A survey of the prescription of the novel anticonvulsants lamotrigine, vigabatrin, gabapentin and topiramate was carried out in an outpatient setting to investigate the outcome of prescription and factors affecting response. One hundred and fifty randomly selected patient notes were analysed retrospectively. Drug continuation and seizure freedom were used as measures of response. Twenty-nine percent of patients had a brain lesion, 14% a psychiatric disorder, 7% neonatal seizures, 21% a family history of epilepsy and 15% a learning disability. On average at the moment of initiation of the novel anticonvulsant the patients had had a diagnosis of epilepsy for 18 years, were 33 years old, had 19 seizures per month and had previously used two drugs which failed to control their epilepsy. The first novel anticonvulsant was in 55% of cases lamotrigine, in 43% vigabatrin and in 1 % gabapentin. The overall percentage of patients who stayed on their first novel anticonvulsant was 55%, and 17% became seizure-free. No factors were found to influence the response in terms of drug continuation. For seizure freedom, the presence of a psychiatric disorder and partial seizures predicted a significantly poorer response. Length of seizure disorder, seizure frequency at initiation, the number of previously used failed drugs and the total number of drugs previously used were all significantly lower in the seizure-free group.