Comparative Analysis of Antimicrobial Stewardship Impact Across Hospitals

Thesis Abstract

Antimicrobial resistance remains a critical global health threat, and antimicrobial stewardship (AMS) programs are pivotal in optimizing antimicrobial use across healthcare facilities. This study addresses the variability in AMS impact across hospitals, investigating how program structure, implementation fidelity, and contextual factors influence antimicrobial consumption, resistance patterns, and patient outcomes. The aim is to compare AMS effectiveness across hospitals and identify determinants of successful stewardship that can inform policy and practice. Specific objectives include (1) quantify and compare antimicrobial consumption metrics (defined daily doses per 1,000 patient-days) across a representative sample of tertiary and secondary care hospitals; (2) examine differences in adherence to AMS core interventions (prospective audit with feedback, preauthorization, formulary restrictions, guidelines, and education) and their association with clinical outcomes; (3) evaluate trends in antimicrobial resistance patterns for sentinel pathogens over a 24-month period; (4) assess patient-level outcomes including length of stay, in-hospital mortality, and readmission rates in relation to AMS intensity; and (5) develop a contextualized framework of determinants of AMS success applicable to diverse hospital settings. The study adopts a mixed-methods, cross-sectional comparative design underpinned by the Institutional Theory of Change to explore how organizational structures, leadership, and resource allocation shape AMS effectiveness. The population comprises all in-patient admissions from five hospitals in a metropolitan healthcare network over two consecutive years. A stratified random sample of 1,200 patient records, proportionally allocated to tertiary and secondary care facilities, will be analyzed for quantitative outcomes. In parallel, 40 key informants comprising antimicrobial stewards, pharmacists, infectious disease physicians, and hospital administrators will participate in semi-structured interviews. Quantitative data will be triangulated with qualitative insights to illuminate mechanisms driving observed associations. Data collection will involve (i) extraction of antimicrobial consumption data from pharmacy dispensing and formulary systems, (ii) extraction of resistance surveillance data from microbiology laboratories, (iii) extraction of clinical outcomes from electronic health records, and (iv) semi-structured interviews guided by an AMS implementation framework. Instrument validity will be enhanced through pilot testing, triangulation, and expert panel review. Reliability will be established via inter-rater checks for record abstraction and Cronbach’s alpha for survey or interview guides where applicable. Quantitative analysis will proceed with descriptive statistics to profile AMS interventions and outcomes, followed by multilevel mixed-effects regression models to account for clustering at hospital and ward levels. Primary analyses will test associations between AMS intensity (composite score of adherence to core interventions) and antimicrobial consumption, resistance rates, and patient outcomes, adjusting for confounders such as patient comorbidity, severity of illness, and case mix. Time-series analyses will assess resistance trend trajectories. Qualitative data will be analyzed using thematic analysis, with coding anchored in the theoretical framework and emergent themes validated through member checking and investigator triangulation. Integration will follow a convergent design, with joint displays synthesizing quantitative and qualitative findings to identify concordant and discordant results. Expected findings include substantial inter-hospital variation in AMS impact, with higher fidelity to core AMS interventions correlating with reduced defined daily doses, stabilization or decline in resistance rates for key pathogens, shorter lengths of stay, and lower in-hospital mortality in tertiary centers compared with secondary centers after adjusting for confounders. The study anticipates identifying contextual determinants—leadership engagement, dedicated AMS personnel, information technology support, and inter-professional collaboration—that mediate the relationship between AMS practices and outcomes. The contribution to knowledge lies in providing robust, evidence-based cross-hospital benchmarks, a validated framework for assessing AMS effectiveness, and practical recommendations for resource allocation and policy harmonization to optimize antimicrobial use across diverse hospital settings. In conclusion, the research is expected to demonstrate that, beyond the mere existence of AMS programs, the depth, fidelity, and contextual integration of these programs determine their success. Based on the findings, recommendations will include scalable strategies for improving AMS implementation, targeted investment in stewardship personnel and informatics infrastructure, standardized outcome monitoring, and policy measures to promote continuity of AMS practices across hospital networks.