Estimates of the reproduction ratio from epidemic surveillance may be biased in spatially structured populations

TL;DR

Estimating the reproduction ratio from surveillance data can be biased in spatially structured populations, but a spectral correction method can improve accuracy across all epidemic phases.

Contribution

We introduce a spectral correction method to remove bias in reproduction ratio estimates caused by spatial population structure and mobility.

Findings

Spectral correction reduces bias in reproduction ratio estimates.

Method validated with simulated epidemics and COVID-19 case study.

Improves epidemic monitoring accuracy in spatially structured populations.

Abstract

Accurate estimates of the reproduction ratio are crucial to project infectious disease epidemic evolution and guide public health response. Here, we prove that estimates of the reproduction ratio based on inference from surveillance data can be inaccurate if the population comprises spatially distinct communities, as the space-mobility interplay may hide the true epidemic evolution from surveillance data. Consequently, surveillance may underestimate the reproduction ratio over long periods, even mistaking growing epidemics as subsiding. To address this, we use the spectral properties of the matrix describing the spatial epidemic spread to reweigh surveillance data. We propose a correction that removes the bias across all epidemic phases. We validate this correction against simulated epidemics and use COVID-19 as a case study. However, our results apply to any epidemic where mobility is a driver of circulation. Our findings will help improve epidemic monitoring and surveillance and inform strategies for public health response.