PARKA AI: A Sensor-Integrated Mobile Application for Parkinson's Disease Monitoring and Self-Management

Area of Science:

• Digital Health
• Neurodegenerative Disorders
• Human-Computer Interaction

Background:

• Parkinson's disease (PD) affects over 10 million globally, requiring continuous symptom monitoring for effective management.
• Patient-provider communication in PD is often suboptimal due to fragmented data and cognitive challenges.

Purpose of the Study:

• To develop PARKA AI, an iOS application designed to enhance Parkinson's disease self-management and patient-provider communication.
• To leverage wearable sensor data and self-reported logs for comprehensive PD symptom tracking and analysis.

Main Methods:

• A human-centered design approach was employed to create a high-fidelity prototype of the PARKA AI application.
• A large language model (LLM), Google Gemini 1.5 Flash, was utilized for processing and analyzing integrated patient data.
• Data from Apple Watch HealthKit (e.g., mobility, heart rate, sleep) and self-reported logs (e.g., mood, adherence) were incorporated.

Main Results:

• The prototype generates patient-friendly summaries and concise healthcare provider (HCP) reports.
• PARKA AI offers accessible data visualizations and personalized self-management tools.
• The application aims to streamline communication and foster patient engagement in managing Parkinson's disease.

Conclusions:

• LLMs show potential for developing advanced digital health tools for chronic disease management.
• PARKA AI demonstrates a novel approach to integrating objective and subjective data for Parkinson's disease care.
• Future research will focus on real-world usability testing to validate the application's efficacy and accessibility.