The use of goal programming, regression analysis and data envelopment analysis for estimating efficient marginal costs of hospital services

This paper presents a new composite methodology for estimating efficient marginal costs of outputs. The methodology is based on multi-criteria methods involving Data Envelopment Analysis (DEA), Goal Programming (GP) and Regression Analysis (RA) techniques. Firstly, DEA is used to find an empirical frontier onto which each observed dependent variable (cost) is projected. Then GP or RA is applied to the projected data set to yield the parameter estimates of a cost function. That is, firstly, the amounts of technical inefficiency are removed and then the cost function is estimated. The parameters of the cost function provide the efficient marginal costs of outputs. The filtering of the data can be made by means of DEA as it yields target values with frontier rather than average behaviour. The results clearly indicate the superiority of the combination of DEA and GP over the combination of DEA and RA. The methodology was applied to estimate the efficient marginal costs of hospital services (inpatient days, outpatient visits, ancillary services) of public general hospitals in Greece. Concluding, the analysis proves that the estimation of efficient marginal costs of outputs is in author's view particularly useful for hospitals, and can serve as a theoretical basis for determining the prices of different hospital services.