Gender-Neutral Large Language Models for Medical Applications: Reducing Bias in PubMed Abstracts

This paper presents a pipeline for mitigating gender bias in large language models (LLMs) used in medical literature by neutralizing gendered occupational pronouns. A dataset of 379,000 PubMed abstracts from 1965-1980 was processed to identify and modify pronouns tied to professions. We developed a BERT-based model, "Modern Occupational Bias Elimination with Refined Training," or "MOBERT," trained on these neutralized abstracts, and compared its performance with "1965Bert," trained on the original dataset. MOBERT achieved a 70% inclusive replacement rate, while 1965Bert reached only 4%. A further analysis of MOBERT revealed that pronoun replacement accuracy correlated with the frequency of occupational terms in the training data. We propose expanding the dataset and refining the pipeline to improve performance and ensure more equitable language modeling in medical applications. Introduction Background Large language models (LLMs) are now widely used for a range of applications, from creating customer service chatbots to advertising that targets specific clients to predicting financial outcomes from potential economic indicators. LLMs have also increased in presence in the medical sector, ranging from accessible diagnostics to comprehensive literature retrieval, where they hold the promise of leading to a more informed level of care. Given the critical nature of these uses, it is essential to ensure that such LLMs remain free from biases that could potentially impact patient treatment and outcomes.