Research

Quantifying biomarkers in wastewater to monitor public health—known as wastewater-based epidemiology (WBE)—rose to prominence during the COVID-19 pandemic. But data are very noisy, and the potential of WBE as a cost-effective, equitable public health monitoring tool remains unrealized.

Realizing WBE’s potential to monitor public health and inform policy requires that we build models for principled uncertainty quantification of epidemiologically relevant quantities. I both develop foundational methods for scalable Bayesian inference and apply them to address concrete WBE challenges together with interdisciplinary collaborators.

In my methodological work, I developed model-based time series clustering algorithms (Hoffmann, Peel, Lambiotte, and Jones, 2020), a high-performance inference library for Gaussian processes using Stan (Hoffmann and Onnela, 2023), and neural networks to extract summary statistics from complex data for simulation-based inference (Hoffmann and Onnela, 2022).

In my applied work, I initiated a collaboration to monitor SARS-CoV-2 RNA in wastewater to estimate community prevalence, securing more than $850,000 in funding across three grants. I developed models for fecal shedding of SARS-CoV-2 RNA which allowed us to resolve tension between WBE and clinical data, demonstrate that all people likely shed viral RNA in their feces, and show that WBE can be a leading indicator (Hoffmann and Alsing, 2023). I also advised the UK Health Security Agency (akin to the CDC) on the management and analysis of WBE data (Hoffmann, McIntyre-Nolan et al., 2021), contributed to a characterization of sources of uncertainty in WBE (Wade et al., 2022), and published the first comprehensive dataset of wastewater catchment areas in Great Britain (Hoffmann, Bunney, Kasprzyk-Hordern, Singer, 2022).