International Journal of Research in Medical Sciences

Owamagbe E M. et al. Int J Res Med Sci. 2022 Mar;7(10):3125-3137

pISSN 2320-6071 | eISSN 2320-6012

DOI: http://dx.doi.org/.ijrms20224320
Original Research Article
Diabetes Treatment Adherence and Associated factors in Sub-Saharan
Africa: A Systematic Review and Meta-Analysis
1*Owamagbe, E. M., 1Wala, K. T. 2Zike K. E.,
1Department of Chemical Pathology, College of Medicine, Rivers State University, Port Harcourt, Nigeria.
2School of Public Health, University of Port Harcourt, Rivers state, Nigeria

Received: 01 December 2021

Revised: 08 February 2022
Accepted: 12 February 2022

*Correspondence:
Dr. Owamagbe E. M.,
E-mail: dr.owamagbe@yahoo.com

Copyright: © the author(s), publisher and licensee Medip Academy. This is an open-access article distributed
under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-
commercial use, distribution, and reproduction in any medium, provided the original work is properly cited

ABSTRACT
Diabetes mellitus (DM) is one of the four growing pandemic non-communicable diseases (NCD) ravaging
the modern world according to the WHO global report on Diabetes.1 The prevalence of the disease has
been rising globally especially in low- and middle-income countries (LMIC) especially in Sub-Saharan
Africa. The estimated prevalence of diabetes in the World Health Organization (WHO) was African
Region ranges from 1.1% - 15% and is rising steadily. Also, less than 10% of patients with diabetes in
Africa are covered by any kind of medical insurance, as a result of the high cost of treatment, ethnomedical
and alternative healing systems constitute primary and complementary health care for most Nigerians as
in other African populations. The study performed a systematic review and meta-analysis of the adherence
to DM treatment with associated factors as reported across Sub-Saharan Africa. In this meta-analysis,
educational status, duration, marital status and complexity of the drugs were not significantly associated,
whereas patients age and residency were significantly associated with adherence of anti-diabetic
medications at p < 0.05.
Keywords: Diabetes, Treatment, Adherence, Africa.

1. INTRODUCTION

Diabetes is a chronic, metabolic disease characterized by elevated levels of blood glucose (or
blood sugar), which leads over time to serious damage to the heart, blood vessels, eyes, kidneys
and nerves. The most common is type 2 diabetes, usually in adults, which occurs when the
body becomes resistant to insulin or doesn't make enough insulin. The World Health
Organization estimates that about 422 million people worldwide have diabetes, the majority

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3125
living in low-and middle-income countries, and 1.6 million deaths are directly attributed to
diabetes each year[1]. Diabetes mellitus (DM) can potentially lead to multiple difficulties in
various areas of life, ranging from quite well-known physical complications to psychological
symptoms, which include disturbances in sexual function. However, this impact on sexual
function has long been investigated mainly in men, and although there are some relevant papers
in women dating back to the 1970s and 1980s, this issue has not been specifically addressed
until more recently[2–5]. Anti-diabetic drug therapy plays a pivotal role in the glycemic control
of patients with diabetes, which depends on patient’s adherence to anti-diabetics that realized
by controlling the raised glycemic and early inhibiting its consequences [6].
Adherence to ordered anti-diabetic is one of the main dimensions of health service excellence,
which is described as the proportion of the ordered doses of the drug actually taken by a patient
over a specified period of occasions or the extent to which an individual is taking their
medication as instructed by a health care professional. The World Health Organization
emphasized that “rising the effectiveness of anti-diabetic adherence intervention may have a
huge impact on the health of the population than any improvement with definite medical
treatment”[7]. People with diabetes have a high risk of developing a lot of serious health
problems; in fact of that poor glycemic control level can affect multiple organs. Uncontrolled
blood glucose is the leading causes of cardiovascular diseases, blindness, kidney failure, and
lower limb amputation [8]. Poor glycemic control is due to lack of health insurance, using two
or more hypoglycemic agents, and having normal body mass index[9]. Intensive anti-diabetic
therapy, early screening, and diagnosis, encouraging a healthy diet, performing a regular
checkup and diabetic health education are vital in reducing diabetic complications[10].
Adherence to anti-diabetic medication is a major challenge in diabetic treatment. It is affected
by the nature of the treatment or the complexity of the treatment, out of pocket costs, perceived
medication side effect, and hypoglycemia. Anti-diabetic medications are integral for glycemic
control in diabetes. Non-adherence to the drug can alter blood glucose levels, resulting in the
short term and long term complications[11]. Potential barriers for anti-diabetic medication
adherence are diseases related knowledge, health literacy, patient-provider relationship and
drug-related factors such as a drug side effect, polypharmacy and various logistical barriers to
obtain medications[9]. Good diabetic medication adherence improved glycemic control and
leads to less hospitalization. Adherence to antidiabetic medication is the challenge in diabetes
treatment. However, from Sub-Saharan Africa, there are some studies with variability and

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3126
inconsistency findings. Therefore, the aim of this study was to estimate the pooled prevalence
of anti-diabetic medication adherence among diabetic patients in Sub-Saharan Africa.
2. METHODS

The development and reporting of this protocol was in accordance with the Preferred Reporting
Items for Systematic Review and Meta-Analyses Protocols (PRISMA-P) 2015 statement[12].
This protocol was registered with the International Prospective Register for Systematic
Reviews (PROSPERO).

2.1 Study selection criteria

A summary of the participants, interventions, comparators and outcomes considered, as well
as the type of studies included according to PICOS strategy, is provided in Table 1.

Table 1: Criteria for study selection

Population Persons Living with Diabetes
Intervention Insulin, Medication
Comparison Non-Adherence
Outcome Glycaemic Control

2.1.1 Inclusion criteria

Those studies that had ethical approval and reported anti-diabetic medication adherence or
compliance among diabetic patients in the Sub-Saharan African Region (SSA) were included
in the study. Peer-reviewed studies with cross-sectional study design and the primary outcome
of interest was included in the study.

2.1.2 Exclusion criteria

Those studies which showed an unclear prevalence of anti-diabetic medication adherence

among diabetic patients were excluded from this study.

2.2 Search Strategy

To identify appropriate and relevant studies for the review, the following databases were
searched; MEDLINE, Embase, Cumulative Index of Nursing and Allied Health Literature
(CINAHL), Global Health, PsycINFO, African Journals Online and African Index Medicus.
We will also hand search reference lists of relevant studies to identify further literature of

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3127
interest. Additionally, we will use ProQuest Dissertation & Theses, Web of Science, Google
Scholar and Google search engine to identify grey literature such as government and
institutional reports, theses and dissertations, as well as to track citations. Key authors for
studies that meet eligibility criteria will be contacted if there is some missing information or
full texts of their studies cannot be accessed. A comprehensive search strategy and terms was
developed by the team in collaboration with a medical librarian. We conducted a broad search,
adopting some search terms from previous systematic reviews on the topic[13–15]. Search
terms will include “adherence”, “diabetes”, “prevalence”, and their synonyms. Medical subject
headings (MeSH) and key texts words was developed and combined with Boolean operators
“AND” and/ “OR” across and within categories. A full search strategy for MEDLINE database
is provided in Table 2, and was tested and adapted to other databases.

Table 2: Full search strategy for MEDLINE

1 adherence*mp
2 Diabetes treatment*mp
3 prevalence$diabetes$treatment adherence*mp
4 1 OR 2 AND 3
5 type ii diabetes$treatment*mp]
6 risk$diabetes$adherence*mp
7 risk$adherence*mp
8 5 AND 6 AND 7
9 4 AND 8
10 4 OR 8
11 4 AND 8 AND 9 AND 10

2.3 Data Collection and Analysis

2.3.1 Selection Process

For the first stage of the review, titles and abstracts of potentially eligible studies identified
through electronic database searches, was extracted to Endnote Library (EndnoteX9).
Duplicates was removed (both via Endnote function, and manual checking). Two reviewers
then independently reviewed the titles and abstracts of the remaining studies to identify studies

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3128
that potentially meet the inclusion criteria. In the event of discrepancies, the third reviewer
provided input. Selection of full text studies against the eligibility criteria was carried out
independently by the two reviewers. The third reviewer also randomly checked 10% of selected
studies to check for consistency. Reasons for excluding studies was documented throughout
the process. A PRISMA flow diagram was used to outline the literature search and selection
process as shown in Figure 1.

1167 records identified through 58 additional records identified

database searching through other sources

258 of records after duplicates removed

967 records screened 650 records excluded

317 of full-text articles 287 of full-text articles

excluded, for poor quality and
assessed for eligibility no pooled prevalence of
adherence

30 of studies included in
qualitative synthesis

22 of studies included in
quantitative synthesis
(meta-analysis)

Figure 1: PRISMA-P Flowchart[16]

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3129
2.4 Data Extraction
Data extraction from selected studies was carried out independently by two reviewers. In the
event that there are disagreements, there was discussions between the two reviewers and if a
consensus is not reached the third reviewer adjudicated.

2.5 Quality Assessment

We evaluated the overall quality of individual studies using the Joanna Briggs Institute
Prevalence Critical Appraisal Tool[17]. This tool appraises external and internal validity of
each individual study by addressing issues of representativeness of sample, recruitment of
participants, identification of the condition, its measurement and statistical analysis among
others. Two reviewers independently assessed the quality of included studies and
disagreements was resolved through discussions with a third reviewer.

2.6 Data Analysis

A systematic review of the pooled prevalence of anti-diabetic medication adherence was

carried out using a random effect model, generating a pooled prevalence with 95% CIs using
STATA software version 11. All indicators used to determine the potential publication bias
were Bag’s test for subjective assessment and Egger’s test (a p-value less than 0.05) for
objective evaluation[18]. Lastly, sensitivity analysis was done to estimate whether the pooled
effect size was affected by single studies.

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3130
3. RESULTS

A total of 22 studies with 6461 participants included in this metaanalysis is summarized in

Table 3.

Table 3: Characteristics of studies included in a meta-analysis of the prevalence of

adherence to anti-diabetic medication
S/N Author Year Type of DM Study Type Sample Adherence (95% CI)
Size
1 Nasir T et al. 2011 II Cross-sectional 384 41.80 (36.87, 46.73)
2 Ali et al. 2017 I&II Cross-sectional 146 54.80 (46.73, 62.87)
3 Belayneh KG et al. 2014 I&II Cross-sectional 270 72.20 (66.86, 77.54)
4 Kalayou KB et al. 2017 II Cross-sectional 320 97.00 (95.13, 98.87)
5 Gebrehiwot T et al. 2013 II Cross-sectional 267 75.70 (70.56, 80.84)
6 Ashebir K et al. 2016 II Cross-sectional 285 68.80 (63.42, 74.18)
7 Mastewal A et al. 2016 II Cross-sectional 288 85.10 (80.99, 89.21)
8 Ayele K et al. 2012 I&II Cross-sectional 222 78.40 (72.99, 83.81)
9 Jemal et al. 2017 II Cross-sectional 196 70.40 (64.01, 76.79)
10 Bonger et al. 2018 II Cross-sectional 619 95.70 (94.10, 97.30)
11 Asres T et al. 2014 II Cross-sectional 113 72.00 (63.72, 80.28)
12 Tessema T et al. 2016 II Cross-sectional 322 66.80 (61.66, 71.94)
13 Kassahun et al. 2016 II Cross-sectional 309 37.20 (31.81, 42.59)
14 Abate 2019 I&II Cross-sectional 416 31.20 (26.75, 35.65)
15 Abebe et al. 2014 I&II Cross-sectional 391 45.90 (40.96, 50.84)
16 Gerada et al. 2017 I&II Cross-sectional 378 66.90 (62.16, 71.64)
17 Tilahun T et al. 2017 I Cross-sectional 182 59.90 (52.78, 67.02)
18 Sorato et al. 2016 II Cross-sectional 194 84.00 (78.84, 89.16)
19 YohannesTekalegn 2018 II Cross-sectional 412 87.60 (84.42, 90.78)
20 Kalayou KB et al. 2017 II Cross-sectional 300 83.70 (79.52, 87.88)
21 Fseha B et al. 2017 II Cross-sectional 200 61.00(54.24,67.76)
22 D. J. Tesfaye et al. 2015 I& Cross-sectional 247 91.90 (88.50 95.30)

3.1 Prevalence of adherence to anti-diabetic medication among DM adult patients

(meta-analysis)
The pooled prevalence using the fixed effect model was shown that significant heterogeneity
between the studies. Hence, we analysed using the random effects model. Using a random
effects model, the estimated pooled prevalence of anti-diabetic medication adherence among
adult DM patients reported by the 22 studies was 69.533% (95% CI (61.083%, 77.982%)) with
significant heterogeneity between studies (I2=97.8%, p≤0.001). The pooled prevalence of anti-
diabetic adherence presented using the forest plot (Fig. 2).

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3131
Fig. 2. Forest plot showing the pooled prevalence of adherence to anti-diabetic medication among DM adult patients

3.2 Investigation of heterogeneity

Heterogeneity in systematic reviews and meta-analysis results of studies is inevitable due to
the difference in study quality, methodology, sample size, and inclusion criteria for participants
[18–20]. In this meta-analysis the value of I2 is a definite indication of significant high
heterogeneity, so we conducted the analysis with a random effects model to adjust for the
observed variability. Then we further try to investigate the sources of heterogeneity using meta-
regression model by using publication year and sample size as covariates. The result of the
meta-regression analysis showed that both covariates were not statistically significant for the
presence of heterogeneity (Table 4).

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3132
Table 4: A meta-regression analysis of factors with the heterogeneity of the prevalence of
adherence to anti-diabetic medication among DM adult patients

Heterogeneity Source Coefficients Std. error t P>t 95% CI

Publication year 0.03878 0.277 0.14 0.890 -0.5, 0.6
Sample Size 0.00027 0.003 0.08 0.935 -0.01, 0.01

3.3 Publication bias

Occurrence of publication bias was observed using funnel plots and tests (Egger’s and begs).
In this meta-analysis funnel plots and tests demonstrating evidence of publication bias. Each
point in funnel plots represents a separate study and asymmetrical distribution is evidence of
the presence of publication bias[17]. First, each study’s effect size was plotted against the
standard error and visual inspection of the funnel plot suggests asymmetry, as eight studies lay
on the left side and fourteen studies on the midline of the line representing the pooled
prevalence (Fig. 3). We also performed Egger’s, and Bag’s tests to investigate publication bias.
The result of these tests showed significant evidence of publication bias (p value < 0.05) (Table
5).

Table 5: Publication bias of the adherence to anti-diabetic medication among DM adult

Std_Eff Coefficients Std. Error T P>t 95% C.I.

Slope Bias 4.745 0.10 43.97 0.0000 4.52,4.97
-0.216 0.05 -3.91 0.001 -0.33,-0.10

Fig 3: Funnel plots to test, the publication bias of the 22 studies,

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3133
3.4 Sensitivity analysis
The result of sensitivity analyses using a random effects model suggested that no single study
unduly influenced the overall prevalence estimate of adherence to anti-diabetic medication
among DM adult patients in Sub-Saharan Africa (Fig 4).

Fig. 4. Result of Sensitivity analysis of the 22 studies

3.5 Factors assessed for Meta Analysis

There were 6 studies described determinant factors with anti-diabetic medication adherence of
the 22 studies. Among them, (3) studies were mentioned about the age of the patient, (4)
educational status, (4) marital status, (4) patients’ residency, (2) complexity of drugs, and (3)
durations of the DM, were found to be correlated with adherence. In this meta-analysis,
educational status, duration, marital status and complexity of the drugs were not significantly
associated, whereas patients age and residency were significantly associated with adherence of

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3134
anti-diabetic medications at P < 0.05. The pooled effect sizes for the factors in these studies
were shown in (Table 6).

Table 6: Results of meta-analyses and test for heterogeneity of anti-diabetic adherence

Variable OR (95%CI) Heterogeneity between studies
Q stat p-value I2 test for overall effect (p)
Age 1.82(1.27, 2.60) 1.88 0.39 0.00% 0.001
Marital status 0.66 (0.22, 2.00) 3.14 0.004 77.20% 0.463
Educational status 1.10(0.49, 2.50) 9.71 0.021 69.10% 0.814
Residency 4.88 (1.71, 13.95) 30.42 0.001 90.10% 0.003
Complexity of drug 0.58 (0.22, 1.51) 0.92 0.338 0.00% 0.265
Duration of the DM 0.576 (0.197, 1.685) 41.48 0.001 95.20% 0.313

4. DISCUSSION
This study is a meta-analysis and systematic review of Anti-diabetic medication adherence and
determinant factors. The pooled prevalence of anti-diabetic medication adherence in this study
was 69.53% (95% CI (61.083%, 77.982%)). This study was somehow in line with a survey
conducted on A Systematic Review of Adherence with Medications for Diabetes 36–93% [10].
However, this a Systematic Review and meta-Analysis pooled prevalence was higher than
studies done in Japan 58% and A systematic mixed studies review in the Middle East and North
Africa region 61% [9]. Based on the regional subgroup analysis, the study area was conducted
to assess the potential heterogeneity between studies. Of the 22 studies, the highest estimated
adherence to anti-diabetic medication prevalence found in to be, 88.19% (95% CI: 80.47 to
95.92%), I2=84.1%. Age is one of the determining factors for the good practice of antidiabetic
medication adherence. Patients who have older age better adhere than the other age groups. In
addition, residency also another determining factor, those patients from urban had better
adherence than patients from rural residents of towards anti-diabetic medications. This was
supported by A systematic mixed studies review in the Middle East and North Africa
region[21].

5. CONCLUSION

The result of the review suggests that the anti-diabetic medication adherence is low. The
combined results of the six studies about the determinants of the good practice of anti-diabetic
medications indicated that two factors: age and residency were the main determinants of
adherence of the anti-diabetic medications in Sub-Saharan Africa (SSA). Hence, we
recommend that health organizations Diabetic care service should incorporate diabetic

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3135
medication adherence services. Additionally, health educations and early screening of poorly
adherent as well as training of the hospital health professional on the adherence of the anti-
diabetic medications are highly recommended.

Limitations

The bias may be there because of the search was only in English language. Furthermore,
Scarcity of data in some regions of Africa may make a problem to generalize the findings.

Authors’ contributions

OEM, WKT and ZKE developed the concept of this study. OEM and WKT performed the
searches, extracted the data and wrote the first draft of the manuscript. WKT and ZKE provided
important intellectual input to revise the draft. All authors read and approved the final
manuscript.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.

Supplementary Material: www.shorturl.at/svBT5

References
[1] WHO. Diabetes. Diabetes, https://www.who.int/health-topics/diabetes#tab=tab_1 (2021,
accessed 3 September 2021).
[2] Bargiota A, Dimitropoulos K, Tzortzis V, et al. Sexual dysfunction in diabetic women.
Hormones 2011; 10: 196–206.
[3] Heinemann J, Atallah S, Rosenbaum T. The impact of culture and ethnicity on sexuality and
sexual function. Curr Sex Heal Reports 2016; 8: 144–150.
[4] Shadman Z, Akhoundan M, Poorsoltan N, et al. Factors associated with sexual function in
Iranian women with type 2 diabetes mellitus: Partner relationship as the most important
predictor. Iran Red Crescent Med J 2014; 16: e14941.
[5] Maiorino M, Bellastella G, Esposito K. Diabetes and sexual dysfunction : current perspectives.
Diabetes, Metab Syndr Obes Targets Ther 2014; 7: 95–105.
[6] Abaza H, Marschollek M. SMS education for the promotion of diabetes self-management in
low & middle income countries: a pilot randomized controlled trial in Egypt. BMC Public
Health 2017; 17: 962.
[7] Adisa R, Fakeye TO. Treatment non-adherence among patients with poorly controlled type 2
diabetes in ambulatory care settings in southwestern Nigeria. Afr Health Sci 2014; 14: 1–10.
[8] SE O, JJK K, SES T, et al. Assessing the influence of health systems on Type 2 Diabetes
Mellitus awareness, treatment, adherence, and control: A systematic review. PLoS One 2018;
13: e0195086.

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3136
[9] Ng’ang’a L, Ngoga G, Dusabeyezu S, et al. Implementation of blood glucose self-monitoring
among insulin-dependent patients with type 2 diabetes in three rural districts in Rwanda: 6
months open randomised controlled trial. BMJ Open 2020; 10: e036202.
[10] AT J, YE Y, Tilahun B, et al. Access to mobile phone and willingness to receive mHealth
services among patients with diabetes in Northwest Ethiopia: a cross-sectional study. BMJ
Open 2019; 9: e021766.
[11] NS L, Puoane T, CA D, et al. Referral outcomes of individuals identified at high risk of
cardiovascular disease by community health workers in Bangladesh, Guatemala, Mexico, and
South Africa. Global health action 2015; 8: 26318.
[12] Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for systematic review and
meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 2015 41 2015; 4: 1–9.
[13] Rahmanian E, Salari N, Mohammadi M, et al. Evaluation of sexual dysfunction and female
sexual dysfunction indicators in women with type 2 diabetes: a systematic review and meta-
analysis. Diabetol Metab Syndr 2019; 11: 73.
[14] Barbagallo F, Mongioì LM, Cannarella R, et al. Sexual Dysfunction in Diabetic Women: An
Update on Current Knowledge. Diabetol 2020, Vol 1, Pages 11-21 2020; 1: 11–21.
[15] Sharifiaghdas F, Azadvari M, Shakhssalim N, et al. Female sexual dysfunction in type 2
diabetes: a case control study. Med Princ Pract Int J Kuwait Univ Heal Sci Cent 2012; 21:
554–559.
[16] Moher D, Liberati A, Tetzlaff J, et al. Preferred Reporting Items for Systematic Reviews and
Meta-Analyses: The PRISMA Statement. PLOS Med 2009; 6: e1000097.
[17] Munn Z, Moola S, Riitano D, et al. The Development of a Critical Appraisal Tool for Use in
Systematic Reviews: Addressing Questions of Prevalence. Int J Heal Policy Manag 2014; 3:
123–128.
[18] Egger M, Smith GD, Schneider M, et al. Bias in meta-analysis detected by a simple, graphical
test. Br Med J 1997; 315: 629–634.
[19] Azharuddin M, Adil M, Sharma M, et al. A systematic review and meta-analysis of non-
adherence to anti-diabetic medication: Evidence from low- and middle-income countries. Int J
Clin Pract; 75. Epub ahead of print 1 November 2021. DOI: 10.1111/IJCP.14717.
[20] Baker WL, Michael White C, Cappelleri JC, et al. Understanding heterogeneity in meta-
analysis: The role of Meta-regression. Int J Clin Pract 2009; 63: 1426–1434.
[21] KB K, JK E, Baruani B, et al. Medication Adherence Clubs: a potential solution to managing
large numbers of stable patients with multiple chronic diseases in informal settlements. Trop
Med Int Health 2015; 20: 1265–1270.

International Journal of Research in Medical Sciences | Mar 2022 | Vol 7 | Issue 10 Page| 3137