Measuring the growth of infectious disease modelling publications and their impact on policymaking: A large language model-assisted bibliometric review

Background Infectious disease modelling (IDM) is increasingly used to understand disease transmission and inform public health policy. Its growth and influence on policy have not been quantified, partly due to the large volume of literature. Using large language model (LLM)-assisted review, we quantified the expansion of IDM publications, trends in policy citations, and regional disparities in research contributions and uptake. Methods An LLM-assisted bibliometric review was conducted in Embase, Medline, and Web of Science, identifying IDM publications to December 2024 via GPT-4o. Eligible studies employed mathematical, statistical, or mechanistic models for infectious disease outcomes. LLM accuracy was iteratively refined through human review. We extracted publication metadata, geographic scope, and policy citations via Overton, a global database of policy documents. Growth trends were analysed using negative binomial regression; geographic disparities were assessed by World Bank income classifications. Results We identified 33,255 IDM publications over 44 years, with distinct growth phases. Compared with human search, LLM selection and extraction achieved 98% and 100% accuracy. Publication volume rose with the emergence of HIV/AIDS, expanded through successive outbreaks (Ebola, SARS, H1N1, MERS, Zika), surged just before COVID-19, then declined after 2021. Policy citations accounted for 1.7% of IDM publications, mirroring overall growth and peaking during periods of heightened public health attention. Citations largely reflected national research outputs, with some cross-regional adoption of IDM evidence. Conclusion Strengthening IDM’s policy impact may require reducing geographic disparities, fostering collaboration, and improving uptake mechanisms. Post-COVID-19, policy citations declined despite continued IDM growth, suggesting a lag or shift in priorities.