Evaluating Hospital Efficiency and Productivity: A Bootstrap-Corrected DEA and Malmquist Analysis

Background/Objectives: Efficient resource allocation is essential for the sustainability of public hospitals operating under financial and capacity constraints. This study evaluated the volume-based technical efficiency and productivity of public hospitals within a Greek Health Region and explored associations between efficiency and selected contextual characteristics. Methods: Of 15 public hospitals in the study setting, two were excluded on production-comparability grounds, leaving 13 hospitals analyzed over 2015-2021 using input-oriented Data Envelopment Analysis under constant and variable returns to scale. Bootstrap bias correction was applied to CRS efficiency estimates, while scale efficiency was derived from CRS and VRS scores. Exploratory associations between 2021 bias-corrected efficiency and population ageing, population density, and unemployment were examined using Simar-Wilson double-bootstrap truncated regression. Productivity change was assessed using the Malmquist Productivity Index. Results: Mean bias-corrected CRS efficiency ranged from 0.7486 to 0.8556, and no hospital achieved a bias-corrected score of 1. Mean scale efficiency ranged from 0.8981 to 0.9416. Population ageing, population density, and unemployment were negatively associated with efficiency in the primary second-stage specification, with coefficient directions remaining broadly consistent across sensitivity models. Malmquist analysis indicated an overall decline in productivity, driven predominantly by shifts in the estimated production frontier, with substantial year-to-year variation. Conclusions: Hospital efficiency and productivity varied substantially across hospitals and over time. DEA can support regional performance review, but results should be interpreted alongside case mix, quality, accessibility, population need, and local service responsibilities rather than as stand-alone performance rankings.