Predictive value of death certification in the case ascertainment of epilepsy.

OBJECTIVES Death certificates are an unreliable source of information on cause of death, and mortality due to epilepsy can thus be underestimated. METHODS We investigated people with epilepsy who had died, and attempted to identify factors that influence inclusion of epilepsy on the death certificate; eight factors were hypothesised and entered into a univariate logistic regression model. RESULTS Epilepsy was on the death certificate of 16/243 (7%) people who had had epilepsy. Factors that influenced whether or not epilepsy appeared on the certificate were seizure frequency, antiepileptic drug treatment, cause of death, and certifying physician. Factors that did not seem to influence the inclusion of epilepsy were presence of convulsive seizures, occurrence of seizures during follow up, and age at death. CONCLUSIONS We have estimated the degree of unreliability of death certificates (as currently used in the UK) as a source of information on cause of death in epilepsy. We have found that epilepsy may not appear on death certificates even if people had active epilepsy.