Abstracts

Rationale: Nonadherence to antiseizure medications (ASMs) is associated with adverse consequences. Still, people with epilepsy frequently have indications for a large number of medications beyond ASMs. Understanding whether ASM adherence is distinct from non-ASM adherence would inform whether future adherence interventions should target specific medication classes (i.e. ASMs) as opposed to individuals most likely to be poorly adherent across their medications.

Methods: This was a retrospective cohort study using Medicare (2013-2015). We included beneficiaries with epilepsy (≥1 ASM, plus International Classification of Diseases, Ninth Revision, Clinical Modification diagnostic codes), and a 20% random sample without epilepsy. Adherence for each medication class was measured by the proportion of days covered (PDC) in 2013-2015. We used Spearman correlation coefficients, Cohen’s kappa statistics, and multilevel logistic regressions.

Results: There were 82,200 beneficiaries with epilepsy, and 638,280 beneficiaries without epilepsy. Spearman correlation coefficients between ASM PDCs and each of the 4 non-ASM PDCs among beneficiaries with epilepsy ranged 0.44-0.50, Cohen’s kappa ranged 0.33-0.38, and within-person differences between each ASM’s PDC minus each non-ASM’s PDC were all statistically significant (p < 0.01) though median differences were all very close to 0. Fifty-four percent of variation in adherence across medications in beneficiaries with epilepsy was due to differences between individuals. Adjusted predicted probabilities of adherence were: ASMs 74% (95% confidence interval [CI] 73%-74%), proton pump inhibitors 74% (95% CI 74%-75%), antihypertensives 77% (95% CI 77%-78%), statins 79% (95% CI 78%-79%), and levothyroxine 82% (95% CI 81%-82%). Adjusted predicted probabilities of adherence to non-ASMs were 81% (95% CI 81%-81%) for beneficiaries with epilepsy versus 77% (77%-77%) for beneficiaries without epilepsy.