Carol Bennett

Mathew Thomson, Jean-David Therrien, Nikho Hizon et al.

Wastewater surveillance (WWS) has emerged as a valuable tool for public health surveillance, particularly since the COVID-19 pandemic. Its long-term utility is constrained, however, by fragmented data systems, inconsistent metadata practices, and poor interoperability. The Public Health and Environmental Surveillance Open Data Model (PHES-ODM) was developed as an open, collaborative framework to standardize WWS data and support transparent, ethical data use aligned with FAIR principles. Adopted by the Public Health Agency of Canada and adapted by the EU Sewage Sentinel System, the model is now used in over 25 countries. This paper introduces version 3 of the model, which addresses persistent barriers to interoperability and data utility. Key enhancements include new tables for public health actions, external repository linkages (e.g., GISAID, GenBank), and analytical workflow documentation, as well as support for complex relational linkages across sites, samples, measures, and populations. Tools for mapping across other data formats, including PHA4GE and the US CDC National Wastewater Surveillance System, and for supporting long and wide data formats are also introduced. We compare PHES-ODM against six other WWS data standards across 25 features. Balancing robustness with usability, PHES-ODM v3 provides a scalable, modular infrastructure adaptable to diverse WWS and environmental surveillance programs.

André M. Carrington, Douglas G. Manuel, Paul W. Fieguth et al.

Optimal performance is critical for decision-making tasks from medicine to autonomous driving, however common performance measures may be too general or too specific. For binary classifiers, diagnostic tests or prognosis at a timepoint, measures such as the area under the receiver operating characteristic curve, or the area under the precision recall curve, are too general because they include unrealistic decision thresholds. On the other hand, measures such as accuracy, sensitivity or the F1 score are measures at a single threshold that reflect an individual single probability or predicted risk, rather than a range of individuals or risk. We propose a method in between, deep ROC analysis, that examines groups of probabilities or predicted risks for more insightful analysis. We translate esoteric measures into familiar terms: AUC and the normalized concordant partial AUC are balanced average accuracy (a new finding); the normalized partial AUC is average sensitivity; and the normalized horizontal partial AUC is average specificity. Along with post-test measures, we provide a method that can improve model selection in some cases and provide interpretation and assurance for patients in each risk group. We demonstrate deep ROC analysis in two case studies and provide a toolkit in Python.