The PopGrouper as a tool for morbidity adjustment in regional comparisons of health care: an analytical framework

Analyzing regional variations can help improve equity, efficiency, and quality in health care provision. The PopGrouper is a population-based classification system which classifies persons with similar health care needs into distinct groups. It exhibits a high degree of morbidity differentiation. We present an analytical framework to use the PopGrouper in examining regional variations across different outcomes and populations using routine patient-level data.

We develop a two-step empirical strategy to examine the relative regional performance on a set of efficiency and quality outcomes (e.g., hospital bed days, cost of care, mortality). First, we propose PopGroup-standardized observed-to-expected ratios to compare regional performance. Second, we develop a multilevel regression model to separately estimate regional variation related to patient need measured by the PopGroup and variation related to regional characteristics.

We provide an analytical framework that demonstrates the PopGrouper's application as a tool for morbidity adjustment in the assessment of relative regional performance in efficiency and quality outcomes and the regional characteristics that explain this performance. We provide suggestions for empirical notation, interpretation of results, and graphical analyses of findings. The developed framework will be applied in subsequent empirical papers.

This paper sets the analytical foundations to be applied in regional comparative analyses using the PopGrouper allowing for conclusions about unexplained variations in quality and efficiency of health care.