SEARCH

SEARCH BY CITATION

Keywords:

  • antibiotics;
  • acute respiratory tract infections;
  • clinical trial;
  • physician practice patterns

Objectives: To measure and improve antibiotic use for acute respiratory tract infections (ARIs) in the elderly.

Design: Prospective, nonrandomized controlled trial.

Setting: Ambulatory office practices in Denver metropolitan area (n=4 intervention practices; n=51 control practices).

Participants: Consecutive patients enrolled in a Medicare managed care program who were diagnosed with ARIs during baseline (winter 2000/2001) and intervention (winter 2001/2002) periods. A total of 4,270 patient visits were analyzed (including 341 patient visits in intervention practices).

Intervention: Appropriate antibiotic use and antibiotic resistance educational materials were mailed to intervention practice households. Waiting and examination room posters were provided to intervention office practices.

Measurements: Antibiotic prescription rates, based on administrative office visit and pharmacy data, for total and condition-specific ARIs.

Results: There was wide variation in antibiotic prescription rates for ARIs across unique practices, ranging from 21% to 88% (median=54%). Antibiotic prescription rates varied little by patient age, sex, and underlying chronic lung disease. Prescription rates varied by diagnosis: sinusitis (69%), bronchitis (59%), pharyngitis (50%), and nonspecific upper respiratory tract infection (26%). The educational intervention was not associated with greater reduction in antibiotic prescription rates for total or condition-specific ARIs beyond a modest secular trend (P=.79).

Conclusion: Wide variation in antibiotic prescription rates suggests that quality improvement efforts are needed to optimize antibiotic use in the elderly. In the setting of an ongoing physician intervention, a patient education intervention had little effect. Factors other than patient expectations and demands may play a stronger role in antibiotic treatment decisions in elderly populations.