Automated spike and seizure detection: Are we ready for implementation?

Area of Science:

• Neurology
• Medical Informatics
• Artificial Intelligence in Medicine

Background:

• Automated detection of spikes and seizures in electroencephalography (EEG) has been researched for decades.
• Despite advances, automated detection is not standard practice in Epilepsy Monitoring Units (EMUs).
• Implementation of automated detection software in EMUs faces adoption challenges.

Purpose of the Study:

• To identify barriers and enablers for implementing automated spike and seizure detection software in an EMU setting.
• To understand user perspectives on adopting new automated EEG analysis tools.
• To inform strategies for successful integration of automated detection technology.

Main Methods:

• A qualitative study involving 22 semi-structured interviews.
• Participants included 14 technicians and neurologists involved in EEG recording/reporting and 8 neurologists receiving EEG reports.
• Study adherence to Consolidated Criteria for Reporting Qualitative Studies (COREQ).

Main Results:

• 14 barriers and 14 enablers for implementation were identified.
• Technicians reported most barriers, primarily lack of trust in software accuracy (seizure detection, false positives).
• Concerns about skill degradation and job loss were noted; enablers included workflow efficiency, quantification, and willingness to adopt.

Conclusions:

• User perspectives are crucial for successful implementation of automated spike and seizure detection in EMUs.
• Addressing technician trust and job security concerns is vital.
• Leveraging potential efficiency gains and quantification benefits can facilitate adoption.