Electronic Clinical Quality Measures of Antibiotic Use in Hospitalized Adults With Uncomplicated Community-Acquired Pneumonia

BACKGROUND: Community-acquired pneumonia (CAP) is the top indication for inpatient antibiotic use and overuse. We describe the development of 2 electronic clinical quality measures (eCQMs) assessing the percentage of hospitalized patients with uncomplicated CAP who had unsupported antibiotic duration beyond 5 days and empiric antibiotics targeting methicillin-resistant Staphylococcus aureus and/or Pseudomonas aeruginosa. METHODS: Both eCQMs were developed based on clinical guidelines, literature, and expert feedback and then specified electronically. The eCQMs were applied to patient data from 2 academic medical centers and 109 Department of Veterans Affairs (VA) healthcare systems to evaluate importance, reliability, validity, and feasibility, as defined by the Centers for Medicare and Medicaid Services. RESULTS: Across all hospitals, the mean percentage of eligible patients with unsupported excess duration and inappropriate empiric antibiotic use ranged from 40% to 55% and from 17% to 56%, respectively. Within the VA cohort, reliability (median [first-tenth decile, n]) for duration and empiric eCQMs was high: 93% (63%-97%, 28 238) and 91% (63%-96%, 47 034), respectively. When compared to chart review at 1 academic hospital, both eCQMs had high sensitivity (96%, both) and specificity (92% and 93% for duration and empiric eCQMs, respectively). Feasibility was maximized by using only data elements in structured electronic health record fields. Some technically complex elements (eg, clinical stability) were retained for face validity. CONCLUSIONS: Two eCQMs identifying unsupported excess antibiotic duration and inappropriately broad empiric antibiotic use in patients with uncomplicated CAP had high importance, reliability, validity, and feasibility. These eCQMs could be benchmarked to assist antibiotic stewardship programs in understanding and improving their antibiotic use.