Automated, machine learning-based alerts increase epilepsy surgery referrals: A randomized controlled trial

Area of Science:

• Neurology
• Medical Informatics
• Surgical Outcomes

Background:

• Epilepsy surgery offers a potential cure for drug-resistant epilepsy.
• Referral rates for epilepsy surgery evaluations are often suboptimal.
• Identifying eligible candidates for epilepsy surgery requires specialized clinical assessment.

Purpose of the Study:

• To evaluate the effectiveness of automated, electronic alerts in increasing referrals for epilepsy surgery.
• To assess the impact of a clinical decision support system on surgical referrals for pediatric epilepsy patients.

Main Methods:

• A prospective, randomized controlled trial was conducted across 14 pediatric neurology clinics.
• A natural language processing-based clinical decision support system within the electronic health record (EHR) screened patients.
• Potential surgical candidates were randomized to receive an alert or standard care, with referral for neurosurgical evaluation as the primary outcome.

Main Results:

• Automated alerts significantly increased the likelihood of referral for epilepsy surgery evaluation (9.8% vs. 3.1%).
• Patients receiving alerts were more than three times as likely to be referred for presurgical evaluation (adjusted HR=3.21).
• Nine patients (4.4%) in the alert group underwent epilepsy surgery, compared to none in the control group.

Conclusions:

• Machine learning-based automated alerts can enhance the utilization of referrals for epilepsy surgery evaluations.
• Implementing clinical decision support systems may improve access to surgical treatment for epilepsy.
• Automated alerts show promise in optimizing surgical referral pathways for pediatric epilepsy management.