Wearables detect seizures, help prevent related injuries

The use of wearable devices is feasible for the automated seizure detection in the home environment, leading to better care outcomes, such as fewer seizure-related injuries and more accurate seizure diaries, according to a recent study.

“The vast majority of the users were satisfied overall with the wearable device, considered it was easy to use the device, and agreed that the use of the device improved their quality of life,” the researchers said.

A structured online questionnaire was administered to 242 respondents who had direct experience with using wearable devices for the automatic detection of seizures in the home environment. Of the participants, 175 were caregivers while 67 were persons with epilepsy (PWEs). Of note, the questionnaire was adaptive, such that previous answers determined the following set of questions.

Majority of the participants either had or were caring for people with generalized tonic-clonic seizures. The median age at seizure onset was 9 years, while that for first device use was 17 years. Most participants used clinically validated devices, such as the NightWatch (36.36 percent), Empatica (30.16 percent), and Epi-Care (19.01 percent). On the other hand, 13.63 percent used nonvalidated devices such as smartwatches. [Epilepsia 2022;doi:10.1111/epi.17189]

The vast majority of participants were highly satisfied with their devices, scoring a median of 6 points on a 7-point Likert scale. Those who used validated devices were significantly more satisfied with its performance than counterparts using nonvalidated devices (median, 5.5 vs 5 points; p=0.011).

At the time of the survey, 204 respondents (84.30 percent) were still using their devices, with a median duration of 14 months. Meanwhile, 38 participants (15.70 percent) had already terminated use after a median of 6 months. Kaplan-Meier analysis showed that retention was significantly lower for users of nonvalidated devices (p=0.038).

Common reasons for device discontinuation included too many false alarms (52.94 percent), missed seizures (29.41 percent), difficulty of use (8.82 percent), and the stigma associated with the use of such devices (5.88 percent).

Nevertheless, most users considered these automated detection devices easy to use, with a median score of 6 points on the Likert scale. Usability was significantly worse for nonvalidated devices (median, 6 vs 5 points; p=0.029).

Similarly, validated devices showed significantly better sensitivity, as perceived by the participants (p=0.003). These devices also reached the target sensitivity of ≥95 percent more frequently than nonvalidated comparators (p<0.001). Daytime and nocturnal false alarm rates were low and comparable among devices.

Notably, participants agreed that using such devices led to better quality of life, though such an effect appeared to be significantly weaker for nonvalidated devices (p=0.009). In particular, 30 percent of all respondents said they noticed a decrease in seizure-related injuries due to the device alarms. Moreover, 65.41 percent noted that the devices helped improve the accuracy of their seizure diaries.

“A potential responder bias toward users experiencing good performance may imply an overly optimistic overall assessment of the devices,” said the researchers, acknowledging the potential impacts of this limitation on the accuracy of the findings.

Nevertheless, “bearing these limitations in mind, our results support that validated seizure detection wearable devices are feasible and useful in the home environment, and can be recommended for clinical application,” they added.