Abstracts

Rationale:
Electronic seizure diaries (e-diaries) are increasingly used by patients to monitor and manage their seizures. Understanding biases in how users report seizures is important, as clinicians may use these data for decision making. The ways in which parents report infantile spasms in e-diaries is not well described.
Method:
We analyzed a 2007-2018 extract of data from Seizure TrackerTM, an electronic seizure diary available worldwide on mobile devices and on the internet. We reviewed all children with seizures classified by users as “infantile spasms”. We report demographic information (sex) and multiple seizure characteristics: time/date of seizure, percentage of spasms occurring as individual spasms (versus in clusters), cluster duration, number of spasms per cluster, and percentage of spasms from 11pm-7am (i.e., “nighttime”). We simultaneously analyzed a sample of infants recorded on overnight video electroencephalogram (vEEG) for diagnosis of infantile spasms on the same variables, coded by a child neurologist with additional training in clinical neurophysiology. Using chi-square tests, independent-samples t-tests, and Mann-Whitney U tests, we compared the two cohorts on demographic and seizure characteristics to understand potential biases in e-diary reporting.
Results:
There were 314 children in Seizure Tracker with infantile spasms and 9 patients diagnosed with infantile spasms on overnight vEEG. In the Seizure Tracker cohort, there were 15,225 seizure events with one or more reported spasms, with a total of 226,932 reported spasms. In the clinical cohort there were 93 seizure events with one or more spasms, with a total of 535 spasms. The seizure characteristics of each group are displayed in Figure 1. The two groups were similar in sex (52.6% vs. 55.6% female, p = 0.86). The clinical cohort had a significantly greater percentage of spasms as individual spasms relative to the Seizure Tracker group (8.2% vs. 0.7%, p < 0.001). The clinical cohort had a significantly lower number of spasms per cluster (10.8 vs. 16.5, p < 0.001) and a significantly lower cluster duration (2.8 minutes vs. 6.8 minutes, p < 0.001) relative to the Seizure Tracker cohort. Finally, the timing of spasms was compared between cohorts over the course of a time period (9:19pm to 9am) during which all children on vEEG were being recorded and thus could potentially have a recorded seizure. Over the course of this time period, the Seizure Tracker cohort had a significantly lower percentage of spasms that occurred between 11pm-7am relative to the clinical cohort (55.8% vs 66.7%, p < 0.001). These results are summarized in Table 1.
Conclusion:
There are several biases in how parents report infantile spasms in an electronic seizure diary. Importantly, parents may miss individual spams and, instead, may be more likely to report long clusters. In addition, parents may under report spasms occurring at nighttime. Pediatric epileptologists should be aware of these reporting biases when making decisions for children with infantile spasms based on e-diary data.
Funding:
:N/A