The study involves a systematic review and meta-analysis of randomized controlled trials (RCTs) and was designed in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) statements (Multimedia Appendix 1) [18].

Electronic data searches of PubMed, Embase, Scopus, and Web of Science were conducted between May and October 2022, with an update in October 2023 to identify RCTs focusing on eHealth interventions to improve pharmacotherapy adherence in patients who had an ACS incident. Articles published from January 2000 to October 2023 were considered for screening. The following search keywords or concepts and relevant synonyms words were used combined via appropriate Boolean operators: Mobile health, text messages, smartphones, eHealth, and mobile applications in combination with cardiovascular disease, secondary prevention, adherence, medication adherence, medication nonadherence, coronary artery disease, acute coronary syndrome, myocardial infarction, and cardiac rehabilitation. The detailed keywords and search strategies used in different databases are presented in Multimedia Appendix 2.

The Population Intervention Comparison Outcome(s) Study design statement for this systematic review was as follows: participants: patients with post-ACS; intervention: eHealth technology (telehealth, eHealth, smartphone, texting, mobile health [mHealth], phone apps, etc); comparisons: standard or usual care; outcomes: adherence to guidelines-recommended post-ACS pharmacotherapy; and study design: RCTs.

For this systematic review, studies were included if they investigated the effectiveness of eHealth technologies (eg, mobile phone app, a web-based app, a smartphone app, an electronic device, or texting) in an RCT for a duration of at least 12 weeks. The studies were focused on improving adherence to various classes of guideline-recommended cardioprotective medications, such as aspirin, statins, β-blockers, P2Y12 inhibitors, and ACEIs or ARBs, among patients after ACS. The studies, to be included in the meta-analysis, had to provide information on the number of subjects or proportion of use for the 5 classes of medications indicated for secondary prevention after ACS in both the intervention and control groups at baseline as well as at follow-up periods. Studies that examined eHealth interventions for the treatment of obesity, hypertension, dyslipidemia, secondary lifestyle factors, and smoking cessation in patients without ACS were excluded. Articles published in non-English language results were also excluded. Additionally, abstracts, case reports, editorials, and conference presentations were not considered for this systematic review.

Titles and abstracts found through the electronic database searches were imported into the Covidence systematic review software (Veritas Health Innovation Ltd). A total of 2 reviewers (ASB and WT) independently screened the abstracts. Studies not meeting the predetermined selection criteria were excluded. After reviewing the abstracts of identified studies, the full texts of eligible publications were subsequently evaluated by the same 2 independent reviewers for potential inclusion in the final analysis. In addition, bibliographies of relevant publications were also examined to identify any articles missed during the original database searches. Any disagreements regarding article inclusion and exclusion of the article were resolved through collaborative review. The final results were reviewed by all authors.

A total of 2 authors (ASB and WAA) independently extracted the data into a predetermined Excel (Microsoft Corp) spreadsheet, and the third author (WT) evaluated the collected data. Study characteristics (authors, year of publication, country, registration, design, and duration of the trial protocol); patient demographics (mean age, sex, size, and type of control condition); design parameters (type of intervention and length of intervention); eHealth intervention features (type of electronic device, messaging frequency, and a web-based app), and outcomes (method and frequency of assessment for adherence and type of medication) were extracted.

A total of 2 independent authors evaluated the risk of bias within individual studies using the modified Cochrane risk-of-bias tool for RCTs [19], a 7-item instrument that assesses selection bias, allocation concealment, implementation bias, measurement bias, follow-up bias, reporting bias, and others. Any discrepancies were resolved by consensus.

The primary outcome includes overall adherence to cardioprotective medication classes (statins, aspirin, P2Y12 inhibitors, ACEI or ARBs, and β-blockers) in patients with ACS following eHealth interventions. The secondary goal is to evaluate the effectiveness of eHealth interventions on medication adherence at 6-month and 12-month follow-up periods.

The included trials reported the extent of medication nonadherence, including to specific drug classes, using both self-report questionnaires and objective methods like medication possession ratio or prescription claims data at different time intervals (3-, 6- or 12-months post intervention). When trials reported multiple follow-up assessments, we pooled the relevant data to determine the effect of the intervention at 6-month and 12-month time points. To analyze the data, effect sizes were estimated using standardized risk differences (RD) between the intervention and control groups, as well as RD for each cardiovascular medicine in the study and weighted pool estimates. We considered variations in outcome measurement across studies by applying appropriate statistical methods (ie, fixed effects for heterogeneity level of ≤25% and random effects when heterogeneity level was >25%) to generate meta-analytic estimates of intervention effect and presented as RD [19]. A chi-square test was used to assess homogeneity between studies and a homogeneity P value of less than .10 was considered statistically significant [20]. Subgroup analyses were performed using self-report and objective adherence assessment methods to determine the effectiveness of intervention on adherence to individual or all drug classes both at 6-month and 12-month follow-ups. Due to the small number of studies meeting our inclusion criteria, publication bias was not assessed [21]. All statistical analyses were performed using STATA MP statistical software (version 16.1; StataCorp). The results were expressed as an “RD” with 95% CI. A P value less than .05 was considered a statistically significant difference between the intervention and the controls.

The database search resulted in 831 titles and abstracts, while the bibliographic search of these articles did not result in any other articles. Of the 831 articles screened, 779 articles were considered for screening and 40 articles met the eligibility criteria, but most of these RCTs did not have data on medication adherence or follow-up (n=24), no specific data on ACS (n=5), were not an eHealth intervention (n=4), and other reasons (n=4). Finally, 5 peer-reviewed journal articles met the inclusion criteria (Figure 1) [13-17] and their features are summarized in Table 1.

All RCTs were published in English between the years 2020 and 2022. A total of 3 of the included studies were from China, while 1 was conducted in Australia and another 1 in New Zealand. The 5 studies included 4100 participants, with 2093 and 2007 people included in the intervention and control groups, respectively. The majority of the targeted participants were male, with their average age ranging from 57 to 64.8 years.

Adherence to statins, aspirin, P2Y12 inhibitors, ACEIs or ARBs, and β-blockers was evaluated using text message, WeChat, and smart phone–based app interventions. Both self-report [13-15,17] and objective [13,15] adherence measurement methods were used by the included studies to determine the level of nonadherence to prescribed medications. For example, 2 studies used the 8-item Morisky medication adherence scale (MMAS-8) to determine the proportion of medication nonadherence [15,17], while other studies used self-reports along with objective methods like prescription claims data [13], or a medication possession ratio [15]. Patients were deemed adherent, for example, if at different time points, they had a medication possession ratio of ≥80% (determined based on the number of days patients are in possession of the dispensed drugs divided by the number of days used for follow-up) [15] or self-report of medication use >80% of the time in the previous 30-days (24/30 days) [13]. People were also considered adherent based on having a score of ≥6 using the MMAS-8 scale, which was obtained mainly from “yes” and “no” responses and 1 Likert-based question [17].

Figure 2 shows the weighted RD between the intervention and the control groups for statins, P2Y12 inhibitors, ACEIs or ARBs, β-blockers, and aspirin. A total of 4 studies provided data for intervention on adherence to statins, showing that eHealth intervention was not associated with significant improvement in adherence both at 6 months (RD –0.00, 95% CI –0.03 to 0.03; I²=0.0%) and 12 months (RD –0.02, 95% CI –0.05 to 0.02) post intervention. Similarly, no improvements in the medication adherence following eHealth intervention were observed for other classes of medications, such as P2Y12 inhibitors (RD –0.01, 95% CI –0.04 to 0.02; I²=0.0% at 6 months and RD –0.01, 95% CI –0.05 to 0.03; I²=0.0% at 12 months), aspirin (RD 0.00, 95% CI –0.06 to 0.07; I²=0.0% at 6 months and RD 0.00, 95% CI –0.07 to 0.06; I²=0.0% at 12 months), ACEIs or ARBs (RD –0.01, 95% CI –0.04 to 0.02; I²=0.0% at 6 months and RD 0.01, 95% CI –0.04 to 0.05; P=.18; I²=42.4% at 12 months), and β-blockers (RD 0.00, 95% CI –0.03 to 0.03; I²=0.0% at 6 months and RD –0.01, 95% CI –0.05 to 0.03; P=.31; I²=15.4% at 12 months).

A subgroup analysis of the effectiveness of eHealth interventions on medication adherence at 6 months’ and 12 months’ time points post intervention was conducted based on the adherence assessment methods used (self-report vs objective). The findings showed no significant combined RDs in overall medication adherence between intervention and control groups both at 6-month and 12-month follow-up periods, irrespective of the type of medication adherence assessment method applied (see detailed analyses in Multimedia Appendix 4). Heterogeneity among studies was high for most analyses, except when using the objective adherence assessment method. Furthermore, the studies that provided data for pooled estimation of the effect of the intervention on adherence to all drug classes also revealed no improvement in adherence both at 6 months (RD 0.01, 95% CI –0.00 to 0.02; I²=92.3%) and 12 months (RD 0.00, 95% CI –0.01 to 0.01; I²=1.00%).

Overall, all studies showed a low risk of bias, and only 1 study did not provide a description of the blinding of outcome assessment. In Table 2, we reported the methodological quality assessment of individual studies as per the modified Cochrane risk of bias tool for RCTs.

This work presents a systematic review and quantitative analysis of randomized trials evaluating the effect of eHealth interventions on adherence to medications acting on the cardiovascular system in people who experienced ACS. The findings of this review revealed that eHealth-based intervention was not associated with significant improvements in adherence to guideline-recommended medications such as statins, P2Y12 inhibitors, aspirin, or hypertension medications or overall medication classes after ACS incidents. At 6 and 12 months’ time points after the intervention, no significant improvements in cardioprotective medication adherence were observed, irrespective of the methods of adherence assessment used.

Previous evidence from a meta-analysis of 16 randomized controlled trials found that sending medication-related reminders in the form of SMS doubled medication adherence across a range of chronic conditions [22]. This is in contrast to the findings of our review, which focused on patients with an acute medical condition on top of preexisting chronic diseases. ACS is a life-threatening condition that may require a specialized intervention plan. Consistent postdischarge assistance and education for long-term secondary prevention of ACS remain a health care service challenge. eHealth-based interventions are hypothesized to have the potential to improve the implementation and accessibility of nonpharmacological modalities for promoting the long-term secondary prevention of ACS [23]. These low-cost, easy-to-implement, and scalable eHealth interventions may have used as a consolidated source of information. Neither do they require extra investment or extensive clinician involvement.

Despite being low-cost and scalable, however, the contents of eHealth messages may not be enough to elicit behavioral changes in patients with ACS. This is particularly difficult in contexts where there are competing health goal priorities affecting medication adherence, like in case of patients experiencing acute illnesses superimposed on chronic conditions.. Similar initiatives that were published in the past, that addressed a single behavior, for example, such as exercise or smoking cessation, reported the success of eHealth intervention [24,25]. Targeting multiple behaviors simultaneously may be burdensome for patients demanding both acute and chronic medical care. SMS text messaging program alone may not satisfy participants’ needs for more personalized interventions [26].

Prior research demonstrates that eHealth technologies can address extrinsic social and economic factors to improve medication adherence after ACS [27-29]. For example, the study by Redfern et al [27] highlighted how an iterative, theory-based development process for texting interventions can inform optimized eHealth tools to enhance post-ACS pharmacotherapy adherence. Their user-centered approach integrating behavior change techniques suggests that future eHealth interventions could benefit from participatory design tailored to patient needs and preferences.

Applying such standardized methodology grounded in guidelines and tailored content may advance the effectiveness of eHealth adherence promotion after ACS. In addition, another pilot study by Riegel et al [29] tested a personalized telehealth intervention using behavioral economics and financial incentives to promote maintained aspirin adherence after ACS hospitalization, representing a participatory approach tailored to patient motivations that showed promising trends versus declining adherence in the controls.

From our pooled analysis, the findings revealed that the implemented eHealth programs were not significantly more effective than usual care in improving cardioprotective medication adherence in people who experienced recent ACS, regardless of the adherence assessment method used. This may point to the need for a redesign of eHealth interventions’ contents or components as well as their delivery. Based on the literature, for interventions to be more effective, the design should be theory-driven, iteratively designed, and culturally tailored to provide educational and motivational information to patients who experienced ACS. A successful intervention outcome may result from incorporating personalized intervention during a design process [29]. Most of the studies included in our analyses did not use theory-driven interventional materials, which are found to be effective in improving self-management practices and adherence in certain health conditions and patient populations [30-32]. The application of theory-based interventions provided in the right context may lead to improved adherence to treatments. However, it is important to note that the given patients, who were included in the targeted studies, were enrolled at or soon after their hospital discharge, they are more likely to have higher adherence and lifestyle adjustments, regardless of the implementation of an intervention.

This review synthesized evidence from published studies of relatively high quality, with all included studies being RCTs. Due to the rigor of their study designs, the majority of articles included in the meta-analysis had a minimal risk of bias. However, our evidence from the systematic review should be interpreted in the context of several major limitations. Variations in study participants’ characteristics, definitions of medication adherence, and sample sizes may account for the wide range of reported risk reductions. More importantly, the small number of studies from only 3 countries (Australia, China, and New Zealand) limits the generalizability of the findings to other settings. The methods used to assess adherence were different among the included studies, making it difficult to fully understand the effectiveness of the implemented interventions on adherence. This is particularly important given the absence of a gold-standard adherence measurement or adequate level of adherence that could predict health outcomes. In the literature, there are different types, measurement tools, and definitions of what constitutes medication adherence, bringing substantial heterogeneity between clinical studies [33,34]. Generating evidence from this meta-analysis may have been affected by these differences, depending on the amount to which these variations were present. Therefore, owing to the limited number of studies with strict inclusion criteria, variations in the ways interventions were delivered, and differences in health service delivery, the overall generalizability of our findings is limited. Finally, as only English-language papers were selected, it is possible that some important studies may have been missed during the literature search.

The current meta-analysis revealed that there is limited evidence on the effectiveness of eHealth interventions on adherence to guideline-recommended medications after ACS compared to controls. Although prior studies demonstrate the use of eHealth tools for modifying cardiovascular risk factors, our pooled analysis of RCTs specific to people who had ACS did not find a significant association between eHealth interventions and medication adherence across 5 standard drug classes at 6- and 12-month follow-ups. This highlights the need for further studies to better understand the role of different eHealth approaches, including those beyond text messaging, to enhance post-ACS pharmacotherapy adherence and potentially resultant cardiovascular outcomes.