Development of a high-performing, cost-effective and inclusive Afrocentric predictive model for stroke : a meta-analysis approach

Abstract

BACKGROUND : Predicting stroke risk is critical for preventive interventions. Most validated prediction models do not include data from African populations and may not be appropriate for the region. Relying solely on statistical significance to identify predictors may compromise algorithm performance. Also, some of the existing models include expensive biomarkers that are unsuitable for resource-limited settings. This study aims to develop a cost-effective and inclusive Afrocentric predictive model for stroke (CAPMS). METHODS : We conducted a meta-analysis following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses protocol and searched the PubMed, Scopus, African Journal, Medline, Cochrane Library, Web of Science, and Cumulative Index for Nursing and Allied Health Literature databases. We included case‒control and cohort studies reporting stroke risk factors and their estimates among African populations. Titles and abstracts were independently screened. Meta-analyses were performed using Comprehensive Meta-analysis version 3. RESULTS : More than 50% of the eligible studies examined both ischemic and hemorrhagic stroke. More than 20 stroke risk factors were identified in Africa, with 18 eligible for meta-analysis. Homocysteine (risk weight [Rw] = 13.9, risk stability index [Ri] = 0.67), hypertension (Rw = 5.6, Ri = 0.94), and cardiac events (Rw = 3.1, Ri = 0.8) were the strongest independent predictors. Low green vegetable consumption (Rw = 2.4, Ri = 1.0), stress (Rw = 1.76, Ri = 1.0), and hypertension were the most clinically responsive risk factors. All risk factors/biomarkers except homocysteine cost between $2.8 and 12.2, indicating cost-effectiveness. A critical risk point of 12.7 was set at the 90th percentile. The cumulative Rw and costs for CAPMS 1 (20 and $1.2–4.6) and CAPMS 2 (22.4 and $6.5–17.3) indicate high performance and cost-effectiveness. CONCLUSIONS : Targeted screening via the CAPMS 1 and CAPMS 2 models offers a cost-effective solution for stroke screening in African clinics and communities. Immediate validation of the CAPMS is needed to evaluate its performance, feasibility, and acceptability in the region. REGISTRATION : The study protocol is registered with PROSPERO (ID: CRD42023430437).