Wastewater surveillance overcomes socio-economic limitations of laboratory-based surveillance when monitoring disease transmission: The South African experience during the COVID-19 pandemic

Abstract

Wastewater and environmental surveillance has been promoted as a communicable disease surveillance tool because it overcomes inherent biases in laboratory-based communicable disease surveillance. Yet, little empirical evidence exists to support this notion, and it remains largely an intuitive, though highly plausible hypothesis. Our interdisciplinary study uses WES data to show evidence for underreporting of SARS-CoV-2 in the context of measurable and statistically significant associations between economic conditions and SARS-CoV-2 incidence and testing rates. We obtained geolocated, anonymised, laboratory-confirmed SARS-CoV-2 cases, wastewater SARS-CoV-2 viral load data and socio-demographic data for Gauteng Province, South Africa. We spatially located all data to create a single dataset for sewershed catchments served by two large wastewater treatment plants. We conducted epidemiological, persons infected and principal component analysis to explore the relationships between variables. Overall, we demonstrate the co-contributory influences of socio-economic indicators on access to SARS-CoV-2 testing and cumulative incidence, thus reflecting that apparent incidence rates mirror access to testing and socioeconomic considerations rather than true disease epidemiology. These analyses demonstrate how WES provides valuable information to contextualise and interpret laboratory-based epidemiological data. Whilst it is useful to have these associations established for SARS-CoV-2, the implications beyond SARS-CoV-2 are legion for two reasons, namely that biases inherent in clinical surveillance are broadly applicable across pathogens and all pathogens infecting humans will find their way into wastewater albeit in varying quantities. WES should be implemented to strengthen surveillance systems, especially where economic inequalities limit interpretability of conventional surveillance data.

Fig 1. Diagram indicating the direction of influence of socio-economic status, population structure and mixing on the true burden of SARS-CoV-2 cases, reported burden of SARS-CoV-2 cases and levels of SARS-CoV-2 in wastewater.

The green arrow indicates the research question posed by this work.

Fig 2. Map showing the spatial location of the two sewersheds and locations of health care facilities within the Gauteng Province [20].

Fig 3. SARS-CoV-2 testing rate (per 100 000 population), incidence rate (per 100 000 population), and 4 week moving average proportion test positive (%) by epidemiological week 22, 2021 to week 10, 2022, for (a) sewershed D and (b) sewershed O.

Fig 4. Truncated correlation matrices between socio-economic and demographic parameters against testing, cumulative incidence and mean proportion test positive for sewersheds D and O, annotated as a ‘heat map’ to represents the Spearman’s correlation coefficient (r) between socio-economic, demographic variables vs clinical variables (1 June 2021 to 18 March 2022).

Fig 5. PCA biplots displaying socioeconomic and demographic status parameters, cumulative incidence rate, testing rate and mean positivity rate within (a) sewershed D and (b) sewershed O.

In the biplots, the magnitude and colouring of the vectors are related to the variable loading scores, while the vector direction and quadrant location is informed by the interrelationship between variables and their contribution to dimensions 1 and 2.

Fig 6. SARS-CoV-2 concentrations in wastewater in log-transformed genome copies per millilitre (right axis) and the number of laboratory-confirmed cases (top figures, green bars) or incidence per 100,000 persons (bottom figures, blue bars) of SARS-CoV-2 geolocated to a residential address in the sewershed by epidemiological week from week 22, 2021 to week 10, 2022 for sewersheds D (left, figures a and c respectively) and O (right, figures b and d respectively).

Fig 7. A scatter plot with inset illustrating the relationship between laboratory-confirmed cases of SARS-CoV-2 identified in sewersheds D and O during the study period (epidemiological week 22, 2021 to week 10, 2022) compared to estimated SARS-CoV-2 infections during the same epidemiological week, calculated using wastewater levels of SARS-CoV-2, wastewater flow rates and assumed per person excretion rates.