The Influence of Context on Antimicrobial Prescribing for Febrile Respiratory Illness:
A Cohort Study
Courtney Hebert, MD; Jennifer Beaumont, MS; Gene Schwartz, MD; and Ari Robicsek, MD
Ann Intern Med. 7 August 2012;157(3):160-169
Background: Little is known about the influence of contextual factors on a physician’s likelihood to prescribe antimicrobials for febrile respiratory illness (FRI). Context includes epidemiologic context (for example, a pandemic period) and personal context (for example, recent exposure to other patients with FRI).
Objective: To examine the association between contextual factors and antimicrobial prescribing for FRI.
Design: 5.5-year retrospective cohort study.
Setting: A network of Midwestern primary care providers.
Patients: All patients presenting with FRI during influenza seasons between 2006 and 2011.
Measurements: Antimicrobial prescribing for FRI during pandemic and seasonal influenza periods.
Results: 28 301 unique patient encounters for FRI with 69 physicians in 26 practices were included. An antibiotic was prescribed in 12 795 (45.2%) cases. The range of prescribing among physicians was 17.9% to 83.7%. Antibiotics were prescribed in 47.5% of encounters during the seasonal period and 39.2% during the pandemic period (P < 0.001). After multivariable adjustment for patient and physician characteristics, antibiotic prescribing was lower in the pandemic period (odds ratio [OR], 0.72 [95% CI, 0.68 to 0.77]) than in the seasonal period. The likelihood of prescribing an antibiotic decreased as the number of FRI cases that a physician had seen in the previous week increased (OR, 0.93 [CI, 0.86 to 1.01] for 2 to 3 patients with FRI seen in the previous week; OR, 0.84 [CI, 0.77 to 0.91] for 4 to 6 patients; OR, 0.71 [CI, 0.64 to 0.78] for 7 to 11 patients; and OR, 0.57 [CI, 0.51 to 0.63] for ?12 patients compared with the reference range of 0 to 1 patients). Pandemic season and recent personal context were also associated with antiviral prescribing.
Limitation: Retrospective study in a single geographic area.
Conclusion: Epidemiologic context and the number of cases of FRI that a physician had recently seen were associated with his or her likelihood to prescribe antimicrobials for FRI. Interventions that enhance a physician’s contextual awareness may improve antimicrobial use.