Making waves: Defining the lead time of wastewater-based epidemiology for COVID-19

Abstract

Individuals infected with SARS-CoV-2, the virus that causes COVID-19, may shed the virus in stool before developing symptoms, suggesting that measurements of SARS-CoV-2 concentrations in wastewater could be a "leading indicator" of COVID-19 prevalence. Multiple studies have corroborated the leading indicator concept by showing that the correlation between wastewater measurements and COVID-19 case counts is maximized when case counts are lagged. However, the meaning of "leading indicator" will depend on the specific application of wastewater-based epidemiology, and the correlation analysis is not relevant for all applications. In fact, the quantification of a leading indicator will depend on epidemiological, biological, and health systems factors. Thus, there is no single "lead time" for wastewater-based COVID-19 monitoring. To illustrate this complexity, we enumerate three different applications of wastewater-based epidemiology for COVID-19: a qualitative "early warning" system; an independent, quantitative estimate of disease prevalence; and a quantitative alert of bursts of disease incidence. The leading indicator concept has different definitions and utility in each application.

Keywords: COVID-19; SARS-CoV-2; Surveillance; Wastewater-based epidemiology.

Fig. 1

Correlations between measured wastewater SARS-CoV-2 concentrations and reported COVID-19 case counts (y-axis) vary depending on lead time (x-axis), correlation metric (panel a), incidence metric (b), and time (c) in Boston, Massachusetts during April 2020 to March 2021. Wastewater data were collected using methods previously described (Wu et al. 2020a; Wu et al. 2020b; data available at https://www.mwra.com/biobot/biobotdata.htm, North system). Case data includes cases in Suffolk and Middlesex Counties, MA, which are served by the wastewater plant (USA Facts; data available at https://usafacts.org/visualizations/coronavirus-covid-19-spread-map). Zero lead time refers to the correlation between wastewater and the case counts on the day of the wastewater sampling. Positive lead times refer to wastewater correlated with later case counts (e.g., a lead time of +3 days refers to the correlation between wastewater and cases 3 days later). Negative lead times refer to wastewater correlated with earlier case counts. Colors are only used to distinguish curves; the same color in different subplots are not necessarily related. (a) Spearman: Spearman correlation between 7-day average case counts (i.e., mean number of new case counts over the day of wastewater sampling and the previous 6 days) and wastewater virus concentrations. This is the baseline analysis. (a) Pearson: Pearson correlation between 7-day average case counts and wastewater virus concentrations. (a) Pearson (log): Pearson correlation between 7-day average case counts and base-10 logarithms of wastewater virus concentrations. (b) 7-day average (lagging): Baseline analysis; same as (a) Spearman. (b) 7-day average (center): Spearman correlation between centered 7-day average case counts (i.e., mean number of new case counts over the day of wastewater sampling, the preceding 3 days, and the following 3 days) and wastewater virus concentrations. (b) Raw case counts: Spearman correlation between daily number of new case counts (without smoothing) and wastewater virus concentrations. (c) Baseline analysis like in (a) Spearman, but using data only from each calendar quarter (e.g., 2020 Q2 is April-June 2020).