Effectiveness of medication self-management education programs for older adults: A systematic review

1. Rationale for the Study

Over the past decades, pharmaceutical consumption has continuously increased due to aging of the population and the rise in chronic diseases [1]. According to Organisation for Economic Co-operation and Development (OECD) data, 48.6% of patients aged 75 and over are taking five or more medications as of 2021, with notably higher rates in some countries including Portugal (73.0%), Italy (64.7%), and South Korea (64.2%) [1].

Older adults typically take one or more prescribed medications due to chronic conditions such as hypertension, high cholesterol, and diabetes [2]. A study conducted in Italy analyzing more than 1.48 million older adults revealed that 54.2% experienced adverse effects from inappropriate medication use, resulting in a 16% increase in hospitalization risk [3]. According to the Centers for Disease Control and Prevention (CDC), approximately one-third of American adults in their 60s and 70s regularly use five or more prescription medications, which may increase the risk of overdose [4].

Inappropriate medication use, coupled with physical decline in older adults, reduces medication adherence and effectiveness. Older individuals experiencing cognitive decline may have difficulty managing medications due to issues such as forgetting dosages [5]. Various adverse effects may occur, including decreased attention, confusion, cognitive problems, falls and accidents, depression, and anxiety [2]. This represents a significant challenge in healthcare for older adults, emphasizing the need for careful adjustment of medication self-management and administration methods.

Differences in socioeconomic status and medical knowledge among older adults can lead to a high prevalence of potentially inappropriate medication [5]. This suggests the need for regular medication reviews and systematic, comprehensive, ongoing education to mitigate medication therapy-related risks in older adults. Such education can play a crucial role in helping older individuals perform safer and more appropriate medication self-management in relation to their health conditions.

Previous systematic reviews have highlighted the positive impact of medication self-management interventions for older adults. However, they also reported substantial variation in intervention design, duration, and outcome measures [6,7]. Various studies and efforts are ongoing worldwide to improve inappropriate medication use among older adults. For example, research using interventions such as nurse-led education on medication administration, medication self-management strategies, patient-tailored counseling, and follow-up phone calls to improve medication adherence in older patients with multiple conditions has successfully demonstrated significant effects on participants’ medication self-management [8]. Another study showed significant improvements in clinical outcomes as the number of older adults who reduced or discontinued sedative-hypnotic use increased through pharmacist-led medication education [9].

Previous research showed that a culturally tailored education program for Haitian immigrants diagnosed with hypertension demonstrated significant improvements in medication adherence and hypertension knowledge [10]. Six weeks after the intervention, participants showed a notable reduction in systolic and diastolic blood pressure, highlighting the potential impact of culturally sensitive educational approaches on managing chronic conditions effectively [10]. A study that developed and implemented a medication self-management education program (MSEP) significantly improved medication awareness and reduced medication misuse behavior in hypertensive patients [11]. Medication therapy management services have shown potential in reducing inappropriate medication use. However, the effectiveness of these services varies due to differences in design and short follow-up periods. Scoping reviews also emphasized the lack of a standardized educational approach for medication self-management among older adults [6,7].

A review of individual studies reveals that interventions for medication self-management in older adults vary widely in terms of delivery providers, educational content, frequency, and methods. Despite these differences, many studies reported positive changes, such as improved medication adherence and reduced misuse among older participants. This supports the potential benefit of education programs, even in the absence of a standardized approach. In particular, the significant effects shown by older recipients of interventions, such as improved medication adherence and reduced medication misuse behaviors, indicate the importance of medication self-management education. Therefore, this study aims to systematically review the existing literature on MSEP for older adults. By identifying and synthesizing the characteristics and outcomes of these interventions, this review seeks to inform the development of more effective and practical education programs tailored to the needs of older adults.

Previous systematic reviews have highlighted the positive impact of medication self-management interventions for older adults [6,7,9]. However, these studies also reveal significant gaps in the current understanding. For instance, some reviews found that the effects of certain programs, like medication therapy management, are often limited to short-term outcomes, leaving their long-term sustainability in question [6]. Others have mapped a wide array of available interventions without identifying a “gold standard,” which creates challenges for healthcare providers in selecting the most appropriate approach [7]. Furthermore, reviews focused on specific settings such as community pharmacies have shown inconsistent effects across different clinical outcomes, indicating that the success of an intervention is highly context-dependent [9]. A common limitation across these reviews is the substantial variation in intervention design and outcomes, often presented without a theoretical framework to explain why certain approaches might be more effective or how collaborative processes influence success.

This review addresses this gap by uniquely applying Imogene King’s Goal Attainment Theory as an analytical lens. This theoretical framework moves beyond a descriptive summary of intervention characteristics. It allows for a deeper analysis of the underlying dynamics of interaction and transaction between providers and older adults, offering a process-oriented explanation for achieving mutual goals. By systematically examining how program scope, provider type, and duration relate to specific clinical and psychosocial outcomes, this study offers novel insights into the mechanisms of successful self-management education. This approach provides a theory-informed understanding that helps to interpret the heterogeneity reported in prior reviews and to build a more robust evidence base for practice.

2. Research Objectives

This study aims to systematically evaluate the effectiveness of MSEP for older adults. The specific objectives are to: (1) synthesize the evidence on clinical and psychosocial outcomes of these programs; (2) analyze the characteristics of effective interventions, examining factors such as program scope (comprehensive vs. medication-specific), provider type (nurses, pharmacists, multidisciplinary teams), and duration (single-session vs. multi-session); and (3) identify key elements that contribute to successful goal attainment in medication self-management, guided by King’s Goal Attainment Theory.

3. Theoretical Framework

This study employs Imogene King’s Goal Attainment Theory [12] as an analytical framework to guide the systematic review of MSEP for older adults. King’s theory posits that goals are achieved through a dynamic process involving action, reaction, interaction, and transaction between healthcare providers and patients.

King’s Goal Attainment Theory is built upon interacting systems, primarily the personal system (individual characteristics such as perception, self-concept, and capacity for growth) and the interpersonal system (provider-patient dynamics involving communication, interaction, and collaborative transactions), which operate within a broader social system (healthcare organizational contexts). This review focuses specifically on how the personal and interpersonal systems contribute to the transaction process that leads to goal attainment.

To apply this framework to the systematic review, we operationalized these theoretical components by mapping the key characteristics of included studies as follows:

• The interpersonal system: The structure and process of provider-patient interactions were analyzed through program scope (comprehensive vs. medication-specific approaches), provider type (nurses, pharmacists, multidisciplinary teams), and program duration (single vs. multi-session interventions). These elements collectively define how the interaction and transaction phases unfold between providers and patients.

• The personal system: Individual patient factors and their influence on goal attainment were examined through the lens of patient characteristics and individual capacity factors as reported in the included studies.

• Goal attainment: The successful outcome of the transaction process was evaluated through both clinical outcomes (e.g., medication adherence, hospitalization rates) and psychosocial outcomes (e.g., self-efficacy, quality of life), representing the achievement of mutually established health goals.

By structuring the analysis through these theoretically-grounded dimensions, the framework allows for systematic interpretation of how different program characteristics contribute to the process of achieving medication self-management goals in older adults, moving beyond descriptive categorization to provide theoretical insights into the mechanisms of effective interventions.

1. Study Design

This systematic review examines MSEP for older adults, identifying key characteristics and synthesizing implementation and outcome findings. It adheres to the Cochrane Handbook for Systematic Reviews of Interventions (version 6.5.0) [13] and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [14].

2. Literature Inclusion and Exclusion Criteria

Specific inclusion criteria were established according to the PICOTS-SD framework (Participants, Intervention, Comparison, Outcomes, Time, Setting, and Study Design). However, in this review, no specific restrictions were imposed on ‘outcomes’, ‘time’, or ‘setting’. The ‘participants’ criterion included adults aged 65 and over [15] (or studies explicitly describing participants as older adults). For the ‘intervention’, only studies involving medication self-management programs or education addressing adherence, polypharmacy, misuse, or abuse were considered. In terms of ‘comparison’, eligible studies had to include a control group (receiving usual care or alternative interventions). The ‘study design’ was limited to randomized controlled trials (RCTs) or quasi-experimental studies with pre-post measurement.

Studies were excluded if they did not target older adults—defined as studies that did not explicitly state that participants were older adults or had a mean/median age below 65 (except when the study explicitly designated the target population as older adults despite including younger participants). Studies were also excluded if the intervention was not primarily a medication self-management (education) program, if there was no control group, or if the study was not an experimental study designed to evaluate effectiveness (e.g., proceedings or procedure studies). In addition, studies that did not report analyzable data (such as statistical outcomes), studies from the gray literature, case reports, qualitative studies, literature reviews, and meta-analyses were excluded. Finally, studies for which the full text was unavailable or that were duplicate publications were also excluded.

3. Literature Search and Selection

The literature search period was set to include research literature indexed up to September 3, 2024. Electronic databases searched included PubMed, Embase, MEDLINE, Cochrane Library, CINAHL, PsycInfo, and Web of Science Core Collection. Search keywords are listed in Supplementary Table 1. Although this study targeted only randomized controlled pre-post experimental designs and quasi-experimental studies, searches were conducted specifying that keywords should be included in the title and abstract. When simultaneous title and abstract searches were not possible, searches were performed based on abstract conditions.

4. Data Assessment

Quality assessment was independently conducted using checklists by two researchers experienced in systematic reviews and meta-analyses, along with one additional researcher. To enhance assessment consistency, two practice sessions were conducted using the assessment tools. When disagreements occurred, additional quality assessments were performed through meetings until consensus was reached among researchers to derive agreed-upon results.

For quality assessment of literature, randomized comparative studies were evaluated using the Cochrane Collaboration’s Revised Risk of Bias tool for randomized trials (RoB 2.0), assessing five items: bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome, and bias in selection of the reported result. Non-randomized comparative studies were evaluated using the Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool, assessing seven items: confounding bias, selection bias, intervention classification bias, bias due to deviations from intended interventions, bias due to missing data, outcome measurement bias, and selective reporting bias.

The overall risk of bias for RCTs was judged according to the Cochrane Handbook for Systematic Reviews of Interventions. A study was rated as ‘low risk’ only if all domains were assessed as ‘low risk.’ It was rated as ‘some concerns’ if at least one domain was assessed as ‘some concerns,’ with no domains rated as ‘high risk.’ The overall judgment was ‘high risk’ if at least one domain was rated as ‘high risk’ or if the study was judged to have ‘some concerns’ for multiple domains. Our assessment reported in Supplementary Figure 1 accurately follows this principle.

5. Data Analysis

Although meta-analysis is a common method for quantitative synthesis in systematic reviews, it was deemed inappropriate for this study due to the significant heterogeneity observed across the included articles. The sources of heterogeneity were substantial, including: (1) diverse participant populations with varying chronic conditions; (2) wide variations in intervention design, content, and delivery methods (e.g., comprehensive vs. medication-specific); (3) a lack of consistency in reported outcome measures (e.g., medication adherence, clinical indicators like blood pressure, psychosocial indicators like quality of life); and (4) different follow-up durations. Given these conditions, statistically pooling the results could have produced misleading conclusions. Therefore, a structured narrative synthesis, guided by King’s Goal Attainment Theory, was chosen as the most appropriate method to meaningfully synthesize the findings and explore the characteristics of effective programs across different contexts.

Data synthesis was conducted through a theory-driven structured narrative synthesis, systematically analyzing how the four key intervention dimensions related to King’s theoretical framework. Program scope (comprehensive vs. medication-specific) was examined as reflecting the action and reaction phases of the transaction process, representing how healthcare systems recognize and respond to medication management needs. Provider type (nurses, pharmacists, multidisciplinary teams, others) and program duration (single-session vs. multi-session) were analyzed as indicators of the interaction and transaction phases, revealing how different interpersonal dynamics and relationship investments facilitate collaborative goal-setting. Outcome type (clinical vs. psychosocial) served to evaluate goal attainment, assessing whether interventions achieved the health objectives established between providers and patients.

The content aligned with the research objectives was extracted and organized from the finally selected studies using Microsoft Excel. Extracted information included: authors, publication year, country, research design, details of educational interventions (e.g., content, delivery method, number and duration of sessions, frequency of education), participant characteristics (e.g., age, race/ethnicity, gender), authors’ definitions of medication self-management education, and outcome measures (e.g., measurement tools, timing and frequency of measurement).

For the purpose of this review, the primary outcome was defined as medication adherence, which was the most commonly reported clinical indicator of effectiveness. Secondary outcomes included drug-related problems, hospitalization rates, and self-efficacy. The risk of bias assessment for each study was conducted based on the reported results for the primary outcome. If medication adherence was not reported, the assessment was based on the main clinical outcome presented by the study authors.

Data were synthesized descriptively by categorizing studies according to the theoretical framework, with the included studies grouped along the four operationalized dimensions: (1) Program scope (comprehensive vs. medication-specific interventions), (2) outcome type (clinical vs. psychosocial indicators), (3) provider type (multidisciplinary teams, nurses, pharmacists, or other healthcare providers), and (4) program duration (single-session vs. multi-session interventions). This theory-guided categorization enabled analysis of how different intervention characteristics aligned with the personal and interpersonal systems within King’s framework, and how these characteristics contributed to successful transaction processes and goal attainment.

The narrative synthesis examined how each program facilitated the progression from problem recognition (action) through provider-patient communication and collaboration (interaction) to goal achievement (transaction and attainment). Trends, patterns, and gaps were identified within each dimension and across their interactions to provide theoretical insights into the mechanisms of MSEP and inform future research and program development for older adults.

1. Results of Literature Search and Selection

Literature searches were conducted using advanced search options, with two researchers independently performing searches and comparing search strategies until results matched across databases. Initial search results yielded 684 articles: PubMed (87), Embase (106), MEDLINE (32), Cochrane Library (274), CINAHL (22), PsycInfo (29), and Web of Science (134). After removing 206 duplicate articles using EndNote, 478 articles underwent primary screening of titles and abstracts. At this stage, 361 articles were removed: 144 non-older adult targeted studies, 103 studies not focusing on medication self-management education, 52 studies without control groups, 42 non-experimental studies, two qualitative studies, seven literature reviews and meta-analyses, three articles without full text, and eight duplicates, leaving 117 articles for full-text review.

During full-text review, 101 articles were excluded (Supplementary Table 2): 47 non-older adult targeted studies, 18 studies not focusing on medication self-management education, six single-group studies without control groups, 19 proceedings/procedure studies without research results, eight studies lacking statistical results, two gray literature items without research results, and one without full text. Additionally, 18 articles were added through reference lists and direct searches. Finally, 34 articles were included in the systematic review (Figure 1). The full bibliographic list of included studies (A1~A34) is provided in Appendix 1.

Figure 1.

PRISMA flow diagram.

2. Results of Data Assessment

Of 34 studies, 28 RCTs were assessed using RoB 2.0 [A1-A28], with 11 showing “some concerns” in bias, particularly in D2 (deviations from interventions) and D4 (outcome measurement) (Supplementary Figure 1). Other RCTs had low bias risk. Six non-randomized studies, assessed via ROBINS-I [A29-A34], exhibited moderate bias overall, with low risk in D3 (intervention classification) and D7 (result selection, except [A32]) and moderate bias in D4 for most studies. Other domains showed low to moderate bias (Supplementary Figure 1).

3. Results of General Characteristics of Selected Literature

The general characteristics of the 34 selected studies are shown in Table 1. Furthermore, Table 2 presents a brief summary of the characteristics for each included study, systematically categorized according to King’s theoretical framework operationalization. The studies demonstrated diverse approaches to medication self-management education, reflecting different strategies within the interpersonal and personal systems that contribute to goal attainment.

Table 1.

General Characteristics of the Selected Studies (N=34)

Table 2.

Summary of Intervention Characteristics for Included Studies According to King’s Goal Attainment Theory Framework (N=34)

Regarding publication countries, the United States had the highest number with 12 studies (30.0%), followed by South Korea and Australia with four studies (10.0%) each. Concerning research methodology, RCTs were more prevalent with 28 studies (82.4%) compared to quasi-experimental studies with six studies (17.6%). Study duration varied, with 14 studies (41.2%) lasting 1-2 years and 10 studies (29.4%) lasting less than 1 year. Regarding funding, 27 studies (79.4%) received financial support.

4. Results of Program Scope: Comprehensive vs. Medication-Specific Interventions

Reflecting the action and reaction phases of King’s transaction process, the 34 studies demonstrated two distinct approaches to recognizing and responding to medication management needs in older adults. Twenty studies implemented comprehensive interventions that integrated medication self-management education with broader health-focused components, such as geriatric assessments, chronic disease management, or health counseling. This comprehensive approach represents a broad systemic reaction to complex health needs, as seen in D’Agostino et al. [A1], who employed a multidimensional approach combining geriatric assessment and motivational counseling for older adults in a substance abuse program, and Leveille et al. [A17], who collaborated with social workers to address disability and chronic illness management.

In contrast, 14 studies delivered medication-specific interventions, focusing on adherence, misuse, or specific drug regimens. Examples include Almutairi et al. [A2], which targeted psychotropic medication reduction in aged care residents, and Kim [A15], which provided hypertension-focused education via remote methods. Furthermore, detailed information regarding the medication self-management status of each study is presented in Supplementary Table 3.

5. Results of Outcome Type: Clinical vs. Psychosocial Indicators

The outcome patterns observed across studies reflect different dimensions of goal attainment within King’s framework. Clinical outcomes, observed in 28 studies ([A1,A2,A4-A7,A9-A19,A21,A23-A31,A34]), included measurable health indicators such as medication adherence, blood pressure control, hospitalization rates, and drug-related problems. These clinical indicators represent tangible goal attainment in the physical health domain, as demonstrated by Gohil et al. [A12], who measured reductions in antibiotic days of therapy, and Varas-Doval et al. [A24], who evaluated uncontrolled health problems.

Psychosocial outcomes, reported in 15 studies ([A3,A4,A6,A8,A14,A15,A20,A22,A23,A25-27,A30,A32,A33]), encompassed quality of life, self-efficacy, medication knowledge, and self-management behaviors. These psychosocial indicators reflect goal attainment within the personal system, showing changes in self-concept, knowledge, and capacity for self-management, as seen in Ballard et al. [A3] (quality of life in dementia patients) and Lee [A32] (medication awareness in hypertensive older adults). Nine studies ([A6,A8,A14,A15,A23,A25-27,A30]) assessed both clinical and psychosocial outcomes, demonstrating comprehensive goal attainment across multiple domains. Detailed information on the characteristics of the interventions in MSEP is provided in Supplementary Table 4.

6. Results of Provider Type: Nurses, Pharmacists, and Multidisciplinary Teams

Provider variation represents different patterns of interpersonal system dynamics and transaction processes. Intervention providers varied, with 13 studies led by pharmacists, seven by nurses, six by multidisciplinary teams, and eight by other providers (e.g., physicians, therapists). Pharmacist-led interventions typically embodied technical transaction patterns, with studies such as Lin et al. [A18] and Harlow et al. [A30] focusing on medication reviews and adherence, leveraging pharmacists’ specialized drug expertise for precise, medication-focused interactions.

Nurse-led interventions demonstrated holistic transaction patterns, with studies including Elpida et al. [A8] and Farahmand et al. [A29] emphasizing patient education and self-management through sustained relationship building and continuous engagement. Multidisciplinary teams facilitated complex transaction processes, as seen in Woodham et al. [A34] and Wang et al. [A25], which integrated diverse professional expertise to address multifaceted needs, such as cardiovascular disease management. Other providers included geriatric care managers (D’Agostino et al. [A1]) and physiotherapists (Rios et al. [A23]), representing specialized interaction approaches tailored to specific patient populations. Detailed information on the characteristics of the interventions in MSEP is provided in Supplementary Table 4.

7. Results of Program Duration: Single-Session vs. Multi-Session Interventions

Program duration reflects the depth and intensity of the interaction and transaction phases within the interpersonal system. Of the 34 studies, four implemented single-session interventions, 28 delivered multi-session interventions, and two had unspecified intervention counts (“NI”). Single-session interventions represent concentrated transaction processes, as demonstrated by Almutairi et al. [A2] and Lalonde et al. [A16], which involved focused, one-time education or consultation, often delivered remotely or in specialized settings for specific, well-defined goals.

Multi-session interventions facilitated extended transaction processes, ranging from 2 to 12 sessions and including Ballard et al. [A3] (9 sessions) and Nir and Weisel-Eichler [A22] (12 sessions), emphasizing repeated engagement and relationship development over time. These extended programs enabled iterative goal-setting and achievement cycles, with multi-session programs being particularly prevalent in comprehensive interventions addressing complex conditions like dementia or cardiovascular disease. Detailed information on the characteristics of the interventions in MSEP is provided in Supplementary Table 4.