The abstract is read at the publisher; the summary is JClub's.
ProbeFit qualifies and personalises multimodal wearable seizure detectors from the first recorded seizure, predicting later sensitivity ≥0.5 with AUC 0.74 in 27 held-out patients.
- Why it matters: Missed seizures limit seizure-frequency assessment, and wearable performance varies by patient; ProbeFit addresses the need to select suitable patients and personalise detectors without collecting multiple seizures.
- What they did: Using multicentre SeizeIT2 data from 125 patients, two generic detectors’ peak responses to the first seizure produced a qualification probability, and the same seizure plus preceding background personalised the detector; it was developed on 96 and tested unchanged on 29.
- The result: Qualification correlated with later sensitivity (Spearman rho 0.38, one-sided p=0.033), outperformed a heart-rate rule (AUC 0.76 vs 0.59; difference 0.17, 95% CI 0.05–0.29), and personalisation increased later sensitivity by 2.2 percentage points in held-out patients and 6.3 points in development patients at equal false-alarm rates.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.