1813. Development and Validation of Novel Ambulatory Antibiotic Stewardship Metrics

Background Over 260 million antibiotic courses are prescribed in ambulatory settings per year in the United States: 41% of which are for acute respiratory tract infections (ARTI). Over 50% of these antibiotic courses are inappropriate. However, interventions to improve ambulatory prescribing are little studied, and metrics to track antibiotic use are not well validated. Methods To validate metrics for ARTIs in adults, we conducted a retrospective cohort study from January 1, 2016 to December 31, 2016 at 32 primary care practices. We randomly selected 1,200 office visits with a coded respiratory tract diagnosis and determined by medical record review the proportion of visits in which antibiotic prescription was inappropriate using modified Infectious Diseases Society of America treatment guidelines. We determined clinic and provider characteristics associated with inappropriate prescribing. By linear regression, we also determined the aggregate metrics best correlated with inappropriate antibiotic prescribing. Results An antibiotic was prescribed in 37% of visits in which a respiratory tract diagnosis was coded. Of these prescriptions, 69% were inappropriate. Demographics associated with inappropriate prescribing included advance practice provider vs. physician (72% vs. 58%, P = 0.02), family medicine vs. internal medicine (75% vs. 63%, P = 0.01), board certification after vs. before 1997 (75% vs. 63%, P = 0.02), and practice in a non-teaching vs. teaching clinic (73% vs. 51%, P < 0.001). Rate of antibiotic prescribing in visits where any respiratory tract diagnosis was coded (R2 = 0.23, P < 0.001) and rate of antibiotic prescribing in visits where a respiratory tract diagnosis that almost never requires an antibiotic was coded (R2 = 0.24, P < 0.0001) were most strongly correlated with inappropriate prescribing. Conclusion Rate of antibiotic prescribing in visits where any respiratory tract diagnosis was coded and rate of antibiotic prescribing in visits where a respiratory tract diagnosis that almost never requires an antibiotic was coded may be useful proxies to estimate the rate of inappropriate prescribing for ARTIs. This study could inform ambulatory antibiotic benchmarking metrics and interventions to decrease inappropriate antibiotic prescribing for ARTIs in ambulatory settings. Disclosures All authors: No reported disclosures.