Patient reporting of suspected adverse drug reactions to antiepileptic drugs: Factors affecting attribution accuracy

This study was aimed to assess the frequency and number of suspected ADRs reported by patients taking antiepileptic drugs (AEDs) and to explore the factors that may affect patients' symptom attribution accuracy. A validated questionnaire containing an extensively checklist of symptoms was distributed to outpatients prescribed one or more AEDs. Data on concomitant drugs and diseases were obtained from outpatient records. All symptoms identified were assessed for causality. Of 1388 questionnaires distributed to 1214 patients, 830 completed questionnaires were returned (59.8%) from 727 patients. In total, 7815 symptoms were identified on 757 questionnaires (91.2%). Symptom severity ratings were positively related to the number of symptoms reported (p = 0.003). Causality assessment found that 71.9% of the symptoms were 'true' ADRs and 28.1% were 'false' ADRs. Attribution accuracy was primarily influenced by the number of symptoms identified and indication for AED therapy, fewer symptoms and use for non-epilepsy indications being associated with greater attribution accuracy.

► Direct patient reporting of suspected ADRs is increasing worldwide. ► No data on the quality of patient reports have been published. ► The majority of reported symptoms were classed as probable and possible ADRs. ► Number of symptoms and indication of AED therapy were related to accuracy attribution. ► Patient self-reporting is a useful method for screening ADRs to AEDs.