A Comparative Study of Various Data Envelopment Analysis Models with Applications to Primary Healthcare Efficiency in Bahia, Brazil

Abstract

This thesis investigated how scale size should be handled in DEA models analyzing healthcare delivery in Brazil. Both empirical and theoretical results were used to determine the effect that the use of per capita data had on the DEA model. It was found that in the constant returns to scale setting , per capita data and total value data give the same efficiency scores and rankings in the DEA model. It was also found that results vary slightly between per capita and total value DEA models in the variable returns to scale setting. Future research should give more attention to the value of comparing per capita value DEA models with total value DEA models.