The use of a computer-assisted medical record audit (CAMRA) for reviewing antibiotic use in ambulatory patients was evaluated. A random sample of 40 medical records documenting treatment of streptococcal pharyngitis, otitis media, or acute, uncomplicated urinary tract infections at two family practice clinics was used to evaluate the accuracy and efficiency of computerized prescription screening and CAMRA relative to medical record audit (MRA) alone. Accuracy was the ability to correctly classify antibiotic therapy as appropriate or inappropriate. The initial computerized prescription screening criteria were modified to reduce the proportion of false positives and negatives and a second random sample of 40 medical records was audited. The computerized prescription screening was the most efficient method, requiring less than one hour of professional time to audit 80 medical records. MRA and CAMRA took 17.4 and 3.0 hours, respectively. By definition, MRA was 100% accurate. Computerized prescription screening and CAMRA correctly classified 73% and 78% of the medical records, respectively. The results of this study are similar to a previous study reviewing antihypertensive therapy, but this study showed CAMRA less favorably. This is primarily because of the many diagnoses for which a particular antibiotic can be prescribed and the wide dosage ranges for antibiotics based on body weight. CAMRA could be more useful for evaluating antibiotic therapy if diagnostic information were available before doing the computerized prescription screening and if the computerized prescription screening criteria included patient diagnosis and body weight.