Analysis of X-Inefficiency in U.S. Hospitals. Abstract, Executive Summary andFinal Report

摘要

The study examined the correlates of X-inefficiency in 1,996 U.S. urban hospitalsin 1997. Stochastic frontier analysis was used to estimate inefficiency scores for each hospital in the study. The estimated level of X-inefficiency ranged from 3 percent to 84.2%, with a mean of 12.6%. Preliminary results confirm many of the hypotheses developed from X-inefficiency Theory. Results suggest that X-inefficiency is inversely associated with financial pressures and competition. Further, not-for-profit hospitals were more X-inefficient than their for-profit counterparts. The estimated inefficiency scores were found to be robust over a variety of model specifications. The inefficiency estimates also had anticipated correlations with a number of variables expected to be associated with inefficiency, including teaching status, size, cost per adjusted admission, for-profit status, FTE staff per patient day, and profitability. These correlations serve to cross validate the X-inefficiency estimates.