Abstracts

Rationale: Retrospective self-report surveys (e.g., PHQ9, QOLIE-89) can be limited by recall bias, and ignore how emotions and experiences may change over time and across contexts. Ecological momentary assessment (EMA) involves repeatedly sampling a subject’s experiences in real time, and in naturalistic settings. In this study, we explored the utility, feasibility and psychometric properties of a six-question EMA survey administered via a smartphone app. We selected questions that would sample seizure experience, as well as key aspects of interictal wellbeing.


Methods: Adults with epilepsy (English-speaking, owning smartphones) were recruited during routine clinic visits and asked to download the study app (Expiwell). A daily survey (“EMA-daily”) was administered between 1000 and 1500 with a single yes/no question (have you had a seizure in the past 24h?) and five 7-point Likert-scaled questions: how satisfied are you with your life today? how sad (or depressed or blue) do you feel today? how anxious (or nervous or worried) do you feel today? are you feeling dizzy today? are you feeling sleepy (or sedated) today? Participants also received a longer biweekly survey (“EMA-14d”), serially presenting all items of the PHQ9 (depression), GAD7 (anxiety) and the ESS (sleepiness) questionnaires. Subjects were asked to engage with the app for up to a year. Those who did not respond within the first week were reminded via telephone once. No financial reimbursements were provided.


Results: 100 eligible subjects were approached. 19 declined to participate, and another 35 subjects consented, but never engaged with the app despite reminders. The remaining 46 subjects provided 3167 complete EMA-daily responses (including 227 seizure days), with a median of 71 responses/patient (range 1-265) over a median duration of 114 days (1-312). Surveys were completed in a median duration of 28.9s. Responses to individual EMA-daily items pertaining to sadness, anxiety and sleepiness were strongly correlated with contemporaneously provided PHQ9, GAD7 and ESS scores respectively (r ~0.4-0.8, p< 0.0001). Low satisfaction ratings were most strongly linked with high sadness scores (OR 1.93), and less so to anxiety, sleepiness or dizziness (OR ~1.01-1.09, multiple linear regression). Seizure days were associated with significantly higher self-ratings of anxiety, sleepiness and dizziness, but also significantly lower sadness/depression (multiple logistic regression). High satisfaction scores