Evaluating ChatGPT's Utility in Addressing Socioeconomic Disparities in Burn Patients: A Comparative Study With...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Health Disparities Research

Background:

• Patients from low socioeconomic status (SES) backgrounds encounter significant barriers to quality burn care, including limited access and follow-up.
• Online health information seeking is common, but resources like Google may offer overwhelming or irrelevant content, exacerbating disparities.

Purpose of the Study:

• To compare the accuracy, readability, and socioeconomic status (SES)-relevance of burn care information provided by ChatGPT and Google.
• To evaluate AI's potential in addressing health information disparities for vulnerable populations.

Main Methods:

• A standardized set of burn care questions was developed based on clinical guidelines.
• Responses from ChatGPT (v4.0) and Google's top result were analyzed for accuracy (Global Quality Score), readability (Flesch-Kincaid grade level), and SES-relevance (inclusion of affordability/access themes).
• A Wilcoxon signed-rank test was used for statistical comparison.

Main Results:

• ChatGPT yielded significantly higher quality responses (GQS 4.35) compared to Google (GQS 2.25, p<.01).
• Both platforms had similar readability (grade levels 8-9), but ChatGPT addressed SES issues in 74% of responses versus Google's 33%.
• ChatGPT was unanimously preferred for 50% of the evaluated questions.

Conclusions:

• ChatGPT demonstrates superior performance over Google in delivering accurate and SES-relevant burn care information.
• AI tools like ChatGPT show promise in mitigating health information disparities by providing tailored, accessible guidance for low SES patients.
• Further research is recommended to validate these findings across diverse clinical areas and patient demographics.