Natural Language Processing in Environmental Health Research

Large Language Models (LLMs) are utilized across multiple disciplines for summarizing information and generating domain relevant responses. The Puerto Rico Testsite for Exploring Contamination Threats (PROTECT) Center at Northeastern University aims to address environmental health issues, particularly focusing on the impact of exposure to environmental contaminants on preterm birth rates in Puerto Rico. We explore the application of RAG in creating an AI assistant tailored to the needs of PROTECT for information dissemination and community engagement. The virtual assistant is designed to interact with community members, providing them with information about environmental health risks and preventive measures. We assess the performance our model and its efficacy in environmental health research compared to other Large Language Models (LLMs). Our findings suggest that equipping an LLM with a RAG model in the domain of environmental health studies can enhance outreach efforts.