Comparing case-mix systems for nursing home payment.

Abstract

Case-mix systems for nursing homes use resident characteristics to predict the relative use of resources. Seven systems are compared in structure, accuracy in explaining resource use, group homogeneity, and ability to identify residents receiving heavy care. Resource utilization groups, version II (RUG-II), was almost uniformly the best system, although management minutes and the Minnesota case-mix system were also highly effective. Relative weights for case-mix groups were sensitive to cost differences and should be recomputed for new applications. Multiple criteria should be used in choosing a case-mix system, including consideration of inherent incentives and how residents' characteristics are defined.