An individual level infectious disease model in the presence of uncertainty from multiple, imperfect diagnostic tests.

The article focuses on the development and evaluation of a novel Bayesian spatio-temporal individual-level infectious disease modeling framework, termed the diagnostic ILM, which explicitly incorporates uncertainty from imperfect diagnostic tests. This model extends traditional compartmental models (e.g., SEIR) by jointly estimating latent disease status and transmission dynamics while accounting for diagnostic sensitivities and specificities, applied to data from the 2006 Iowa mumps epidemic involving multiple diagnostic tests with varying accuracy. Through simulation studies, the diagnostic ILM demonstrated improved estimation of transmission probabilities and credible interval coverage compared to heuristic approaches that classify infectious individuals based solely on positive test results, particularly when test sensitivities are low. The application to the Iowa mumps data revealed nuanced transmission patterns influenced by age, spatial proximity, and testing status, highlighting the model's capacity to incorporate negative test results and untested individuals via a sparks term. Limitations include computational complexity, assumptions of constant test accuracy over time, and challenges in estimating intervention effects, but the framework offers a valuable tool for infectious disease inference in settings with diagnostic uncertainty.