Evaluation of the HCFA model for the analysis of mortality following hospitalization.

摘要: From 1987 through 1990, the Health Care Financing Administration (HCFA) evaluated variations in mortality rates experienced by patients admitted to hospitals participating Medicare program. This study was conducted evaluate adequacy of model used for that purpose. Detailed clinical data were gathered on 42,773 84 statistically selected hospitals. The effect risk adjustment using HCFA model, which is based claims data, compared a risk-adjustment physiologic and data. Models include markedly superior those containing only demographic characteristics predicting probability patient death, addition resulted further improvement. correlation ranks uses one uses, addition, .91. As screen identification "high (mortality) outlier" hospitals, had moderate sensitivity (81 percent) specificity (79 percent), high negative predictive value (90 low positive (64 when model. two models gave similar results determine structural related rates: with higher proportion registered nurses or board-certified physician specialists, greater level access high-technology equipment lower risk-adjusted rates. These suggest current claims-based procedure may satisfactorily be characterize associated hospitalization. could also as basis epidemiological analyses factors affect death. However, it does not positively identify outlier providers problematic care.