How does date-rounding affect phylodynamic inference for public health?

Abstract

Phylodynamic analyses enable the inference of epidemiological parameters from pathogen genome sequences for enhanced genomic surveillance in public health. Pathogen genome sequences and their associated sampling times are the essential data in every analysis. However, sampling times are usually associated with hospitalisation or testing dates and can sometimes be used to identify individual patients, posing a threat to patient confidentiality. To lower this risk, sampling times are often given with reduced date-resolution to the month or year, which can potentially bias inference of epidemiological parameters. Here, we characterise the extent to which reduced date-resolution biases phylodynamic analyses across a diverse range of empirical and simulated datasets. We develop a practical guideline on when date-rounding biases phylodynamic inference and we show that this bias is both unpredictable in its direction and compounds with decreasing date-resolution, higher substitution rates, and shorter sampling intervals. We conclude by discussing future solutions that prioritise patient confidentiality and propose a method for safer sharing of sampling dates by translating them uniformly by a random number.