Geography in Frontier-Based Efficiency Measurement of Hospitals and Subnational Health Systems: A Methodological Critique and the GEO-EFF Reporting Framework
Year: 2026
Type: article
Abstract: Frontier-based efficiency measures increasingly influence health-economic benchmarking, resource allocation, and service-configuration decisions, yet hospitals and subnational health systems operate under markedly different geographic production conditions. This methodological paper argues that acknowledging geography is not equivalent to incorporating it into an efficiency model and that this distinction can alter the estimand, comparator set, ranking, and policy meaning of efficiency scores. To ground the critique and develop reporting guidance, we purposively assembled a source-verified maximum-variation set of 20 peer-reviewed empirical frontier studies spanning DEA- and SFA-family methods, spatial and meta-frontier designs, quality-sensitive models, and rural, regional, and national settings. Fifteen studies used DEA-family methods and six used SFA-family methods, with one applying both. Eight treated geographies through spatially explicit frontier or spatial-econometric models, six through second-stage contextual or environmental adjustment, three through group or meta-frontiers, one through multiple formal approaches, and two primarily as descriptive or implicit context. Eleven used a second-stage analysis, five directly modeled quality or undesirable outputs, and seven reported robustness procedures. These counts characterize the exemplar set by design and are not prevalence estimates. We distinguish four mechanisms often compressed into the term geography: regional technology heterogeneity, spatial dependence, network and access connectivity, and geographic cost conditions. We propose GEO-EFF, a six-domain, non-ordinal reporting framework for geography-responsive efficiency studies. GEO-EFF is a reporting prompt rather than a validated appraisal or scoring instrument and requires external validation
Author: Dariimaa Ganbat
Institution: Mongolian National University of Medical Sciences
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W7213316620