Journal Design Emerald Editorial
African Food Systems Research (Interdisciplinary - incl Agri/Env) | 02 September 2011

A Quasi-Experimental Evaluation of Systems Optimisation for Improved Diagnostic Yield in Tanzanian District Hospitals

J, u, m, a, M, f, i, n, a, n, g, a, ,, G, r, a, c, e, M, w, a, k, y, u, s, a
Health SystemsDiagnostic YieldQuasi-ExperimentalTanzania
Intervention linked to 2.45x higher odds of conclusive diagnosis (95% CI: 1.78–3.37).
Absolute diagnostic yield improved by 18.7 percentage points post-implementation.
Integrated approach combined workflow, decision support, and staff training.
Quasi-experimental design provides robust causal estimates in real-world setting.

Abstract

{ "background": "Diagnostic yield in sub-Saharan African district hospitals is often suboptimal, constrained by fragmented and under-resourced clinical systems. Systems optimisation interventions are frequently implemented, but robust methodological evaluations of their impact on diagnostic outcomes are lacking.", "purpose and objectives": "This study aimed to evaluate the impact of a multi-component systems optimisation package on diagnostic yield in a real-world hospital setting, using a quasi-experimental design to estimate causal effects.", "methodology": "A controlled interrupted time-series analysis was conducted across four matched district hospitals. Two hospitals received the intervention, which integrated laboratory workflow re-engineering, clinical decision support algorithms, and targeted staff training. The primary outcome was the monthly proportion of conclusive diagnoses from admitted patients. The effect was estimated using a generalised linear mixed model: $\\logit(\\pi{it}) = \\beta0 + \\beta1 Tt + \\beta2 X{it} + \\beta3 (Tt \\times X{it}) + ui + \\epsilon{it}$, where $\\pi{it}$ is the probability of a conclusive diagnosis in hospital $i$ at time $t$, $Tt$ is the post-intervention period, and $X{it}$ is the intervention indicator. Robust standard errors were clustered at the hospital level.", "findings": "The intervention was associated with a significant increase in diagnostic yield. The adjusted odds ratio for a conclusive diagnosis was 2.45 (95% CI: 1.78 to 3.37). The absolute improvement in the intervention hospitals was 18.7 percentage points (from a baseline of 54.3% to 73.0%) post-implementation, a change not observed in control sites.", "conclusion": "A structured systems optimisation package can substantially improve diagnostic conclusiveness in resource-constrained hospital settings. The findings demonstrate that synergistic adjustments to laboratory, clinical, and human resource subsystems can effectively mitigate diagnostic delays and uncertainties.", "recommendations": "Health policymakers