Maximizing quality of care data through a health center data warehouse

Background: Federally qualified health centers (HC) have been leaders in ambulatory care quality improvement (QI) through the Health Disparities Collaboratives (HDC) which created a venue for exchange of best practices among health centers. The HDC created national benchmarks for health center performance; however, it lacked a systematic way to assess and compare the quality performance of HCs based on their practice characteristics (e.g. size, geography) and populations served (e.g. race/ethnicity, migrant workers, homeless). Furthermore, since practice level QI activities are often specific to the HC environment or population served, identifying QI practices from HCs with similar characteristics is crucial. Creating detailed quality benchmarks and performance targets based on characteristics by population served, health services provided, and health systems environment, may provide HCs with more useful information for planning QI strategies and partnering with local public health agencies. This project is a partnership between an academic institution and a primary care association which hosts quality of care data for 150 health centers in 24 states in a data warehouse (DW). The health centers collectively serve over one million patients. The main objectives of this project is to 1) assess data warehouse function and capacity for collecting, storing, and analyzing quality data, 2) describe the quality performance of HCs including creating HC-specific benchmarks and comparisons, and 3) identify best practices by profiling high performing health centers. Methods: All patient information stored in the data warehouse in PECS format were combined for analysis. Summary statistics were generated at the data warehouse and health center level. Detailed quality of care reports were generated for major chronic diseases (diabetes, cardiovascular disease, asthma, depression). Process measures and outcomes measures were selected and benchmarks were created based on health center location (state, rural/urban), practice characteristics (size, scope of services, financial indicators), and population characteristics. Results/Outcomes: The DW was able to combine patient data from 110 health centers using PECS registry for analysis (N=182,177). The combined dataset contains patients with diabetes (90,076), hypertension (69,987), cardiovascular disease (121,266), asthma (10,601), and depression (22,644). Patient characteristics include demographics, insurance status, and over 150 variables to generate process and outcomes measures. Quality measures were successfully coupled with various HC characteristics to create differential benchmarks. Conclusions: The HC DW was successful in differentiating quality performance based on HC characteristics. The DW and its analytical approach show great promise for creating useful information to improve HC quality of care.