Abstract Details

Title Neuromodulation and Chronic Pain: Utilizing Artificial Intelligence to Guide Self-management of Pain

Topic Pain

Presentation(s) P8 - Poster Session 8 (11:45 AM-12:45 PM)

Poster/Presentation
Number 14-003

Objective To develop and validate a minimum viable product (MVP) of a mobile application utilizing artificial intelligence to help users navigate chronic pain utilizing neuroadaptive and functional recovery strategies.

Background Pain is a complex human experience involving emotional, cognitive, and physiological pathways. Neuroadaptive strategies have been shown to reduce self-reported pain and improve movement. Repetitive training is associated with strengthening neural pathways and weakening unwanted pain pathways. However, studies have shown that adherence to rehabilitation programs is 63-82% in the hospital and drops to 21-47% after discharge due to lack of a routine and forgetfulness. This platform was designed to address these barriers by incorporating multimodal rehabilitation strategies with AI integration to tailor programs to the individual with daily reminders and adaptive strategies.

Design/Methods The MVP uses a closed-loop, large language model (LLM) to analyze user-input and provide tailored plans characterized into four adaptive modes. Recommendations are generated utilizing neuroadaptive strategies. Additionally, the platform provides trends analysis, daily logs, reminders, and an interactive chat feature. A total of 5 simulated entries of common pain archetypes with standardized characteristics were tested with the MVP for a proof-of-concept preliminary trial. A five-day interaction sequence was trialed to assess usability of the app and appropriateness of responses. Responses were assessed utilizing the Health Information Technology Usability Evaluation Scale (Health-ITUES).

Results The MVP was overall rated to be a user-friendly app (4.3 ± 0.6) with sub-category scores of impact (4.5 ± 0.5), perceived usefulness (4.7 ± 0.5), perceived ease-of-use (4.2 ± 0.2), and user control (4.0 ± 1.0). No unsafe or contradicting recommendations were provided in testing.

Conclusions This proof-of-concept demonstrates the feasibility of a closed-loop LLM application as a tool for self-management of pain incorporating neuroadaptive and functional recovery strategies. Preliminary findings reinforce the usability of the application to reinforce self-regulatory behaviors and promote adherence.

Authors/Disclosures
Siva S. Krothapalli, BS PRESENTER	Mr. Krothapalli has nothing to disclose.
Snigdha Kosuri, MPH	Miss Kosuri has nothing to disclose.
Ethan Hulewicz	Mr. Hulewicz has received personal compensation for serving as an employee of MerchLogix. Mr. Hulewicz has received personal compensation for serving as an employee of Georgia Institute of Technology.

��ɫ��

��ɫ��