Geographical aggregation of asynchronous wastewater-based surveillance programmes, with application to Canada

A Bayesian hierarchical model estimates continuous latent wastewater concentrations at provincial and national scales from asynchronous site-level respiratory virus data in Canada, matching clinical positivity with R2 = 0.65–0.93.

• Why it matters: Aggregating heterogeneous, asynchronous wastewater observations into broader-scale indicators is methodologically challenging, limiting real-time, population-level surveillance.
• What they did: The framework uses Bayesian hierarchical inference to account for measurement precision, catchment population size, and changing site participation, validated on simulated data and applied to 25 sites in six provinces (Jan 2024–Dec 2025).
• The result: Applied to observational data, it produced continuous provincial and national concentration estimates with moderate to strong agreement to clinical test positivity (R2 = 0.65–0.93), and sensitivity analyses showed robustness to shorter fitting windows and site additions/removals.