Variation in Capabilities of Electronic Health Records for Quality Reporting

Background: Quality improvement (QI) programs often rely on data from electronic health records (EHRs) to measure clinician performance on clinical quality measures. H3: Healthy Hearts in the Heartland, a quality improvement program focused on cardiovascular quality of care in primary care practices in the Midwest, uses Physician Quality Reporting System (PQRS) reports to measure quality over time. During study implementation, we encountered differences between EHRs in PQRS reporting functionalities. We report here on a sub-study to identify PQRS reporting functions that can support QI work at participating practices. Methods: In the larger H3 trial, the n=226 practices used 13 different EHRs certified by the Office of the National Coordinator (ONC). We reached out to practice facilitators who had expertise working with individual EHRs and had them complete a brief survey about each EHR’s reporting capabilities. The survey focused on: (1) PQRS reporting time frame (e.g., calendar year, custom time frame, etc.) (2) whether results were reported at the practice level or by individual clinician (3) costs of generating reports, if any, and (4) whether lists of individual patients could be generated from report results. In cases where there were multiple content experts and there were disagreements, these were reconciled via discussion. For the purposes of QI work in the H3 study, we identified whether or not EHRs could generate PQRS reports for custom 1-year time frames (i.e., rolling 12-month periods). We report here on interim results. Results: All but 2 of the EHRs under study could produce reports for calendar years. Most (n=9, 69%) had the ability to produce reports for rolling 12-month periods. About half (n=6) could generate reports at both the practice level and by individual clinician. All EHRs offered at least some pre-set, standardized reports at no cost; also, 2 were known to allow custom reports at no additional cost. Report generation costs varied across sites using the same EHR depending on sites’ individual contracts with EHR vendors and subscriptions for optional EHR features. The majority of EHRs (n=9) had the option to produce patient lists from quality measure numerators or denominators. Conclusions: We observed wide variability in PQRS reporting functions among certified EHRs used at primary care practices. In addition, there was variation between practices using the same EHR vendor due to variable purchasing of optional EHR functions. This variability is not often reported, and has implications in certified EHRs’ ability to support QI work that requires repeated quality measurement efforts over time. Although our results need to be validated in other settings, they highlight potential challenges of using certified EHRs to improve the quality of care in practice.