A Toolbox for Surfacing Health Equity Harms and Biases in Large Language Models

TL;DR

This paper introduces a comprehensive framework and datasets for identifying biases in large language models used in healthcare, aiming to improve health equity by surfacing potential harms in model-generated medical answers.

Contribution

It presents a multifactorial human assessment framework and the EquityMedQA dataset, enabling more effective detection of biases in LLMs for medical applications.

Findings

Our approach uncovers biases missed by narrower evaluations.

Diverse assessment methods and raters improve bias detection.

The methodology highlights the importance of participatory review processes.

Abstract

Large language models (LLMs) hold promise to serve complex health information needs but also have the potential to introduce harm and exacerbate health disparities. Reliably evaluating equity-related model failures is a critical step toward developing systems that promote health equity. We present resources and methodologies for surfacing biases with potential to precipitate equity-related harms in long-form, LLM-generated answers to medical questions and conduct a large-scale empirical case study with the Med-PaLM 2 LLM. Our contributions include a multifactorial framework for human assessment of LLM-generated answers for biases, and EquityMedQA, a collection of seven datasets enriched for adversarial queries. Both our human assessment framework and dataset design process are grounded in an iterative participatory approach and review of Med-PaLM 2 answers. Through our empirical study, we find that our approach surfaces biases that may be missed via narrower evaluation approaches. Our experience underscores the importance of using diverse assessment methodologies and involving raters of varying backgrounds and expertise. While our approach is not sufficient to holistically assess whether the deployment of an AI system promotes equitable health outcomes, we hope that it can be leveraged and built upon towards a shared goal of LLMs that promote accessible and equitable healthcare.