Assessing Potentially Preventable Hospitalizations at the County Level: A Comparison of Measures Using Medicare Data and State Hospital Discharge Data.

Abstract

This study examined the representativeness of the County Health Rankings and Roadmaps (CHR) measure of potentially preventable hospitalizations, which is derived from Medicare inpatient claims data, as an indicator of potentially preventable hospitalizations for adults aged ≥18 years. Potentially preventable hospitalizations were evaluated using rates of ambulatory care sensitive conditions (ACSCs). CHR rates of hospitalization for ACSCs based on Medicare data for 2010, Agency for Healthcare Research and Quality Prevention Quality Indicator #90 Overall Composite (PQI #90 Composite) rates of ACSCs based on hospital inpatient data for adults aged ≥18 years for 2011, and 2011 total mortality rates for adults aged ≥18 years for 212 counties in 3 US states were evaluated. Pearson correlation analyses were used to assess the linear association between the PQI #90 Composite and CHR rates of hospitalization for ACSCs as well as associations of these measures with total mortality. Steiger's Z-test was conducted to examine whether the PQI #90 Composite and CHR measures of health care quality were similarly correlated with total mortality. The age- and sex-adjusted PQI #90 Composite for adults ≥18 years was moderately correlated with the CHR rate of hospitalization for ACSCs. The PQI #90 Composite and CHR measures of hospitalization for ACSCs were similarly correlated with mortality. These findings suggest that in the absence of easily accessible, high-quality data for adults aged ≥18 years, the CHR measure of potentially preventable hospitalizations provides a modest but acceptable approximation of county-level disparities in potentially preventable hospitalizations for the US adult population.