Application of publicly available nationally collected data for regional hospital quality improvement and disparity reduction initiatives

The study compared the nation as a whole and fourteen regions selected for participation in the RWJF Aligning Forces for Quality initiative (Cincinnati, Cleveland, Detroit, Humboldt County, Kansas City, Maine, Memphis, Minnesota, Seattle, Willamette Valley, Western Michigan, Western New York, Wisconsin, and York County, PA). We compiled regional statistical profiles from a systematic review of all major public data sources (AHRQ, CMS, AHA, AHRQ, NCHS, HRSA, JCAHO and US Census Bureau). The CMS MEDPAR was used to develop racial comparisons of hospital outcomes (medical and surgical death rates, incidents rates of safety problems).We used the CMS/JCHO HospitalCompare consensus process indicators (AMI, Heart Failure, Pneumonia, and Surgical Infection Prevention) and computed “structural disparity” indicators by region weighting a summary process indicator for each hospital in a region by the proportion of white and black discharges in the region as identified in the MEDPAR. The principal findings were: (1) While there were substantial variations between AF4Q regions on demographic, disease and health care measures, on the average these areas had somewhat lower poverty rates, more health care resources, lower age adjusted death rates and less racial disparities in death rates than the nation as a whole. (2) The degree of racial segregation in terms of hospital use by blacks and whites varied but was relatively low. The Overall Index of Dissimilarity was .333, and the most segregated regions were Wisconsin (.637), Minnesota (.543) and Detroit (.540). (3) There was no difference in the HospitalCompare process summary measure adjusted for where blacks and whites received care in the 14 regions (black/white ratio .99). (4) While black medical admission death rates were lower than whites (.70), surgical death rates were higher (1.21) as were incidents of safety problems (1.39).

We conclude that: (1) The moderate degree of segregation among hospital providers and the lack of variation between hospitals on HospitalCompare indicators contributed to the lack of regional racial disparity on these measures of hospital quality. (2) While there were significant differences between blacks and whites on the MEDPAR indicators, the low incidence rates make it impossible to compare the performance of individual hospitals and the complex factors contributing to these differences don't lend themselves easily to corrective action. (3) Hospitals are in the early stages of developing the internal capacity to analyze performance by race and ethnicity and the HospitalCompare process and MEDPAR outcome indicators have limited usefulness in comparing individual hospital performance.