ABSTRACT
BackgroundWastewater-based epidemiology is a promising approach but robust epidemiological models to relate wastewater to community prevalence are lacking. Assessments of SARS-CoV-2 infection rates have relied primarily on convenience sampling, which does not provide reliable estimates of community prevalence because of inherent biases. MethodsFrom August 2020 to February 2021, we conducted a serial stratified randomized samplings to estimate the prevalence of anti-SARS-CoV-2 antibodies in 3,717 participants, and weekly sampling of community wastewater for SARS-CoV-2 concentrations in Jefferson County, KY. With the use of a Susceptible, Infected, Recovered (SIR)-type model, we obtained longitudinal estimates of prevalence and compared these with wastewater concentration, using regression analysis. FindingsModel analysis revealed significant temporal differences in epidemic peaks; the average incidence rate based on serological sampling in some areas was up to 50% higher than health department rates based on convenience sampling. The model-estimated average prevalence rates correlated well with wastewater (correlation=0{middle dot}63). In regression analysis, a weekly unit increase in wastewater concentration of SARS-CoV-2 corresponded to an average increase of between 1-1{middle dot}3 cases of SARS-CoV-2 infection per 100K residents. InterpretationPublicly available health department incidence rates vastly underestimate true community incidence and wastewater has a high potential to provide robust estimates of community spread of infection. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSAdministratively reported clinical case rates of coronavirus disease 2019 (COVID-19) infected individuals are biased due to a wide range of factors from testing access to concerns about missing low and non-symptomatic and self-tested individuals. Wastewater estimates offer an alternative to support community monitoring based on fecal shedding of the virus but are difficult to interpret when compared with the available public health data sets of infection rates. We examined all English literature until February 24, 2022, on Web of Science and PubMed with the terms ["seroprevalence" or "antibody"] AND ["COVID-19" or "SARS-CoV-2"] AND ["wastewater"]. We identified six studies. None of these studies considered randomized COVID-19 community anti-SARS-CoV-2 antibody testing paired with wastewater data. Added value of this studyThe study demonstrates how results from serial stratified randomized serological sampling of the community can be used to build a longitudinal model that can interpolate and extrapolate community levels of infection beyond specific testing dates. Such a model correlates well with wastewater concentrations indicating its utility as a surrogate for infection prevalence. The testing data used in the study were collected before wide availability of COVID-19 vaccines and are therefore unique as they are unlikely to include a significant number of false positive results. Implications of all the available evidenceThe study demonstrates that convenience sampling obtained data from health department reporting seriously underestimates community-wide prevalence of infection. In contrast, wastewater-based epidemiology may be a faster, cost-effective, and more robust method of estimating the prevalence of viral infections within specific urban areas.
ABSTRACT
The majority of sewer systems in the United States and other countries, are operated by public utilities. In the absence of any regulation, public perception of monitoring wastewater for population health biomarkers is an important consideration for a public utility commission when allocating resources for this purpose. In August 2021, we conducted a survey as part of an ongoing COVID-19 community prevalence study in Louisville/Jefferson County, KY. The survey comprised of seven questions about awareness of and privacy concerns and was sent to 32,000 households randomly distributed within the county. A total of 1,220 sampled adults participated in the probability sample, and 981 were used in analysis. A total of 2,444 adults additionally responded in the convenience sample, and 1,751 were used in analysis. The samples were weighted to produce estimates representative of all adults in the county. Public awareness of tracking COVID-19 virus in the sewers was low. Opinions about how data from this activity are shared strongly supported public disclosure of monitoring results. Responses showed more support for measuring the largest areas (>30,000 to 50,000 households) typically representing population levels found in a community or regional wastewater treatment plant. Those who had a history of COVID-19 infection were more likely to support highly localized monitoring. Understanding wastewater surveillance strategies and thresholds of privacy concerns requires in-depth, comprehensive analysis of public opinion for continued success and efficacy of public health monitoring. Graphic for Table of Contents (TOC)/Abstract Art O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=44 SRC="FIGDIR/small/21264954v1_ufig1.gif" ALT="Figure 1"> View larger version (25K): org.highwire.dtl.DTLVardef@8774aforg.highwire.dtl.DTLVardef@fdbeaborg.highwire.dtl.DTLVardef@f0fc3forg.highwire.dtl.DTLVardef@14097dd_HPS_FORMAT_FIGEXP M_FIG C_FIG
