Abstracts

Rationale: Medication non-adherence (MNA) is the leading cause of breakthrough seizure, which subsequently causes increased health care costs. MNA also is a risk factor for Sudden Unexplained Death in Epilepsy. In our underserved population at The University of Illinois epilepsy clinic we perceived a high rate of MNA to antiepileptic drugs (AEDs) and sought to understand which factors contribute to this. Our goal was also to determine if a possible improvement of adherence was seen after receiving a phone call from the epilepsy nurse. Adherence was measured objectively through AED levels measured during normal follow up clinic visits. Patients self-report of medication adherence was also evaluated. Methods: Retrospective chart review of 166 patients seen in the adult epilepsy clinic between December 2016 and December 2018 was conducted. Patient demographic information included date of birth, sex, race, education, employment status, and zip code. Data about medication included names of AEDs, number of tablets/capsules per day of AEDs and other medications. The number of emergency department visits and hospital admissions for seizures was also abstracted, along with AED levels if drawn during that visit. Up to four most recent clinic visits were evaluated for reported seizure frequency, reported adherence, reasons for MNA, current AED doses, and AED levels if available. MNA was defined as having undetectable serum level of AED while probable MNA was defined as a subtherapeutic level based on dose. Nurse phone calls encounters were looked at for MNA patients to determine if adherence improved at subsequent visits. Results: Overall MNA rate was 41.6 %: of those only 4.8 % had subtherapeutic levels ('probable MNA'). The remaining 36.8 % had undetectable levels in the blood. No significant differences were found for sex, race/ethnicity, education, employment status, address crime index, between the group of adherent and MNA patients. Majority of our entire patient population (68.7 %) lives in an area with average income <50,000; 73% of patients living in areas with income 50.000-79,999, showed medication adherence (30 patients vs 11 patients). No significant differences were found for type of epilepsy, duration of epilepsy, seizure type, or number of AEDs. However, a trend towards Focal Epilepsy being more frequent among adherent patients, and generalized epilepsy, being more frequent among MNA epilepsy was seen. 59 % of patients who reported having weekly seizures were MNA. Whereas, 68.3 % of patients having yearly seizures and 67.7% of patients seizure free, were adherent with their AEDs (P=0.03). Among those who were MNA to AED by serum level, 73.9 % of patients had reported being adherent to their physician. Nurse calls were completed for 70% of patients with undetectable levels. Of those who received a call, 43% became adherent, and 57% stayed MNA. Of those of did not receive a phone call, 67% remained MNA [OR=1.5 (CI: 0.22-9.42)].  Conclusions: Our findings show 41.6 % of MNA in our clinic population. Only two of the recent studies of adherence in epilepsy have used AED level measurement, and levels were measured during hospitalizations due to breakthrough seizures (showing MNA of 39%). Our study gives a more accurate evaluation of MNA as AEDs levels were measured during follow up visits. 73.9 % of those patients with undetectable levels self-reported to the physician that they were adherent, raising concerns that self-reported adherence should be verified, especially when patients are still experiencing frequent seizures. We did not find any specific predictors of MNA. However, trend showing patients with generalized epilepsy are more likely to be MNA. Data also showed that patients in higher income areas were more likely to be adherent. In conclusion, in this very specific underserved patient population, a large proportion of patients are not taking their medications, but this cannot be easily linked to one specific factor. Our study has shown nurse phone calls reminding patients about adherence increases the chances of becoming adherent by 50%. The true impact of a systematic method relying on nurses' calls encouraging adherence will be better measured in a future prospective trial.  Funding: No funding