Assessing the Performance of GPT-4o in Evaluating the Safety of Home Environment for Older Adults

What this Trial Is About

Researchers are studying how well the multi-modal large model GPT-4o can evaluate the safety of home environments for older adults living in Beijing communities. The study focuses on older adults aged 65 and above, examining different areas of their homes to identify potential safety concerns. A professional occupational therapist will also participate to compare the evaluations provided by GPT-4o with those of a human expert. Participants will have photos taken of seven specific areas in their homes, including pathways or entrances, hallways, stairs or steps, living rooms, kitchens, restrooms, and bedrooms. These photos will then be assessed using the "Home Environment Assessment" scale by both GPT-4o and a professional occupational therapist. The therapist who is not involved in the initial evaluation will analyze and compare the results and recommendations from both evaluations to identify any significant differences. Throughout the study, participants will be involved in providing access to their home environments for photography. Researchers will collect data through the photos and assessments, focusing on the safety and modification suggestions for the home setting. The main outcome measure is the Home Environment Assessment Scale, which will be completed on a single day. The study aims to provide insights into the accuracy and usefulness of GPT-4o in assessing home safety for the elderly, with monitoring continuing until the study's end in October 2025.

Assessing the Performance of GPT-4o in Evaluating the Safety of Home Environment for Older Adults