Unmasking and Quantifying Racial Bias of Large Language Models in Medical Report Generation

TL;DR

This paper investigates racial biases in large language models used for medical report generation, revealing disparities that mirror real-world healthcare inequalities and emphasizing the need for bias mitigation.

Contribution

It provides the first comprehensive qualitative and quantitative analysis of racial biases in GPT models within medical contexts, highlighting specific bias patterns and their implications.

Findings

Models project higher costs for White populations.

Models suggest longer hospital stays for White patients.

Biases reflect real healthcare disparities.

Abstract

Large language models like GPT-3.5-turbo and GPT-4 hold promise for healthcare professionals, but they may inadvertently inherit biases during their training, potentially affecting their utility in medical applications. Despite few attempts in the past, the precise impact and extent of these biases remain uncertain. Through both qualitative and quantitative analyses, we find that these models tend to project higher costs and longer hospitalizations for White populations and exhibit optimistic views in challenging medical scenarios with much higher survival rates. These biases, which mirror real-world healthcare disparities, are evident in the generation of patient backgrounds, the association of specific diseases with certain races, and disparities in treatment recommendations, etc. Our findings underscore the critical need for future research to address and mitigate biases in language models, especially in critical healthcare applications, to ensure fair and accurate outcomes for all patients.