A1693
Title: Uncovering public health priorities at the World Health Organization
Authors: Xiao Hui Tai - University of California, Davis (United States) [presenting]
Lauren Peritz - University of California Davis (United States)
Katheryn Russ - University of California Davis (United States)
Nick Ulle - University of California Davis (United States)
Carl Stahmer - University of California Davis (United States)
Abstract: Government delegates convene each year in the World Health Assembly, the decision-making body of the World Health Organization, to set a global Health agenda. They allocate scarce resources for crisis response and long-term programs. With thousands of diplomatic statements on diverse topics, the systematic analysis of WHA meetings has remained a challenge for researchers. This gap is addressed using natural language processing. Thousands of delegate statements across seven major Health issues are extracted and classified using fine-tuned BERT models, demonstrating superior results over traditional word frequency-based classifiers and a chat-based large language model. A longstanding model interpretability problem is addressed by demonstrating how classification determinations are made. The analysis highlights tensions between industrialized and developing countries, particularly around infectious disease mitigation.
