When Japanese hospitals try to adopt a benchmarking method, we will find three unique aspects related to hospital management. Firstly, it has not been confirmed what critical factors are for managing hospitals. Secondly, there is difficulty when grouping various hospitals on equal footing. Thirdly, we cannot designate priority among various performance indicators in hospital administration.
This study tries to solve these problems related to benchmarking in hospital management and develop a benchmarking method for Japanese hospitals using a multivariate statistical technique.
First of all, we selected five important performance indicators from 15 various indicators using principal component analysis. Then, we classified sample hospitals in five groups using cluster analysis by two exogenous factors among five factors. Thirdly, we made the assumption that the objective function would be to break even on the balance sheet of hospitals. Under this assumption we did logistic regression analysis to find important performance indicators for breaking even.
According to these analyses, we recognize that the performance indicators, which effect the balance statistically, would differ among hospital groups on equal footing.