Hierarchical Performance and Unobservable Heterogeneity in Health: A Dual-Level Efficiency Approach Applied to NHS Pathology in England

Understanding the source of inefficiency within health system organisational structures is a key aspect of performance measurement and management; and is of increasing importance to policy makers. This study uses a unique panel dataset to study the efficiency performance of pathology services in the National Health Service (NHS) in England for the first time. We apply a dual-level stochastic frontier (DLSF) model (J Prod Anal 37(1):1–27, 2012) to isolate the source of inefficiency at two vertically distinct organisational levels: an upper level of Strategic Health Authorities (SHAs); and a lower level of laboratories grouped within SHAs. Wedevelop the DLSF framework—in line with recent developments in the wider panel data literature—to control for the influence unobserved heterogeneity, which is a key issue for healthcare performance analysis. We find statistically significant variation in inefficiency performance at both organisational levels in pathology services. Weuse these measures to compute overall inefficiency for NHS pathology services, and corresponding savings estimates. Finally we comment on the wider modelling implications of our research with respect to the separation of inefficiency and unobserved heterogeneity when applied to multi-level data structures.