Using composite indicators to evaluate the efficiency of health care system

This article produces a system-level composite indicator on population health in publicly provided primary health care. We first summarize information from various indicators of care by principal component analysis (PCA). We then regress the costs of care against these indicators to evaluate whether they are related. The existing health status indicator provides a point of comparison. Our results suggest that PCA can be used to extract information efficiently and thus to reduce the dimensionality of data. Furthermore, the use of the existing health status indicator to estimate cost-efficiency leads to invalid inference on overall efficiency, while the use of composite indicator significantly reduces the bias.