Enhancing access to medical literature through an LLM-based browser extension

Abstract

Medical literature is a great source of validated and recent information available through many resources, such as the Internet. Yet, the complex terminology poses a challenge in comprehending the text and makes this verified information not accessible to individuals with less knowledge of the specific area. However, large language models (LLMs) and artificial intelligence (AI) have proven to play a role in addressing these challenges.

This thesis introduces SimpliMed, a web-based browser extension that utilizes an LLM instance (GPT-4). It simplifies medical article abstracts and aims to improve the accessibility of medical literature to audiences with varying educational backgrounds. SimpliMed's efficacy was evaluated by conducting a user study. The participants used SimpliMed for five days to read simplified medical abstracts based on given topics. During this time, they filled out open-ended and scaled questionnaires. The results demonstrate that the extension effectively simplified medical article abstracts and proved beneficial in aiding comprehension, facilitating the learning of new knowledge, and developing critical thinking abilities among participants.

The findings suggest that SimpliMed advances our knowledge of LLM-based tools for simplifying medical literature and could contribute to improving health literacy among users and empowering them to make informed decisions about their health. Also, by fostering motivation and interest in learning, it can stimulate deeper engagement with medical literature.