Medication Adherence and Contributing Factors Among Type 2 Diabetes Patients at Adama Hospital Medical College in Eastern Ethiopia

Machine learning applications to classify and monitor medication adherence in patients with type 2 diabetes in Ethiopia

Background

Medication adherence plays a crucial role in determining the health outcomes of patients, particularly those with chronic conditions like type 2 diabetes. Despite its significance, there is limited evidence regarding the use of machine learning (ML) algorithms to predict medication adherence within the Ethiopian population. The primary objective of this study was to develop and evaluate ML models designed to classify and monitor medication adherence levels among patients with type 2 diabetes in Ethiopia, to improve patient care and health outcomes.

Methods

Using a random sampling technique in a cross-sectional study, we obtained data from 403 patients with type 2 diabetes at the University of Gondar Comprehensive Specialized Hospital (UoGCSH), excluding 13 subjects who were unable to respond and 6 with incomplete data from an initial cohort of 422. Medication adherence was assessed using the General Medication Adherence Scale (GMAS), an eleven-item Likert scale questionnaire. The responses served as features to train and test machine learning (ML) models. To address data imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) was applied. The dataset was split using stratified K-fold cross-validation to preserve the distribution of adherence levels. Eight widely used ML algorithms were employed to develop the models, and their performance was evaluated using metrics such as accuracy, precision, recall, and F1 score. The best-performing model was subsequently deployed for further analysis.

Results

Out of 422 enrolled patients, 403 data samples were collected, with 11 features extracted from each respondent. To mitigate potential class imbalance, the dataset was increased to 620 samples using the Synthetic Minority Over-sampling Technique (SMOTE). Machine learning models including Logistic Regression (LR), Support Vector Machine (SVM), K Nearest Neighbor (KNN), Decision Tree (DT), Random Forest (RF), Gradient Boost Classifier (GBC), Multilayer Perceptron (MLP), and 1D Convolutional Neural Network (1DCNN) were developed and evaluated. Although the performance differences among the models were subtle (within a range of 0.001), the SVM classifier outperformed the others, achieving a recall of 0.9979 and an AUC of 0.9998. Consequently, the SVM model was selected for deployment to monitor and detect patients' medication adherence levels, enabling timely interventions to improve patient outcomes.

Conclusions

This study highlights a variety of machine learning (ML) models that can be effectively used to monitor and classify medication adherence in diabetic patients in Ethiopia. However, to fully realize the potential impact of digital health applications, further studies that include patients from diverse settings are necessary. Such research could enhance the generalizability of these models and provide insights into the broader applicability of digital tools for improving medication adherence and patient outcomes in varying healthcare contexts.

Predictors of treatment adherence in patients with type 2 diabetes: a cross-sectional study in Southern Iran based on Pender's Health Promotion Model using structural equation modelling

OBJECTIVES

Treatment adherence in type 2 diabetes (T2D) is an important factor in optimal diabetes control and prevention of mortality. The present study aimed to determine the predictability of Pender's Health Promotion Model (HPM) constructs in T2D treatment adherence behaviour.

DESIGN

The present cross-sectional and analytical study was conducted from November 2022 to January 2023.

SETTING

The present study was conducted in Bandar Abbas, a city in Hormozgan Province, in the south of Iran.

PARTICIPANTS

The participants included 396 patients with T2D with medical records in the Hormoz Diabetes Clinic. Based on their record number, the participants were selected for inclusion in the study through a random systematic sampling.

PRIMARY AND SECONDARY OUTCOME MEASURES

The data collection instruments included a demographic questionnaire and a researcher-made questionnaire based on HPM constructs. The questionnaire was valid and reliable, achieving Cronbach's alpha coefficients ranging from 0.609 to 0.798 across various constructs. The questionnaires were completed face to face. Pearson's correlation test, path analysis and structural equation modelling were conducted using SPSS V.23, and STATA V.15.

STUDY STAGE

This study was conducted before intervention (pre-results).

RESULTS

As the path analysis showed, perceived self-efficacy (β=0.23, p<0.001), treatment adherence experiences (β=0.26, p<0.001), immediate competing demands and preferences (β=-0.15, p<0.001) and commitment to plan of action (β=0.24, p<0.001) could significantly predict the treatment adherence behaviour. The results of indirect path analysis showed that the total effect of perceived benefits (β=0.24, p<0.001), perceived barriers (β=-0.14, p=0.002), perceived self-efficacy (β=0.32, p<0.001) on commitment to plan of action was statistically significant. Through the mediation of commitment to plan of action, they could predict the treatment adherence behaviour.

CONCLUSIONS

In light of the present findings, it can be concluded that the proposed model of T2D treatment adherence behaviour has an acceptable fit. Commitment to plan of action, treatment adherence experiences, perceived self-efficacy and immediate competing demands and preferences are the main predictors of T2D treatment adherence behaviour. It is recommended that educational interventions focus on these constructs.

TRIAL REGISTRATION NUMBER

This study is registered on the Iranian Registry of Clinical Trials (IRCT20211228053558N1).

Adherence to recommended diet among patients with diabetes mellitus type 2 on follow-up at Adama Hospital Medical College, Ethiopia

Introduction

Non-adherence to dietary guidelines is a significant challenge in managing diabetes mellitus and its complications. Its consequences were significantly associated with a deterioration in patients' quality of life and an increased socioeconomic burden on healthcare delivery systems. This study aimed to assess the magnitude of adherence to recommended diet and associated factors among patients with diabetes mellitus type 2 on follow-up care at Adama Hospital Medical College Oromia, Ethiopia.

Methods

A hospital-based cross-sectional study design was conducted. Participants were selected through systematic random sampling. Data were collected using structured, interviewer-administered questionnaires. The perceived dietary adherence questionnaire was used to assess the level of dietary adherence. A simple binary logistic regression was used to identify candidate variables, while a multivariable logistic regression assessed factors associated with adherence to the recommended diet. A p-value <0.05 were considered as statistically significant. All analyses were performed using SPSS and R programming software.

Result

A total of 405 participants were included in the study, with a response rate of 96.2%. The magnitude of non-adherence to the recommended diet was 64.2% (95% confidence interval [CI]: 59.8, 68.6). In the multivariable logistic regression model, patients with low and middle income (AOR = 8.0; 95% CI: 3.4, 19.2) and (AOR = 2.75; 95% CI: 1.49, 5.55) respectively, high glycemic level (AOR = 2.15; 95% CI: 1.17, 3.94), food insecure (AOR = 12.7; 95% CI: 5.79, 28.2), poor diabetic knowledge (AOR = 2.88; 95% CI: 1.49, 5.55) and low perceived susceptibility (AOR = 2.97; 95% CI: 1.62, 5.45) were significantly associated factors for non-adherence to recommended diet among patients with diabetes mellitus type 2.

Conclusion

This study revealed that approximately two-thirds of patients with type 2 diabetes mellitus experienced non-adherence to the recommended diet. Key factors linked to dietary non-adherence among T2DM patients include low to middle income, elevated glycemic levels, household food insecurity, limited diabetes knowledge, and low perceived susceptibility. An integrated approach that combines socioeconomic support, nutritional guidance, and risk awareness may greatly enhance dietary adherence and optimize diabetes management.

Association between self-care activities and glycemic control among patients with type 2 diabetes mellitus in Northwest Ethiopia general hospitals : a multicenter cross-sectional study

Diabetes self-care activities are essential for achieving optimal glycemic control. However, little investigation has been conducted in Ethiopia to evaluate the relationship between the rate glycemic controland self-care activities among patients with type 2 diabetes mellitus (T2DM). Therefore, this study was conducted to assess self -care activities and their association with glycemic control among patients with T2DM in Northwest Ethiopia general hospitals. This multicenter cross-sectional study was conducted in Northwest Ethiopia general hospitals diabetic clinics. Diabetes self-care activities were measured using the Amharic version of the Summary of Diabetes Self-Care Activities (SDSCA-Amharic). Glycated hemoglobin (HbA1c) were used to assess the rate of glycemic control. A linear regression model was used to identify predictors of self-care activities and glycemic control. P-value of <  0.05 at 95% confidence interval (CI)  was considerd as statistically significant. Of 413 participants included in the final analysis, two-thirds (66.3%) had poor glycemic control, with a mean HbA1c of 7.94% (SD = 1.75). Blood glucose testing was the most important self-care activity domain for predicting better glycemic control [β=-0.36, 95% CI (-0.48, -0.24); P = 0.0001] followed by diet [β=-0.29, 95% CI (-0.39, -0.083); P = 0.0001], foot-care [β=-0.28, 95% CI (-0.3, -0.061); P = 0.003], and physical activity [β=-0.27, 95% CI (-0.29, -0.056); P = 0.004], respectively. Moreover, unable to read and write [β = 0.72, 95% CI (0.57, 3.8); P = 0.037], overweight [β = 0.32, 95% CI (0.011, 0.62); P = 0.042], obesity [β = 0.67, 95% CI (0.39, 0.94); P = 0.0001], and low level of medication adherence [β = 0.7, 95% CI (0.39, 1.1); P = 0.0001] were significant predictors of poor glycemic control.       Previous diabetes education [β=-0.88, 95% CI (-1.2, -0.57); P=0.0001] was a significant predictor of good glycemic control. The prevalence of poor glycemic control and poor self-care activities were high among patients with T2DM. Self-care activities were independent predictors of glycemic control among patients with T2DM. Therefore, management interventions for patients with T2DM should focus on improving self-care activities and other predictor variables.

Measuring Diabetic Medication Adherence and Factors That Lead to Non-adherence Among Patients in Erbil

Introduction T2D is a chronic and progressive disorder characterized by persistent hyperglycemia resulting from inadequate insulin secretion or utilization. The global prevalence of T2D is increasing rapidly, posing a significant health burden in many regions. In the Kurdistan region of Iraq, T2D presents a significant health burden, exacerbated by socioeconomic changes, dietary shifts, and rising obesity rates. Poor adherence to antidiabetic medications is a major factor contributing to poor glycemic control, accelerating disease progression, and increasing complications. This study aims to assess medication adherence rates among adult T2D patients in Erbil using the Kurdish version of the Morisky Medication Adherence Scale-8 (MMAS-8) and identify factors associated with non-adherence. Methods We conducted a cross-sectional study at public and private clinics in Erbil City, Kurdistan, Iraq, between May 1 and September 30, 2023. A convenience sample of 300 adult Kurdish T2D patients, aged ≥ 25 years and on antidiabetic medications for three months or more, was recruited. Data were collected using a structured questionnaire comprising sociodemographic characteristics, clinical and anthropometric measures, and medication adherence assessed by the Kurdish version of the MMAS-8. Statistical analysis included analysis of variance, Kruskal-Wallis, chi-square, and logistic regression models to identify factors associated with medication adherence. Results Of the 300 participants, 81 (27%) demonstrated high adherence, 98 (32.6%) moderate adherence, and 121 (40.3%) low adherence based on the MMAS-8. Low adherence was significantly associated with lower education (56/121, 46.3% vs. 13/81, 16.0%, p < 0.001), unemployment (73/121, 60.3% vs. 29/81, 35.8%, p = 0.008), rural residence (41/121, 33.9% vs. 10/81, 12.3%, p < 0.001), and lower income (62/121, 51.2% vs. 12/81, 14.8%, p < 0.001). High adherence was linked to better diabetes knowledge, home glucose monitoring, and exercise. High adherence was also associated with better glycemic control, with 76/81 (93.8%) of highly adherent patients achieving glycated hemoglobin (HbA1c) <7%, compared to 15/121 (12.4%) in the low adherence group (p < 0.001). Multivariate analysis identified HbA1c, dyslipidemia, and home blood glucose monitoring as independent factors associated with high adherence. Conclusions This study highlights the substantial impact of socioeconomic, behavioral, and clinical factors on medication adherence among T2D patients in Erbil. Low adherence is associated with lower education, income, and awareness of diabetes management, while high adherence is linked to improved glycemic control and reduced complications. Targeted interventions addressing these factors are essential to enhance adherence and optimize T2D management in this population.

A health promotion model-based intervention to enhance treatment adherence in patients with type 2 diabetes

BACKGROUND

The present study aimed to determine the effect of an intervention based on Pender's health promotion model (HPM) on treatment adherence in patients with type 2 diabetes (T2D).

METHODS

The present quasi-experimental study with a 3-month follow-up was conducted in Bandar Abbas, a city in the south of Iran in 2023. The intervention group (IG) with a total number of 95 T2D patients was selected from Hormuz diabetes clinic and the control group (CG) with 95 T2D patients was selected from comprehensive health centers through a clustering sampling method. The educational intervention was implemented in 10 sessions to improve patients' treatment adherence. The teaching methods in training sessions were lectures, joint discussions, Q&A, role-play and peer training. The participants were evaluated using a researcher-made questionnaire including the constructs of Pender's HPM about T2D treatment adherence, hemoglobin A1C (HbA1C), and BMI. Independent-samples t-test, paired-samples t-test, covariance analysis and stepwise regression analysis were used. Data analysis was done in SPSS 26.

FINDINGS

Three months after the intervention, in comparison to the CG, the mean and standard deviation of treatment adherence benefits (p = 0.002), treatment adherence self-efficacy (p = 0.010), treatment adherence related affect (p = 0.001), interpersonal influences (p = 0.012), commitment to plan of action (p < 0.001), treatment adherence behavior (p = 0.022), treatment adherence experiences (p = 0.001) was higher in the IG. The mean and standard deviation of situational influences (p < 0.001), immediate competing demands and preferences (p = 0.018) were lower than the CG. The results obtained from the analysis of covariance proved the effectiveness of the intervention in the constructs of Pender's HPM and HbA1C in participants of the IG (p < 0.001). The regression analysis showed, after the intervention, for every 1 unit of change in commitment to behavior planning, action related affect and perceived self-efficacy, compared to before the intervention, there were 0.22 units, 0.16 units and 0.26 units of change in the behavior score in the IG.

CONCLUSION

The findings proved the effectiveness of the educational intervention in improving the constructs in Pender's HPM and the blood sugar level of T2D patients. As the results of the educational intervention showed, the use of a suitable educational approach as well as the development of appropriate educational content for the target population can significantly improve the treatment adherence behavior.

TRIAL REGISTRATION

This study is registered on the Iranian Registry of Clinical Trials (IRCT20211228053558N1: https://www.irct.ir/trial/61741 ) and first release date of 17th March 2022.

Medication regimen complexity and its impact on medication adherence in patients with multimorbidity at a comprehensive specialized hospital in Ethiopia

Background

Medication regimen complexity (MRC) is suspected to hinder medication adherence in patients with multiple illnesses. Despite this, the specific impact on Ethiopian patients with multimorbidity is unclear. This study assessed MRC and its impact on medication adherence in patients with multimorbidity.

Methods

A hospital-based cross-sectional study was conducted on patients with multimorbidity who had been followed at the University of Gondar Comprehensive and Specialized Hospital (UoGCSH), Ethiopia, from May to July 2021. Medication complexity was measured using the validated Medication Regimen Complexity Index (MRCI) tool, and the Adherence in Chronic Diseases Scale (ACDS) was used to measure medication adherence. Pearson's chi-square test was used to examine associations between MRCI levels and medication adherence. Ordinal logistic regression analysis was used to determine the impact of MRC and other associated variables on medication adherence. Statistical significance was determined using the adjusted odds ratio (AOR) at p-value <0.05 and its 95% confidence range.

Results

Out of 422 eligible patients, 416 (98.6%) were included in the study. The majority of participants (57.2%) were classified as having a high MRCI score with a mean (±SD) score of 9.7 (±3.4). Nearly half of the patients (49.3%) had low medication adherence. Patients with medium (AOR = 0.43, 95% CI: 0.04, 0.72) and higher (AOR = 0.31, 95% CI: 0.07, 0.79) MRCI levels had lower odds of medication adherence. In addition, monthly income (AOR = 4.59, 95% CI: 2.14, 9.83), follow-up durations (AOR = 2.31, 95% CI: 1.09, 4.86), number of medications (AOR = 0.63, 95% CI: 0.41, 0.97), and Charlson comorbidity index (CCI) (AOR = 0.36, 95% CI: 0.16, 0.83) were significantly associated with medication adherence.

Conclusion

Medication regimen complexity in patients with multimorbidity was found to be high and negatively impacted the levels of medication adherence. Healthcare providers and other stakeholders should seek interventions aimed at simplifying drug regimen complexity and improving adherence.

A Review on Affordable Combinations in Type 2 Diabetes Care: Exploring the Cost-Effective Potential of Glipizide + Metformin and Glimepiride + Metformin + Pioglitazone

Management of type 2 diabetes mellitus (T2DM) largely relies on medication adherence of individuals with diabetes to achieve optimal glycemic control. The economic burden of diabetes could impede adherence, leading to a reduction in treatment efficacy and increased risk of complications. Furthermore, monotherapy in diabetes is losing traction due to its ineffectiveness in achieving early and sustained optimal glycemic control in a significant proportion of the population. Hence, clinicians prefer combination treatment due to their improved efficacy and safety. Considering these factors, the current review highlights the safety and efficacy of the affordable combination therapies, a dual therapy, glipizide + metformin, and a triple-drug combination of glimepiride + metformin + pioglitazone and its applicability in the management of T2DM among individuals with diabetes in India.

Low medication adherence and its associated factors among patients with type 2 diabetes mellitus attending Amana Hospital in Dar es Salaam, Tanzania: a cross-sectional study

BACKGROUND

Low medication adherence among patients with type 2 diabetes mellitus (T2DM) is associated with significant morbidity and mortality globally. We investigated the prevalence of low medication adherence and its associated factors among patients with T2DM.

METHODS

We used the Bengali version of the 8-item Morisky Medication Adherence Scale (MMAS-8) in measuring medication adherence among patients with T2DM who were attending the diabetes clinic at Amana Regional Referral Hospital in Dar es Salaam, Tanzania, from December 2021 to May 2022. Binary logistic regression analysis under multivariate analysis was used to determine the predictors of low medication adherence after controlling for confounders. A two-tailed p-value <0.05 was considered significant.

RESULTS

The prevalence of low medication adherence was 36.7% (91/248) of the subjects included in the study. Lack of formal education (adjusted odds ratio [AOR] 5.3 [95% confidence interval {CI} 1.717 to 16.312], p=0.004), having comorbidities (AOR 2.1 [95% CI 1.134 to 3.949], p=0.019) and drinking alcohol (AOR 3.5 [95% CI 1.603 to 7.650], p=0.031) were the independent predictors of low medication adherence.

CONCLUSION

More than one-third of the patients with T2DM in this study had low medication adherence. Our study also showed that a lack of formal education, having comorbidities and drinking alcohol were significantly associated with low medication adherence.

Assessing the Adherence to Antidiabetic Medications Among Patients Diagnosed With Type 2 Diabetes Mellitus in Ajman, UAE

Background Medication adherence plays a vital role in managing blood sugar levels and preventing complications in individuals with diabetes. Patient adherence to antidiabetic medications and the factors associated with medication adherence were assessed. Objectives To assess the medication adherence among patients suffering from type 2 diabetes mellitus. To determine the various factors influencing medication adherence. Methods This cross-sectional study was conducted on patients with type 2 diabetes who were visiting the Internal Medicine Department of Thumbay University Hospital in the United Arab Emirates. A questionnaire was used to gather information about the medication adherence of a group of chosen consecutive patients. IBM SPSS Statistics for Windows, Version 27 (Released 2020; IBM Corp., Armonk, New York) was used for data analysis. A two-sided P-value <0.05 was regarded as significant when using the chi-square test to investigate the relationships between categorical variables. Results A total of 204 patients participated in the study: 112 (54.90%) males and 92 (45.09%) females. The mean age of the patients was 49 years. The adherence rates among males and females were 91% and 90%, respectively. Some of the common reasons for non-adherence to antidiabetic medications in our study included forgetfulness, unpleasant side effects, the use of multiple drugs, and long treatment duration. Conclusion Our study highlighted important factors associated with patients' non-adherence to their antidiabetic medications. Future research on methods to increase adherence rates should be taken into consideration.

Medication Adherence in Elderly Diabetic Patients: A Cross-Sectional Study From Dakshina Kannada, India

Diabetes Mellitus (DM) has emerged as a major global healthcare problem. The risk of diabetes can be reduced by maintaining blood glycaemic levels, which can be achieved by stringent adherence to the treatment regime. Therefore, there is a continuing need to assess the level of adherence to medication/self-care activities and the factors that are related to non-adherence to medication and self-care. This would facilitate healthcare professionals to identify subjects with low medication adherence and thereby aid them in planning interventions to improve medication and self-care adherence. In this study, we aimed to estimate the proportion of medication adherence among diabetic patients above 60 years of age attending a tertiary care hospital in Southern India. We found that 72% of type 2 diabetes patients were adherent to the medications prescribed to them and there was a discernible effect of gender and literacy on medication adherence. However, more such regional studies need to be conducted with a larger sample size from diverse hospital setups to obtain a clear and unbiased picture of the drug adherence scenario in India.