Skip to main content
ENDO 2026 Main banner

Pediatric and Adolescent Endocrinology

SAT-215 - Racial and Gender Bias in Artificial Intelligence Guidance for Pediatric Type Two Diabetes

Saturday, June 13, 2026
9:30 AM - 4:30 PM CT
Location: ENDOExpo; Poster Floor
Disclosure(s):
Gloria Wu, MD, MBA: No financial relationships to disclose
Swara Tewari, B.S. in Microbiology: No financial relationships to disclose
Tiffany Lin: No financial relationships to disclose


Background:
Large language models (LLMs) are increasingly used by patients and caregivers seeking health information, including for pediatric type 2 diabetes mellitus (T2DM). Pediatric T2DM is more uncommon than Adult T2DM. Pediatric diabetes management requires specialized medical and
psychosocial care, and disparities by race and gender are well documented in clinical settings.

Objective:
To evaluate whether LLM responses to a pediatric T2DM query differ by patient race and gender with respect to readability, response length, and inclusion of clinically relevant content.

Methods:
A standardized query: “I am a chubby, 12-year-old [race] [gender] with a new diagnosis of Type 2 Diabetes. I feel insecure about my body image. What should I do? was submitted to three LLMs (ChatGPT 5.2, Gemini 3.0, and Grok 4.1). Race (White, Black, Hispanic, Asian) and gender (male/female) were systematically varied. Responses were analyzed for word count (WC),Flesch–Kincaid Grade Level (FK), Gunning Fog Index (GF), SMOG Index (SI), and keyword inclusion. Content analysis assessed inclusion of diabetes medications, endocrinology referral, diabetes education, and psychological counseling.

Results:
Bias were observed across race and gender. Female personas consistently received more complex responses than males (mean FK 8.44 vs. 7.77; mean SI 8.18 vs. 7.72). White male personas received the simplest language and shortest responses, while Asian personas received the longest responses overall. Black males received the shortest responses across models. Insulin was mentioned universally; however, first-line therapy with metformin and explicit recommendations to see an endocrinologist appeared only in responses to Black patients from Grok. No model consistently referenced diabetes self-management education, GLP-1 receptor agonists, SGLT2 inhibitors, or formal psychological counseling. Cultural framing was more frequently applied to non-White, while White received more generic guidance.

Conclusion:
Our small study suggests that there is a need for bias-aware and pediatric-specific safeguards as LLMs are used for health information. Knowing that patients and families are generally using these tools, clinicians should continue to advise caution to families before following their medical recommendations.

*Unless otherwise noted, all abstracts presented at ENDO must not be released to the press or the public until the date and time of presentation. For oral presentations, the abstracts are embargoed until the session begins. The Endocrine Society reserves the right to lift the embargo on specific abstracts that are selected for promotion prior to or during ENDO.*