Abstracts

Rationale: Adherence to treatment in many long-term conditions is recognized to be poor. Inadequate seizure control in epilepsy may therefore in some circumstances relate to poor adherence rather than the efficacy of anti-epileptic drugs (AED). There are limited data available on adherence to AEDs in pediatric populations and there are considerable methodological problems with the existing data. Using a novel dataset for community pharmacy-dispensed medication we wished to assess the adherence to AED treatment both for individual children and for a whole population. We also aimed to correlate adherence with a series of clinical variables to identify factors that might predict good or poor adherence. Methods: All children (≤16 years of age) from the Tayside region of Scotland with a diagnosis of epilepsy who were receiving AED treatment during one year epochs in 2000-01 and 2005-06 were identified and a clinical database constructed from case records. Data were collected on seizure frequency,AEDs and their dosage, whether taken as monotherapy or polytherapy, and the presence of other chronic illness. The Health Informatics Centre at the University of Dundee has a database which captures every encashed community pharmacy prescription in the Tayside region (population 450,000). An adherence index [AI] was calculated by cross-linking and then calculating the ratio of the quantity of AED collection by the patient from a local pharmacy to the expected quantity from the physician-directed prescription of AED. This data was available both cross-sectionally and longitudinally in 2 annual cohorts. Results: A total 320 children were studied. The overall adherence was 69%. However only 30% of the study population encashed more than 90% of the physician-directed AED dose, whereas 25% of them encashed less than 50%. For the group of children with the best adherence (>90%),34% were on monotherapy and 18% were on polytherapy. Adherence to the three most commonly used AEDs given as monotherapy was 73% for Valproate, 67% for Carbamazepine and 52% for Lamotrigine. Seizure frequency had no effect upon adherence whether or not a child was seizure free or continued to have multiple daily seizures. The presence of an additional chronic illness did not adversely affect adherence. For children on long term AED treatment, mean adherence fell from 73% to 64% over a 5 year period and for the children with best adherence (>90%) in 2000 there was a 50% fall in adherence when assessed in 2006. Conclusions: We have described a novel population-based technique for assessing adherence to AED medication in children with epilepsy. Overall adherence to treatment is poor and it is of note that this appears to be independent of seizure frequency. Adherence falls over time and is worse with polytherapy than monotherapy. Identification of these and other factors that might predict poor adherence for individual children may be helpful in the design and evaluation of interventions to improve adherence to AED treatment.