoses and medications, medication allergies,
treatments, surgeries, summaries of hospital admissions and discharges, lab test results, and
summaries of any previous doctor visits. All of this information could be retrieved in timely
fashion which helps healthcare providers make more appropriate decisions (Pope et al, 2013).
EHR systems can also help providers from different units of a hospital, or even from different
hospitals, to coordinate the care-giving process by knowing the medical history of the patient
they are treating (Cordell et al, 1998). Additionally, prior research suggested that the use of EHR
systems by healthcare providers can enhance quality of patient care and reduce medical errors
(Hillestad et al, 2005). These benefits cannot be achieved if providers avoid EHR systems
However, despite the great potential EHR systems has to generate substantial value to
healthcare providers, evidence of avoidance behaviors from healthcare providers towards EHR
are still reported (Kane and Labianca, 2011). We also find a dearth of IS research in this area and
the findings are mixed and unclear (Kellermann and Jones, 2013). Prior research has paid some
attention to the consequences of resistance to information systems such as (Markus, 1983; Poon
et al, 2004; Lapointe and Rivard, 2005; Bhattacherjee and Hikmet, 2008), however very little
109
attention has been given on IS avoidance (Liang and Xue, 2009) and particularly in the
healthcare context (Kane and Labianca, 2011). As a result, we know little concerning why
Health Information Technology (HIT) is being avoided and specifically why healthcare
providers avoid EHR systems in the postadoption stages. In this study we aim to understand the
We tested the complete conceptual research model from the Technology Threat
Avoidance Theory (TTAT) (Liang and Xue, 2009) to explain how healthcare providers develop
threat perceptions, evaluate safeguard measures, and engage in avoidance behavior towards the
EHR system. We undertake our study in the context of an international hospital that has enforced
the use of an EHR system. We used survey data to test the TTAT model. Findings and
discussions on how well these findings support the original conceptual model are provided in this
paper. This work is one of the very few attempts to provide empirical explanations about the
phenomenon of EHR avoidance. Plus, to our knowledge, this is one of the first studies to
empirically validate and test the entire TTAT model in the healthcare context. Doing so will not
only advance the theoretical understanding of EHR avoidance behaviors, but also offer practical
TECHNOLOGY AVOIDANCE
In this paper we define EHR avoidance as the process in which individuals aim to enlarge
the distance between their current safe state and the undesired end state caused by using the EHR
system (Carver and Scheier, 1982; Edwards, 1992). In our study we distinguish avoidance
behaviors from the acceptance and the resistance behaviors when theorizing EHR avoidance.
Avoidance and acceptance behaviors may coexist. For example, behaviors such as indirect usage
and delayed usage are considered avoidance behaviors while the technology is still being
110
accepted/used. We believe that EHR acceptance and EHR avoidance constructs are different and
one cannot be explained as the opposite of the other (Liang and Xue, 2009; Carver and White,
1994; Elliot, 2006; Elliot and Covington, 2001). The distinction between adoption and avoidance
constructs is more important in this study because we aim to explain avoidance in a healthcare
setting which has unique characteristics that makes it different from the more general “corporate
environments” in which technology adoption models have been successfully tested in prior
studies. Therefore, applying adoption models within the healthcare context will have less
explanatory power (Smith et al, 2014). Moreover, the adoption theories are designed to explain
voluntary IT usage and have limited capabilities in explaining mandated IT usage; which is the
case for most EHR implementation initiatives (Bhattacherjee and Hikmet, 2013). Nevertheless,
technology acceptance theories can provide useful findings when assessing the adoption of
safeguards which individuals use to avoid threats associated with using the EHR systems, yet
this approach does not provide a complete explanation of the technology avoidance
phenomenon. For example, physicians may be influenced to avoid the EHR system if they
considered the threat of losing valuable time using the EHR system while performing a medical
task. Healthcare providers must first perceive the loss in time as a threat and then can take
several actions such as using a paper-based system or asking someone else to fill in the EHR
while continue doing their medical task. If we only study adoption of the paper-based system, we
will form partial understanding of the avoidance phenomenon because we ignored the evaluation
of threat and alternative actions. Individual’s aim to avoid threats associated with the EHR
system not to adopt safeguards. The safeguards are only tools which help healthcare providers
111
Additionally, we believe that EHR avoidance behaviors differ from EHR resistance
behaviors mainly in terms of the stage in which each is manifested (Kane and Labianca, 2011).
Healthcare providers may first resist the change from working with the Paper-Based Recording
System (PBRS) to the newly implemented EHR system; and overtime, when the EHR system is
completely implemented and enforced, avoidance behaviors may become present (Kim and
Kankanhalli, 2009). Therefore, we focus our study on avoidance behaviors towards the EHR
system because the EHR system has already been implemented and mandated for all employees
LITERATURE REVIEW
Prior studies in the IS literature has focused on the resistance behaviors prior to system
implementation of an IS (example: Kim and Kankanhalli, 2009). However, relatively little effort
has been spent to investigated how avoidance occurs in post-implementations stages, especially
in the healthcare settings (Kane and Labianca, 2011). One of the few attempts to theorize
technology avoidance was the work of (Liang and Xue, 2009) who introduced the Technology
Threat Avoidance Theory (TTAT). It suggests that when users perceive a threat associated with
the technology they are using (threat appraisal), they become motivated to actively avoid the
threat and start to find safeguard that would help them avoid the threats associated with the
system (coping appraisal). Liang and Xue derived a measurement model from the TTAT and
empirically tested it in the context of spyware threats (Liang and Xue, 2010), another model was
derived from TTAT and tested in the context of telemedicine (Xue et al, 2014; and Xue et al,
2015). Additionally, (Arachchilage and Love, 2013) tested the TTAT in the context of game
design framework for avoiding phishing attacks. Similar to the TTAT, the work of (Rho and
112
Ryu, 2011) examined the cognitive process of avoidance through coping behavior, they
explained how users first perceive a threat such as the exposure of personal information, which
then influences emotion-focused coping that in turn influences technology avoidance. (Hwang,
and tested its influence on the technology acceptance variables. Another study on HIT avoidance
was conducted by (Kane and Labianca, 2011), they examined HIT avoidance as a predictor of
In summary, the review of prior IS literature revealed a number of research gaps. First,
very few studies aimed to investigate avoidance in general and specifically avoidance towards
EHR systems. Second, there is an evident lack of empirically tested unified model of HIT
avoidance. Third, majority of the prior research was conducted in the North-American context
with very limited insights from global perspectives. In this paper we will empirically test the
complete conceptual model of TTAT, including all of its constructs, with a focus on the post
INVESTIGATIVE CONTEXT
Our study was conducted in the same hospital used for chapter two. It is a large public
hospital located in the capital city of Jordan, Amman. The hospital consists of multiple health
departments and 33 specialists’ clinics. The total inpatient capacity is 450 beds, with a plan of
adding 100 more beds by early 2016. The hospital has recently implemented an EHR system that
is mandated to all employees. It is one of two healthcare organizations in the country that have
completely implemented the EHR system. The project started in 2009, implemented in 2014, and
was enforced early 2015. No more paper records are being used at the hospital. The new EHR
system is named “Hakeem”, the Arabic word for “Doctor”. Hakeem was implemented through a
113
governmental program which is the first national e-health initiative in Jordan. The system was
Communications Technology, Royal Medical Health Awareness Society and Private Hospitals
Association.
Understanding avoidance behaviors towards the EHR system requires special attention to
the unique characteristics of the healthcare context. Healthcare organizations differ from other
corporate environments in many ways. First, healthcare providers, in most cases, are not
employees of the hospital but rather work as private contractors delivering care through different
hospitals (Halamka, 2011). This allows them to decide whether to use or avoid the EHR system
even if it was mandated by the healthcare organization. Second, healthcare providers usually
perceive entering data into the EHR system as an unnecessary task that hurdles them from
performing their original job of providing care to patients, and some consider this task
inappropriate for the physicians because it requires doing tasks that represents threat to their
professional status (Lapointe and Rivard, 2005). Third, some healthcare providers have concerns
about how EHR systems affects their communication with patients such as loss of eye contact
while using the computer (Cotea, 2010). Finally, healthcare providers fear that the EHR systems
will take time away from actually caring for the patient and make the art of their job impersonal
(Sassen, 2009).
The TTAT (Liang and Xue, 2009) explains individuals’ behavior of avoiding threats
associated with the use of technology. It suggests that threats are influenced by the levels of its
114
severity and individual’s believes of their susceptibility to these threats. If a threat is realized,
individuals begin to aim to avoid the threats by using a chosen safeguarding measure. The
safeguarding measure does not necessarily have to be an IT source but also could be an
alternative way to complete the task (e.g. using the paper-based record system (PBRS) as a
safeguard to avoid using the EHR system). TTAT explains that the safeguarding measure would
how effective could it be in avoiding the threat. For example, the individual assessment
regarding how effectively the PBRS can be to avoid a threats associated with using the EHR
system. Safeguard cost refers to the physical and cognitive costs associated with performing the
using the safeguard measure. Self-efficacy is individuals’ confidence in being able to take the
safeguard measure. However in the context of our study, safeguarding measures are reported as
using PBRS when avoiding the ERR system. Because the PBRS requires no special skill and has
been the original way of performing tasks at the hospital before the implementation of the EHR
system, we did not include self-efficacy of using paper records in this study, and only measured
for the effect of safeguard costs and safeguard effectiveness on perceived avoidability.
According to TTAT, after perceiving a threat users would react by performing some type
of coping. Users would be motivated to avoid the system (problem-focused coping) if they knew
how to avoid the system and had the ability to do so, otherwise if users believe that the threat
cannot be fully avoided by taking safeguarding measures, they would passively avoid the system
by engaging in emotion-focused coping in which they create a false reality about the threat
environment (Liang and Xue, 2009). However, (Liang and Xue, 2010) explained how emotion-
115
focused coping reduces users’ awareness of the threat which affects their perceptions of reliable
safeguard mechanisms. This would result in less impact of the safeguarding measure on
think about the threat to escape from the situation. Although users know that the threat is present,
they would block it from their rational thinking. Therefore, the relationship between safeguard
effectiveness and avoidance motivation will be weaker as threat levels get higher. Thus, in this
study we only included the problem-focused coping mechanism. This did not affect the
theoretical model because the emotion-focused coping construct is an outcome of the model
rather than a predictor. TTAT also included two influential variables: Risk Tolerance and Social
Influence. Risk tolerance was defined as the “minimum discrepancy between the undesired end
state and the current state that users are able to tolerate” (Liang and Xue, 2009). This means that
users with different levels of risk tolerance would endure different levels of threats associated
with using the EHR system. Thus, it is suggested that risk tolerance has direct influence on
perceived threat. Social influence in this study is suggested to affect how employees perceive
threats associated with the EHR system, their perceptions on safeguarding measures (i.e. costs
and effectiveness), and their motivation to avoid IT threats. Social influence on perceptions of
threat associated with the EHR system and social influence on the safeguarding measures (i.e.
variables: age, gender, and level of education. It is argued that employees with different positions
would perceive different threats and would tolerate risk differently. Thus, we included position at
hospital as a control variable. Addition, because social influence may have indirect effect on
avoidance behaviors and a direct effect on perceived threats, safeguarding measures, and
116
avoidance motivation, it was appropriate to control for the number of employees in a unit. We
also controlled for the size of units at the hospital based on the number of beds in a unit.
The main constructs of the conceptual model of TTAT (Liang and Xue, 2009) were
adopted in this study. Table 4.1 presents a summary of all the constructs in the model, their
117
Individuals’ subjective assessment that having the
Perceived
PEFF PBRS can effectively avoid the threats associated with
Effectiveness
using the EHR system.
The individuals’ assessment of the levels of control
Perceived
PAVO they believe they have over avoiding the EHR system
Avoidability
by using the PBRS.
The degree to which users of the EHR system are
Avoidance
AVMO motivated to avoid associated threats by using the
Motivation
PBRS.
The process of aiming to avoid the use of the EHR
Avoidance system to enlarge the distance between a secure state
AVBH
Behavior and the unsafe state associated with the use of the EHR
system.
TTAT explored avoidance behavior from a threat perspective. It suggests that once a
threat is perceived users would evaluate safeguarding measures to be taken in an aim to avoid the
threat associated with the use of the system. Levels of susceptibility, severity, risk tolerance, and
social influence impact users’ threat perceptions. Once a threat is perceived, users evaluate
safeguarding measures by how effective these measures are in terms of avoiding the threat
associated with the use of the system, and by how costly is it to take these measures as an
alternative. Both effectiveness and cost of the safeguard along with social influence predict
users’ perceived avoidability. In turn, perceived threat and perceived avoidability influence
avoidance motivations which has a direct effect on the avoidance behavior. The conceptual
model is illustrated in Figure 4.1 and the model hypothesis are listed in Table 4.2.
118
Figure 4.1: TTAT Conceptual Model (Liang and Xue, 2009)
119
METHODOLOGY
Instrument Development
Existing validated scales were adopted for this study. We adopted the instrument items
used for the model derived from TTAT (Linag and Xue, 2010). We also adopted items from
different sources to capture all of the constructs of TTAT that were not previously used in
empirical models, these include: First, Perceived Avoidability which is defined as the
individuals’ assessment of the levels of control they believe they have over avoiding the EHR
system by using the PBRS. Perceived Avoidability is formulated by considering how the
safeguarding measure effectively avoids the associated threat and what costs are associated with
the safeguarding measure. The more a user believes the safeguard is effective and with low costs
the more the likelihood of feeling they have control over avoiding the threat associated with
Hakeem. In the context of our study, the suggested safeguarding measure is using the PBRS to
avoid Hakeem. Items for Perceived Avoidability were adopted from the perceived behavioral
control scale (Ajzen, 2006). Second, Social Influence items were adopted from the scale used by
(Venkatesh et al, 2003) which includes social factors, subjective norms, and image items. First,
subjective norms: these are based on the concept of what do the important others think about
using the system. The source of these norms could be coming from peers and colleagues in the
form of knowledge transfer by sharing past experiences and opinions (informational influence)
(Burnkrant and Cousineau 1975; Deutsch and Gerard 1955), or could be from superiors
(Kelman, 1974). Compliance is affected by the need for getting approval, rewards, or the fear of
punishment. Internalization is realized by harmonizing user’s goals and values about the system
with others’. Identification is shaped by users self-define their relationship with others. Second,
120
social factors: just like the internalization of the normative influence, these are individuals’
internalization of others subjective culture and the interpersonal agreements made with others in
specific social situations (Thompson et al, 1991). Third, image: this is the degree to which the
use of the system is perceived as an enhancement to individuals’ status in the social system
(Moore and Benbasat, 1991). We captured two types of social influence for this study: Social
Influence on using the system (Hakeem), and Social Influence on using the safeguarding
measure (PBRS). Third, Risk Tolerance items were adopted from (Cable and Judge, 1994; and
Judge et al, 1999). Finally, we added 2 items (AVBH3 & AVBH4) to capture Avoidance
Behavior, (Liang and Xue, 2010) had only 2 items for the construct and we needed at least 3 to
run Confirmatory Factor Analysis (CFA) successfully. All items were modified to fit the context
of our study. Measurement items were anchored on five-point Likert scales (1 = strongly
disagree, 5 = strongly agree). The instrument was reviewed by IS researchers before the study.
The final version of the questionnaire was developed for the main study as shown in Table 4.3 in
the Appendix.
The total estimated number of employees who work at the hospital is 5,500. Majority of
these employees are potential users of the new system. However administrative staff are less
likely to use the system as they do not participate in the care giving process, however they still
have access to the system and would be able to print out reports as needed. The sample is a
combination of all stakeholders of the system (i.e. physicians, nurses, lab specialists, radiologists,
pharmacists, admin staff, and resident medical students). Data was collected in 2 phases. First,
we collected data for the pilot study from employees who works only for the Dermatology unit at
121
After conducting Explanatory Factor Analysis (EFA) and Confirmatory Factor Analysis
(CFA) of the data collected for the pilot test, it was necessary to alter some of the items to
address a number of issues related to the psychometric properties of the study’s constructs.
Additionally, some new items were added to better capture the concepts of some constructs.
First, we dropped 5 items from the Perceived Severity construct because they had low loadings
range (0.40 – 0.60) and kept the other 5 items which had acceptable loadings range of (0.80 –
0.99), we also dropped one item from the Perceived Effectiveness construct that had a low
loading of (0.69). Second, Perceived Threats items were reworded to include the word “using”
instead of “switching”. At this stage the EHR system has been completely implemented and the
switch from the PBRS has already happened. Thus, it was appropriate to change the items to
better reflect reality. Third. We reworded the items of Perceived Costs to include the word
“previous” when referring to the PBRS to be clear that we refer to the old way of working. We
aim to capture the costs associated with going back to the old way of working with the PBRS as
a safeguarding measure to avoid threats associated with Hakeem. Fourth, for both types of Social
Influence (PBRS and Hakeem) we included “people in my hospital” in the items to eliminate
healthcare providers mistaking the word “people” for their friends or relatives. Fifth, we dropped
2 items from the Risk Tolerance construct because they were conflicting with the cultural context
of this study. This first item was “I like (or would like) to play the lottery”, and this was dropped
for our survey because gambling is considered as sinful action within the Jordanian culture. The
second items was “I generally hold out for the best price on something, even if it means waiting
a long time”, and this was dropped from our survey because consumer behaviors and market
strategies in Jordan are different and sale prices does not normally change with time, so this item
122
did not make sense to the Jordanian population. Finally, we found that item PAVO2 from
Perceived Avoidability was reverse codded so we fixed it for the main study survey.
After refining the survey items based on the pilot data’s psychometric properties, we
collected data for the main study from the entire hospital, but employees who filled the first
survey (i.e. pilot study) were specifically asked not to fill in the survey for the second time. We
collected a total of (220) responses, descriptive statistics are reported in Table 4.4.
Instrument Validation
To validate the survey instrument, the psychometric properties of the survey were
assessed by conducting Explanatory Factor Analysis (EFA) and Confirmatory Factor Analysis
(CFA) using MPlus 7.1 (Muthén and Muthén, 1998-2006). Cronbach’s α was performed using
SPSS 20 (IBM Corp. 2013). After the pilot data was collected, we refined the items to keep those
with statistical significant loading larger of (0.7). Cronbach’s α reliability tests for all constructs
exceeded (0.8). After the main data was collected, we again conducted CFA analysis and again
all items had a significant loading greater than (0.7). All constructs had Cronbach’s α values
123
exceeding (0.8). The CFA analysis provided strong support for our measurement model, which
suggested that the items under each of the constructs were adequately measuring the constructs.
Table 4.5 in the Appendix lists the items’ loadings and the model fit statistics for the
CFA.
The survey responses came in (2) waves: (W1: N=161; W2: N=59). We checked for
nonresponse bias by testing the difference in means between waves (See Table 4. 6). We found
no significant difference between the two respondent groups based on the sample attributes
RESULTS
We tested the TTAT theoretical model by applying Structural Equation Modeling (SEM)
using Mplus version 7.1 (Muthén and Muthén, 1998-2006). We applied the maximum likelihood
estimator with robust standard errors (MLR). Because our model is not saturated (i.e., not all
124
possible regression paths were included) we evaluated the model fit indicators depicted in Table
4.7.
The model fit indices showed ill fit of the model. First, according to (Bollen, 1998)
suggested that NC value outside the range (2.0 and 6.0) indicates poor fit. Second, the model had
a very low (CFI = 0.593), this indicates illness in the model fit. The rule of thumb for the CFI
and other incremental indices is that values greater than roughly (.90) may indicate reasonably
good fit of the researcher’s model (Hu & Beltler, 1999). Third, the Root Mean Square Error of
Approximation (RMSEA) is (0.258). According to the rules of thumb by (Browne and Cudeck,
1993) RMSEA value between (0.05) and (0.08) suggest reasonable error of approximation and
values greater than (0.1) suggests poor fit. We conclude that the model’s goodness of fit is
indices test using MPlus. This suggests possible paths to be added to or dropped from the model
to improve its goodness-of-fit. Also it would suggest alternative theoretical models that would
best fit our data. The suggested model and its new improved model fit indices are included in the
next section of this paper. The standardized path coefficients, its level of significance, and the
standard errors (S.E) are provided in Table 4.8, and results are depicted in Figure 4.2.
125
*p<0.05; **p<0.01; ***p<0.001
126
Our study found a number of hypotheses from TTAT that were not confirmed. First, the
interaction effect of perceived susceptibility and perceived severity on perceived threat was not
significant. This is an interesting finding that implies that if either variable takes a value of zero,
the other variable would still have an impact on perceived threat. For example, if healthcare
providers know that the EHR system will cause severe harm to their job, they will perceive the
EHR system as a threat even if they see no possibility for them to be harmed. Also, if healthcare
providers know that they are likely to be harmed from using the EHR system, they will perceive
it as a harm even if they do not view that harm as a serious problem. Second, Social influence on
using Hakeem had a positive direct effect on perceived threat. This means the more positive
influence about Hakeem the more users perceive it as a threat. This is a counter-intuitive finding
that contradicts the original theory which hypothesizes that social influence would have a
negative effect on perceived threat. This could be explained by the psychological reactance
theory (Brehm, 1966) which explains how psychological reactance occurs in response to threats
to freedoms and is used to deliberately taunt the authority who enforces these threats regardless
of the utility or disutility that the activity confers. Given that social influence includes the
opinions of healthcare providers’ superiors, then it is possible that healthcare providers manifest
there reactance behaviors in the shape of disliking and resenting the EHR system which was
reflected in their responses to perceiving Hakeem as a threat. We believe that this would require
further investigation. We suggest conducting interviews with users from our sample to better
explain this finding. Third, perceived cost of using the PBRS as a safeguarding measure had no
associated with the EHR system and begin to evaluate the possibility of using the PBRS as an
alternative way to accomplish their tasks, they would consider going back to the PBRS
127
regardless of how much trouble would be involved as long as the PBRS is perceived to be
effective in eliminating threats associated with using the EHR system. This is related to the loss
aversion concept from economics (Kahneman and Tversky, 1984) which suggests that people
tend to strongly prefer avoiding losses to acquire gains. When healthcare providers perceive the
EHR system to be a source of threat that will result in losses in their jobs, they would strongly
prefer avoiding it to eliminate these threats. However, healthcare providers find that enduring
costs related to using the PBRS is acceptable because it results in gaining a threat-relief situation.
This means that they valued the gains, in this situation, more than their losses. However, in
reality they are weighing the losses of using the PBRS much less than the losses stemming from
the threats associated with the EHR system. Forth, Perceived avoidability had no significant
impact on avoidance motivation. It is sufficient for healthcare providers to perceive the EHR
system as a threat to be motivated to avoid the system regardless of there was a known
safeguarding measure available for them to use. This implies having a non-usage behavior or a
resistance behavior where healthcare providers are motivated to simply not use the EHR system
before evaluating any other safeguarding measure such as using the PBRS. Our findings
indicates the limited explanation perceived avoidability has in this model. It serves as
endogenous independent variable rather than an exogenous predictor. Probably this explains why
studies that derived theoretical models from the TTAT have dropped perceived avoidability from
their model (Liang and Xue, 2010; Arachchilage and Love, 2013) and only tested the direct
effect of perceived avoidability antecedents on avoidance motivation. Fifth, the interaction effect
motivation. This confirms the former argument that healthcare providers are motivated by threats
alone to avoid the EHR system. Sixth, risk tolerance had no significant influence on perceived
128
threat. Barsky et al. (1997) suggest that individuals tend to show similar tolerant responses to
risky situations across settings. This implies that the no matter how much risk healthcare
providers are able to tolerate, it will not affect their perceptions of threats associated with the
EHR system. Evidence from the literature demonstrates that risk tolerance is a personal trait
related to demographic variables including age, gender, marital status, race, religion, education,
and income (Filbeck et al. 2005; Grable 2000; Hallahan et al. 2004). There is the possibility of
having some response bias in our sample in terms of the variables we did not control for (i.e.
marital status, race, religion, and income). Finally, we found that 3 of the control variables had
positive influence on avoidance behavior: number of employees in the unit, number of beds in
the unit, and age. However, gender, levels of education and position at the hospital had no effect
on the outcome variable. All other hypothesis in the original theory were confirmed. The model
avoidance motivation and (17%) in perceived avoidability. However, the main concern from
these findings is the lack of the model fit. In the following section we provide a suggested model
Suggested model
In an aim to find an alternative model with better fit indices. We had to make appropriate
changes to the original model while still being aligned with the theory. We first dropped
perceived avoidability form the new model because it has no contribution in mediating the
effects of its antecedent on perceived avoidance. Next, we ran a modification indices test using
MPlus to evaluate what additional paths to the model would provide better fit. The new model
included the paths between perceived avoidability antecedents (perceived cost, perceived
effectiveness, and social influence) and avoidance motivation. We believe this is an appropriate
129
addition to the model since the mediator (perceived avoidability) no longer exists. Additionally
modification indices test suggested including the direct effect of social influence on avoidance
behavior. This is also an appropriate addition to the model since the control variable (number of
employees in unit) has a direct positive effect in avoidance motivation. The more employees in
the unit the more avoidance behaviors are reported. Thus, we believe social influence, including
influence from employees at the unit, would have a direct impact on avoidance behavior. Finally,
the modification indices test suggested adding the paths of the direct effects of avoidance
motivations’ antecedents (i.e. perceived effectiveness, perceived cost, and perceived threat) on
avoidance behavior. This also fall in line with the originally theory, it only suggests that there
would be a direct effect on avoidance behavior as well as an indirect effect through the mediator
perceived avoidance. The new suggested model results are shown in Figure 4.3 and the new
130
(*p<0.05; **p<0.01; ***p<0.001)
CFI became (0.903) larger than the minimum cutoff (0.9), and SRMR is less than 0.05.
However, although RMSEA become much closer to the recommended range (0.0 – 0.1), it is still
131
a little bit higher than the maximum recommended value of. However, this would be due to a
number of reasons. First, the model has relatively low degrees of freedom, (Kenny et al, 2014)
argue to not even compute the RMSEA for low df models. Second, RMSEA is a parsimony-
adjusted index and its formula includes a built-in correction for model complexity (i.e. It favors
simpler models) (Hooper et al, 2008). Thus, the RMSEA would improve if the model had less
parameters. However, the suggested model is the most parsimonious while still being aligned
with the original theory. Finally, RMSEA is sensitive to sample size, (Hu and Bentler, 1999)
suggested that RMSEA does not perform well and is less preferable when the sample size is less
than 250. In conclusion, the overall model fit indices of the suggested model are considerably
better than the original model and shows a more reasonable fit.
The suggested model resulted in new significant relations between the constructs. First,
perceived threat has a positive direct effect on the avoidance behavior. This implies that the more
healthcare providers perceive the EHR system as a threat to their job the more they would avoid
the system. This is related to our argument on resistance behaviors. Healthcare providers are
willing to avoid the system without them knowing of any other safeguarding measure. This
would result in simply not using the EHR system. Second, perceived effectiveness had a direct
effect on both of avoidance motivation and avoidance behavior. Whereas perceived cost did not.
This confirms our former argument that healthcare providers are willing to use the PBRS if they
perceive it to be effective in eliminating the threats associated with the EHR system regardless of
the costs they may endure in the process. Finally, the suggested model shows that social
influence of using the PBRS had a positive effect on the avoidance behavior. However, there was
no significant direct effect from social influence about Hakeem on the avoidance behavior. This
implies that positive opinions about using the PBRS would influence healthcare providers to
132
avoid the EHR system. While positive opinions about using the EHR system does not influence
their decision of avoiding the EHR system. We can conclude that healthcare providers are
leaning more towards avoiding the EHR system and using the PBRS. Healthcare providers’
avoidance behaviors are not effected by what other people think of the EHR system nor how
much trouble it will cause to use the PBRS. On the other hand, in order increase healthcare
providers’ avoidance behaviors it is sufficient to perceive the EHR system as a threat, to perceive
the PBRS as an effective alternative, or to have positive opinions from others on using the PBRS.
The suggested model explained (81%) of the variance in avoidance behavior, (83%) in
CONTRIBUTIONS
Contributions to Research
Our study makes several contributions to IS research. First, it tested the full conceptual
model of TTAT for the first time. This includes the effects of risk tolerance and social influence.
Our findings explains EHR avoidance phenomenon based on threat perceptions which are
effected by levels of susceptibility, severity, and social influence. Additionally, EHR avoidance
behaviors are determined by avoidance motivations which are effected by the effectiveness of
using the PBRS and levels of threats associated with the EHR system. We also found direct
effects of social influence of using the PBRS and the effectiveness of its use on the EHR
avoidance behavior. These findings contributes to the IS research which still requires
advancements in understanding EHR avoidance behaviors. Second, our model explained 81% of
the variance in avoidance behavior, this is much better than previous attempts (Liang and Xue,
2010; Arachchilage and Love, 2014) which explained at most 21% of the variance in the
outcome variable. Third, our study shows that TTAT could be adapted to explain technology
133
avoidance behaviors in different contexts. The TTAT first empirical test was conducted by
(Liang and Xue, 2010) to explain users’ avoidance of IT security threats such as malware and
cybercrimes, we find that its basic solid foundations are also applicable in the healthcare context
involving threats induced by using EHR systems. Fourth, we confirmed that perceived threat,
and social influence and effectiveness of the safeguarding measure have significant direct effect
on avoidance behaviors. Fifth, our study reveals a counterintuitive finding that positive social
influence about Hakeem has a positive direct effect on perceived threat. This interesting finding
should motivate future work to better investigate different possibilities surrounding this
Contributions to Practice
Our study provides a number of contributions to practice. First, because this study is
conducted in a developing country, which is still at the first stages of digitizing organizations, it
becomes very important to understand what predicts technology avoidance behavior and what do
new users consider as avoidable threats. Answers to these questions would help system
developers and system designers to build improved systems that overcomes sources of threat.
Second, understanding HIT avoidance and especially towards EHRs would help practitioners get
the appropriate support to reduce avoidance motivations. Governments in many countries have
largely invested in HIT systems. Thus, being able to explain causes of HIT avoidance would help
in finding solutions to make these systems more adopted and with higher success rates. Finally,
our work shows that TTAT could be applied successfully in different contexts. Thus, the theory
could be replicated using different technologies across industries. Findings would help
practitioners better understand avoidance in the context of which it was tested in, and that would
serve as the first step towards solving problems associated with avoidance behaviors.
134
LIMITATIONS
This study has its limitations. First, in order to empirically test the model, we selected the
implications on job productivity and performances as the threats and the PBRS as a safeguarding
measure. This does not necessarily mean that threats are only directed on the productivity and
performances of employees, and that the only way to cope with a threat associated with an EHR
system is resorting to the PBRS. Threat could be related to different sources such as the design of
the system and the safeguarding measure could be manifested in different behaviors such as
delaying the use of the system. Research can be conducted with different threat sources and
safeguards to examine whether the findings of this study will change. Second, to fully understand
the positive effect between social influence on the use of the EHR system and perceived threat,
further investigations are required. It is interesting to study how social norms, social factors, and
image effect ones’ perception of threats associated with the EHR system. Third, although
TTAT’s conceptual model suggested the mediation effect of perceived avoidability, there was no
empirical evidence for its mediation and therefore it was dropped from our suggested model.
Future research may be directed towards testing perceived avoidability using antecedents other
than the ones related to the safeguarding measures (i.e. cost and effectiveness). Finally, because
our survey data is cross-sectional and collected from a single source, common method bias may
be a concern (Xue et al, 2014). Future research can apply a longitudinal study or collect data
135
REFERENCES
Ajzen, I., 2006. "Constructing a theory of planned behavior questionnaire." Available at people.
umass. edu/aizen/pdf/tpb. measurement.pdf.
Arachchilage, Gamagedara, N., and Love, S., 2013. "A game design framework for avoiding
phishing attacks." Computers in Human Behavior 29, no. 3, 706-714.
Bhattacherjee, A., Davis C., Hikmet, N., and Kayhan, V., 2008. "User reactions to information
technology: Evidence from the healthcare sector." Association for Information Systems.
Bhattacherjee, A., Davis, C., and Hikmet N., 2013. "Physician Reactions to Healthcare IT: An
Activity-Theoretic Analysis." System Sciences (HICSS), 46th Hawaii International
Conference on. IEEE.
Browne, M., and Cudeck R., 1993. "Alternative ways of assessing model fit." Sage Focus
Editions 154, 136-136.
Burnkrant, Robert E., and Cousineau, A., 1975. "Informational and normative social influence in
buyer behavior." Journal of Consumer research, 206-215.
Callen, J., Paoloni, R., Li, J., Stewart, M., Gibson, K., Georgiou, A., Braithwaite, J. &
Westbrook, J. 2013. Perceptions of the effect of information and communication
technology on the quality of care delivered in emergency departments: A cross-site
qualitative study. Annals of Emergency Medicine, 61, 131-144.
Carver, C. S. 2006. "Approach, Avoidance, and the Self-Regulation of Affect and Action,"
Motivation and Emotion (30), pp. 105-110.
Carver, C. S., and White, T. L. 1994. "Behavioral Inhibition, Behavioral Activation, and
Affective Responses to Impending Reward and Punishment: The BIS/BAS Scales,"
Journal of Personality and Social Psychology (67), pp. 319-333.
Clark, J. K., Wegener, D. T. & Fabrigar, L. R. 2008. Attitudinal ambivalence and message-based
persuasion: motivated processing of pro-attitudinal information and avoidance of
counter-attitudinal information. Personality and Social Psychology Bulletin, 34, 565-577.
Cordell, W. H., Overhage, J. M. & Waeckerle, J. F. 1998. Strategies for improving information
management in emergency medicine to meet clinical, research, and administrative needs.
Annals of Emergency Medicine, 31, 172-178.
136
Cotea, C. (2010). Electronic health record adoption: Perceived barriers and facilitators. Centre
for Military and Veterans' Health: The University of Queensland.
Deutsch, Morton, and Gerard H., 1955. "A study of normative and informational social
influences upon individual judgment." The journal of abnormal and social psychology 51,
no. 3, 629.
Edwards, J. (1992) "A cybernetic theory of stress, coping, and well-being in organizations,"
Academy of Management Review (17) 2, pp. 238-274.
Elliot, A. J., and Covington, M. V. 2001. "Approach and Avoidance Motivation," Educational
Psychology Review (13:2), pp. 73 92.
Gefen, D., D. Straub, and M. Boudreau (2000) "Structural equation modeling and regression:
guidelines for research practice," Communications of the AIS (4) 7.
Hare, K., Whitworth, B., Deek, F., and Norris, T., 2006. “Identifying Barriers to IT Adoption in
Clinical Health Care: A Pilot Study”, Health Care and Informatics Review Online, ISSN
1176-4201, http://hcro.enigma.co.nz/website/index.cfm
Hillestad, R., Bigelow, J., Bower, A., Girosi, F., Meili, R. & Scoville, R. 2005. Can Electronic
Medical Record Systems Transform Health Care? Potential Health Benefits, Savings, and
Costs. Health Affairs, 24, 1103-1117.
Hooper, Daire, Coughlan J., and Mullen M., 2008. "Structural equation modelling: Guidelines
for determining model fit." Articles, 2.
Hu, L.T. and Bentler, P.M. (1999), "Cutoff Criteria for Fit Indexes in Covariance Structure
Analysis: Conventional Criteria versus New Alternatives," Structural Equation Modeling,
6 (1), 1-55.
Hwang, Y., 2005. "Investigating enterprise systems adoption: uncertainty avoidance, intrinsic
motivation, and the technology acceptance model". European Journal of Information
Systems 14, no. 2, 150-161.
137
Kane, G. C. & Labianca, G. 2011. IS Avoidance in Health-Care Groups: A Multilevel
Investigation. Information Systems Research, 22, 504-522.
Kellermann, A. L. & Jones, S. S. 2013. What it will take to achieve the as-yet-unfulfilled
promises of health information technology. Health Affairs, 32, 63-68.
Kenny, David A., Kaniskan B., and McCoach D., 2014. "The performance of RMSEA in models
with small degrees of freedom." Sociological Methods & Research, 0049124114543236.
Kim, H.W., and Kankanhalli, A., 2009. "Investigating user resistance to information systems
implementation: a status quo bias perspective”, MIS quarterly 33, 3.
Lapointe, L., and Rivard, S., 2005. “A Multiple Model of Resistance to Information Technology
Implementation”, MIS Quarterly, 29, 3, pp. 461-491.
Liang, H., and Xue, Y., 2009. "Avoidance of information technology threats: a theoretical
perspective." MIS quarterly, 71-90.
Liang, H., and Xue, Y., 2010. "Understanding security behaviors in personal computer usage: A
threat avoidance perspective." Journal of the Association for Information Systems 11, no.
7, 394-413.
Markus M. L., 1983. “Power, Politics, and MIS Implementation”, Communications of the ACM,
26, 6, pp. 430-444.
Moore, G. C., and Benbasat, I. 1991. “Development of an Instrument to Measure the Perceptions
of Adopting an Information Technology Innovation,” Information Systems Research
(2:3), pp. 192-222.
Muthén, L. K., & Muthén, B. O. (1998-2011). Mplus User's Guide. Sixth Edition. Los Angeles,
CA: Muthén & Muthén.
Poon, E., Blumenthal, D., Jaggi, T., Honour, M., Bates, D.W., and Kaushal, R., 2004.
“Overcoming barriers to adopting and implementing computerized physician order entry
systems in U.S. hospitals”, Health Affairs 23, 4, pp. 184–190.
Pope, C., Halford, S., Turnbull, J., Prichard, J., Calestani, M. & May, C. 2013. “Using computer
decision support systems in NHS emergency and urgent care: ethnographic study using
normalization process theory,” BMC Health Services Research, 13, 1-13.
Rho, Ock, H., and Ryu I., 2011. "The impact of information technology threat on IT
appropriation and IT avoidance,” In Computational Intelligence, Communication
138
Systems and Networks (CICSyN), Third International Conference on, pp. 195-200. IEEE,
2011.
Romanow, D., Cho, S. & Straub, D. 2012. “Riding the wave: Past trends and future directions for
health IT research,” MIS Quarterly, 36, iii-x.
Sassen, E.J. (2009). Love, hate, or indifference: How nurses really feel about the electronic
health record system. Computers, Informatics, Nursing, 27(5), 281–287.
Smith, Tim, Grant, G., and Ramirez, A., 2014. "Investigating the influence of psychological
ownership and resistance on usage intention among physicians”, System Sciences
(HICSS), 47th Hawaii International Conference on. IEEE.
Stream, G.R. (2009). Trends in adoption of electronic health records by family physicians in
Washington State. Informatics in Primary Care, 17(3), 145–152.
Thompson, R. L., Higgins, C. A., and Howell, J. M., 1991. “Personal Computing: Toward a
Conceptual Model of Utilization,” MIS Quarterly (15:1), pp. 124-143.
Venkatesh, V., M.G. Morris, G.B. Davis, and F.D. Davis, 2003. “User Acceptance of
Information Technology: Toward A Unifying View,” MIS Quarterly, 27, 3, pp. 425-478.
Xue, Y., Liang, H., Mbarika, V., Hauser, R., & Schwager, P. (2014). Understanding Healthcare
Professionals' Resistance of Telemedicine: an Empirical Study in Ethiopia. In PACIS (p.
225).
139
APPENDIX
Measurement Items
140
Using Hakeem is dreadful to my job productivity and
PTHR4
performances.
Using Hakeem is a risk to my job productivity and
PTHR5
performances.
The paper based system would be useful for removing threats
PEFF1
associated with Hakeem.
The paper based system would increase my productivity and
PEFF2
performances by removing threats associated with Hakeem.
Perceived The paper based system would enable me to perform my job
Effectiveness PEFF3
faster. Liang and
PEFF4 The paper based system would make it easier to do my job. Xue, 2010
The paper based system would enhance my effectiveness in
PEFF5
removing threats associated with Hakeem.
PCOS1 Going back to the paper based system is difficult.
PCOS2 Going back to the paper based system is costly.
Perceived Going back to the paper based system is time consuming. Liang and
PCOS3
Cost Xue, 2010
I am not going back to the previous paper based system
PCOS4
because it is too much trouble.
People in my hospital who influence my behavior think that I
SOIP1
should use the paper based system.
People in my hospital who are important to me think that I
SOIP2
should use the paper based system.
I use the paper-based system because of the proportion of
SOIP3
coworkers who use the paper based system.
My managers have been helpful in the use of the paper-based
SOIP4
system.
Social
My supervisor is very supportive of the use of the paper-based
Influence SOIP5 Venkatesh
system.
(PBRS) et al, 2003
In general, the hospital has supported the use of the paper-
SOIP6
based system.
People in my hospital who use the paper-based system have
SOIP7
more prestige than those who do not.
People in my hospital who use the paper-based system have a
SOIP8
high profile.
Having the paper-based system is a status symbol in my
SOIP9
hospital.
For me to avoid Hakeem in the future would be:
PAVO1
(Very Difficult, Difficult, Neutral, Easy, Very Easy)
If I wanted to I could avoid Hakeem in the future
PAVO2 (Definitely false, Probably false, Neither True nor False,
Perceived Probably true, Definitely true) Ajzen, 2006
Avoidability How much control do you believe you have over avoiding
Hakeem?
PAVO3
(Far too Little, Too Little, About Right, Too Much, Far too
Much)
PAVO4 It is mostly up to me whether or not I avoid using Hakeem.
If I have access to the paper-based system, I intend to go back
AVMO1 to the paper based system to avoid threats resulting from using
Avoidance Hakeem.
Motivation If I have access to the paper based system, I predict I would go
AVMO2 back to the paper based system to avoid threats resulting from
using Hakeem.
141
If I have access to the paper based system, I plan to go back to Liang and
AVMO3 the paper based system to avoid threats resulting from using Xue, 2010
Hakeem.
I avoid using Hakeem to mitigate my job threats associated
AVBH1
with Hakeem.
Avoidance I prefer to improve the previous paper based system rather
AVBH2
Behavior than switch to Hakeem. Liang and
AVBH3 I avoid using Hakeem in order to do my job well. Xue, 2010
AVBH4 I avoid Hakeem.
142
Instrument Validation
143
CI: 0.000 - 0.174
SRMR: 0.019
SOIP1 0.777 X2: 928.862
SOIP2 0.789 df: 27
SOIP3 0.775 P < 0.001
Social SOIP4 0.857 CFI: 0.985
Influence SOIP5 0.881 2.47 1.00 0.952 RMSEA: 0.390
(PBRS) SOIP6 0.867 CI: 0.368 - 0.411
SOIP7 0.829 SRMR: 0.102
SOIP8 0.846
SOIP9 0.847
PAVO1 0.815 X2: 16.441
PAVO2 0.840 df: 3
PAVO3 0.876 P < 0.001
Perceived
3.06 0.96 0.877 CFI: 0.930
Avoidability
RMSEA: 0.143
PAVO4 0.856 CI: 0.080 - 0.214
SRMR: 0.233
AVMO1 0.994 X2: 9.546
AVMO2 0.973 df: 2
P < 0.05
Avoidance
3.03 1.06 0.977 CFI: 0.966
Motivation
AVMO3 0.974 RMSEA: 0.131
CI: 0.056 - 0.219
SRMR: 0.011
AVBH1 0.934 X2: 7.107
AVBH2 0.929 df: 2
AVBH3 0.977 P < 0.05
Avoidance
2.81 1.29 0.971 CFI: 0.993
Behavior
RMSEA: 0.108
AVBH4 0.945 CI: 0.030 - 0.198
SRMR: 0.006
144
CHAPTER FIVE:
ABSTRACT
We use the Revealed Causal Mapping (RCM) methodology to better understand the phenomenon
of Electronic Health Records (EHR) avoidance. Little research has aimed to explain avoidance
behavior towards technology in general and towards EHR in particular. We use the threat
avoidance feedback loop as a theoretical lens for this study. We revealed the main constructs
and the concepts shaping each of these constructs by analyzing data collected from interviewing
healthcare providers working at a large international hospital. The hospital has completely
implemented the EHR system and the system has been enforced to all healthcare providers. Our
findings show that healthcare providers decide to avoid the EHR system when they consider the
system to be a threat to their job and when there is an available option other than using the
system. Healthcare providers can avoid threats associated with the EHR system by conducting
145
CHAPTER FIVE:
INTRODUCTION
The implementation of Electronic Health Records (EHR) can bring number of benefits to
healthcare providers. It helps them access patients’ medical history which helps them assess and
diagnose patients faster and more accurately (Pope et al, 2013). Additionally, EHR systems has
the potential of reducing medical errors (Hillestad et al, 2005). However, evidence show that
healthcare providers tend to avoid using the EHR systems when possible (Kane and Labianca,
2011). It is important to understand this phenomenon to help healthcare providers take better
advantage of all of the benefits EHR systems bring to the healthcare organization if used
meaningfully. IS research has very limited work in this area and findings are still mixed and
unclear (Kellermann and Jones, 2013). Prior research has mainly focused on EHR adoption (e.g.
Gan and Cao, 2014; Hewitt and McLeod, 2011; Hung, 2013; Jha et al, 2008; and Kemper et al,
2006), and some work focused on resistance (e.g., Markus, 1983; Poon et al, 2004; Lapointe and
Rivard, 2005; Bhattacherjee and Hikmet, 2008). However, limited attention has been given to IS
avoidance and specially EHR avoidance. Healthcare context has unique characteristics which
makes it difficult to generalize findings that explain avoidance towards technologies within
different industries to those technologies used in the healthcare industries. Furthermore, the
complexity of the avoidance behavior and ways of its manifestations does not allow us to simply
explain avoidance by lower levels of adoption. Thus, we argue for the need of a distinctive
theoretical perspective that aims to explain avoidance behaviors towards EHR systems. In this
study we aim to reveal the main concepts shaping the antecedents of the avoidance construct. We
146
base our study on the underpinning concepts of the threat avoidance feedback loop (Liang and
Xue, 2009. We conducted qualitative interviews with healthcare providers from a large public
hospital, located in Jordan, to capture their perceptions of the implemented EHR system. To
analyze their responses, we used Revealed Causal Mapping (RCM), a qualitative methodology
commonly used to identify constructs and linkages revealed from respondents’ statements
(Nelson et al., 2000). RCM has proven to be useful in studying emerging phenomena that require
rich and contextualized understanding (Narayanan and Armstrong, 2005). Given that our goal is
to understand the perceptions of individuals interacting directly with a newly implemented EHR
system, we believe that the RCM method is appropriate for our investigation of these
perceptions. We base our study on the conceptual process of threat avoidance presented by
THEORETICAL BACKGROUND
loop (see Figure 5.1). It posits three main stages. First, the threat appraisal; this is triggered when
users perceive a potential harm that has negative consequences in their environment associated
with the technology they are using. Once the harm is confirmed, users develop a perception of
their current state. They set being harmed as the undesired end state and compare it with their
current state. The assessment of threat depends mainly on how close both states are. The distance
between the current state and the undesired end state is inversely proportional to the strength of
the potential negative consequences (Liang and Xue, 2009). A threat is perceived when the
distance decreases to a certain value. Second, the coping appraisal; only when a threat is
perceived and confirmed users become motivated to actively avoid the threat and start to find
safeguarding measures that would help them avoid the threats associated with the system. The
147
threat must occur before the coping appraisal starts (Lazarus and Folkman, 1984). Users assess
available action options and decide what safeguarding measures to be taken to cope with the
threat. The assessment may be based on a number of factors such as the cost of the safeguarding
measure, the perceived effectiveness of the safeguarding measure, and the levels of self-efficacy
users believe they have in terms of adapting the safeguarding measure (Liang and Xue, 2009).
Third, the coping behavior; individuals can manifest two types of coping: problem-focused
coping and emotion-focused coping (Lazarus and Folkman 1984). Problem-focused coping is
performed by handling the source of the threat directly. It can be achieved by taking
safeguarding measures such as using the Paper Based Recording System (PBRS) instead of the
EHR system to avoid the threats associated with the EHR system. After taking the safeguarding
measure users perceive the distance between the current state and the undesired end state to be
larger, thus they perceive reduced threat. On the other hand, emotion-focused coping is related to
the psychological mindsets that users create to handle threats. It could be achieved by passively
creating false perceptions of the environment but without trying to actively change it. For
example, a common emotion-focused coping mechanism is to not think about the threat to
escape from the situation. Although users know that the threat is present, they would block it
from their rational thinking (Liang and Xue, 2009). If users fail to find a safeguarding measure
that can help them avoid the threat, they need to practice emotion-focused coping so that their
psychological well-being is maintained (Beaudry and Pinsonneault, 2005). When both appraisals
coping) to be performed to avoid threats associated with the technology. These behaviors will
continue until the difference between the current state and the undesired end state are sufficiently
148
Figure 5.1: Threat Avoidance Feedback Loop
RESEARCH METHODOLOGY
We used RCM for this study, the same methodology as the one we used for the study in
chapter three. We also collected data from the same healthcare providers at the same hospital.
For information about the RCM methodology, the hospital and the participants please refer to the
Data Analysis
respondent’s explicit statements. We used a four-step process: (1) data elicitation, (2)
149
construction of revealed causal maps, (3) validation of the maps, and (4) interpretation of the
maps.
Step 1: Data Elicitation; we conducted the open-ended questions using Qualtrics. Table
3.2 in the Appendix of chapter three includes the questions used for this study Fifty-nine
respondents typed their responses online to the open-ended questions. In applying the RCM
method, researchers often rely on two major types of data: text-based data and interview data
(Narayanan and Armstrong, 2005). As our research objective was to reveal the cognitive
structure pertaining to the resistance behavior domain, it was appropriate to use open-ended
questions, similar to an interview approach (Nelson et al., 2000), and to analyze the narratives of
individual participants for their experiences, perceptions, and beliefs with regard to the new EHR
system.
derived revealed causal maps of perceptions step by step. First, we identified causal statements
from an informant’s narrative by looking for key words such as “because,” “therefore,” “so,”
“while,” and “if–then.” Then, we coded each causal statement into cause and effect and replaced
the “cause” and “effect” in a statement with the concepts (the key words or phrases) and linked
“cause” and “effect” with an arrowed line to obtain a causal map at the concept level. We
highlighted repeated words and phrases and grouped them into constructs. Finally, we
aggregated all concept-level maps across all informants to obtain a causal map at the construct
level, Table 5.1 shows an example of the described process. In total, 11 concepts and 227
Table 5.1: An Illustration of the Four–Step Procedure for Constructing Causal Maps
Step 1: Identifying causal statement Example of causal statement:
“I think that data manipulation by hackers, it is serious
problem that I can think of it as a threat”
150
Step 2: Constructing raw causal maps Cause:
1. Data manipulation by hackers.
Effect:
I can think of it as a threat
Step 3: Coding Raw Phase (Coded Concept)
1. Cyberattacks
2. Considered a threat
Step 4: Recasting raw causal maps into
concept-level revealed causal maps
Step 3: Validation; we reviewed the avoidance literature to validate the concepts and
constructs identified in the RCMs. Similar to (Nelson et al., 2000), we used the point of
redundancy to evaluate the convergence of concepts elicited from the responses (Axelrod, 1976).
In our analysis, the 17 concepts converged at the 18th respondent--meaning that no new concept
was revealed in the map of the 19th respondent or beyond-- indicating the sufficiency of our
151
sample size. The additional relations yielded by adding the revealed causal map of each
Step 4: Interpretation of the Maps; we examined the key constructs and linkages
uncovered in the map (see Figure 5.3). The constructs and their associations demonstrate what
RESULTS
Our data analysis revealed six constructs that are essential to understanding avoidance
towards the EHR System (EHRS). Five are initially posited in the avoidance feedback loop
process: Perceived Harm, Threat Appraisal, Coping Appraisal, Problem-Focused Coping, and
Emotion-Focused Coping; Our analysis also revealed the underlying concepts associated with
each revealed construct. The list of constructs and concepts identified in the study with the
percentage of each of these concepts occurrence in relation to all of the other concepts within the
152
Table 5.2: Revealed Constructs of EHR Avoidance and Its Concepts
Construct Definition Concepts/Dimensions Definition %
Perceived The awareness of the Technical Problems Any technical 0.14
Harm emergence of a problem healthcare
harmful event in the providers encounter
environment that when using the
results in becoming EHR.
closer to the undesired Potential Loss of Data The probability of 0.24
end state (anti-goal) losing medical data
(Liang and Xue, because of technical
2009) errors or any other
difficulty when
interacting with the
EHR.
Unnecessary Efforts Perceptions that 0.33
efforts required to
complete tasks using
the HER is not
needed and that the
task may be
completed without
these extra efforts.
Threat The sense of threat Lower Productivity Low effectiveness 0.67
Appraisal developed from of productive effort.
finding the current Evaluated in terms
state too close to the of the rate of
undesired end state accomplishing tasks
(anti-goal) per unit of input (i.e.
(Liang and Xue, patients).
2009) Cyber Attacks Attempts by hackers 0.33
to damage or
destroy the HER
system, the
computer it runs on,
the hospital’s
network, or the data
stored on the
system.
Coping The level of Ability to take appropriate The levels of
Appraisal motivation to engage safeguarding measures confidence in the 1.00
in a coping behavior to ability of using
avoid the threat by some other option to
increasing the accomplish the task
discrepancy between without the need to
the current state and use the EHR system.
the undesired end state
(anti-goal)
(Liang and Xue,
2009)
Emotion Cognitively regulating Accepting that the
Focused negative emotions Acceptance problem had 0.73
Coping arising from threat occurred but that
(Lazarus and nothing could be
Folkman1984). done about it (Tsai
et al, 2007).
153
Distraction Diverting attention 0.27
away from the
problem by thinking
about other things or
engaging in some
activity (Tsai et al,
2007).
Problem- Behaviors that directly Indirect Usage Using the EHR via 0.16
Focused address and/or change an alternative
Coping the problem causing system or person.
threat (Lazarus and Delayed Usage Using the EHR after 0.37
Folkman 1984). a period of time by
which tasks become
late or postponed.
Partial Usage Using part of the 00.47
EHRS functionality
and resisting other
parts. For example:
using the EHRS to
retrieve data and
resist to enter the
data in the EHRS
forms.
Perceived Harm
When healthcare providers realize any harm associated with using Hakeem, they develop
a perception of their current state. They then set being harmed by using Hakeem as the anti-goal
(undesired end state) and compare it with their current state (Liang and Xue, 2009). Our study
shows that perceived harm influences healthcare providers’ perceptions of the threat associated
with Hakeem. The more harm was realized from using Hakeem the higher perceptions of threat
associated with Hakeem will be realized. We found three main dimensions of what healthcare
providers perceive as harm: Technical Problems, Potential Loss of Data, and Unnecessary
Efforts.
Technical Problems: responses revealed that one of the perceived harm definitions is
having to deal with difficulties related to technical issues such as frequent power outages. The
hospital has backup power generators that automatically regenerates power to medical equipment
and patients’ rooms in a case of a power outage. However, Hakeem is not connected to these
154
generators. Thus, the system will be ideal and healthcare providers will not be able to interact
“Only when I face technical difficulties in the case I mentioned [power outage] it could take some time to
complete the forms after that” (Male, 33, Nurse, Dermatology).
Potential Loss of Data: the analysis revealed that healthcare providers consider the
potential of having the digital data lost is detrimental to their jobs. Although this applies to the
paper files as well, but none of the respondents revealed their concerns of having a paper file
being lost.
“…our increased dependency on software material…which causes the risk of information being lost…”
(Male, 24, Pre-Med, General Surgery).
Unnecessary Efforts: most of the responses identified the process of redundantly entering
patients’ data, and being asked unnecessary questions as impediments to their job. They
considered doing the same task repeatedly and being forced to answer questions that are
“I wish you can be able to skip through unnecessary data… It keeps asking us the same questions every
time the patients is following up, sometimes things do not change and I wish we can skip through that, also
in some cases it asks us unnecessary questions that are not relevant to the case of the patient… we are
doing more unnecessary work now” (Female, 42, Nurse, Dermatology).
Threat Appraisal
Healthcare providers develop a sense of threat when they find that their current state is
too close to the anti-goal (i.e. harms associated with Hakeem) (Liang and Xue, 2009). The
evaluation of the threat based on the discrepancy between the safe current-state and the
dangerous end-state takes place within the threat appraisal stage. If a threat is confirmed,
healthcare providers will begin evaluating safeguarding measures within the coping appraisal.
We found 2 main threats identified by respondents: Lower Productivity, and Cyber Attacks.
155
Lower Productivity: Healthcare providers’ responses revealed that the drop in their
explained productivity in terms of the number of files being completed per shift and the number
‘’I find Hakeem a threat to my job productivity…I do not accomplish as many reports as I did using the
paper system’’ (Female, 29, Lab Specialist, Medical Lab).
“…It takes so long to fill in reports which is valuable time I could spend with other patients… I usually end
up seeing less patients” (Male, 38, Doctor, ER).
Cyber Attacks: healthcare providers revealed that a major threat of having patients’
records online via Hakeem is their fear of potential cyber-attacks. They explained how patient
data should be kept private and no unauthorized access should exist, and by having the records
“What I consider a threat is related to having our data online which makes it easy for attackers [Cyber
Attacks/Hackers] to view and delete or do anything with the data” (Male, 24, Pre-Med, General Surgery).
Coping Appraisal
When healthcare providers confirm threats associated with Hakeem they begin to
evaluate appropriate safeguarding measures that would help in increasing the discrepancy
between their pre-identified safe status and the threat occurring from the use of Hakeem.
Healthcare providers would be motivated to engage in a coping behavior if the options they
considered are suitable, effective, feasible, and they are capable of performing related activities
“I want a data entry specialist to interact with the system, doctors and nurses would only [implying should
be only required to] submit verbal or written requests…” (Male, 55, Doctor, Dermatology).
If healthcare providers believe that the threats associated with Hakeem could be avoided
be taking a safeguarding measure that they are capable of mastering, is effective, and is feasible,
156
then they would take the problem-solving approach to change their reality (Liang and Xue,
2009). Healthcare providers would deal directly with the source of the threat to eliminate it. We
threats associated with Hakeem: Indirect Usage, Delayed Usage, and Partial Usage.
Indirect Usage: Our analysis shows that healthcare providers may avoid Hakeem by
having someone else interact with the system on their behalf. For example, a doctor may ask his
appointed nurse to enter the data of the patient being checked by orally dictating the required
data to be entered.
“I usually ask our nurse to enter data into the system while I am checking on the patient… by doing this I
have my main focus on the patient not the computer.” (Male, 43, Doctor, ENT).
Delayed Usage: Our analysis revealed that some healthcare providers may aim to avoid
threats associated with Hakeem by procrastinating the usage of Hakeem. By doing so, they are
“…I have Hakeem as a second choice with the old system [PBRS]. Today we cannot finish the patient visit
without finishing all forms and reports, this takes time and most of this information at my clinic is not
applicable because I care for kids. So I try to write all inputs on the patients’ folder and by the end of the
day we enter the data, which will make the flow of patients faster.” (Female, 46, Doctor, Pediatric Unit).
“I usually complete all the forms at the end of my shift. In this way I guarantee to have my time focused on
the patient during each visit…sometimes this means to have leftover work after the end of my shift” (Male,
Nurse, 31, ENT).
Partial Usage: we found healthcare providers who reported that they are using the system
partially by performing tasks they like and avoiding the ones they dislike. For example, some
healthcare providers revealed that they enjoyed using Hakeem when retrieving patients’ medical
history, whereas they considered entering the data into the system a hassle that needs to be
avoided.
“I like Hakeem because it helps in viewing all of the history of the patient quickly, I do not need to go and
search for files in a cabinet any more…entering data into the system takes so much time…I would like to
use the paper system when I need it…” (Male, 56, Doctor, ER).
157
If healthcare providers believe that the threat associated with Hakeem cannot be entirely
avoided by taking any safeguarding measure available to them, they would passively avoid the
threat by performing emotion-focused coping (Liang and Xue, 2009). Emotion-focused coping is
achieved by creating false reality of the environment without actually changing it. Our analysis
revealed two main dimensions of emotional focused coping performed by healthcare providers at
Acceptance: this is identified when healthcare providers are aware of the threats but they
“… I only fill in the forms as expected…we are following the hospitals policy and only using Hakeem”
(Female, 44, Nurse, Pediatric Unit).
Distraction: this is identified when healthcare providers divert their attention from the
“I do not do anything different. I use the system without thinking of threats” (Male, 39, Pharmacist,
Outpatient Pharmacy).
158
Figure 5.3: Aggregated Revealed Causal Map of EHR Threat Avoidance
Our findings revealed that the more healthcare providers perceive harms associated with
Hakeem, the more they would consider Hakeem as being a threat to their job.
“When we have technical problems with the system…this means my work will be delayed and I will be less
productive.” (Male, 40, Doctor, Urology).
The more Hakeem is considered as a threat the more healthcare providers will start to
evaluate safeguarding measures to seek elimination of the threats associated with Hakeem.
“…may be using a paper based system, then entering the data into Hakeem later… or may be make some of
the fields optional so we do not need to completely fill in the forms before submitting them…” (Female, 27,
Nurse, Dermatology).
“I would like to be able to send hard copies of the imaging when the system is down” (Male, 30,
Radiology, Imaging).
which type of coping to engage in. If an appropriate safeguarding measure was selected then they
159
are more likely to engage in a problem-focused coping mechanism by objectively applying an
“I use both electronic and paper systems. If I have no time to fill in the data in the system or the patient
needs immediate care I only use the paper system” (Male, 33, Doctor, ER).
that threats will not be eliminated completely by any of the safeguarding measures, they would
“To cope with the threats…I follow instructions and complete the task as required” (Female, 24, Clinical
Assistant, Dermatology)
DISCUSSION
Our findings revealed a number of interesting insights about EHRS avoidance. First, we
found that healthcare providers aim to avoid the threat associated with the EHRS rather than
avoiding the EHRS itself. The emotion-focused coping mechanism indicates that healthcare
providers adapt to threats without actually avoiding the EHRS. Further, the avoidance behavior
is only triggered when a threat is realized. If no threat was confirmed then the likelihood of
avoiding the EHRS becomes too low. Second, we found that Hakeem’s design related issues is a
major contributor to healthcare providers perceptions of the EHRS. EHRS design was not
considered as a harm neither a threat, but rather as a source of negative feelings which healthcare
providers shape about using the EHRS. It is reported that Hakeem has a poor interface that is not
user-friendly. Also healthcare providers complained about how each form must be fully
completed before they can move to the next step, and in some cases the fields on the form are
perceived as irrelevant to the situation which causes irritation to the person required to fill in all
of the data into the forms. Another common issue reported was the access limitation to the
EHRS. The way Hakeem was designed made it only accessible from pre-identified computers
160
within the hospital. Healthcare providers are not able to access the EHRS remotely from outside
of the hospital or by using mobile devices even within the hospital. We believe that it is essential
to plan to eliminate these revealed design problems in order to mitigate avoidance behaviors.
Third, our analysis highlighted an interesting phenomenon related to avoidance; that is the
concept of “work arounds”. Prior work in healthcare defined work-arounds as "clever methods
for getting done what the system does not let you do easily" (Ash et al, 2003; Ash et al, 2004).
Similarly, (Kobayashi et al, 2005) defined work-arounds as "informal temporary practices for
handling exceptions to normal workflow". Additionally, (Morath and Turnbull, 2005) define
crucial work goal within a system of dysfunctional work processes that prohibits the
technology and regulatory demands with the need to provide adequate care to patients, through
this process healthcare provider may see a greater need to improvise or work around intended
work practices (Halbesleben et al, 2008). Our study revealed that healthcare providers have in
common that they desire completing their tasks in the best possible way. They thrive to perform
their job using best practices. However, when using the EHRS becomes a barrier to them doing
their job as they believe it must be done, then EHRS would be avoided (work-around it). Our
study revealed that healthcare providers in the ER unit are the most to avoid the EHRS. Mainly
because the healthcare providers in the ER deal with very time-critical medical emergencies
more than all other providers at different units of the hospital. An example was reported by an
ER doctor who explained why he considers using PBRS when dealing with critical cases:
“…If we admit a person with a gun-shot wound or a stab wound which requires immediate care, and I sit
down and start typing information into Hakeem, people who are accompanying the patient will perceive
that as me being careless and only wasting time “playing” with a device rather than immediately devote
my attention to the patient” (Male, 38, Doctor, ER).
161
In cases similar to the one reported here, it seems that healthcare providers agree that the
best practice would be to avoid the EHRS. Many times, the solutions devised to address a block
involve bypassing safety features in medical technology (Agency for Healthcare Research and
Quality (AHRQ, 2006). Finally, our analysis of responses did not reveal the linkage between
coping behaviors and threat appraisal. When healthcare providers engage in a coping behavior
they implicitly believe that the behavior is sufficient to reduce threats associated with Hakeem.
We also found that healthcare providers tend to engage in the same coping behavior for longer
periods of time after the threat has been realized. This allows them to worry less about the status
of the threat as they continuously engage in the same coping behavior which gives them more
comfort doing the task without the need to reevaluate the threat one more time. Thus, the linkage
CONTRIBUTIONS
For Research
Our study has a number of implication to the IS research. First, the study provides
qualitative insights on the EHRS avoidance phenomenon which is not well understood in IS.
Second, our study revealed that when healthcare providers perceive Hakeem as a threat to their
job they begin to aim to avoid the threats rather than avoiding the EHRS itself. In the majority of
the cases this means to actually avoid using the EHRS itself by performing one of the three
revealed dimensions of problem focused coping behaviors. However, other cases shows that the
avoidance was achieved by performing emotional focused coping which implies the usage of the
EHRS rather than avoiding it. This gives a broader and more enriched understanding of how the
EHRS is being avoided and what is actually being avoided. Third, the constructs of the RCM has
some consistency with the avoidance feedback loop, however the interactions of these constructs
162
were not previously tested or confirmed. Our finding is the first study to explore the linkages
between constructs of the avoidance feedback loop. We did not find any evidence from the
responses that confirms the loopback from the coping stage to the threat appraisal stage. We
rather revealed that healthcare providers engage in a coping behavior and believe it is sufficient
to eliminate threats associated with Hakeem. Fifth, we provided the dimensions within all
construct influencing the EHRS threat avoidance behavior. This gives us a richer understanding
of each construct. Prior theorizations use these constructs abstractly with very limited insights
on the dimensions shaping each construct. Seventh, our findings is the first in IS to account for
all possible shapes of avoidance. Prior theories relied merely on the conceptualization of
avoidance as “not using”. Finally, our findings provided a model that can be the basis for
For Practice
Our study provides a number of implications to practice. First, we provide one of the few
qualitatively detailed explanations to the EHRS avoidance phenomenon. Our rich findings would
help system developers and system designers understand what users actually perceive when
interacting with the EHRS. Second, we revealed that one of the main contributors to avoidance
are the perceived threats associated with the EHRS. We also identified these threats. Therefore,
healthcare providers would be able to take advantage of the EHRS benefits by eliminating these
threats. Third, revealing main dimensions of harm and threats would help healthcare managers
identify these problems and eliminate them at early stages. Finally, revealing the main problem-
focused coping mechanisms would help healthcare managers better understand what healthcare
163
Our study has a number of limitations. First, the sample size used for this study may be a
concern. We recommend that empirical studies on a larger scale be conducted to validate and
modify the RCM developed in this study. Second, because our data was collected from a single
source, validation concerns may arise. Future research can apply a longitudinal study or collect
data from multiple sources to validate the findings. Third, the dimensions of the constructs from
our study are limited to our sample. Further research is encouraged to explore other possible
diminutions that may emerge from different samples. Finally, responses analyzed in this study
did not show evidence of the linkage between the coping behaviors and the threat appraisal. We
encourage future research to attempt to capture this either by conducting a more extensive
qualitative interview or by developing appropriate survey items that can be used in an empirical
164
REFERENCES
Beaudry, A., & Pinsonneault, A. (2005). Understanding user responses to information
technology: A coping model of user adaptation. MIS quarterly, 493-524.
Bhattacherjee, A., and Hikmet, N. 2007. "Physicians' resistance toward healthcare information
technologies: a dual-factor model." System Sciences, 2007. HICSS 2007. 40th Annual
Hawaii International Conference on, IEEE, 2007.
Bhattacherjee, Anol, Christopher J. Davis, Neset Hikmet, and Varol Kayhan, 2008. "User
reactions to information technology: Evidence from the healthcare sector." Association
for Information Systems.
Deng, X., and Chi, L. (2012). Understanding Post-adoptive Behaviors in IS Use: A Longitudinal
Analysis of System Use Problems in the Business Intelligence Context. Journal of
Management Information System, 29(3), 305-340.
Gan, Q., & Cao, Q., 2014. “Adoption of Electronic Health Record System: Multiple Theoretical
Perspectives,” System Sciences (HICSS), 47th Hawaii International Conference on
System Science, (pp. 2716-2724), IEEE.
Hewitt, B., & McLeod, A., 2011. “Modeling Security in Acceptance of Electronic Health Record
Systems,” Journal of Information Privacy and Security, 7(3), 23-45.
Hillestad, R., Bigelow, J., Bower, A., Girosi, F., Meili, R., Scoville, R., & Taylor, R. 2005. “Can
electronic medical record systems transform health care? Potential health benefits,
savings, and costs,” Health Affairs, 24(5), 1103-1117.
Hung, S. Y., Yu, W. J., Tsai, M. F., & Yen, D. C., 2013. “What Drives Physicians' Intention to
Use Electronic Medical Record System: The Roles of Perceived Service Level, Computer
Self-Efficacy, and Perceived Risk,” PACIS, (p. 194).
Jha, A. K., Doolan, D., Grandt, D., Scott, T., and Bates, D. W., 2008. “The use of health
information technology in seven nations,” International journal of medical informatics,
77, 12, pp. 848-854.
Kane, G. C. & Labianca, G. 2011. IS Avoidance in Health-Care Groups: A Multilevel
Investigation. Information Systems Research, 22, 504-522.
Kellermann, A. L. & Jones, S. S. 2013. What it will take to achieve the as-yet-unfulfilled
promises of health information technology. Health Affairs, 32, 63-68.
Kemper A.R., Uren R.L., Clark S.J., 2006. “Adoption of Electronic Health Records in Primary
Care Pediatric Practices,” Pediatrics, 118(1):20-24.
Lapointe, L., and Rivard, S., 2005. “A Multiple Model of Resistance to Information Technology
Implementation”, MIS Quarterly, 29, 3, pp. 461-491.
165
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. Springer publishing
company.
Liang, H., and Xue, Y., 2009. "Avoidance of information technology threats: a theoretical
perspective." MIS quarterly, 71-90.
Liang, H., and Xue, Y., 2010. "Understanding security behaviors in personal computer usage: A
threat avoidance perspective." Journal of the Association for Information Systems, 11, no.
7, 394-413.
Markus M. L., 1983. “Power, Politics, and MIS Implementation”, Communications of the ACM,
26, 6, pp. 430-444.
Muthén, L. K., & Muthén, B. O. (1998-2011). Mplus User's Guide. Sixth Edition. Los Angeles,
CA: Muthén & Muthén
Narayanan, V.K.., and Armstrong, D. J. (2005). Causal Mapping for Research in Information
Technology. Hershey, PA: Idea Group Publishing.
Nelson, K.M., Nadkarni, S., Narayanan, V.K., and Ghods, M. (2000). Understanding Software
Operations Support Expertise: A Revealed Causal Mapping Approach. MIS Quarterly,
24(3), 475-507.
Poon, E., Blumenthal, D., Jaggi, T., Honour, M., Bates, D.W., and Kaushal, R., 2004.
“Overcoming barriers to adopting and implementing computerized physician order entry
systems in U.S. hospitals”, Health Affairs 23, 4, pp. 184–190.
Pope, C., Halford, S., Turnbull, J., Prichard, J., Calestani, M. & May, C. 2013. “Using computer
decision support systems in NHS emergency and urgent care: ethnographic study using
normalization process theory,” BMC Health Services Research, 13, 1-13.
Tsai, H. Y., Compeau, D., & Haggerty, N. (2007). Of races to run and battles to be won:
technical skill updating, stress, and coping of IT professionals. Human Resource
Management, 46(3), 395-409.
166
APPENDIX A:
SURVEY FORMS
Cover Page
Welcome to the survey on attitudes towards the Electronic Health Record system
“Hakeem”. We thank you for participating and hope you will find this survey interesting and
useful. In this survey you will be asked questions about your beliefs and attitudes about Hakeem.
The purpose of this research is to help us understand attitudes and behaviors towards Health
Information Technology (HIT) in general and toward Electronic Health Records (EHR) in
particular.
HAKEEM.
On this survey, Hakeem refers to the Electronic Health Record system implemented
recently at the hospital. The survey takes approximately 30 minutes to complete. The findings
from this research will help us to develop more effective understanding about attitudes and
Please understand that your answers are completely PRIVATE and CONFIDENTIAL.
No one other than the researchers will ever see your individual responses. This project was
reviewed and approved by Washington State University’s Institutional Review Boards. The
information in this consent form is provided so that you can decide whether or not you wish to
participate in this study. Your participation is considered voluntary. Even if you agree to
participate, you are free to withdraw from the study at any time. To do so, simply exit from the
Web Browser.
167
In the event of any publication or presentation resulting from the research, no personally
identifiable information will be shared because your name is in no way linked to your responses.
Your confidentiality will be maintained to the degree permitted by the technology used.
However, no guarantees can be made regarding the interception of data sent via the Internet by
If you have any questions, complaints or concerns about this research, please contact
can stop at any time. You do not have to answer any questions you do not want to answer.
You will be granted a certificate of completion for participating in taking this survey. At
the end of the survey you will be asked to enter your name as you want it to appear on your
certificate.
168
Informed Consent
"I have read the preliminary description of this study and agree to participate. I
understand that there are no anticipated risks, and I am free to discontinue my participation at
IMPORTANT: By clicking on the NEXT button below you are giving your consent to
participate in this study. If you do not give your consent, simply exit from the browser.
If you have already completed this survey in the past, then please STOP and do not complete this
one.
169