Antibiotic prescriptions for respiratory infections in VA patients high, varied by clinician

An 8-year Veterans Affairs study showed that veterans with acute respiratory infections (ARIs) commonly receive antibiotics—a practice discouraged by guidelines—regardless of patient or setting characteristics and with substantial variation at the clinician level.

The study, published on July 21 by Annals of Internal Medicine, found a “persistently high prevalence” of outpatient antibiotic prescriptions for ARIs in 1,044,523 patient visits. From Jan. 1, 2005, to Dec. 31, 2012, the proportion of visits with ARI diagnoses that resulted in antibiotic prescriptions increased from 67.5% to 69.2% (P<0.001).

The researchers noted substantial variation in antibiotic prescribing at the clinician level. The 10% of clinicians who prescribed the most antibiotics did so during at least 95% of their ARI visits. The 10% who prescribed the least did so in 40% or fewer of their ARI visits, according to the study. This clinician-level variation remained after the researchers adjusted for patient characteristics, and clinicians differed widely within clinics, according to the study. Prescribing was slightly higher at VA medical center-based clinics than at community-based outpatient clinics (70% vs. 64%), the study found.

“Variation in ARI management seems to be strongly influenced by the prescribing patterns of individual providers,” the authors wrote. “This is a ripe target for further research, quality improvement, and antibiotic stewardship interventions.”

Macrolides have become the predominant antibiotic class prescribed, making up 43.4% of the antibiotics prescribed, according to the study. The proportion of macrolides prescribed rose from 36.8% in 2005 to 47% (P<0.001) in 2012, the study found.

“This trend is concerning given the lack of additional benefit of macrolides over narrow-spectrum antibiotics for ARI treatment; the increase in macrolide-resistant pneumococcal disease; and potential cardiotoxicity, especially considering the large number of veterans in our study who had cardiovascular comorbid conditions,” the study's authors wrote. In contrast, prescriptions of penicillins (36% to 32.1%, P<0.001) and fluoroquinolones (15% to 12.7%, P<0.001) decreased over time.

Antibiotic prescribing was highest for sinusitis (adjusted proportion, 86%) and bronchitis (85%) and varied little according to age, setting, or comorbid conditions. The researchers noted that antibiotic prescribing was associated with many factors, such as temperature, distance to clinic, setting type, and geographic region, although these associations were small.

The authors noted several limitations to their study, including exclusion of patients with comorbid conditions and definitively diagnosed infections in an attempt to identify cases with low likelihood of benefit from antibiotics. The study also lacked certain additional data on factors contributing to the prescribing decision, such as duration of symptoms, physical exam findings, or clinician specialty, the authors stated.