Discontinuations of drugs used in the treatment of chronic conditions such as hyperlipidemia often signal adverse drug effects or therapeutic ineffectiveness. Discontinuations of antihyperlipidemic drug therapy among 2369 patients were evaluated using the computerized files and written medical records of two health maintenance organizations (HMOs) in Massachusetts for the period 1988 to 1990. Three methods were assessed for the identification of potential discontinuations of drug therapy using the automated databases. Overall, 75% of discontinuations flagged by the automated databases were confirmed in the medical charts. (1) In new users, 618 of the 635 (97%) drugs identified as drug switches by the computerized HMO pharmacy files were confirmed as discontinuations by the medical charts. (2) Of 620 eligible HMO members 332 (54%) identified with greater than 6 months between the last refill for an antihyperlipidemic drug and the end of the study stopped the drug therapy according to the medical chart. (3) Of 219 eligible drug therapies flagged with an 'inactive' or with an 'omit' status in the clinical encounter files 198 (90%) were discontinued according to the medical chart. This study demonstrates the utility of clinical automated databases to facilitate the study of drug discontinuations in primary care settings, reducing the cost and the amount of time required for the evaluation.