Hospital efficiency in Türkiye via spatial Stochastic Frontier Analysis

One of the primary challenges in efficiency literature arises from dealing with cases involving datasets with multiple outputs for decision-making units. While Data Envelopment Analysis naturally handles multiple outputs in efficiency calculation, Stochastic Frontier Analysis also effectively accommodates multiple outputs under certain modelling conditions. To estimate a spatial Stochastic Frontier model, we utilized a distinct dataset with multiple outputs, which were consolidated into a single output using a newly developed approach. We applied the methodology to Turkish hospitals and found statistically significant spatial dependence in hospitals’ technical efficiency. The median of efficiency estimates is 79.3%. The mean total output elasticities are 0.1943 for beds, 0.1186 for physicians, 0.3221 for specialist physicians, and 0.3417 for nurses. For the hospitals under consideration, the returns to scale is very close to 1. Our results show the influence of regional factors on hospital performance. These findings highlight the importance of considering spatial interactions and regional policies to improve hospital efficiency.