Medicare Reimbursement and the Quality of Hospital Care

This study examines the relationship between hospital reimbursement per discharge and the clinical quality of care received by Medicare patients before and after implementation of the Prospective Payment System (PPS) in 1983/84. The objectives are to evaluate the link between program payments and quality, and to identify characteristics of higher and lower quality hospitals in the period 1981-1986. Hospital quality was assessed by an existing measure of clinical process and by patient mortality rates for five specific diseases. The financial, market and organizational variables that measure hospital structure were used to predict clinical process, changes over time in clinical process and mortality rates. The level of quality amongst hospitals and changes over time were examined in relation to average payment, average costs, and profitability per Medicare discharge at 297 hospitals in five states. The relationship between Medicare payment per patient discharge and the clinical quality of care was relatively weak. It was statistically significant after PPS, although confounded by the payment formula factors for major teaching and rural hospitals. While most hospitals increased their cost per discharge after PPS and quality improved, many hospitals cut cost per discharge and also improved clinical process. Market variables accounted for the majority of explained variation in clinical process. When reimbursement and county level market variables were included in the model, the organization variables were not significantly associated with quality before or after PPS. There was a significant negative association between a high proportion of hospital Medicaid discharges and clinical process scores throughout the study period. Over time rural hospitals improved their quality the most. Ten percent of the hospitals were observed to decline in quality, measured by clinical process, yet there was no consistent pattern amongst them in terms of cutting costs, profitability, ownership or location. At the hospital level there was a wide degree of variation observed in clinical process provided to patients, cross-sectionally and in the degree of improvement in process during the study period. This variation requires further study to assist in better targeting of Medicare program activity to monitor the quality of hospital services provided to beneficiaries and to improve the reimbursement incentives for better quality patient care.