The impact of upcoding, cream skimming and readmissions on hospitals’ efficiency: the case of Lombardy

This paper investigates the efficiency of Italian hospitals during the period 1998–2007 and consider the impact of some DRG distortions like uocoding, cream skimming and readmission. We apply both a stochastic frontier approach and a multilevel model. We show that readmissions are the most relevant distortion. Moreover, cream skimming has a negative impact for public and private hospitals and upcoding has a clear positive but little impact only for public hospitals. These results imply that the reaction of the different hospital types to the public policy aimed to reduced the DRG distortions is different. Private hospitals tend to reduce the DRGs while the public hospitals increase the readmissions. Last, private hospitals are less efficient than public and not–for–profit ones, once we take the DRG distortions (and other covariates) into account, both for the SF approach and the ML model.