Production Frontier Analysis and Organizational Performance Assessment in St. Petersburg Healthcare System

Abstract

The paper deals with measurement of organizational performance in healthcare. It critically analyses the measurement procedure of healthcare organizations’ performance introduced in the Order of St. Petersburg Healthcare Committee № 439-p of 05.11.2013 “On approval of indicators and criteria for performance assessment of St. Petersburg public healthcare organizations, their administration and personnel”. The main shortcomings of the performance measure introduced by this order can be summarized as follows: weak justification of the list of attributes to be assessed; poor metrics for the performance attributes employed; additivity of the overall organizational performance measure; weak incentives for improving performance provided to healthcare organizations. An alternative approach to measure the performance of healthcare organizations that relies on the production frontier analysis methodology allows to overcome the weaknesses of the organizational performance measure setup by the order. The production frontier for St. Petersburg public in-patient hospitals is constructed with nonparametric Data Envelopment Analysis (DEA) model. It estimates the DEA efficiency scores provided by the BCC-output model with two inputs (medical staff salary and materials costs paid from the Obligatory Medical Insurance Fund) and two outputs (the number of properly treated and improperly treated patients). The distinctive feature of the proposed model is explication of the undesirable output (the number of improperly treated patients) achieved through normalization of the variables with polarized maxmin rule. The empirical base of the study includes the input-output data on 47 St. Petersburg public in-patient hospitals reported in 2013–2015. Analysis of the St. Petersburg public in-patient hospitals’ performance estimates in 2013–2015 argues in favor of their consistency and, hence, applicability of the proposed measurement model.