An analysis of the Prospective Payment System's labor-nonlabor share by diagnosis-related group.

When Congress in 1983 legislated a new Prospective Payment System (PPS) for Medicare hospital payment, the payment algorithm was founded on a simplifying assumption of a constant 80-20 percentage share of labor and nonlabor costs across all diagnosis-related groups (DRGs). Using Medicare claims data and hospital cost reports, this study examines the accuracy of this assumption. While a few DRGs are found to vary significantly from the norm, a systematic cancelling out of high and low labor-intensive DRGs results in no material PPS payment bias at the hospital level. Indeed, rural hospitals, if anything, benefit by the assumption. A very small number of outlier DRGs and hospitals are troublesome, nonetheless, implying fine-tuning of the algorithm.