{ "background": "The adoption of health systems in district hospitals is critical for improving service delivery, yet robust methodological frameworks for measuring and analysing adoption rates over time are lacking, particularly in resource-constrained settings. ", "purpose and objectives": "This study aimed to develop and assess a novel Bayesian hierarchical model to evaluate the adoption rates of integrated health systems within district-level hospitals, providing a methodological tool for longitudinal analysis. ", "methodology": "We constructed a Bayesian hierarchical model where the log-odds of system adoption for hospital i at time t, , is given by = \ + \ t + ui + vt, with ui \ N (0, \²u) and vt \ N (0, \²v) representing hospital-specific and temporal random effects, respectively. Model parameters were estimated using Markov chain Monte Carlo simulation with weakly informative priors, applied to longitudinal administrative data. ", "findings": "The model identified a positive temporal trend in adoption, with a posterior mean for \ of 0. 23 (95% credible interval: 0. 17 to 0. 29). This indicates that, on average, the odds of adoption increased by approximately 26% per year. Hospital-level heterogeneity (\ᵤ) was substantial, suggesting variable institutional capacity. ", "conclusion": "The proposed Bayesian hierarchical model provides a statistically robust framework for quantifying adoption dynamics, effectively capturing both temporal trends and institutional heterogeneity. ", "recommendations": "Health policymakers should utilise this modelling approach to identify hospitals with lagging adoption for targeted support. Future research should integrate covariates on funding and staffing to explain the observed heterogeneity. ", "key words": "Bayesian inference, health systems adoption, hierarchical modelling, health policy evaluation, longitudinal analysis", "contribution statement": "This paper introduces a novel Bayesian hierarchical modelling framework for health systems research, providing a method to formally account for uncertainty and heterogeneity in adoption
Mfinanga et al. (Mon,) studied this question.