Empiric antibiotic prescribing for suspected acute cystitis may lead to unnecessary prescriptions when urine cultures are negative. This study assessed whether physician overestimation of the likelihood of bacterial infection contributed to unnecessary antibiotic prescriptions.
This was a cross-sectional study in Toronto, Canada, from 1998 to 2000 of 231 women 16 years and older who underwent standardized clinical assessments and urine culture testing. The main outcome was an unnecessary antibiotic prescription, defined as a prescription where the urine culture was negative. The difference between physician estimates of the likelihood of a positive urine culture and the measured culture rate for women with similar symptoms was used to measure overestimation error. Logistic regression was used to assess associations between unnecessary prescriptions and clinical factors or overestimation error. Multiple logistic regression was used to adjust for the effect of clinical factors.
Of 230 women assessed, 186 (80.9%) were prescribed antibiotics and 74 (32.2%) were prescribed an unnecessary antibiotic where the urine culture was negative. When an overestimation error above the median value (14.75%) was present, the odds of an unnecessary antibiotic prescription were increased (adjusted odds ratio = 3.72; 95% confidence interval = 1.75-7.89). A high overestimation error was associated with the symptoms of urinary frequency or suprapubic tenderness and costovertebral angle tenderness on examination.
Physician overestimation of the likelihood of a positive urine culture in women with symptoms of acute cystitis was associated with unnecessary antibiotic prescribing. Antibiotic overuse may be reduced by developing treatment strategies that deemphasize nonspecific clinical findings that contribute to physician overestimation error.