An audit of therapeutic drug monitoring (TDM) of anticonvulsants was performed to assess both its use and misuse in the management of patients with epilepsy. Over a four week period all samples received for phenytoin, carbamazepine, sodium valproate and phenobarbitone assays were included in the audit. The aims were to establish the source of the specimens, the reasons for the requests and to ascertain what action, if any, would be taken when the result of the assay was provided. A total of 163 separate assays were performed over the four week period (43 phenytoin, 74 carbamazepine, 41 valproate, 5 phenobarbitone). Only 18.7% of all requests originated from the adult neurology department. The vast majority of tests had been ordered by junior medical staff (only 10% by consultants) and approximately 50% were 'routine' with no satisfactory clinical reason for the request offered. There was a tendency to manipulate prescribed doses on the basis of drug levels alone without taking the clinical picture into consideration. These results demonstrate a general ignorance, especially amongst junior medical staff, of the value of TDM of anticonvulsants, and reinforce the need for both an educative and interpretive service to be provided by the Chemical Pathology Department.