How efficient are Kenyan hospitals? : an application of frontier analysis techniques

Efficiency is an important concern for health systems. This includes delivery of the health care, health financing, and investment on hospital facilities and the management of health facilities. Measurement of efficiency in health facilities is important to ensure maximum allocation and utilization of limited resources. The aim of this study was to estimate efficiency in level IV Kenyan hospitals using data envelopment analysis and stochastic frontier analysis. Panel data were collected from 27 public and faith-based hospitals between 2008 and 2012. Data envelopment analysis (DEA), a non-parametric approach and stochastic frontier analysis (SF A) a parametric approach were applied to the data. Ownership as a factor of efficiency was assessed from the collected samples. The results show evidence of technical inefficiencies across the hospitals. Based on DEA bootstrapped model, the efficiency scores was 0.7597 and 0.7751 for 2011 and 2012 data respectively. Using the cross sectional data set, SFA values were comparable to DEA with an average of 0.7919 and 0.7701 for the 2011 and 2012 data sets respectively. Based on the panel data, the SF A model gave a range of scores that were between 0.62 (Pitt and Lee and Battese and Coelli) and 0.85 (true effect models). There was no evidence of patterns in efficiency scores over time based on both DEA and SF A approaches. This data did not suggest a significant effect on efficiency based on hospital ownership. In conclusion, this study shows presence of technical inefficiencies in Kenyan hospitals. It also provides a platform in exploring further the frontier techniques and incorporating ownership when measuring efficiency in Kenyan hospitals