Accounting for heterogeneity in the measurement of hospital performance

With prospective payment of hospitals becoming more common, measuring their performance is gaining in importance. However, the standard cost frontier model yields biased efficiency scores because it ignores technological heterogeneity between hospitals. In this paper, efficiency scores are derived from a random intercept and an extended random parameter frontier model, designed to overcome the problem of unobserved heterogeneity in stochastic frontier analysis. Using a sample of 100 Swiss hospitals covering the years 2004 to 2007 and applying Bayesian inference, significant heterogeneity is found, suggesting rejection of the standard cost frontier model. Estimated inefficiency decreases even below the 14 percent reported by Hollingsworth (2008) for European countries. Accounting for unobserved heterogeneity would make hospitals rated below 85 percent efficiency according to the standard model gain up to 12 percentage points, serving to highlight the importance of heterogeneity correction in the estimation of hospital performance.