The association between the quality of physician prescribing and the degree of prescription documentation in the medical record was studied in four family-practice offices. Carbon-copy prescriptions were collected for five months at each office, and 438 were randomly selected for evaluation. For each prescription, a case abstract was prepared from the medical record. The abstract summarized the patient's previous care related to the problem for which the prescription was issued. The abstracts were blinded and randomly assigned for review by a panel of physicians and pharmacists working in family practice, resulting in a total of 1866 evaluations. Each panelist evaluated appropriateness of drug choice, potential benefit, and degree of prescription documentation. Actual prescription documentation of drug name, strength, quantity, directions, and refills was determined by comparing the carbon-copy prescription with the medical record. Multiple regression analysis revealed a significant correlation between panel evaluations for drug choice and actual documentation of drug name, strength, and directions. Likewise, the mean rating of potential benefit was significantly associated with actual documentation of drug name, strength, directions, and refills. The mean rating of appropriateness of documentation was significantly associated with all five independent documentation variables. Drug name, strength, and directions were recorded in 94%, 60%, and 75% of the medical records, respectively. These data demonstrate a positive association between the quality of prescribing and actual prescription documentation in the medical record.