Minimum clinical utility standards for wearable seizure detectors: A simulation study.

Area of Science:

• Neurology and Biomedical Engineering
• Epilepsy Monitoring and Device Technology

Background:

• Current epilepsy management relies on seizure diaries, which suffer from underreporting.
• Wearable and implantable seizure detection devices offer potential improvements but lack defined accuracy standards.
• Establishing sufficient accuracy levels is critical for clinical adoption and patient benefit.

Purpose of the Study:

• To simulate clinical use cases for seizure detection devices.
• To identify the necessary accuracy levels (sensitivity and false alarm rate) for different applications.
• To guide the development and validation of wearable and implantable seizure detection technologies.

Main Methods:

• Utilized a realistic seizure simulator (CHOCOLATES) to generate ground truth data.
• Simulated seizure detector signals with varying capabilities (sensitivity, false alarm rate, device type).
• Evaluated five distinct clinical use cases: RCTs, medication adjustment, injury prevention, SUDEP prevention, and seizure cluster treatment.

Main Results:

• Randomized clinical trials (RCTs) showed efficiency with various wearable parameters, favoring implantable devices for higher accuracy and lower false alarm rates.
• Medication adjustment in clinic demonstrated robustness across a wide range of device accuracies.
• Injury prevention, SUDEP prevention, and seizure cluster treatment scenarios critically required high sensitivity, with minimal impact from false alarm rates.

Conclusions:

• The specific clinical use case dictates the required accuracy for wearable seizure detection devices.
• Simulation provides a framework for verifying device utility based on specific use cases and device parameters.
• Tailoring accuracy requirements to use cases is essential for effective epilepsy management technology.