Development and Health System Deployment of an Electronic Health Record-Integrated Chatbot Intervention for...

Area of Science:

• Gerontology
• Health Informatics
• Public Health

Background:

• Emergency departments (EDs) often screen for fall risk but fail to notify patients or connect them to preventive resources.
• A gap exists in automated interventions for fall risk notification and referral to prevention programs.
• Artificial intelligence (AI) chatbots offer a scalable solution for patient education and resource dissemination.

Purpose of the Study:

• To describe the development and iterative improvement of an automated fall risk notification and referral intervention.
• To detail the integration of an AI chatbot (Livi) with electronic health records (EHRs) for fall prevention.
• To share end-user feedback and its impact on intervention refinement.

Main Methods:

• Leveraged existing EHR fall risk screening data to identify high-risk patients.
• Developed an EHR workflow to provide a QR code at ED discharge, linking to the Livi chatbot.
• Conducted iterative usability testing with community members to refine the chatbot and intervention.

Main Results:

• Iterative testing enhanced the chatbot with features like larger font size, Spanish language option, and expanded resource locations.
• The EHR-integrated workflow enabled system-wide deployment and rapid updating of resources.
• Clinicians were not burdened with manual referrals, as the system automated notification and linkage to resources.

Conclusions:

• A scalable, EHR-integrated intervention was developed for automated fall risk notification and personalized resource referral.
• This pragmatic approach improves population health by utilizing existing workflows and reducing clinician burden.
• Future research will involve a randomized controlled trial to assess the intervention's impact on healthcare utilization.