664 Antimicrobial Stewardship: Lessons from Pandemic Influenza

Saturday, March 20, 2010: 2:30 PM
International North (Hyatt Regency Atlanta)
Ari Robicsek, MD , NorthShore University HealthSystem, Evanston, IL
Gene Schwartz , NorthShore University HealthSystem, Evanston, IL
Rumi Semer , NorthShore University HealthSystem, Evanston, IL
Lance R. Peterson , NorthShore University HealthSystem, Evanston, IL

Background: Overuse of antibiotics is a common problem in inpatient and outpatient settings. The occurrence of pandemic influenza offered an opportunity to observe how the context of a patient's presentation affected a physician's prescribing decision.

Objective: To compare antibacterial and antiviral prescribing for patients presenting to primary care physicians with influenza-like illness (ILI) during usual seasonal flu and pandemic flu seasons.

Methods: A large physician network in Illinois shares an electronic medical record (EMR). A query was written that used chief complaint, encounter diagnoses and office temperature measurements to identify patients with ILI. All primary care office encounters from September 3, 2007 through November 1, 2009 were evaluated. Physicians who saw at least 250 cases of ILI during this time period, and worked in our network in 2007 through 2009, were included. For all patients with ILI, EMR abstraction was used to determine treatments prescribed. Four timeframes were compared: Usual flu season 1 (1/28 through 4/6, 2008); Usual flu season 2 (2/9 through 3/29, 2009); Pandemic flu season (4/27 through 6/14 and 10/12 through 11/1, 2009) and Non-flu season. Flu seasons were periods when, for 2 consecutive weeks, ≥ 6% of visits were for ILI.

Results: 44 physicians practicing at 27 sites were eligible for the study. Over the study period, these physicians had 333,740 patient encounters, 23,455 (7.0%) of which were for ILI. The percentage of ILI encounters in which antibiotics were prescribed was consistent for individual physicians between the two usual flu seasons (Pearson correlation coefficient = 0.81), but varied widely across the 44 physicians (17.0% 91.0%). The percentage of ILI encounters in which antivirals and antibiotics were prescribed during non-flu season, usual flu seasons, and pandemic flu season was 1.9% and 50.5%; 6.4% and 44.8%; 16.2% and 33.3% respectively (P < 0.01 for all differences). All but 3 of 44 physicians were less inclined to prescribe antibiotics in pandemic flu season than in usual flu seasons, and all but 5 of 44 physicians were more likely to prescribe antivirals for ILI during pandemic flu season (Figure). Despite the significant changes in prescribing patterns, the overall percentage of ILI encounters in which any agent (i.e. an antibiotic or an antiviral) was prescribed changed little, from 50.1% during usual flu season to 47.5% (P < 0.01) during pandemic flu season.

Conclusions: Physicians were less likely to prescribe antibiotics for ILI during flu season than the rest of the year, and were further less inclined to prescribe antibiotics during pandemic than usual flu season. This suggests the possibility that physicians base prescribing decisions not only on the presentation of an individual patient, but on community contextual information. This may provide an avenue through which physician education could drive more rational antimicrobial use.