Data modelling recipes for SARS-CoV-2 wastewater-based epidemiology

Visualizing Neighborhood COVID-19 Levels, Trends, and Inequities in Wastewater: An Equity-Centered Approach and Comparison to CDC Methods

CONTEXT

Monitoring neighborhood-level SARS-CoV-2 wastewater concentrations can help guide public health interventions and provide early warning ahead of lagging COVID-19 clinical indicators. To date, however, U.S. Centers for Disease Control and Prevention's (CDC) National Wastewater Surveillance System (NWSS) has provided methodology solely for communicating national and state-level "wastewater viral activity levels."

PROGRAM

In October 2022, the Boston Public Health Commission (BPHC) began routinely sampling wastewater at 11 neighborhood sites to better understand COVID-19 epidemiology and inequities across neighborhoods, which vary widely in sociodemographic and socioeconomic characteristics. We developed equity-centered methods to routinely report interpretable and actionable descriptions of COVID-19 wastewater levels, trends, and neighborhood-level inequities.

APPROACH AND IMPLEMENTATION

To produce these data visualizations, spanning October 2022 to December 2023, we followed four general steps: (1) smoothing raw values; (2) classifying current COVID-19 wastewater levels; (3) classifying current trends; and (4) reporting and visualizing results.

EVALUATION

COVID-19 wastewater levels corresponded well with lagged COVID-19 hospitalizations and deaths over time, with "Very High" wastewater levels coinciding with winter surges. When citywide COVID-19 levels were at the highest and lowest points, levels and trends tended to be consistent across sites. In contrast, when citywide levels were moderate, neighborhood levels and trends were more variable, revealing inequities across neighborhoods, emphasizing the importance of neighborhood-level results. Applying CDC/NWSS state-level methodology to neighborhood sites resulted in vastly different neighborhood-specific wastewater cut points for "High" or "Low," obscured inequities between neighborhoods, and systematically underestimated COVID-19 levels during surge periods in neighborhoods with the highest COVID-19 morbidity and mortality.

DISCUSSION

Our methods offer an approach that other local jurisdictions can use for routinely monitoring, comparing, and communicating neighborhood-level wastewater levels, trends, and inequities. Applying CDC/NWSS methodology at the neighborhood-level can obscure and perpetuate COVID-19 inequities. We recommend jurisdictions adopt equity-focused approaches in neighborhood-level wastewater surveillance for valid community comparisons.

Machine learning in wastewater: opportunities and challenges - "not everything is a nail!"

This paper highlights the potential of machine learning (ML) for wastewater applications, with a focus on key applications and considerations. It underscores the need for simplicity in ML models to ensure their interpretability and trustworthiness, cautioning against the use of overly complex 'black box' models unless absolutely necessary, especially with limited data. Not all modelling problems should be considered nails for which the ML hammer is the best-available tool. We emphasise the critical role of thorough data collection, including metadata, given its scarcity in some areas. Future research is encouraged to develop benchmark hybrid models to bridge the educational gap for environmental engineers and to establish best practices for managing data and model metadata, thereby improving ML's accessibility and utility in wastewater applications.

Estimates of SARS-CoV-2 Infections and Population Immunity After the COVID-19 Pandemic in Austria: Analysis of National Wastewater Data

BACKGROUND

Postpandemic surveillance data on coronavirus disease 2019 (COVID-19) infections may help inform future public health policies regarding severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) testing, vaccinations, or other COVID-19 measures. We estimate the total SARS-CoV-2 infections in Austria after the end of the pandemic from wastewater data and utilize these estimates to calculate the average national levels of SARS-CoV-2 infection protection and COVID-19 death protection.

METHODS

We estimated the total SARS-CoV-2 infections in Austria after the end of the pandemic (5 May 2023, per World Health Organization) up to May 2024 from wastewater data using a previously published model. These estimates were used in an agent-based model (ABM) to estimate average national levels of SARS-CoV-2 infection protection and COVID-19 death protection, based on waning immunity estimates of infections and vaccination in previous literature.

RESULTS

We estimate approximately 3.2 million infections between 6 May 2023 and 23 May 2024, with a total of 17.8 million infections following 12 May 2020. The ABM estimates that the national average death protection was approximately 82% higher in May 2024 than before the pandemic. This represents a relative decrease of 8% since May 2023. It also shows that 95% of people in Austria were infected with SARS-CoV-2 at least once by May 2024. National infection protection remained relatively low after the onset of Omicron.

CONCLUSIONS

These findings should be considered for public health decisions on SARS-CoV-2 testing practices and vaccine booster administrations.

A multivariate analysis to explain residue errors in pathogen concentration in wastewater-based epidemiology

With the beginning of the COVID-19 pandemic, wastewater-based epidemiology (WBE), which according to Larsen et al. (2021), describes the science of linking pathogens and chemicals found in wastewater to population-level health, received an enormous boost worldwide. The basic procedure in WBE is to analyse pathogen concentrations and to relate these measurements to cases from clinical data. This prediction of cases is subject to large errors, due to various factors such as dilution effects, decay or wastewater matrix and inhibitors. In this study we used different models to identify the most important, what we call, wastewater-based epidemiologically relevant parameters (WBERP) to describe these errors. We used linear regression and random forest regression as base models for predicting cases and random forest regression also to analyse the importance of different WBERP. Two catchments, one with a large proportion of combined sewers and one with separate sewers, served as study areas. Our results show that the most important information to be included in any model are the variants of concern (VOCs), a time-variable parameter. The performance for both catchments is improved by ~30 % in terms of root mean square error when the VOCs are used as additional information. For practical applications, this is a real drawback as it means that every time a new pathogen variant becomes dominant, we need to know the specific behaviour of the variant in the wastewater and its detection in order to interpret the WBE data correctly. This limits the predictive capabilities of such systems, perhaps not in terms of dynamics but for quantitative statements. The addition of other physicochemical parameters and faecal markers only marginally improved the results. Furthermore, there were differences in the importance of the parameters between the catchments, which limits the generalisability of the conclusions. The results show that more complex wastewater matrices (high proportion of combined sewer system) influence the relationship between pathogen concentration and medical cases more than those of less complex wastewater matrices (separate sewer system).

Epidemiological model can forecast COVID-19 outbreaks from wastewater-based surveillance in rural communities

Wastewater has emerged as a crucial tool for infectious disease surveillance, offering a valuable means to bridge the equity gap between underserved communities and larger urban municipalities. However, using wastewater surveillance in a predictive manner remains a challenge. In this study, we tested if detecting SARS-CoV-2 in wastewater can forecast outbreaks in rural communities. Under the CDC National Wastewater Surveillance program, we monitored the SARS-CoV-2 in the wastewater of five rural communities and a small city in Idaho (USA). We then used a particle filter method coupled with a stochastic susceptible-exposed-infectious-recovered (SEIR) model to infer active case numbers using quantities of SARS-CoV-2 in wastewater. Our findings revealed that while high daily variations in wastewater viral load made real-time interpretation difficult, the SEIR model successfully factored out this noise, enabling accurate forecasts of the Omicron outbreak in five of the six towns shortly after initial increases in SARS-CoV-2 concentrations were detected in wastewater. The model predicted outbreaks with a lead time of 0 to 11 days (average of 6 days +/- 4) before the surge in reported clinical cases. This study not only underscores the viability of wastewater-based epidemiology (WBE) in rural communities-a demographic often overlooked in WBE research-but also demonstrates the potential of advanced epidemiological modeling to enhance the predictive power of wastewater data. Our work paves the way for more reliable and timely public health guidance, addressing a critical gap in the surveillance of infectious diseases in rural populations.

Wastewater-based epidemiology: deriving a SARS-CoV-2 data validation method to assess data quality and to improve trend recognition

Introduction

Accurate and consistent data play a critical role in enabling health officials to make informed decisions regarding emerging trends in SARS-CoV-2 infections. Alongside traditional indicators such as the 7-day-incidence rate, wastewater-based epidemiology can provide valuable insights into SARS-CoV-2 concentration changes. However, the wastewater compositions and wastewater systems are rather complex. Multiple effects such as precipitation events or industrial discharges might affect the quantification of SARS-CoV-2 concentrations. Hence, analysing data from more than 150 wastewater treatment plants (WWTP) in Germany necessitates an automated and reliable method to evaluate data validity, identify potential extreme events, and, if possible, improve overall data quality.

Methods

We developed a method that first categorises the data quality of WWTPs and corresponding laboratories based on the number of outliers in the reproduction rate as well as the number of implausible inflection points within the SARS-CoV-2 time series. Subsequently, we scrutinised statistical outliers in several standard quality control parameters (QCP) that are routinely collected during the analysis process such as the flow rate, the electrical conductivity, or surrogate viruses like the pepper mild mottle virus. Furthermore, we investigated outliers in the ratio of the analysed gene segments that might indicate laboratory errors. To evaluate the success of our method, we measure the degree of accordance between identified QCP outliers and outliers in the SARS-CoV-2 concentration curves.

Results and discussion

Our analysis reveals that the flow and gene segment ratios are typically best at identifying outliers in the SARS-CoV-2 concentration curve albeit variations across WWTPs and laboratories. The exclusion of datapoints based on QCP plausibility checks predominantly improves data quality. Our derived data quality categories are in good accordance with visual assessments.

Conclusion

Good data quality is crucial for trend recognition, both on the WWTP level and when aggregating data from several WWTPs to regional or national trends. Our model can help to improve data quality in the context of health-related monitoring and can be optimised for each individual WWTP to account for the large diversity among WWTPs.

Effective Inhibitor Removal from Wastewater Samples Increases Sensitivity of RT-dPCR and Sequencing Analyses and Enhances the Stability of Wastewater-Based Surveillance

Wastewater-based surveillance (WBS) is a proven tool for monitoring population-level infection events. Wastewater contains high concentrations of inhibitors, which contaminate the total nucleic acids (TNA) extracted from these samples. We found that TNA extracts from raw influent of Berlin wastewater treatment plants contained highly variable amounts of inhibitors that impaired molecular analyses like dPCR and next-generation sequencing (NGS). By using dilutions, we were able to detect inhibitory effects. To enhance WBS sensitivity and stability, we applied a combination of PCR inhibitor removal and TNA dilution (PIR+D). This approach led to a 26-fold increase in measured SARS-CoV-2 concentrations, practically reducing the detection limit. Additionally, we observed a substantial increase in the stability of the time series. We define suitable stability as a mean absolute error (MAE) below 0.1 log10 copies/L and a geometric mean relative absolute error (GMRAE) below 26%. Using PIR+D, the MAE could be reduced from 0.219 to 0.097 and the GMRAE from 65.5% to 26.0%, and even further in real-world WBS. Furthermore, PIR+D improved SARS-CoV-2 genome alignment and coverage in amplicon-based NGS for low to medium concentrations. In conclusion, we strongly recommend both the monitoring and removal of inhibitors from samples for WBS.

Estimating rates of change to interpret quantitative wastewater surveillance of disease trends

BACKGROUND

Wastewater monitoring data can be used to estimate disease trends to inform public health responses. One commonly estimated metric is the rate of change in pathogen quantity, which typically correlates with clinical surveillance in retrospective analyses. However, the accuracy of rate of change estimation approaches has not previously been evaluated.

OBJECTIVES

We assessed the performance of approaches for estimating rates of change in wastewater pathogen loads by generating synthetic wastewater time series data for which rates of change were known. Each approach was also evaluated on real-world data.

METHODS

Smooth trends and their first derivatives were jointly sampled from Gaussian processes (GP) and independent errors were added to generate synthetic viral load measurements; the range hyperparameter and error variance were varied to produce nine simulation scenarios representing different potential disease patterns. The directions and magnitudes of the rate of change estimates from four estimation approaches (two established and two developed in this work) were compared to the GP first derivative to evaluate classification and quantitative accuracy. Each approach was also implemented for public SARS-CoV-2 wastewater monitoring data collected January 2021-May 2023 at 25 sites in North Carolina, USA.

RESULTS

All four approaches inconsistently identified the correct direction of the trend given by the sign of the GP first derivative. Across all nine simulated disease patterns, between a quarter and a half of all estimates indicated the wrong trend direction, regardless of estimation approach. The proportion of trends classified as plateaus (statistically indistinguishable from zero) for the North Carolina SARS-CoV-2 data varied considerably by estimation method but not by site.

DISCUSSION

Our results suggest that wastewater measurements alone might not provide sufficient data to reliably track disease trends in real-time. Instead, wastewater viral loads could be combined with additional public health surveillance data to improve predictions of other outcomes.

Machine learning surveillance of foodborne infectious diseases using wastewater microbiome, crowdsourced, and environmental data

Clostridium perfringens (CP) is a common cause of foodborne infection, leading to significant human health risks and a high economic burden. Thus, effective CP disease surveillance is essential for preventive and therapeutic interventions; however, conventional practices often entail complex, resource-intensive, and costly procedures. This study introduced a data-driven machine learning (ML) modeling framework for CP-related disease surveillance. It leveraged an integrated dataset of municipal wastewater microbiome (e.g., CP abundance), crowdsourced (CP-related web search keywords), and environmental data. Various optimization strategies, including data integration, data normalization, model selection, and hyperparameter tuning, were implemented to improve the ML modeling performance, leading to enhanced predictions of CP cases over time. Explainable artificial intelligence methods identified CP abundance as the most reliable predictor of CP disease cases. Multi-omics subsequently revealed the presence of CP and its genotypes/toxinotypes in wastewater, validating the utility of microbiome-data-enabled ML surveillance for foodborne diseases. This ML-based framework thus exhibits significant potential for complementing and reinforcing existing disease surveillance systems.

Normalization of viral loads in Wastewater-Based Epidemiology using routine parameters: One year monitoring of SARS-CoV-2 in urban and tourist sewersheds

In wastewater-based epidemiology, normalization of experimental data is a crucial aspect, as emerged in the recent surveillance of COVID-19. Normalization facilitates the comparison between different areas or periods, and it helps in evaluating the differences due to the fluctuation of the population due to seasonal employment or tourism. Analysis of biomarkers in wastewater (i.e. drugs, beverage and food compounds, microorganisms such as PMMoV or crAssphage, etc.) is complex to perform, and it is not routinely monitored. This study compares the results of alternative normalization approaches applied to SARS-CoV-2 loads in wastewater using population size calculated with conventional hydraulic and/or chemical parameters (i.e. total suspended solids, chemical oxygen demand, nitrogen forms, etc.) commonly used in the routine monitoring of water quality. A total of 12 wastewater treatment plants were monitored, and 1068 samples of influent wastewater were collected in urban areas and in highly touristic areas (summer and/or winter). The results indicated that both census and population estimated with ammonium are effective and reliable parameters with which to normalize SARS-CoV-2 loads in wastewater from urban sewersheds with negligible fluctuating populations. However, this study reveals that, in the case of tourist locations, the population calculated using NH4-N loads can provide a better normalization of the specific viral load per inhabitant.

Wastewater Surveillance to Confirm Differences in Influenza A Infection between Michigan, USA, and Ontario, Canada, September 2022-March 2023

Wastewater surveillance is an effective way to track the prevalence of infectious agents within a community and, potentially, the spread of pathogens between jurisdictions. We conducted a retrospective wastewater surveillance study of the 2022-23 influenza season in 2 communities, Detroit, Michigan, USA, and Windsor-Essex, Ontario, Canada, that form North America's largest cross-border conurbation. We observed a positive relationship between influenza-related hospitalizations and the influenza A virus (IAV) wastewater signal in Windsor-Essex (ρ = 0.785; p<0.001) and an association between influenza-related hospitalizations in Michigan and the IAV wastewater signal for Detroit (ρ = 0.769; p<0.001). Time-lagged cross correlation and qualitative examination of wastewater signal in the monitored sewersheds showed the peak of the IAV season in Detroit was delayed behind Windsor-Essex by 3 weeks. Wastewater surveillance for IAV reflects regional differences in infection dynamics which may be influenced by many factors, including the timing of vaccine administration between jurisdictions.

Model parameter estimation with imprecise information

Model parameter estimation is a well-known inverse problem, as long as single-value point data are available as observations of system performance measurement. However, classical statistical methods, such as the minimization of an objective function or maximum likelihood, are no longer straightforward, when measurements are imprecise in nature. Typical examples of the latter include censored data and binary information. Here, we explore Approximate Bayesian Computation as a simple method to perform model parameter estimation with such imprecise information. We demonstrate the method for the example of a plain rainfall-runoff model and illustrate the advantages and shortcomings. Last, we outline the value of Shapley values to determine which type of observation contributes to the parameter estimation and which are of minor importance.

Estimating the COVID-19 prevalence from wastewater

Wastewater based epidemiology has become a widely used tool for monitoring trends of concentrations of different pathogens, most notably and widespread of SARS-CoV-2. Therefore, in 2022, also in Rhineland-Palatinate, the Ministry of Science and Health has included 16 wastewater treatment sites in a surveillance program providing biweekly samples. However, the mere viral load data is subject to strong fluctuations and has limited value for political deciders on its own. Therefore, the state of Rhineland-Palatinate has commissioned the University Medical Center at Johannes Gutenberg University Mainz to conduct a representative cohort study called SentiSurv, in which an increasing number of up to 12,000 participants have been using sensitive antigen self-tests once or twice a week to test themselves for SARS-CoV-2 and report their status. This puts the state of Rhineland-Palatinate in the fortunate position of having time series of both, the viral load in wastewater and the prevalence of SARS-CoV-2 in the population. Our main contribution is a calibration study based on the data from 2023-01-08 until 2023-10-01 where we identified a scaling factor ( 0.208 ± 0.031 ) and a delay ( 5.07 ± 2.30 days) between the virus load in wastewater, normalized by the pepper mild mottle virus (PMMoV), and the prevalence recorded in the SentiSurv study. The relation is established by fitting an epidemiological model to both time series. We show how that can be used to estimate the prevalence when the cohort data is no longer available and how to use it as a forecasting instrument several weeks ahead of time. We show that the calibration and forecasting quality and the resulting factors depend strongly on how wastewater samples are normalized.

Wastewater-based epidemiology as a public health resource in low- and middle-income settings

In the face of emerging and re-emerging diseases, novel and innovative approaches to population scale surveillance are necessary for the early detection and quantification of pathogens. The last decade has seen the rapid development of wastewater and environmental surveillance (WES) to address public health challenges, which has led to establishment of wastewater-based epidemiology (WBE) approaches being deployed to monitor a range of health hazards. WBE exploits the fact that excretions and secretions from urine, and from the gut are discharged in wastewater, particularly sewage, such that sampling sewage systems provides an early warning system for disease outbreaks by providing an early indication of pathogen circulation. While WBE has been mainly used in locations with networked wastewater systems, here we consider its value for less connected populations typical of lower-income settings, and in assess the opportunity afforded by pit latrines to sample communities and localities. We propose that where populations struggle to access health and diagnostic facilities, and despite several additional challenges, sampling unconnected wastewater systems remains an important means to monitor the health of large populations in a relatively cost-effective manner.

Wastewater-Based Epidemiology for SARS-CoV-2 in Northern Italy: A Spatiotemporal Model

The study investigated the application of Wastewater-Based Epidemiology (WBE) as a tool for monitoring the SARS-CoV-2 prevalence in a city in northern Italy from October 2021 to May 2023. Based on a previously used deterministic model, this study proposed a variation to account for the population characteristics and virus biodegradation in the sewer network. The model calculated virus loads and corresponding COVID-19 cases over time in different areas of the city and was validated using healthcare data while considering viral mutations, vaccinations, and testing variability. The correlation between the predicted and reported cases was high across the three waves that occurred during the period considered, demonstrating the ability of the model to predict the relevant fluctuations in the number of cases. The population characteristics did not substantially influence the predicted and reported infection rates. Conversely, biodegradation significantly reduced the virus load reaching the wastewater treatment plant, resulting in a 30% reduction in the total virus load produced in the study area. This approach can be applied to compare the virus load values across cities with different population demographics and sewer network structures, improving the comparability of the WBE data for effective surveillance and intervention strategies.

From pandemic to endemic: Divergence of COVID-19 positive-tests and hospitalization numbers from SARS-CoV-2 RNA levels in wastewater of Rochester, Minnesota

Traditionally, public health surveillance relied on individual-level data but recently wastewater-based epidemiology (WBE) for the detection of infectious diseases including COVID-19 became a valuable tool in the public health arsenal. Here, we use WBE to follow the course of the COVID-19 pandemic in Rochester, Minnesota (population 121,395 at the 2020 census), from February 2021 to December 2022. We monitored the impact of SARS-CoV-2 infections on public health by comparing three sets of data: quantitative measurements of viral RNA in wastewater as an unbiased reporter of virus level in the community, positive results of viral RNA or antigen tests from nasal swabs reflecting community reporting, and hospitalization data. From February 2021 to August 2022 viral RNA levels in wastewater were closely correlated with the oscillating course of COVID-19 case and hospitalization numbers. However, from September 2022 cases remained low and hospitalization numbers dropped, whereas viral RNA levels in wastewater continued to oscillate. The low reported cases may reflect virulence reduction combined with abated inclination to report, and the divergence of virus levels in wastewater from reported cases may reflect COVID-19 shifting from pandemic to endemic. WBE, which also detects asymptomatic infections, can provide an early warning of impending cases, and offers crucial insights during pandemic waves and in the transition to the endemic phase.

Estimating actual SARS-CoV-2 infections from secondary data

Eminent in pandemic management is accurate information on infection dynamics to plan for timely installation of control measures and vaccination campaigns. Despite huge efforts in diagnostic testing of individuals, the underestimation of the actual number of SARS-CoV-2 infections remains significant due to the large number of undocumented cases. In this paper we demonstrate and compare three methods to estimate the dynamics of true infections based on secondary data i.e., (a) test positivity, (b) infection fatality and (c) wastewater monitoring. The concept is tested with Austrian data on a national basis for the period of April 2020 to December 2022. Further, we use the results of prevalence studies from the same period to generate (upper and lower bounds of) credible intervals for true infections for four data points. Model parameters are subsequently estimated by applying Approximate Bayesian Computation-rejection sampling and Genetic Algorithms. The method is then validated for the case study Vienna. We find that all three methods yield fairly similar results for estimating the true number of infections, which supports the idea that all three datasets contain similar baseline information. None of them is considered superior, as their advantages and shortcomings depend on the specific case study at hand.

Wastewater-based epidemiology for COVID-19 using dynamic artificial neural networks

Global efforts in vaccination have led to a decrease in COVID-19 mortality but a high circulation of SARS-CoV-2 is still observed in several countries, resulting in some cases of severe lockdowns. In this sense, wastewater-based epidemiology remains a powerful tool for supporting regional health administrations in assessing risk levels and acting accordingly. In this work, a dynamic artificial neural network (DANN) has been developed for predicting the number of COVID-19 hospitalized patients in hospitals in Valladolid (Spain). This model takes as inputs a wastewater epidemiology indicator for COVID-19 (concentration of RNA from SARS-CoV-2 N1 gene reported from Valladolid Wastewater Treatment Plant), vaccination coverage, and past data of hospitalizations. The model considered both the instantaneous values of these variables and their historical evolution. Two study periods were selected (from May 2021 until September 2022 and from September 2022 to July 2023). During the first period, accurate predictions of hospitalizations (with an overall range between 6 and 171) were favored by the correlation of this indicator with N1 concentrations in wastewater (r = 0.43, p < 0.05), showing accurate forecasting for 1 day ahead and 5 days ahead. The second period's retraining strategy maintained the overall accuracy of the model despite lower hospitalizations. Furthermore, risk levels were assigned to each 1 day ahead prediction during the first and second periods, showing agreement with the level measured and reported by regional health authorities in 95 % and 93 % of cases, respectively. These results evidenced the potential of this novel DANN model for predicting COVID-19 hospitalizations based on SARS-CoV-2 wastewater concentrations at a regional scale. The model architecture herein developed can support regional health authorities in COVID-19 risk management based on wastewater-based epidemiology.

Estimating COVID-19 cases on a university campus based on Wastewater Surveillance using machine learning regression models

Wastewater Surveillance (WS) is a crucial tool in the management of COVID-19 pandemic. The surveillance is based on enumerating SARS-CoV-2 RNA concentrations in the community's sewage. In this study, we used WS data to develop a regression model for estimating the number of active COVID-19 cases on a university campus. Eight univariate and multivariate regression model types i.e. Linear Regression (LM), Polynomial Regression (PR), Generalised Additive Model (GAM), Locally Estimated Scatterplot Smoothing Regression (LOESS), K Nearest Neighbours Regression (KNN), Support Vector Regression (SVR), Artificial Neural Networks (ANN) and Random Forest (RF) were developed and compared. We found that the multivariate RF regression model, was the most appropriate for predicting the prevalence of COVID-19 infections at both a campus level and hostel-level. We also found that smoothing the normalised SARS-CoV-2 data and employing multivariate modelling, using student population as a second independent variable, significantly improved the performance of the models. The final RF campus level model showed good accuracy when tested using previously unseen data; correlation coefficient of 0.97 and a mean absolute error (MAE) of 20 %. In summary, our non-intrusive approach has the ability to complement projections based on clinical tests, facilitating timely follow-up and response.

A time series based machine learning strategy for wastewater-based forecasting and nowcasting of COVID-19 dynamics

Monitoring COVID-19 infection cases has been a singular focus of many policy makers and communities. However, direct monitoring through testing has become more onerous for a number of reasons, such as costs, delays, and personal choices. Wastewater-based epidemiology (WBE) has emerged as a viable tool for monitoring disease prevalence and dynamics to supplement direct monitoring. The objective of this study is to intelligently incorporate WBE information to nowcast and forecast new weekly COVID-19 cases and to assess the efficacy of such WBE information for these tasks in an interpretable manner. The methodology consists of a time-series based machine learning (TSML) strategy that can extract deeper knowledge and insights from temporal structured WBE data in the presence of other relevant temporal variables, such as minimum ambient temperature and water temperature, to boost the capability for predicting new weekly COVID-19 case numbers. The results confirm that feature engineering and machine learning can be utilized to enhance the performance and interpretability of WBE for COVID-19 monitoring, along with identifying the different recommended features to be applied for short-term and long-term nowcasting and short-term and long-term forecasting. The conclusion of this research is that the proposed time-series ML methodology performs as well, and sometimes better, than simple predictions that assume available and accurate COVID-19 case numbers from extensive monitoring and testing. Overall, this paper provides an insight into the prospects of machine learning based WBE to the researchers and decision-makers as well as public health practitioners for predicting and preparing the next wave of COVID-19 or the next pandemic.

Prediction of hospitalisations based on wastewater-based SARS-CoV-2 epidemiology

Wastewater-based epidemiology is widely applied in Austria since April 2020 to monitor the SARS-CoV-2 pandemic. With a steadily increasing number of monitored wastewater facilities, 123 plants covering roughly 70 % of the 9 million population were monitored as of August 2022. In this study, the SARS-CoV-2 viral concentrations in raw sewage were analysed to infer short-term hospitalisation occupancy. The temporal lead of wastewater-based epidemiological time series over hospitalisation occupancy levels facilitates the construction of forecast models. Data pre-processing techniques are presented, including the approach of comparing multiple decentralised wastewater signals with aggregated and centralised clinical data. Time‑lead quantification was performed using cross-correlation analysis and coefficient of determination optimisation approaches. Multivariate regression models were successfully applied to infer hospitalisation bed occupancy. The results show a predictive potential of viral loads in sewage towards Covid-19 hospitalisation occupancy, with an average lead time towards ICU and non-ICU bed occupancy between 14.8-17.7 days and 8.6-11.6 days, respectively. The presented procedure provides access to the trend and tipping point behaviour of pandemic dynamics and allows the prediction of short-term demand for public health services. The results showed an increase in forecast accuracy with an increase in the number of monitored wastewater treatment plants. Trained models are sensitive to changing variant types and require recalibration of model parameters, likely caused by immunity by vaccination and/or infection. The utilised approach displays a practical and rapidly implementable application of wastewater-based epidemiology to infer hospitalisation occupancy.

Expanding the Pathogen Panel in Wastewater Epidemiology to Influenza and Norovirus

Since the start of the 2019 pandemic, wastewater-based epidemiology (WBE) has proven to be a valuable tool for monitoring the prevalence of SARS-CoV-2. With methods and infrastructure being settled, it is time to expand the potential of this tool to a wider range of pathogens. We used over 500 archived RNA extracts from a WBE program for SARS-CoV-2 surveillance to monitor wastewater from 11 treatment plants for the presence of influenza and norovirus twice a week during the winter season of 2021/2022. Extracts were analyzed via digital PCR for influenza A, influenza B, norovirus GI, and norovirus GII. Resulting viral loads were normalized on the basis of NH4-N. Our results show a good applicability of ammonia-normalization to compare different wastewater treatment plants. Extracts originally prepared for SARS-CoV-2 surveillance contained sufficient genomic material to monitor influenza A, norovirus GI, and GII. Viral loads of influenza A and norovirus GII in wastewater correlated with numbers from infected inpatients. Further, SARS-CoV-2 related non-pharmaceutical interventions affected subsequent changes in viral loads of both pathogens. In conclusion, the expansion of existing WBE surveillance programs to include additional pathogens besides SARS-CoV-2 offers a valuable and cost-efficient possibility to gain public health information.