Dan Larremore, Department of Computer Science, University of Colorado Boulder
Model-informed COVID-19 vaccine prioritization strategies by age and serostatus
Limited initial supply of SARS-CoV-2 vaccine raises the question of how to prioritize available doses. One might reason, intuitively, that doses should be prioritized to directly protect those who are most vulnerable. Yet one might also intuitively argue that we should use vaccination as a means to break chains of transmission by prioritizing early doses to those most responsible for transmission, thereby indirectly protecting the vulnerable by reducing prevalence. Unfortunately, these two intuitive solutions make orthogonal recommendations. Here, we introduce a family of mixed discrete and differential equation models to resolve the tension between these recommendations, and compare five age-stratified vaccine prioritization strategies. By considering the demographics and contact patterns in the country of interest, transmission rates, vaccine properties, and the accumulated immunity in the population due to prior infection with SARS-CoV-2, we show how one can use differential equation models to quantify the tradeoffs between vaccine rollout strategies in a context-specific ways. We also highlight ways in which these models can help ameliorate existing pandemic-related inequities in access to healthcare and protection. In this talk, we will cover both the high-level results and recommendations, as well as vaccine-related modeling choices that complicate the more typical and standard "SIR" type disease model. This work is a collaboration with Kate M. Bubar (CU APPM) and Kyle Reinholt (CU CSCI), among others.