Hindawi

Journal of Diabetes Research

Volume 2020, Article ID 4760624, 10 pages
https://doi.org/10.1155/2020/4760624

Research Article
The Association between Diabetes-Related Distress and
Medication Adherence in Adult Patients with Type 2 Diabetes
Mellitus: A Cross-Sectional Study

Irene A. Kretchy ,1 Augustina Koduah,1 Thelma Ohene-Agyei,1 Vincent Boima,2

and Bernard Appiah3,4
1
Department of Pharmacy Practice and Clinical Pharmacy, School of Pharmacy, College of Health Sciences, University of Ghana,
P.O. Box LG 43, Legon, Ghana
2
Department of Medicine and Therapeutics, School of Medicine and Dentistry, College of Health Sciences, University of Ghana,
P.O. Box GP 4236, Accra, Ghana
3
Centre for Science and Health Communication, PMB M71, Ministries, Accra, Ghana
4
Department of Environmental and Occupational Health, School of Public Health, Texas A&M University Health Science Center,
212 Adriance Lab Rd, 1266 TAMU, College Station, Texas, USA

Correspondence should be addressed to Irene A. Kretchy; ikretchy@ug.edu.gh

Received 1 July 2019; Revised 31 January 2020; Accepted 14 February 2020; Published 2 March 2020

Academic Editor: Akira Sugawara

Copyright © 2020 Irene A. Kretchy et al. This is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.

Background. Type 2 diabetes mellitus (T2DM) is a major public health problem associated with distress. T2DM can affect health
outcomes and adherence to medications. Little is however known about the association between diabetes distress and
medication adherence among patients with T2DM in Ghana. Objective. The objective of the present study is twofold: to estimate
distress associated with T2DM and to examine its association with medication adherence. Methods. A hospital-based cross-
sectional study was conducted among 188 patients with T2DM recruited from a diabetes specialist outpatient clinic at the
Pantang Hospital in Accra, Ghana. Data were obtained using the Problem Areas In Diabetes (PAID) scale and the Medication
Adherence Report Scale. Results. The findings showed that about 44.7% of the patients showed high levels of diabetes-related
distress. Poor adherence to medications was recorded in 66.5% of the patients. Patients who were highly distressed had 68%
lower odds of adhering to their medications compared to those who were not (OR: 0.32, 95% CI: 0.15-0.65). A principal
component analysis revealed four areas of T2DM distress which were conceptualized as negative emotions about diabetes,
dietary concerns and diabetes care, dissatisfaction with external support, and diabetes management helplessness. Conclusion.
Our findings suggest that diabetes distress is a significant determinant of medication adherence behaviour in patients with
T2DM. Thus, incorporating routine screening for distress into the standard diabetes care within the Ghanaian health system
and having health practitioners adopt holistic approaches to diabetes management will be important context-specific
interventions to improve adherence and health outcomes of people living and coping with T2DM.

1. Introduction populations worldwide [2], and it is associated with increas-

ing unhealthy lifestyle such as poor dietary choices, lack of
Type 2 diabetes mellitus (T2DM) is a group of metabolic exercise, and inadequate physical activity [3, 4]. An estimated
diseases characterised by elevated levels of blood glucose, global prevalence of T2DM for age groups 20-79 was 8.8% in
leading to serious damage to other organs over time [1]. It 2015, and it is expected to increase to 10.4% of adults by 2040
is a major cause of morbidity, disability, and mortality among [5]. A greater burden of 75% of persons with diabetes is in
2 Journal of Diabetes Research

developing countries [5]. The prevalence of T2DM in patients with diabetes in a week with approximately 3 new
Ghana is reported to be 6.46%, and this is expected to rise cases reporting within the week.
by 2040 [5, 6].
Type 2 diabetes mellitus is associated with negative 2.2. Participants. A total of 188 patients with T2DM (91.3%)
emotions such as anxiety, depression, and distress, and these were part of this study from the Pantang Hospital between
emotions have been associated with poor clinical conse- May and July 2017, out of the 206 eligible participants using
quences including medication nonadherence [7–10] and gly- a simple random sampling method. The minimum sample
cemic outcomes [8, 11, 12]. Diabetes distress reflects negative size of 105 was obtained using the formula by Cochran
feelings surrounding the disease and refers to the emotional (1963) and the estimated prevalence of diabetes in Ghana at
response to the struggles, concerns, and worries associated 6.46% [6]:
with the broader demands of diabetes [13]. With time, diabe-
tes distress becomes part of the diabetes experience for many ðZ ∝/2 Þ2 pð1 − pÞ
patients and it is usually context-specific [14]. The distress is n0 = ðDeff Þ , ð1Þ
e2
from the daily hassles and demands of the disease manage-
ment [15], worries about poor glycemic control [16], fears
about diabetic complications [16], poor support from signif- where n0 is the minimum sample size, Z ∝/2 is 1.96 at a
icant others [17, 18], stigma [19], and financial difficulties confidence interval at 95%, e is the level of precision, p is
[20]. When diabetes distress becomes protracted and is not the estimated proportion of patients with T2DM, and Deff
identified and managed, patients experience burnout result- is the design effect set at 1.03 and assuming a 10% nonre-
ing in feelings of helplessness, hopelessness, and frustration sponse rate.
with T2DM care [21–23]. Burnout may sometimes be physi- The study participants were adult patients with T2DM
ologically triggered following an acute hyperglycemic crisis aged 18 years and over. Participants with type 1 diabetes, ges-
[24]. Patients with high diabetes-related distress are likely tational diabetes, maturity-onset diabetes of the young, or
to demonstrate poor self-management [21]. latent autoimmune diabetes in adults were excluded from
A systematic review of studies on T2DM in sub-Saharan the study. Participants were also excluded if they had been
Africa reported a lack of studies on the psychosocial aspect of diagnosed of any known psychiatric disorder according to
diabetes management [25]. In Ghana, despite the presence of their medical records. The hospital attends to patients with
some policy drive and programmatic responses to T2DM in mental and physical conditions, and this exclusion criterion
particular and chronic noncommunicable diseases in general, was necessary to avoid any known mental illness from being
the effect of these programmes is yet to reflect in the lives of a possible confounder.
patients [26, 27]. Similar to other sub-Saharan African coun- 2.3. Measures. All participants completed questions on
tries, the psychosocial dimensions of the illness experience demographic characteristics with other clinical data about
have been largely unexplored in Ghana [28, 29]. Previous blood glucose levels obtained from the patient records. Dia-
studies on diabetes in Ghana have focused primarily on the betes distress was assessed using the Problem Areas In Diabe-
prevalence and determinants [6, 30–34]. The primary objec- tes Questionnaire [37]. This is a 20-item measure describing
tive of the study therefore was to estimate diabetes-specific negative emotions related to T2DM such as anger, fear, and
distress and assess its impact on optimising treatment and frustration. It uses a 5-point Likert response scale from 0
adherence to medication in patients with T2DM. Further, representing “no problem” to 4 representing “serious prob-
the study explored the dimensions of diabetes-specific dis- lem.” To obtain the total scores ranging from 0 to 100, initial
tress using principal component analysis of the Problem scores are multiplied by 1.25. Higher scores indicate greater
Areas In Diabetes (PAID) scale to observe variations and distress with T2DM. While participants with scores of 40
emphasize patterns of distress in the patients [7]. Although and above were highly distressed, very low scores of 0–10
the associations between diabetes distress and adherence may be indicative of patients in denial. The measure of reli-
have been assessed previously in other countries [8–12], such ability using Cronbach’s alpha for the PAID scale in this
studies have not been done in Ghana. This has created a gap study was 0.8299.
in knowledge in the context of T2DM from an Afrocentric To estimate the level of adherence to medications, the
perspective. While medications for managing T2DM in study used the Medication Adherence Report Scale (MARS)
Ghana are readily available, adherence to these medications which assesses both intentional and unintentional nonadher-
is still not optimal [35, 36]. Thus, the information from this ence to medicines [38]. Participants responded to five items
study will facilitate context-specific understanding of the (e.g., I forget to take my medicines) on a 5-point scale from
problem from a psychosocial perspective so that culturally “always” to “never,” and a total score of 25 and more indi-
appropriate adherence solutions can be instituted. cated better adherence to medications [38]. The MARS has
been previously used for patients with T2DM in Singapore
2. Methods [39]. In this study, Cronbach’s alpha was 0.6967.
2.1. Study Design. This was a hospital-based cross-sectional 2.4. Ethical Consideration. The study received ethical
study of patients with T2DM who had reported for a clinical approval from the Ethical Review Committee of the Ghana
review at the outpatient clinic at the Pantang Hospital in Health Service with approval number GHS-ERC:130/12/17.
Accra, Ghana. The facility attends to an average of 35 Permission was also obtained from the Administrator of
Journal of Diabetes Research 3

Table 1: Background and clinical characteristics of patients with T2DM receiving treatment at Pantang Hospital.

Frequency Percent
Age
Mean ± SD 59:31 ± 11:94
≤50 41 21.81
51-60 62 32.98
61+ 85 45.21
Sex
Female 136 72.34
Male 52 27.66
Glucose: median (LQ, UQ) 7.8 (6.4, 11)
Comorbidity
No 77 40.96
Yes 111 59.04
Number of medications
Mean ± SD 2:56 ± 1:13
One 31 16.58
Two 69 36.90
Three 54 28.88
Four or more 33 17.65
PAID score
Mean ± SD 36:03 ± 4:28
Not highly distressed (PAID score < 40) 104 55.32
Highly distressed (PAID score ≥ 40) 84 44.68
MARS-5 score
Mean ± SD 21:23 ± 3:64
Low medication adherence (MARS score < 25) 125 66.49
Medication adherent (MARS score of 25) 63 33.51
SD: standard deviation; LQ: lower quartile; UQ: upper quartile.

the Pantang Hospital to conduct the study in the facility. independent variables on the outcome variables. In assessing
Informed consent was obtained from the sampled diabetic the association between the individual item responses of
patients, and confidentiality/privacy was assured before their PAID and medication adherence, the item responses were
participation in the study. dichotomized into patients who regarded the item as “a prob-
lem” and “not a problem” [16, 40]. The statistical test of sig-
2.5. Data Analysis. STATA version 14 was used for the data nificance was set at 5%.
analysis. Frequencies and percentages were reported as
descriptive statistics for categorical variables. Principal 3. Results
component analysis with varimax rotation analysis was used
to explore diabetes distress from the various elements of the 3.1. Patient Characteristics. Table 1 shows the background
PAID scale. For continuous variables, means with standard and clinical characteristics of the patients with T2DM
deviations were reported for normally distributed data while (n = 188). On average, the patients were 59:3 ± 11:9 years
median and interquartile ranges were reported as descriptive old, with females constituting the majority (72.3%). Fifty-
statistics when the normality assumption was violated. nine percent of the patients had at least one comorbidity
Normality assumption of continuous variables was tested and were prescribed a mean of 2:6 ± 1:1 medications for their
with the skewness and kurtosis Shapiro-Francia tests. Chi- conditions. Averagely, participants had a glucose level of less
squared and Fisher exact tests of independence were used than or equal to 7.8 (IQR: 6.4–11.0).
to test for association between categorical independent
variables and the outcome variables (diabetes distress and 3.2. Diabetes Distress. While the average PAID score among
medication adherence). The Wilcoxon rank sum test was the patients was 36:0 ± 4:3, 44.7% (84/188) showed high
used to compare the median of glucose levels across the var- levels of distress with PAID scores ≥ 40. The chi-squared test
ious categories of the outcome variables. Binary logistic showed significant associations between comorbidities
regression models were used to determine the effects of the (p = 0:049), glucose levels (0.006), and high distress.
4 Journal of Diabetes Research

Table 2: Association between background and clinical characteristics and high distress in patients with T2DM receiving treatment at Pantang
Hospital.

Adjusted logistic
High distress
Chi-square p value regression model
No, n (%) Yes, n (%) Odds ratio p value
Sex 0.01 0.939 0.485
Female 75 (55.15) 61 (44.85) ref
Male 29 (55.77) 23 (44.23) 0.78 (0.38–1.57)
Age 2.23 0.327 0.943
≤50 20 (48.78) 21 (51.22) ref
51-60 32 (51.61) 30 (48.39) 0.98 (0.42–2.33)
61± 52 (61.18) 33 (38.82) 0.89 (0.38–2.05)
Glucose level: median (LQ, UQ) 7.3 (6.2, 10) 9.2 (6.7, 12.5) 0.006∗∗ § 1.12 (1.04–1.21) <0.001∗∗∗
Number of medications 4.75 0.191 0.138
One 18 (58.06) 13 (41.94) ref
Two 31 (44.93) 38 (55.07) 2.35 (0.92–6.03)
Three 34 (62.96) 20 (37.04) 1.06 (0.38–2.92)
Four or more 20 (60.61) 13 (39.39) 1.38 (0.44–4.30)
Comorbidity 3.87 0.049 0.130
No 36 (46.75) 41 (53.25) ref
Yes 68 (61.26) 43 (38.74) 0.58 (0.29–1.16)
%: row percentages; n: number of observations; ∗ p < 0:01, ∗∗ p < 0:01, and ∗∗∗ p < 0:001; CI: confidence interval; ref: reference category; LQ: lower quartile;
UQ: upper quartile, §: p value obtained from a Wilcoxon rank sum test.

However, results from the multiple logistic regression model Factor 1 was conceptualized as negative emotions about
showed that the blood glucose level was the only significant diabetes consisting of thirteen variable loadings of items such
predictor of distress (p < 0:001). The odds of high diabetic as the following: “Scared about thoughts of living with diabe-
distress among patients with T2DM increased by 12% with tes” (0.643), “Depressed about thoughts of diabetes” (0.708),
every additional unit increase in the glucose level (OR: 1.12, “Worry about the future and complications” (0.610), “Guilt
95% CI: 1.04-1.21) (Table 2). and anxiety when off-track management of diabetes”
(0.603), and “Diabetes taking too much of mental energy”
3.3. Medication Adherence. The average MARS-5 score was (0.540). The second factor was noted as dietary concerns
21:2 ± 3:6, with one-third of the patients optimally adhering and diabetes care consisting of two variable loadings: “Feel-
to their medications. The blood glucose level was not signif- ings of deprivation regarding food and meals” (0.596) and
icantly associated with medication adherence, and higher “No clear or concrete goals for diabetes care” (0.529). Factor
levels were observed among patients with poor medication 3 was referred to as dissatisfaction with external support.
adherence compared with those who adhered completely to This comprised two variable loadings: “Feeling unsatisfied
their medications (8.4 vs. 7.1, p = 0:056). with diabetes physician” (0.464) and “Friends and family
not supportive” (0.596). Diabetes management helplessness
3.4. High Diabetes Distress and Medication Adherence. The was the fourth factor, constituted by two variables: “Coping
proportion of patients with T2DM who exhibited high dis- with complications of diabetes” and “Feeling burned out by
tress had significantly lower scores on medication adherence constant effort to manage diabetes.” The two variables had
compared with those who were not distressed (20.2 vs. 44.2, 0.510 and 0.489 varimax rotational loadings, respectively.
p = 0:001). Those with high distress had 68% lower odds of
adhering to their medications compared to those who were 4. Discussions
not distressed (OR: 0.32, 95% CI: 0.15-0.65) (Table 3). Ten
out of the twenty items on diabetes distress using the PAID The present study contributes to knowledge of illness experi-
scale showed significant associations with medication adher- ence of patients with T2DM from a psychosocial perspective
ence (p < 0:05) (Table 4). These included discouragement with emphasis on diabetes distress. While studies on diabetes
with diabetes treatment, uncomfortable social situations, distress have mainly been from developed countries, fewer
and feelings of anger, anxiety, guilt, loneliness, and burnout. studies have emerged from Africa [25]. This study therefore
focused on patients with T2DM in a Ghanaian health con-
3.5. Components of Diabetes Distress. Table 5 shows the prin- text, bearing in mind the cultural influences of illness experi-
cipal component analysis of four components with eigen- ences on health outcomes.
values exceeding 1, explaining 27.18%, 7.49%, 6.91%, and In this study, patients who experienced diabetes distress
6.43% of the variance, respectively. tended to poorly adhere to their medications compared with
Journal of Diabetes Research 5

Table 3: Association between background and clinical characteristics and medication adherence status of patients with T2DM receiving
treatment at Pantang Hospital.

Adjusted logistic
Adherent (MARS-5 score of 25)
Chi-square p value regression model
No, n (%) Yes, n (%) Odds ratio p value
Sex 0.3 0.587 0.370
Female 92 (67.65) 44 (32.35) ref
Male 33 (63.46) 19 (36.54) 1.41 (0.67–2.98)
Age 1.19 0.551 0.422
≤50 30 (73.17) 11 (26.83) ref
51-60 39 (62.9) 23 (37.1) 1.86 (0.73–4.75)
61± 56 (65.88) 29 (34.12) 1.41 (0.56–3.52)
Glucose level: median (LQ, UQ) 8.4 (6.6, 11.3) 7.1 (6.2, 10.5) 0.056§ 0.98 (0.91–1.07) 0.721
Number of medications 5.79 0.122 0.161
One 19 (61.29) 12 (38.71) ref
Two 51 (73.91) 18 (26.09) 0.66 (0.24–1.78)
Three 38 (70.37) 16 (29.63) 0.71 (0.25–2.00)
Four or more 17 (51.52) 16 (48.48) 1.83 (0.58–5.81)
Comorbidity 0.01 0.951 0.471
No 51 (66.23) 26 (33.77) ref
Yes 74 (66.67) 37 (33.33) 0.76 (0.36–1.61)
High distress 12.01 0.001∗∗ 0.002∗∗
No 58 (55.77) 46 (44.23) ref
Yes 67 (79.76) 17 (20.24) 0.32 (0.15–0.65)
%: row percentages; n: number of observations; ∗ p < 0:01, ∗∗ p < 0:01, and ∗∗∗ p < 0:001; CI: confidence interval; ref: reference category; LQ: lower quartile;
UQ: upper quartile; §: p value obtained from a Wilcoxon rank sum test.

those who were not distressed about their disease experience factors for physical health and health-related outcomes
and outcome. Diabetes distress is an essential predictor of including medication adherence [16, 47]. This corroborates
clinical outcomes in T2DM care, and it has been linked to our findings from the item analysis of the PAID and adher-
poor self-management, treatment adherence, and blood glu- ence. Due to the chronicity of T2DM and the fact that
cose status in such patients [41]. patients with T2DM will have to manage with living with
Findings from the PCA further showed four areas of dis- the condition for the rest of their lives, our findings suggest
tress by patients with T2DM. These are negative emotions, the need for interventions targeting positive emotions which
dietary concerns, dissatisfaction with external support, and can buffer the effects of the negative emotions. This is
diabetes management helplessness. The concept of diabetes because positive feelings are important in chronic diabetes
distress in this context could be suggested to mean a combi- care and outcomes [48]. In a related study, a meaningful rela-
nation of these four areas of concern. In a related study where tionship between the levels of positive affect and adherence
distress was assessed using the diabetes distress scale, a differ- measures was reported [41]. Thus, support services for
ent measure from what was used in this study, PCA yielded T2DM can address potentially harmful effects of negative
factors that were similar to what was obtained in this study feelings and emphasize benefits of positive emotionality for
[42]. Yet in another study involving southern rural African- better health-related outcomes. Such support services could
American patients with T2DM, a PCA of the PAID yielded include the use of culturally relevant communication inter-
two factors: lack of confidence and negative emotional conse- ventions including mobile phones, mass media, social media,
quences [7]. and face-to-face approaches and social media to address
Living with T2DM can negatively impact on the psycho- emotional needs of patients with T2DM experiencing
logical well-being of patients. Previous studies on negative distress.
emotions with its affective correlates in T2DM have primar- The study also showed that T2DM patients were dis-
ily focused on depression, mainly based on clinical presenta- tressed about the nature of care from their physicians and
tions and diagnosis in the Diagnostic and Statistical Manual support from social relations. In addition, having uncomfort-
for Mental Disorders [43–45]. Yet, the prospective effect of able social situations and the feeling of loneliness with T2DM
negative emotions with subclinical symptoms also has clini- were significantly associated with poor medication adher-
cal implications for T2DM care and adherence [46]. ence. This finding indicates that, from the perspective of
Emotionally negative symptoms like discouragement, the patients, living with T2DM could be synonymous with
anger, anxiety, guilt, and burnout have been indicated as risk feeling alone with poor social support from significant others,
6 Journal of Diabetes Research

Table 4: Association between individual item responses of PAID scale and medication adherence status of patients with T2DM receiving
treatment at Pantang Hospital.

Adherent (MARS-5
Total Unadjusted
Items score of 25)
n (% )
a
No, n (%b) Yes, n (%b) Odds ratio p value
No clear or concrete goals for diabetes care 0.468
No problem 137 (72.87) 89 (64.96) 48 (35.04) ref
Problem 51 (27.13) 36 (70.59) 15 (29.41) 0.77 (0.38–1.55)
Feeling discouraged with diabetes treatment plan 0.007∗∗
No problem 108 (57.45) 63 (58.33) 45 (41.67) ref
Problem 80 (42.55) 62 (77.5) 18 (22.5) 0.41 (0.21–0.78)
Scared about thoughts of living with diabetes 0.134
No problem 61 (32.45) 36 (59.02) 25 (40.98) ref
Problem 127 (67.55) 89 (70.08) 38 (29.92) 0.61 (0.33–1.16)
Uncomfortable social situations related to diabetes 0.002∗∗
No problem 75 (39.89) 40 (53.33) 35 (46.67) ref
Problem 113 (60.11) 85 (75.22) 28 (24.78) 0.38 (0.2–0.7)
Feelings of deprivation regarding food and meals 0.104
No problem 25 (13.3) 13 (52) 12 (48) ref
Problem 163 (86.7) 112 (68.71) 51 (31.29) 0.49 (0.21–1.16)
Feeling depressed about thoughts of diabetes 0.313
No problem 51 (27.13) 31 (60.78) 20 (39.22) ref
Problem 137 (72.87) 94 (68.61) 43 (31.39) 0.71 (0.36–1.38)
Not knowing if mood is related to diabetes 0.009∗∗
No problem 68 (36.17) 37 (54.41) 31 (45.59) ref
Problem 120 (63.83) 88 (73.33) 32 (26.67) 0.43 (0.23–0.81)
Feeling overwhelmed by diabetes 0.177
No problem 40 (21.28) 23 (57.5) 17 (42.5) ref
Problem 148 (78.72) 102 (68.92) 46 (31.08) 0.61 (0.3–1.25)
Worrying about low sugar reactions 0.019∗
No problem 57 (30.32) 45 (78.95) 12 (21.05) ref
Problem 131 (69.68) 80 (61.07) 51 (38.93) 2.39 (1.16–4.95)
Feeling angry about thought of living with diabetes 0.023∗
No problem 60 (31.91) 33 (55) 27 (45) ref
Problem 128 (68.09) 92 (71.88) 36 (28.13) 0.48 (0.25–0.91)
Feeling constantly concerned about food and eating 0.696
No problem 13 (6.91) 8 (61.54) 5 (38.46) ref
Problem 175 (93.09) 117 (66.86) 58 (33.14) 0.79 (0.25–2.53)
Worrying about the future and complications 0.100
No problem 30 (15.96) 16 (53.33) 14 (46.67) ref
Problem 158 (84.04) 109 (68.99) 49 (31.01) 0.51 (0.23–1.13)
Guilty and anxious when off track management 0.022∗
No problem 49 (26.06) 26 (53.06) 23 (46.94) ref
Problem 139 (73.94) 99 (71.22) 40 (28.78) 0.46 (0.23–0.89)
Not accepting “diabetes” 0.085
No problem 115 (61.17) 71 (61.74) 44 (38.26) ref
Problem 73 (38.83) 54 (73.97) 19 (26.03) 0.57 (0.3–1.08)
Feeling unsatisfied with diabetes physician 0.089
No problem 131 (69.68) 82 (62.6) 49 (37.4) ref
Problem 57 (30.32) 43 (75.44) 14 (24.56) 0.54 (0.27–1.1)
Journal of Diabetes Research 7

Table 4: Continued.

Adherent (MARS-5
Total Unadjusted
Items score of 25)
n (% )a
No, n (%b) Yes, n (%b) Odds ratio p value
Diabetes taking too much of mental energy <0.001∗∗∗
No problem 50 (26.6) 21 (42) 29 (58.00) ref
Problem 138 (73.4) 104 (75.36) 34 (24.64) 0.24 (0.12–0.47)
Feeling alone with diabetes <0.001∗∗∗
No problem 84 (44.68) 44 (52.38) 40 (47.62) ref
Problem 104 (55.32) 81 (77.88) 23 (22.12) Six
Friends and family not supportive 0.497
No problem 69 (36.7) 48 (69.57) 21 (30.43) ref
Problem 119 (63.3) 77 (64.71) 42 (35.29) 1.25 (0.66-2.35)
Coping with complications of diabetes <0.001∗∗∗
No problem 76 (40.43) 39 (51.32) 37 (48.68) ref
Problem 112 (59.57) 86 (76.79) 26 (23.21) 0.32 (0.17-0.6)
Feeling “burned out” by constant effort to manage 0.008∗∗
No problem 65 (34.57) 35 (53.85) 30 (46.15) ref
Problem 123 (65.43) 90 (73.17) 33 (26.83) 0.43 (0.23-0.8)
%a: column percentage; %b: row percentages; n: number of observations; ∗ p < 0:01, ∗∗ p < 0:01, and ∗∗∗ p < 0:001; CI: confidence interval; ref: reference category;
LQ: lower quartile; UQ: upper quartile; p values were obtained from an unadjusted binary logistic regression model.

Table 5: Factor loadings, communalities (h2 ), and percent of variance for principal factor extraction and varimax rotation on items of
diabetes distress.

F1 F2 F3 F4 h2
a
Items
Negative emotions about diabetes
Feeling discouraged with diabetes treatment plan 0.472 0.344
Scared about thoughts of living with diabetes 0.643 0.604
Uncomfortable social situations related to diabetes 0.640 0.495
Feeling depressed about thoughts of diabetes 0.708 0.689
Not knowing if mood is related to diabetes 0.709 0.597
Feeling overwhelmed by diabetes 0.466 0.494
Feeling angry about thought of living with diabetes 0.606 0.656
Worrying about the future and complications 0.610 0.519
Guilty and anxious when off-track management 0.603 0.435
Diabetes taking too much of mental energy 0.540 0.500
Feeling alone with diabetes 0.570 0.528
Coping with complications of diabetes 0.603 0.712
Feeling burned out 0.611 0.641
Dietary concerns and diabetes care
No clear or concrete goals for diabetes care 0.529 0.711
Feelings of deprivation regarding food and meals 0.596 0.600
Dissatisfaction with external support
Feeling unsatisfied with diabetes physician 0.464 0.736
Friends and family not supportive 0.596 0.627
Diabetes management helplessness
Coping with complications of diabetes 0.510 0.712
Feeling burned out by constant effort to manage diabetes 0.489 0.641
Percent variance 27.179 7.485 6.906 6.427
a
Factor labels: F 1 : negative emotions about diabetes; F 2 : dietary concerns and diabetes care; F 3 : dissatisfaction with external support; F 4 : diabetes management
helplessness.
8 Journal of Diabetes Research

regardless of evidence that psychosocial support is helpful in 4.2. Implications for Policy. It may also be relevant to recom-
adaptive behaviours by patients with T2DM [49]. Support mend that routine screening for diabetes distress be incorpo-
and care from the social networks of patients including health rated into national standard diabetes treatment guidelines for
care professionals, family, friends, neighbours, colleagues, care within the health care systems in Ghana. This is to
and fellow patients with T2DM could help them take positive ensure that patients with a possible risk of distress can receive
stances, build resilience, relieve distress, and improve on their more comprehensive diabetic care from a psychosocial
well-being [50, 51]. Furthermore, care and support from for- perspective.
mal and familial contacts have positive effects on the medica-
tion adherence behaviour in patients with T2DM due to the 5. Conclusion
encouragement of optimism in such patients [52].
The prevalence of both chronic microvascular and acute This study investigated the link between diabetes distress and
complications of T2DM is much greater in patients with poor medication adherence in patients with T2DM in Ghana. The
glycemic control and poor dietary quality [10, 53, 54]. Con- information suggests that diabetes distress is a significant
cerns about meeting dietary requirements and following determinant of medication adherence behaviour. Thus,
treatment plans were also synonymous to the concept of dia- incorporating routine screening for distress into the standard
betes distress in this study. Psychological distress has been diabetes care within the Ghanaian health system and having
associated with a high risk of T2DM complications, and in health practitioners adopt a biopsychosocial approach to
this study, patients who had difficulty dealing with their com- diabetes management will be important context-specific
plications reported poor medication adherence behaviour. interventions to improve health outcomes of people living
The occurrence of diabetic complications has been proposed and coping with T2DM.
to be significantly higher in nonadherent patients compared
to those who adhere to medications [53].
As indicated in Introduction, patients with high diabetes- Data Availability
related distress show signs of poor self-management of The data that support the findings of this study are available
T2DM [21], and in this study, high blood glucose levels and from the corresponding author, [IAK], upon reasonable
distress were significantly related. Similar to previous studies, request.
this study suggests that assessment and management of dis-
tress in patients with T2DM are crucial in determining health
outcomes. It was observed that participants who felt highly Conflicts of Interest
distressed by the constant effort needed to manage T2DM
All authors have no conflict of interest to declare.
poorly adhered to their medications [55, 56].
This study has some limitations. First, a mixed method
design could have been adopted for this study in order to Acknowledgments
explore the concept of diabetes distress from a more qualita-
tive approach. Second, respondents are from only one hospi- The authors would like to acknowledge the staff and patients
tal; thus, the findings may not be representative of the general at the Pantang Hospital for their support in this study and
population patients with T2DM in Ghana; in addition, cau- during the data collection process. Appreciation also goes
sality could not be established nor could the direction of to the following for their various roles in the preparation of
the effect of diabetes distress on adherence be determined this paper: Prof. Yaa Ntiamoah Badu and the BANGA-
because the data were obtained through a cross-sectional Africa writeshop team, Dr. Rabui Asante, and Mr. Kofi
study approach. Given that this study relied on self-reports, Adjabeng.
recall bias could be a limitation. In spite of these limitations,
this study is among the first to report the association between References
T2DM distress and nonadherence in Ghana, thus identifying
areas of context-specific interventions to improve adherence [1] L. Nalysnyk, M. Hernandez-Medina, and G. Krishnarajah,
in patients with T2DM. “Glycaemic variability and complications in patients with
diabetes mellitus: evidence from a systematic review of the
literature,” Diabetes, Obesity and Metabolism, vol. 12, no. 4,
4.1. Implications for Clinical Practice. These study findings
pp. 288–298, 2010.
are significant in explaining the association between psycho-
[2] J. Bhutani and S. Bhutani, “Worldwide burden of diabetes,”
social interactions and health outcomes in patients with
Indian Journal of Endocrinology and Metabolism, vol. 18,
T2DM. Based on the results, psychological and social no. 6, pp. 868–870, 2014.
context-specific interventions that address diabetes distress
[3] T. Psaltopoulou, I. Ilias, and M. Alevizaki, “The role of diet and
should be considered when patients with diabetes are man- lifestyle in primary, secondary, and tertiary diabetes
aged at health institutions. For example, culturally relevant prevention: a review of meta-analyses,” The Review of Diabetic
communications may need to be developed, pilot-tested, Studies, vol. 7, no. 1, pp. 26–35, 2010.
and implemented to address psychological issues confronting [4] C. Zhang and Y. Ning, “Effect of dietary and lifestyle factors on
patients with T2DM in Ghana. Other cultural dimensions the risk of gestational diabetes: review of epidemiologic evi-
that impact treatment outcomes such as religiosity could also dence,” The American Journal of Clinical Nutrition, vol. 94,
be explored in the management of T2DM in Ghana. suppl_6, pp. 1975S–1979S, 2011.
Journal of Diabetes Research 9

[5] K. Ogurtsova, J. D. da Rocha Fernandes, Y. Huang et al., “IDF problem for an increasing epidemic,” The Patient-Patient-
Diabetes Atlas: global estimates for the prevalence of diabetes Centered Outcomes Research, vol. 6, no. 1, pp. 1–10, 2013.
for 2015 and 2040,” Diabetes Research and Clinical Practice, [20] N. S. Levitt, “Diabetes in Africa: epidemiology, management
vol. 128, pp. 40–50, 2017. and healthcare challenges,” Heart, vol. 94, no. 11, pp. 1376–
[6] M. Asamoah-Boaheng, O. Sarfo-Kantanka, A. B. Tuffour, 1382, 2008.
B. Eghan, and J. C. Mbanya, “Prevalence and risk factors for [21] C. Fritschi and L. Quinn, “Fatigue in patients with diabetes: a
diabetes mellitus among adults in Ghana: a systematic review review,” Journal of Psychosomatic Research, vol. 69, no. 1,
and meta-analysis,” International Health, vol. 11, no. 2, pp. 33–41, 2010.
pp. 83–92, 2018. [22] W. Polonsky, Diabetes Burnout: What to Do When You Can't
[7] S. T. Miller and T. A. Elasy, “Psychometric evaluation of the Take It Anymore, American Diabetes Association, 1999.
Problem Areas in Diabetes (PAID) survey in southern, rural [23] A. K. Symon, S. S. Vargese, E. Mathew, K. R. Akshay, and
African American women with type 2 diabetes,” BMC Public J. Abraham, “Diabetes related distress in adults with type 2
Health, vol. 8, no. 1, 2008. diabetes mellitus: a community-based study,” International
[8] A. U. Pandit, S. C. Bailey, L. M. Curtis et al., “Disease-related Journal Of Community Medicine And Public Health, vol. 6,
distress, self-care and clinical outcomes among low-income no. 1, pp. 151–155, 2018.
patients with diabetes,” Journal of Epidemiology and Commu- [24] A. J. Sommerfield, I. J. Deary, and B. M. Frier, “Acute hyper-
nity Health, vol. 68, no. 6, pp. 557–564, 2014. glycemia alters mood state and impairs cognitive performance
[9] R. R. Rubin, “Adherence to pharmacologic therapy in patients in people with type 2 diabetes,” Diabetes Care, vol. 27, no. 10,
with type 2 diabetes mellitus,” The American Journal of pp. 2335–2340, 2004.
Medicine, vol. 118, no. 5, pp. 27–34, 2005. [25] V. Stephani, D. Opoku, and D. Beran, “Self-management of
[10] J. Silverman, J. Krieger, M. Kiefer, P. Hebert, J. Robinson, and diabetes in sub-Saharan Africa: a systematic review,” BMC
K. Nelson, “The relationship between food insecurity and Public Health, vol. 18, no. 1, p. 1148, 2018.
depression, diabetes distress and medication adherence among [26] A. D.-G. Aikins, “Ghana's neglected chronic disease epidemic:
low-income patients with poorly-controlled diabetes,” Journal a developmental challenge,” Ghana Medical Journal, vol. 41,
of General Internal Medicine, vol. 30, no. 10, pp. 1476–1480, no. 4, pp. 154–159, 2007.
2015. [27] W. Bosu, “A comprehensive review of the policy and program-
[11] J. Aikens and J. Piette, “Longitudinal association between med- matic response to chronic non-communicable disease in
ication adherence and glycaemic control in type 2 diabetes,” Ghana,” Ghana Medical Journal, vol. 46, no. 2, pp. 69–78,
Diabetic Medicine, vol. 30, no. 3, pp. 338–344, 2013. 2012.
[12] K. M. P. Van Bastelaar, F. Pouwer, P. H. L. M. Geelhoed-Duij- [28] A. D.-G. Aikins, “Living with diabetes in rural and urban
vestijn et al., “Diabetes-specific emotional distress mediates the Ghana: a critical social psychological examination of illness
association between depressive symptoms and glycaemic con- action and scope for intervention,” Journal of Health Psychol-
trol in type 1 and type 2 diabetes,” Diabetic Medicine, vol. 27, ogy, vol. 8, no. 5, pp. 557–572, 2003.
no. 7, pp. 798–803, 2010. [29] A. D.-G. Aikins, “Strengthening quality and continuity of dia-
[13] L. Fisher, J. T. Mullan, P. Arean, R. E. Glasgow, D. Hessler, and betes care in rural Ghana: a critical social psychological
U. Masharani, “Diabetes distress but not clinical depression or approach,” Journal of Health Psychology, vol. 9, no. 2,
depressive symptoms is associated with glycemic control in pp. 295–309, 2004.
both cross-sectional and longitudinal analyses,” Diabetes Care, [30] A. G. Amoah, S. K. Owusu, and S. Adjei, “Diabetes in Ghana: a
vol. 33, no. 1, pp. 23–28, 2010. community based prevalence study in Greater Accra,” Diabe-
[14] L. Fisher, J. Gonzalez, and W. Polonsky, “The confusing tale of tes Research and Clinical Practice, vol. 56, no. 3, pp. 197–205,
depression and distress in patients with diabetes: a call for 2002.
greater clarity and precision,” Diabetic Medicine, vol. 31, [31] M. Cook-Huynh, D. Ansong, R. C. Steckelberg et al., “Preva-
no. 7, pp. 764–772, 2014. lence of hypertension and diabetes mellitus in adults from a
[15] L. Fisher, W. H. Polonsky, D. M. Hessler et al., “Understanding rural community in Ghana,” Ethnicity & Disease, vol. 22,
the sources of diabetes distress in adults with type 1 diabetes,” no. 3, pp. 347–352, 2012.
Journal of Diabetes and its Complications, vol. 29, no. 4, [32] I. Danquah, G. Bedu-Addo, K. J. Terpe et al., “Diabetes melli-
pp. 572–577, 2015. tus type 2 in urban Ghana: characteristics and associated fac-
[16] L. M. Delahanty, R. W. Grant, E. Wittenberg et al., “Associa- tors,” BMC Public Health, vol. 12, no. 1, 2012.
tion of diabetes-related emotional distress with diabetes treat- [33] L. K. Frank, J. Kröger, M. B. Schulze, G. Bedu-Addo, F. P.
ment in primary care patients with type 2 diabetes,” Diabetic Mockenhaupt, and I. Danquah, “Dietary patterns in urban
Medicine, vol. 24, no. 1, pp. 48–54, 2007. Ghana and risk of type 2 diabetes,” British Journal of Nutrition,
[17] R. N. Baek, M. L. Tanenbaum, and J. S. Gonzalez, “Diabetes vol. 112, no. 1, pp. 89–98, 2014.
burden and diabetes distress: the buffering effect of social sup- [34] S. M. Gatimu, B. W. Milimo, and M. S. Sebastian, “Prevalence
port,” Annals of Behavioral Medicine, vol. 48, no. 2, pp. 145– and determinants of diabetes among older adults in Ghana,”
155, 2014. BMC Public Health, vol. 16, no. 1, p. 1174, 2016.
[18] B. Karlsen, B. Oftedal, and E. Bru, “The relationship between [35] S. P. Bruce, F. Acheampong, and I. Kretchy, “Adherence to
clinical indicators, coping styles, perceived support and oral anti-diabetic drugs among patients attending a Ghanaian
diabetes-related distress among adults with type 2 diabetes,” teaching hospital,” Pharmacy Practice, vol. 13, no. 1, p. 533,
Journal of Advanced Nursing, vol. 68, no. 2, pp. 391–401, 2012. 2015.
[19] J. Schabert, J. L. Browne, K. Mosely, and J. Speight, “Social [36] V. Mogre, Z. O. Abanga, F. Tzelepis, N. A. Johnson, and
stigma in diabetes : a framework to understand a growing C. Paul, “Adherence to and factors associated with self-care
10 Journal of Diabetes Research

behaviours in type 2 diabetes patients in Ghana,” BMC Endo- mellitus,” International Journal of Environmental Research
crine Disorders, vol. 17, no. 1, p. 20, 2017. and Public Health, vol. 14, no. 12, p. 1522, 2017.
[37] G. W. Welch, A. M. Jacobson, and W. H. Polonsky, “The [53] T. B. Gibson, X. Song, B. Alemayehu et al., “Cost sharing,
Problem Areas in Diabetes Scale: an evaluation of its clinical adherence, and health outcomes in patients with diabetes,”
utility,” Diabetes Care, vol. 20, no. 5, pp. 760–766, 1997. The American Journal of Managed Care, vol. 16, no. 8,
[38] R. Horne and J. Weinman, “Patients' beliefs about prescribed pp. 589–600, 2010.
medicines and their role in adherence to treatment in chronic [54] F. S. Marinho, C. B. M. Moram, P. C. Rodrigues, N. C. Leite,
physical illness,” Journal of Psychosomatic Research, vol. 47, G. F. Salles, and C. R. L. Cardoso, “Treatment adherence and
no. 6, pp. 555–567, 1999. its associated factors in patients with type 2 diabetes: results
[39] C. S. Lee, J. H. M. Tan, U. Sankari, Y. L. E. Koh, and N. C. Tan, from the Rio de Janeiro type 2 diabetes cohort study,” Journal
“Assessing oral medication adherence among patients with of Diabetes Research, vol. 2018, Article ID 8970196, 8 pages,
type 2 diabetes mellitus treated with polytherapy in a 2018.
developed Asian community: a cross-sectional study,” BMJ [55] N. Kumar, B. Unnikrishnan, R. Thapar et al., “Distress and its
Open, vol. 7, no. 9, 2017. effect on adherence to antidiabetic medications among type 2
[40] W. H. Polonsky, B. J. Anderson, P. A. Lohrer et al., “Assess- diabetes patients in Coastal South India,” Journal of Natural
ment of diabetes-related distress,” Diabetes Care, vol. 18, Science, Biology, and Medicine, vol. 8, no. 2, p. 216, 2017.
no. 6, pp. 754–760, 1995. [56] L. A. Nelson, K. A. Wallston, S. Kripalani, L. M. LeStourgeon,
[41] S. S. Jaser, N. Patel, R. L. Rothman, L. Choi, and S. E. Williamson, and L. S. Mayberry, “Assessing barriers
R. Whittemore, “Check it! A randomized pilot of a positive to diabetes medication adherence using the information-
psychology intervention to improve adherence in adolescents motivation-behavioral skills model,” Diabetes Research and
with type 1 diabetes,” The Diabetes Educator, vol. 40, no. 5, Clinical Practice, vol. 142, pp. 374–384, 2018.
pp. 659–667, 2014.
[42] W. H. Polonsky, L. Fisher, J. Earles et al., “Assessing psychoso-
cial distress in diabetes: development of the diabetes distress
scale,” Diabetes Care, vol. 28, no. 3, pp. 626–631, 2005.
[43] A. B. Grigsby, R. J. Anderson, K. E. Freedland, R. E. Clouse,
and P. J. Lustman, “Prevalence of anxiety in adults with diabe-
tes: a systematic review,” Journal of Psychosomatic Research,
vol. 53, no. 6, pp. 1053–1060, 2002.
[44] R. J. Anderson, K. E. Freedland, R. E. Clouse, and P. J.
Lustman, “The prevalence of comorbid depression in adults
with diabetes: a meta-analysis,” Diabetes Care, vol. 24, no. 6,
pp. 1069–1078, 2001.
[45] G. E. Simon and M. Von Korff, “Medical co-morbidity and
validity of DSM-IV depression criteria,” Psychological Medi-
cine, vol. 36, no. 1, pp. 27–36, 2006.
[46] M. M. Skaff, J. T. Mullan, D. M. Almeida et al., “Daily negative
mood affects fasting glucose in type 2 diabetes,” Health Psy-
chology, vol. 28, no. 3, pp. 265–272, 2009.
[47] T. W. Smith, “Personality as risk and resilience in physical
health,” Current Directions in Psychological Science, vol. 15,
no. 5, pp. 227–231, 2006.
[48] S. M. Robertson, M. A. Stanley, J. A. Cully, and A. D. Naik,
“Positive emotional health and diabetes care: concepts, mea-
surement, and clinical implications,” Psychosomatics, vol. 53,
no. 1, pp. 1–12, 2012.
[49] S. Mohebi, M. Parham, G. Sharifirad, Z. Gharlipour,
A. Mohammadbeigi, and F. Rajati, “Relationship between per-
ceived social support and self-care behavior in type 2 diabetics:
a cross-sectional study,” Journal of Education and Health Pro-
motion, vol. 7, no. 1, p. 48, 2018.
[50] G. Spencer-Bonilla, O. J. Ponce, R. Rodriguez-Gutierrez et al.,
“A systematic review and meta-analysis of trials of social net-
work interventions in type 2 diabetes,” BMJ Open, vol. 7,
no. 8, 2017.
[51] J. L. Strom and L. E. Egede, “The impact of social support on
outcomes in adult patients with type 2 diabetes: a systematic
review,” Current Diabetes Reports, vol. 12, no. 6, pp. 769–
781, 2012.
[52] L. Gu, S. Wu, S. Zhao et al., “Association of social support and
medication adherence in Chinese patients with type 2 diabetes