Lindsay D Visnovsky, Brittany Kent, Elizabeth D Rutter, Jorie M Butler, Alec B Chapman, Makoto M Jones, McKenna R Nevers, Peter A Taber, Jian Ying, David C Classen, Barbara E Jones
An eCQM measuring diagnostic quality for hospitalized pneumonia patients was feasible, accurate, and scalable for implementation at two different healthcare systems. Approximately 10% of hospitalizations with a pneumonia diagnosis lack chest imaging confirmation, demonstrating clinically significant variation and opportunities for improvement through real-time measurement and diagnostic decision support.
BACKGROUND: Traditional quality measures in pneumonia require manual chart review and focus on treatment and outcomes, resulting in inconsistent target populations and missed opportunities to improve diagnosis. Advances in electronic health records (EHRs) enable the use of electronic clinical quality measures (eCQMs), which can simultaneously improve diagnosis at the point of care and provide a consistent target population for other quality measures. The authors developed and validated an eCQM for pneumonia diagnostic quality for feasibility, accuracy, and interoperability across systems.
METHODS: Using EHR data from 2015-2022 in two US health systems (federal and academic), the research team developed, validated, and implemented an eCQM that assesses the percentage of patients hospitalized with pneumonia who also have standard confirmatory chest imaging using routinely collected EHR data. Measure accuracy was validated using clinician review of 102 charts as a reference standard. Pneumonia diagnostic quality-the percentage of pneumonia hospitalizations with a positive chest image-was calculated overall and for each facility to examine variation.
RESULTS: The eCQM demonstrated high sensitivity (federal system: 98.3%; academic system: 97.8%) and positive predictive value (100.0% both systems). Among 92,797 hospitalizations for pneumonia at 101 hospitals across both systems, overall diagnostic quality was 90.1%, with variation across facilities (100.0% in the highest-performing hospital vs. 78.3% in the lowest-performing hospital; 83.5% (standard deviation [SD] 2.6%) in the lowest-performing decile vs. 98.3% (SD 1.3%) in the highest-performing decile).
CONCLUSION: An eCQM measuring diagnostic quality for hospitalized pneumonia patients was feasible, accurate, and scalable for implementation at two different healthcare systems. Approximately 10% of hospitalizations with a pneumonia diagnosis lack chest imaging confirmation, demonstrating clinically significant variation and opportunities for improvement through real-time measurement and diagnostic decision support.