Active Learning to Understand Infectious Disease Models and Improve Policy Making

In order to improve confidence in model-based conclusions, it is necessary to gain a thorough understanding of the system and assess how model assumptions and parameters alter the results. Surrogate models are very useful for this purpose since they can be readily explored. We used Pareto-aware symbolic regression to analyze input-response data from an open source individual-based model for pandemic influenza, called FluTE (Chao et al 2010). We made a visualization tool to explore the response surfaces from six parameters on the cumulative clinical attack rate. Every parameter must be chosen and the predicted response plots are shown for each parameter with all other parameters fixed.

Policy makers need clear recommendations on reactive strategies instead of estimates for R0 and vaccine efficacies. We can instantiate our emulator for FluTE with disease and vaccine data and present the effect of social and pharmaceutical measures with a particular selection of the emulator (Example).

Surrogate modeling is relevant for many public health problems. We also analyzed the results from an economic evaluation described in van Hoek et al (2012) to estimate the quality adjusted life year gain of varicella-zoster vaccination (Example).