Abstracts

Rationale: Electronic diaries are increasingly used to record seizures clinically and in research1–6. Key advantages cited over paper diaries are increased accessibility, reminders to track, and timestamps of data entry1. This study investigated variables that influence the duration between when a seizure was reported to occur, and when the participant entered the event in their electronic diary. Methods: Study Design The Human Epilepsy Study (HEP) is a six-year, prospective, observational study whose goal is to identify markers of treatment response in patients with newly-treated focal epilepsy. The HEP seizure diary, an app designed by Irody Inc. was installed on a provided iPod or on participants’ personal smart phone. Participants were instructed to track seizure (sz) status daily. Participants received a daily reminder to enter their sz status. All szs were stamped with date and time of entry, and classified by reported semiology in clinic according to the ILAE classification system (focal aware w/o motor (FA-M), focal aware w/ motor (FA+M), focal w/ impaired awareness (FIA) and focal to bilateral tonic clonic seizures (FBTC)7. Szs were then more broadly grouped as retained awareness (RA = FA-M & FA+M), impaired awareness (IA = FIA & FBTC), or unknown. Diary compliance (“tracker type”) was categorized based on % days with data entered as “Trackers” (>80%), “Moderate Trackers” (20-79%), or “Non-Trackers” (<20%). Lag time was calculated as the difference between time and date of entry to time & date of sz as reported by the participant, in hours.Data AnalysisThe effect of sz type on lag time was modelled using Generalized Estimating Equations (GEE) to account for within-participant correlation using log-transformed hours as the dependent variable and a gamma-distribution. Age at enrolment (years), sex, study site, and tracker type were included as covariates. In the first model, we examined the effect of broad seizure categories (RA vs. IA) on mean tracking lag time. Seizures categorised as unknown were excluded. In the second model, ILAE seizure type (FA-M as reference) was used as the predictor variable. In each case, model-based estimates of adjusted mean and 95% CI tracking lag time were calculated. Analysis was performed using Stata version 15.1 (College Station, TX). Results: A total of 10,575 szs were logged by 232 participants from November 2012 - May 2019. Excluding szs missing either time of sz (1432), or date & time stamp (598), szs reported as occurring after time stamp of entry (296) or duplicate entries (>1 sz reported with same time & date) (128), there were 8121 szs reported by 232 participants. The median tracking lag time was 28 hrs (IQR 10-103, range 0–27,546). In the first model (N=211, n=6953 sz), there was a significant effect of RA vs IA (p<0.001), tracker type (p< 0.001 for moderate tracker and poor tracker vs good tracker), and male sex (p=0.001) on lag. In the second model (N=213, n=7001), compared to FA-M, FBTC sz’s were associated with an increased lag (p = 0.001) whereas FIA (P>.001) and FA+M szs (P=.03) were associated with decreased lag. Tracker type and male sex remained significant predictors of lag. The model-predicted mean tracking lag was 558 hours (95%CI: 445 – 671) for FA-M, 277 hours (95%CI: 223 – 330) for FIA, and 873 hours (95%CI: 650 – 1097) for FBTC sz’s. Conclusions: Participants often do not track szs until substantially after they occurred, a factor which likely affects reliability reporting. Type of seizure, gender, and tracking compliance influence tracking lag. Funding: The HEP study is supported by the Epilepsy Study Consortium (ESCI), a non-profit organization dedicated to accelerating the development of new therapies in epilepsy to improve patient care. The funding provided to ESCI to support HEP comes from industry, philanthropy and foundations (UCB Pharma, Eisai, Pfizer, Lundbeck, Sunovion, The Andrews Foundation, The Vogelstein Foundation, Finding A Cure for Epilepsy and Seizures (FACES), Friends of Faces and others).