Atherosclerotic occlusive disease is common worldwide, increasing in global prevalence, and characterized by arterial stenosis (narrowing) and occlusion secondary to the buildup of lipid-rich plaque within arterial walls [1]. This disease may be further categorized by anatomic location, with atherosclerotic plaque buildup within the arteries of the extracranial and intracranial circulation, heart, and peripheral arteries being referred to as cerebrovascular disease (CVD), coronary artery disease (CAD), and peripheral artery disease (PAD), respectively [1]. An estimated 101 million [2], 197 million [3], and 113 million [4] individuals have CVD, CAD, and PAD worldwide, respectively. CVD, CAD, and PAD are associated with a significantly increased risk of hospitalization; need for coronary, carotid, or lower limb revascularization; major adverse cardiac events (MACEs) and major adverse limb events (MALEs); stroke; and cardiovascular and all-cause mortality [5,6].

International, evidence-based clinical practice guidelines strongly and consistently recommend that adults with CVD, CAD, PAD, and polyvascular disease be prescribed and take antiplatelets and statins (ie, hydroxymethylglutaryl-CoA reductase inhibitors) [7-11]. They also recommend antihypertensives for those patients with concurrent hypertension [7-11]. These recommendations are based on high-quality data from randomized controlled trials (RCTs) and systematic reviews and meta-analyses of RCTs that reported a significantly lower risk of MACE; MALE; need for coronary, carotid, and lower limb revascularization procedures; cardiovascular and all-cause mortality; and health care resource use when these patients are prescribed these medications [12-14].

Despite this, evidence suggests that many patients with atherosclerotic occlusive disease are not prescribed these medications adequately or do not adhere to them after prescription [15-18]. More specifically, only approximately 40% to 90% of patients with any form of atherosclerotic occlusive disease are prescribed any antiplatelet, 30% to 60% any statin, 40% to 60% any antihypertensive, and 10% to 40% all three guideline-recommended cardiovascular medications [15,16]. Studies also report that patient adherence percentages to these medications range from ~70% to 80% for antiplatelets, ~50% to 60% for statins, ~40% to 70% for antihypertensives, and ~20% to 40% for all three guideline-recommended cardiovascular medications [17,18]. Those with PAD are also significantly less likely to be prescribed these medications and adhere to them compared to those with CVD and CAD [19]. The increased frequency at which patients with PAD are underprescribed and nonadherent to guideline-recommended cardiovascular medications may partially explain why studies demonstrate persistently higher risks of cardiovascular-related hospitalization (eg, for atherothrombotic events), MACE, stroke, and cardiovascular mortality in patients with PAD compared to those with CVD or CAD [20].

To improve the use of guideline-recommended cardiovascular medications among patients with atherosclerotic occlusive disease, RCTs [21-48] have evaluated implementation strategies (interventions directed at the health system and health care providers) and adherence-supporting strategies (interventions aimed at patients directly) [49] to improve health care provider prescription of and patient adherence to guideline-recommended cardiovascular medications, respectively. However, the design [21-24], length of follow-up [24,27,31], setting [21,24-26], and implementation or adherence-supporting strategies examined [21,26,27,30-32,36,40-48] varies considerably across published trials. Existing RCTs have also examined implementation and adherence-supporting strategies aimed at improving prescription of or adherence to different medication classes (antiplatelets [25,28,29], statins [27-29], antihypertensives [25,28,42], or all guideline-recommended medications [27,29,43,44]) among patients with different diagnoses (CVD [31-33], CAD [24,26,27], PAD [34,35], and polyvascular disease [36,37]). Further, the implementation and adherence-supporting strategies examined within published trials target different types of health care providers (ie, cardiovascular specialists [38,39] or noncardiovascular specialists [21,23]) and health care provider behaviors [31,33,38,40]. Given this interstudy heterogeneity, evidence-synthesis efforts are needed to interpret the results of the above RCTs collectively.

Two updated, comprehensive, and methodologically rigorous systematic reviews and meta-analyses are needed to determine the effectiveness of implementation and adherence-supporting strategies for improving health care provider prescription of, and patient adherence to, guideline-recommended cardiovascular medications for atherosclerotic occlusive diseases. All forms of atherosclerotic occlusive disease (CVD, CAD, PAD, and polyvascular disease) will be examined within both systematic reviews as the underlying pathophysiology [50-52], and guideline-recommended cardiovascular medications for medical management of these diseases [7-11] are nearly identical. Further, patients with these atherosclerotic occlusive diseases tend to have similar demographics (ie, frequently past or current smokers, often male, and typically over 65 years of age) and comorbidities [53-55]. Finally, these diseases often coexist together [56,57], and similar prescribers are involved in their medical management [58-60]. The results of these systematic reviews will help identify promising implementation and adherence-supporting strategies for each and all classes of cardiovascular medications recommended for patients with atherosclerotic occlusive diseases. The systematic reviews will facilitate an effective comparison of implementation and adherence-supporting strategies by classifying them according to established taxonomies [61,62].

The primary objectives of the two systematic reviews are to synthesize the effectiveness of implementation and adherence-supporting strategies for improving health care provider prescription of, and patient adherence to, guideline-recommended cardiovascular medications among patients with atherosclerotic occlusive disease, including CVD, CAD, PAD, or polyvascular disease. We conceptualized the desired behavior changes for our primary objectives and key actors involved in bringing about these desired behavior changes using the Action, Actor, Context, Target, Time (AACTT) framework [63] (see Table 1 for the underprescription systematic review AACTT framework and Table 2 for the nonadherence systematic review AACTT framework).

The AACTT framework [63] was developed to describe the action (ie, the behavior to change), actor (ie, who is involved in performing the behavior), context (ie, the physical, social, or emotional settings where the behavior is performed), target (ie, the person or people for or with whom the behavior is performed), and time (ie, when the behavior is performed) of behaviors targeted for change in clinical practice. Describing desired behavior changes using the five elements of the AACTT framework [63] provides a concise overview of which individuals at what organizational levels need to do what differently to ensure the effective implementation of desired best practices. Secondary objectives are to determine whether (1) the effectiveness of implementation or adherence-supporting strategies at improving health care provider prescription of or increasing patient adherence to guideline-recommended cardiovascular medications for atherosclerotic occlusive disease varies by patient characteristics, prescriber traits, medication classes, atherosclerotic occlusive disease diagnoses, or clinical settings; and (2) improved health care provider prescription of or patient adherence to guideline-recommended cardiovascular medications improves patient outcomes and reduces health care resource use.

This protocol was created following recommendations from the PRISMA-P (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols) statement [64] (a completed PRISMA-P checklist is provided in Multimedia Appendix 1). The protocols for the systematic reviews of implementation and adherence-supporting strategies for improving health care provider prescription and patient adherence were registered separately on PROSPERO (underprescription systematic review: CRD42023461317; nonadherence systematic review: CRD42023461299). We chose to create a single protocol paper describing the methods of the two planned systematic reviews and meta-analyses as we felt that their background, rationale, and methods overlap significantly. However, as scout searches suggested that one systematic review may include over 100 RCTs, we plan to report their results separately to summarize the results of these evidence syntheses in sufficient detail.

Eligibility criteria for the two systematic reviews were created according to the Population, Intervention, Comparison, Outcome, and Design framework [65]. Their specific inclusion and exclusion criteria are presented in Table 3 (underprescription systematic review) and Table 4 (nonadherence systematic review).

Guideline-recommended cardiovascular medications will be defined as any medications recommended within evidence-based clinical practice guidelines cited within the included studies. In studies where evidence-based clinical practice guidelines are not cited, guideline-recommended cardiovascular medications will be defined as antiplatelets (eg, aspirin or clopidogrel), statins, and antihypertensives (eg, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, β-blockers, calcium-channel blockers, or thiazide diuretics), as these medications are consistently recommended for individuals with CVD, CAD, PAD, and polyvascular disease across multiple evidence-based clinical practice guidelines [7-11].

We will search MEDLINE, Embase, The Cochrane Central Register of Controlled Trials, PsycINFO, and CINAHL from their inception. We will also use the PubMed “related articles” feature and manually search reference lists of included studies and relevant reviews identified during the search.

Our electronic search strategy was designed in consultation with an information scientist. Using combinations of Medical Subject Headings (National Library of Medicine) or Emtree (Elsevier) terms and keywords, we constructed search filters covering the themes “atherosclerotic occlusive disease” and “medication prescription or adherence.” These search filters were then combined with a validated RCT search filter from the Cochrane Handbook for Systematic Reviews of Interventions [67]. The search strategy was subsequently piloted and refined by adding additional thesaurus or indexing terms to the nonsearch filter themes when new, relevant citations were located during iterative pilot searches. The penultimate search strategies were peer-reviewed by another medical librarian in accordance with the Peer Review of Electronic Search Strategies framework [68] (see Multimedia Appendix 2 for final search strategies).

Titles and abstracts of identified citations will be reviewed by two independent reviewers using Rayyan Systematic Review Software [69]. Citations deemed potentially relevant by either reviewer during title/abstract screening will be subject to independent full-text review using relevant inclusion and exclusion criteria (see Tables 3 and 4). Disagreements regarding study inclusion will be resolved by consensus.

The primary outcome for the underprescription systematic review will be the health care provider’s prescription of each (eg, antiplatelets, statins, or antihypertensives) or all classes of guideline-recommended cardiovascular medications between the implementation strategy and control arms. The primary outcome of the nonadherence systematic review will be the patient adherence to each (eg, antiplatelets, statins, or antihypertensives) or all classes of guideline-recommended cardiovascular medications between the adherence-supporting strategy and control arms. Secondary outcomes for both systematic reviews will be differences in patient outcomes and health care resource use between the implementation or adherence-supporting strategy and control arms. Patient outcomes will include MACE [70] (defined as nonfatal myocardial infarction, stroke, or cardiovascular mortality), MALE [70] (defined as limb ischemia requiring revascularization or above-ankle lower limb amputation), stroke, revascularization procedures, amputations, patient-reported outcome measures, cardiovascular mortality, all-cause mortality, and any other patient-important outcomes reported by study authors. Health care resource use outcomes will include emergency department visits, hospital admission or readmission, length of hospitalization, and any other health care resource use outcomes reported by study authors.

Two investigators will independently extract data using a predesigned electronic data extraction spreadsheet (see Tables 5 and 6 for data items to be extracted from the underprescription and nonadherence systematic reviews, respectively).

When potentially relevant articles published in languages other than English are identified, translators will be recruited to review the full texts and extract data if identified studies meet inclusion criteria. Effect estimates and their 95% CIs (or data required to calculate these measures) describing associations between implementation or adherence-supporting strategy use and primary and secondary outcomes will be extracted as reported in studies and may include weighted mean differences or standardized mean differences for continuous outcomes; odds ratios (ORs), relative risks (RRs), and risk differences for binary outcomes; and hazard ratios (HRs) for time-to-event data. Where studies do not provide effect estimates and their 95% CIs (or data required to calculate these measures) within the text or summary tables but summarize this information within figures, data will be extracted independently by two investigators using ImageJ software (National Institutes of Health) [72] and then their results will be averaged across investigators. Where data are not provided in the text, summary tables, or figures, these data will be sought from the supplementary materials of included trials or from trial registration repositories. In cases where data cannot be found using the above methods, study authors will be contacted in a final attempt to obtain all relevant study data. Discrepancies between investigators regarding data extraction will be resolved by consensus.

Two independent investigators will assess the risk of bias for all included trials using the Cochrane risk of bias tool [71]. Discrepancies between investigators regarding risk of bias assessment will be resolved by consensus.

The presence of publication bias will be evaluated using the Egger test and by visually inspected produced funnel plots from primary outcomes [106].

All analyses will be conducted using Stata Standard Edition (version 17.0; Stata Corp). Risk-of-bias graphics will be generated using Review Manager Systematic Review Software (version 5.4; The Cochrane Collaboration).

We will use the GRADE (Grading of Recommendations, Assessment, Development, and Evaluation) [107] to rate certainty in outcome estimates. First, we will assess the risk of bias, imprecision, inconsistency, indirectness, and publication bias in outcome estimates [107]. The overall certainty in these estimates will then be judged as high (“further research is very unlikely to change our certainty in the estimate”), moderate (“further research is likely to have an important impact on our certainty in the estimate and may change the estimate”), or low (“further research is very likely to have an important impact on our certainty in the estimate and is likely to change the estimate”) [107].

These systematic reviews will examine previously published data and are therefore exempt from ethics approval at our institution.

Our systematic reviews will assess all RCTs examining the effectiveness of implementation and adherence-supporting strategies for improving health care provider prescription of, and patient adherence to, guideline-recommended cardiovascular medications among adults with atherosclerotic occlusive diseases. We also aim to determine the effectiveness of implementation strategies at improving medication prescription across different classes of health care providers working across different clinical practice settings. We anticipate that clinician education may be a particularly effective strategy for improving prescription among those working in primary care settings, given that these types of practitioners may have relatively reduced awareness of guideline recommendations for the management of patients with atherosclerotic occlusive disease relative to cardiovascular specialists [108]. As patients treated at tertiary medical centers tend to have complex electronic medical records [109], we anticipate that a facilitated relay of information to clinicians may be particularly effective for improving medication prescription among cardiovascular specialists and clinicians working in tertiary academic centers. Strategies anticipated to be promising for improving guideline-recommended cardiovascular medication prescription across different health care provider types and clinical practice settings include audit and feedback, provider reminder systems, and organizational changes given the promise these interventions have demonstrated for improving medication prescription among patients with varied cardiovascular diseases [87,88].

We will also examine whether the effectiveness of adherence-supporting strategies varies across patients with different types of atherosclerotic occlusive diseases (CVD, CAD, PAD, and polyvascular disease). In particular, identifying promising adherence-supporting strategies for improving medication adherence among patients with PAD is of crucial importance given that research has shown that those with PAD are less likely to be prescribed guideline-recommended medications compared to those with CVD or CAD [19]. As patients with PAD tend to have worse awareness of their condition and relevant secondary medical prevention options compared to those with other atherosclerotic occlusive diseases [110], we anticipate that patient education may be a particularly promising strategy for improving medication adherence among patients with PAD. Adherence-supporting strategies involving fiscal components (eg, medication copayment) may also be a particularly promising strategy for improving medication adherence among patients with PAD given the high costs of guideline-recommended cardiovascular medications [111]. Further, as patients with atherosclerotic occlusive disease tend to be older [53-55], strategies that remind patients to take their prescribed medications may serve as promising adherence-supporting strategies across patients with atherosclerotic occlusive disease in general.

Although systematic reviews have examined the effectiveness of implementation and adherence-supporting strategies for improving health care provider prescription of [87,88], and patient adherence to [89-92], guideline-recommended cardiovascular medications for atherosclerotic occlusive diseases, they have important limitations. One systematic review of implementation strategies for improving health care provider prescription [87] characterized its implementation strategies according to the EPOC taxonomy [112], which was developed to help authors classify implementation strategies in systematic reviews. However, it aggregated implementation strategies together to assess whether they collectively increased health care provider prescription of guideline-recommended medications instead of examining whether strategy effectiveness varied within different EPOC taxonomy [112] levels (ie, professional, organizational, financial, and regulatory). This systematic review [87] also only examined patients with CAD. Although a second systematic review [88] on the effectiveness of implementation strategies for improving health care provider prescription also characterized implementation strategies using the EPOC taxonomy [112], it examined the effectiveness of improving physician adherence to “guideline-recommended care” (eg, use of cardiovascular risk assessment tools or following guidelines for cardiovascular disease screening) in addition to medication prescription [88]. It also included patients with nonatherosclerotic cardiovascular diseases. Limitations of the existing systematic reviews of adherence-supporting strategies for improving patient adherence is that many include patients “at risk” of developing an atherosclerotic occlusive disease or with nonatherosclerotic cardiovascular diseases (eg, hypertension or congestive heart failure) [91,92]. Further, many only examine specific adherence-supporting strategies (eg, polypills, text-message reminders, or patient education) [89,90,92] or adherence to specific cardiovascular medications among patients with CVD or CAD [89-92]. Many also did not categorize their adherence-supporting strategies according to an established taxonomy [91]. Finally, many relevant RCTs have been completed since the publication of prior underprescription [87,88] and nonadherence [89-92] systematic reviews.

Our proposed systematic review has potential limitations. First, we anticipate that it may be difficult to pool the results of some studies due to interstudy heterogeneity [113]. We anticipate that many relevant RCTs [21-48] may have reported different primary and secondary outcomes at different time points and using different outcome measurement scales. Should we be unable to pool our quantitative results, we will summarize our results narratively instead of using the methods outlined above [73]. Second, it may be argued that including patients with all forms of atherosclerotic occlusive disease in our study population may introduce clinical heterogeneity due to small differences in underlying disease pathophysiology [50-52] and epidemiology [53-55]. Nevertheless, the patient populations in our proposed systematic reviews are more similar regarding disease pathology, epidemiology, and medication requirements than others in previously published systematic reviews [87-92]. Further, these conditions commonly coexist within patients [56,57] and similar prescribers are involved in their medical management [58-60]. Finally, although the use of self-reported data to quantify medication adherence introduces the possibility of outcome misclassification bias [94], we plan on performing stratified meta-analysis and meta-regression to determine whether observed differences in medication adherence vary by whether studies quantified medication adherence using objective methods versus participant self-report.

In conclusion, we will perform systematic reviews and meta-analyses of RCTs examining implementation and adherence-supporting strategies designed to improve health care provider prescription of, and patient adherence to, guideline-recommended cardiovascular medications among patients with atherosclerotic occlusive diseases. These systematic reviews will also determine the effect of these interventions on patient-important outcomes and health care resource use. Finally, they will determine whether effectiveness varies by patient characteristics, prescriber traits, and clinical practice setting.