Medication Adherence Interventions Among People Living with Diabetes: A Systematic Review

Introduction

Diabetes mellitus (DM) is a chronic metabolic disease characterized by persistent hyperglycemia and has become a major global public health concern.¹ It is estimated that in 2021, approximately 536.6 million individuals aged 20–79 years worldwide were living with diabetes and this number is projected to rise to 783.2 million by 2045.² Without effective management, chronic hyperglycemia can lead to serious complications such as kidney disease,³ retinopathy⁴ and cardiovascular diseases,⁵ significantly compromising patients’ quality of life and life expectancy. Despite the availability of various effective pharmacological treatments, poor medication adherence remains a major barrier to achieving optimal glycemic control. Studies have reported that the prevalence of poor medication adherence among people living with diabetes ranges from 70% to 80%.^6,7 Suboptimal adherence not only increases the risk of disease progression and the development of complications but also contributes to higher hospitalization rates and healthcare costs.^8–10

In recent years, a variety of intervention strategies have been developed to improve medication adherence in people living with diabetes, including health education programs,¹¹ blood glucose monitoring systems,¹² and digital health technologies.¹³ However, existing systematic reviews have primarily focused on identifying influencing factors and evaluating measurement tools for medication adherence,^14,15 rather than providing a comprehensive analysis of interventions themselves. Some reviews have categorized interventions based on behavior change techniques,¹⁶ yet they have not systematically summarized the specific characteristics and implementation methods of each intervention. Other reviews have narrowly focused on particular intervention types, such as pharmacist-led programs or smartphone applications,^17,18 thereby limiting the generalizability of their findings. Additionally, many available reviews are based on studies conducted more than a decade ago,^19,20 which may not adequately reflect recent advancements in medical technologies and healthcare delivery.

In light of these gaps, the present systematic review aims to comprehensively include studies published in the past decade, covering all types of DM patients. By systematically reviewing and synthesizing the latest evidence on medication adherence interventions, this study seeks to provide an updated, detailed and holistic overview of intervention strategies, implementation approaches and their effectiveness. The findings are expected to fill current gaps in the literature and offer practical insights for the development of more targeted and effective interventions in clinical practice, as well as guide future research directions.

Methods

This review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement²¹ and was prospectively registered in PROSPERO (International Prospective Register of Systematic Reviews, registration number CRD42024545467). The review includes studies published between 1^st January 2015 and 1^st April 2025. A total of five electronic databases were searched for relevant studies: PubMed, Scopus, Embase, Cochrane Library, and ProQuest.

Ethical Considerations

Since this study is a systematic review utilizing publicly available, non-identifiable secondary data from published studies, ethical approval was not necessary in accordance with institutional guidelines.

Search Strategy and Selection Criteria

The keywords used were “(diabetes mellitus OR diabetes OR diabetic OR type 1 DM OR type 2 DM OR T1DM OR T2DM OR glucose intolerance OR glucose metabolism disorders) AND (medication adherence OR medication compliance OR drug compliance OR drug adherence OR medication nonadherence OR nonadherence, medication OR noncompliance, medication OR medication non-adherence OR medication persistence) AND (intervention OR impact OR effect OR effectiveness)”. The inclusion criteria for studies are as follows: (1) study population must consist of individuals diagnosed with DM, (2) intervention must focus on medication adherence strategies, (3) study must be published in a peer-reviewed journal with an English language version, (4) study design must be a Randomized Controlled Trial (RCT) and (5) published between 2015 and 2025. The exclusion criteria are: (1) journals that are not published in peer-reviewed sources, such as conference proceedings, conference books, government reports, grey literature or short communications, (2) review journals or meta-analyses and (3) studies that are not available in full text.

Selection of Article (Screening and Eligibility)

This review excluded unpublished, duplicate, non-English, and irrelevant articles. To ensure objectivity, two researchers (MW and YC) independently screened titles, abstracts, and full texts based on the inclusion and exclusion criteria, cross-checking the extracted data. Disagreements were resolved by a third researcher (PYL). The screening process consisted of two phases: the first involved title and abstract review, followed by full-text assessment in the second phase to confirm relevance and adherence to the criteria.

Data Extraction

The data were extracted using a pre-designed form encompassing the following: (1) first author, publication date, (2) study population and sample size, (3) country, (4) study design, (5) measurement instrument and interventionist, (6) intervention on duration, delivery mode and time points, (7) intervention effect and statistical significance. Due to the high heterogeneity in study designs, interventions, measurement instruments and outcomes, a meta-analysis was not performed.

Risk of Bias Assessment

The quality of the Randomized Controlled Trials (RCTs) included in this review was assessed using the Cochrane Risk of Bias Tool 2 (RoB 2).²² This tool evaluated bias across five domains, ensuring that only studies with a low risk of bias contributed to the evidence on medication adherence interventions for people living with diabetes. Two researchers independently assessed the studies and discrepancies were resolved through discussion among all authors.

Results

Literature Search

Out of 11,711 potential articles, 3,923 duplicate records were removed using Endnote 21 software. Of the remaining 7,788 articles, 7,599 were excluded based on their titles, followed by 189 exclusions based on abstracts and keywords. Ultimately, 38 studies met all eligibility criteria for inclusion in this review. The selection process is illustrated in Figure 1.

Figure 1 Flowchart of the systematic review selection process.

Quality Assessment

Based on the RoB2 standard, the risk of bias for all 38 included RCTs was assessed. The overall risk of bias results showed that 20 studies as having some concerns and 18 studies as low risk of bias. Supplementary Figure 1 and Supplementary Figure 2 present the detailed evaluation of the included studies, covering aspects such as randomization, interventions, missing outcome data, outcome measurement and reporting of results.

Study Characteristics and Diversity

The 38 studies included in this review originated from various countries, with the highest number of studies coming from India, the United States and Iran. Among the 38 studies, 34 (89%) focused on T2DM (Type 2 Diabetes Mellitus) patients, 32 studies (84%) had participants aged ≥18 years, while 6 studies (16%) did not specify the age range. The largest study had a sample size of 63,012 participants, while the smallest included only 41 participants. The total sample size across all studies was 70,457 (excluding one community-based study that did not provide specific numbers). The duration of the studies varied widely, ranging from two weeks to 15 months. The measurement tools used across the included studies were primarily categorized into two forms: self-reported questionnaires and pill counts. The most commonly used adherence assessment instruments were the Morisky Medication Adherence Scale (MMAS), utilized in 14 studies (37%), and the Medication Adherence Report Scale (MARS), used in four studies (11%). Regarding intervention providers, 15 studies (40%) were pharmacist-led, while 19 studies (50%) did not specify the identity of the interventionist and were conducted by research staff. Additionally, three studies (8%) involved nurse-led interventions, and one study (3%) was led by a medical advisor or therapist. Further details are provided in Table 1.

Table 1 Summary of the Selected Studies (N=38 Articles)

Intervention Strategies and Theoretical Models for Medication Adherence in People Living with Diabetes

The 38 studies included in this review utilized a variety of interventions to improve medication adherence among people living with diabetes. 14 studies (37%) employed face-to-face interactions, including oral communication, meetings, educational courses and interviews. 11 studies (29%) utilized multiple combinations of methods. Additionally, five studies (13%) used text messages, three studies (8%) utilized internet-based interventions, two studies (5%) employed telephone interventions, two studies (5%) used mobile applications (apps) and one study (3%) involved fliers. Only eight studies (21%) incorporated theoretical models into their design. The most commonly used models were the theory of planned behavior, self-determination theory, information-motivation-behavioral skills model and the health belief model, with some studies combining multiple models. It is noteworthy that five studies (13%) showed no significant improvement in medication adherence (P > 0.05). Among these, one study used two different questionnaires to measure medication adherence: one questionnaire indicated a positive effect, while the other showed no effect. The remaining studies demonstrated a positive impact of the interventions on medication adherence among people living with diabetes. The detailed results are presented in Tables 2 and 3.

Table 2 Intervention Associated with Medication Adherence Among DM Patients for Selected Studies (n=38)

Table 3 Type of Theory Models and Delivery Mode Used in the Studies (n = 38 Articles)

Discussion

This systematic review shows that over 90% of studies on medication adherence interventions focused on patients diagnosed with T2DM, likely because T2DM accounts for the largest proportion of all DM types, reaching 90–95% of cases.^58,59 The majority of studies were conducted in developing countries, where medication adherence is generally lower, with over 50% of patients exhibiting poor adherence.^60,61 Developing countries frequently encounter substantial constraints in healthcare resources and services, underscoring the importance of implementing cost-effective and efficient strategies, such as enhancing medication adherence. Targeted interventions can enhance adherence, thereby reducing the risks of ineffective treatment and acute complications associated with poor medication adherence.^62,63

Questionnaires are the most commonly used tool for assessing medication adherence, with the MMAS being the most widely employed. This is supported by a systematic review conducted by Clifford et al (2014), which confirms the prevalence of this method.¹⁴ The widespread use of questionnaires may be attributed to their speed, simplicity, and cost-effectiveness, especially in large-scale studies. Compared to other objective measurement methods, questionnaires do not require expensive equipment or specialized personnel and can capture a range of information related to patients’ behaviors, attitudes, and perceptions about medication. They offer insights into patients’ cognitive, attitudinal, and emotional dimensions of medication adherence. In contrast, the pill count method and Proportion of Days Covered (PDC) are used less frequently. These methods, while providing more objective data, are often more complex and difficult to implement on a large scale, with challenges in data collection. Additionally, access to comprehensive medication records may be limited, especially in resource-constrained settings. A systematic review by Cramer (2004) highlighted that electronic monitoring systems can effectively improve medication adherence among people living with diabetes.⁶⁴ However, since the studies included in this review were published before 2003 and were not limited to RCTs, this method was not incorporated into the current review. Future studies could explore the integration of these objective methods with more accessible tools like questionnaires to balance the accuracy of measurement with practical feasibility.

This systematic review included 14 studies that employed face-to-face interventions, primarily in the form of one-on-one education, health education sessions, lectures and workshops, and group discussions. Through direct interaction, patients and healthcare providers (such as physicians, nurses and pharmacists) were able to communicate more effectively, enabling patients to better understand and adhere to their medication regimens. A systematic review by Williams (2014) demonstrated that face-to-face education can significantly improve medication adherence among people living with diabetes, enhancing disease management and treatment outcomes.¹⁹ Although less commonly used, interventions such as text messaging, telephone calls, and printed materials (fliers) have also shown effectiveness in improving medication adherence. A systematic review by Presley (2019) found that these approaches can significantly enhance adherence, with digital materials and printed fliers demonstrating greater effectiveness than telephone-based interventions.¹⁷ With the advancement of technology, apps and internet-based interventions have emerged as promising tools. These technologies offer personalized medication reminders, health education and medication tracking features, thereby supporting patients in better managing their health. In this review, two studies using mobile apps and three using internet-based interventions reported positive outcomes, suggesting that modern digital approaches hold potential for improving medication adherence among patients with DM. A systematic review by Islam (2022), which examined the impact of smartphone applications on medication adherence in DM, emphasized the importance of rigorous evaluation and selection of such apps.¹⁸ The review noted that nearly half of publicly available, free apps failed to demonstrate moderate to high levels of effectiveness. Therefore, implementing only high-quality, evidence-based applications in clinical settings is essential to achieving meaningful health outcomes. However, a systematic review by Williams (2014) reported no significant improvements in adherence with internet-based interventions. This may be attributed to the fact that most of the included studies were conducted over a decade ago, at a time when internet technology was less developed and digital interventions were still in their early stages of implementation.

This systematic review included 11 studies that employed multi-component interventions to improve medication adherence among patients with DM. These interventions combined various strategies with the aim of enhancing patient engagement and adherence through multiple channels. Examples include face-to-face education combined with printed materials (fliers), mobile applications combined with email, educational videos combined with mobile phones and apps, and continuous glucose monitoring devices integrated with daily structured discussions and self-regulation education. Compared with single-mode interventions, multi-component approaches offer advantages in flexibility, personalization and sustainability. For instance, while traditional face-to-face interventions provide personalized education, they are often limited by fixed time and location, making them less adaptable to patients’ evolving needs. In contrast, digital tools such as mobile apps, SMS and emails can deliver support anytime and anywhere, enabling continuous patient engagement and improving adherence outcomes. However, the implementation of multi-component interventions also poses certain challenges. These include patients’ acceptance of new technologies, access to necessary devices and the effective integration of various components within a single intervention framework. Future research should further explore the feasibility and effectiveness of such interventions across diverse populations and settings and focus on optimizing their design to ensure maximum impact.

Notably, only a small number of studies incorporated theoretical models into interventions aimed at improving medication adherence among people living with diabetes. The most commonly used frameworks included the Theory of Planned Behavior (4 studies), the Health Belief Model (2 studies) and the Information-Motivation-Behavioral Skills Model (2 studies). With the exception of one study that applied four theoretical models simultaneously but did not yield significant results, all other theory-based interventions demonstrated improvements in adherence. Among the 30 studies that did not apply any theoretical framework, four reported no significant effects, while the rest showed positive outcomes. This suggests that the effectiveness of an intervention is not solely dependent on the presence of a theoretical model, but rather on the appropriateness of the model selected and the rigor of its application in both the design and implementation phases. The use of theoretical models is not about quantity, but fit. A well-matched theoretical framework can offer a structured understanding of health behaviors, clarify mechanisms of behavior change and guide the selection of targeted strategies and techniques. For example, the Theory of Planned Behavior emphasizes the roles of intention and perceived behavioral control, making it suitable for interventions that aim to enhance motivation and perceived ability to adhere to medication regimens.⁶⁵ The Health Belief Model, on the other hand, focuses on perceived susceptibility, severity and benefits, and is useful in addressing beliefs about disease risk and treatment effectiveness.⁶⁶ Therefore, future studies should prioritize the integration of relevant theoretical frameworks as a foundation for intervention design. Theory-driven approaches not only improve the precision and relevance of interventions, but also facilitate a better understanding of the underlying mechanisms contributing to behavior change. Ultimately, such approaches may lead to higher-quality and more sustainable improvements in medication adherence.

Limitations

Despite conducting an exhaustive search using five databases, this systematic review is not without its limitations. The search was confined to articles published between 1^st January 2015 and 1^st April 2025, exclusively including English articles, potentially influencing the study outcomes. The encompassed studies employed a diverse range of intervention measures, were carried out in various countries, and utilized a wide array of research tools, thereby rendering the conduction of a meta-analysis impractical.

Conclusion

This systematic review provides a comprehensive analysis of various interventions aimed at improving medication adherence among people living with diabetes. The findings suggest that face-to-face health education remains one of the most commonly used and effective intervention strategies, enhancing patient engagement and adherence. However, modern technological interventions, such as mobile applications and internet-based tools, also show great potential for improving adherence, particularly when used in combination with traditional methods. The integration of multi-component interventions, combining approaches like face-to-face interactions, digital tools and continuous monitoring devices, addresses the limitations of single-mode interventions by offering greater flexibility, personalization, and sustainability.

Despite the promising outcomes, challenges remain in ensuring the effective integration of these interventions and in improving patient acceptance of new technologies. Future studies should prioritize the selection of appropriate theoretical models to guide intervention design, as the effectiveness of these models hinges on their relevance to the specific behaviors and characteristics of the target population. The review emphasizes the importance of rigorously evaluating digital tools, with only high-quality validated applications being recommended for clinical use. While multi-component interventions and theoretical frameworks hold significant potential, further research is needed to assess their long-term effects and feasibility, in order to optimize intervention designs and achieve more sustainable improvements in medication adherence. This study provides valuable evidence and recommendations for improving medication adherence in people living with diabetes. Future intervention designs should focus more on comprehensiveness and personalization, providing stronger theoretical support and empirical data for diabetes medication adherence interventions.