Abstract
{ "background": "District hospital systems are critical for food system resilience, as malnutrition and food-borne illnesses directly impact agricultural productivity and community health. However, methodological limitations in measuring system yield—defined as effective service output per unit input—hinder evidence-based resource allocation and intervention planning.", "purpose and objectives": "This intervention study aimed to methodologically evaluate panel-data estimation techniques for quantifying yield improvements in district hospital systems, and to apply the optimal model to generate robust, long-term yield projections.", "methodology": "We conducted a quasi-experimental analysis using a balanced panel of district-level administrative data. The core specification was a two-way fixed effects model: $Y{it} = \\beta0 + \\beta1Intervention{it} + \\mui + \\lambdat + \\epsilon{it}$, where $Y{it}$ is the system yield index. Model robustness was assessed using Driscoll-Kraay standard errors to account for cross-sectional dependence and serial correlation.", "findings": "The methodological evaluation identified the fixed effects estimator as superior for this context. Application of the model projected a significant positive trajectory, with system yield estimated to increase by approximately 18.5% (95% CI: 14.2, 22.8) over the projection period following targeted infrastructural and training interventions.", "conclusion": "Panel-data methods provide a rigorous framework for estimating and forecasting health system yields, offering vital evidence for strengthening the health-agriculture nexus.", "recommendations": "Health and agricultural policymakers should adopt panel-data estimation for routine monitoring of system performance. Investment should prioritise the integrated interventions identified as yield-enhancing, particularly in underperforming districts.", "key words": "health systems, panel data, fixed effects, yield, forecasting, intervention, South Africa", "contribution statement": "This paper provides a novel methodological framework for estimating health system yield, generating the first long-term, district-level projections for the country, which directly inform cross-sectoral food system policy