Many human pathogens are shed within bodily fluids during active infection and make their way into the domestic sanitary sewage system along several routes. Therefore, wastewater sample collection is a viable approach to monitor the prevalence of pathogens [1], including those of pandemic potential and biodefense/biosecurity relevance. Early in the COVID-19 pandemic, researchers identified that SARS-CoV-2 RNA was shed into fecal matter at viral loads high enough to be detected in wastewater [2-4]. Therefore, the preexisting field of wastewater-based epidemiology rallied to transition preexisting methods [5] from academic research into scalable public health surveillance tools. Especially early in the COVID-19 pandemic, traditional swab-based testing could not scale quickly enough to serve as a reliable source of population-level disease transmission data [5-7]. Multiple studies explored the efficacy of wastewater surveillance (WWS) as a stream of epidemiological data to complement case tracking for community transmission monitoring, finding that WWS data tracks trends in clinical case reporting data [8-10]. While the statistical correlation between viral load in sewage and clinical indicators is strong, the exact quantitative relationship between individual-level testing and WWS data is complex and depends on a variety of factors, including the epidemiology of the outbreak, data collection, and processing timelines [11]. Despite these complexities, WWS has been shown to be correlated with community infection dynamics [10] in addition to simply being an effective qualitative detection tool. Implementing WWS within institutional building complexes, such as college campuses, has unique challenges but also enables building-level resolution monitoring and early warning capabilities [12-15]. WWS can be a leading qualitative indicator of disease presence in a community when overall disease prevalence is low, making WWS a good candidate for broad-scale baseline pathogen monitoring [16]. Because WWS is passive and independent of health care seeking behavior, it provides a data stream complementary to active tracking of infections or hospitalizations, both of which have limitations. Additional benefits of WWS include the ability to monitor multiple pathogens, emerging viral variants, and nonbiological hazards [17].

The US Government prioritized WWS to track the spread of COVID-19 and other diseases. For example, environmental monitoring for viral threats via WWS is a key component of pandemic threat early warning systems prioritized in the Biden administration’s “American Pandemic Preparedness: Transforming our Capabilities” plan [18]. In addition, the Under Secretary of Defense for Personnel and Readiness directed the US Department of Defense (DoD) to leverage alternative technologies, including WWS, to supplement existing surveillance strategies in a memorandum titled “Consolidated Department of Defense Coronavirus Disease 2019 Force Health Protection Guidance” [19]. The Centers for Disease Control and Prevention established the National Wastewater Surveillance System [20] to broaden traditional diagnostic test surveillance systems by enabling efficient collection of community-level samples. In addition, the Centers for Disease Control and Prevention is applying WWS within passenger airplanes as part of its Traveler Genomic Surveillance program [21]. Finally, in June 2021, the National Institute of Standards and Technology and the Department of Homeland Security, Science and Technology Directorate convened a virtual workshop entitled “Standards to Support an Enduring Capability in Wastewater Surveillance for Public Health” to identify challenges and solutions for maturing and ensuring WWS capabilities for detecting and monitoring public health threats [22,23].

As a result of practical successes in early research and implementation studies, best practices emerged for how to implement WWS at scale [6]. WWS can be an important tool for epidemiological monitoring and outbreak response if implemented with consideration for various challenges [24,25]; one important aspect to consider is avoiding redundancy with clinical testing by implementing a joint surveillance strategy. The design of a WWS data collection scheme and methods of analysis can have significant impacts on bias and interpretation of the data [26]. When implemented according to best practices, WWS can be a cost-effective part of a public health response system [27]. Pairing WWS with clinical testing allows for both approaches to serve specific needs, thereby enhancing the cost-effectiveness of both [28].

The DoD has installations around the globe with small compact living communities, some of which have overlapping sewersheds with nearby cities. Tens of thousands of military personnel and civilians live and work in these installations. The DoD implements a four-tiered COVID-19 testing scheme. The first three tiers focus on staff at varying levels of critical service and deployment; tier 4 sentinel surveillance is an asymptomatic testing program designed to cover all personnel. Therefore, we focused on tier 4 sentinel surveillance as our point of comparison for WWS cost since WWS is also best suited for broad population monitoring.

Similarities exist between DoD installations and other institutional building complexes like college campuses. Yet implementing a WWS system at DoD sites requires special planning considerations given unique operational constraints and global scale. To address these issues, the DoD commissioned several WWS pilot studies aimed at figuring out the logistical, operational, and financial aspects of implementing a WWS program. One such study demonstrated the effectiveness of wastewater screening of blackwater from Coast Guard vessels [29]. Another study focused on WWS at air force bases (AFBs); the US Air Force (USAF) and Joint Program Executive Office for Chemical, Biological, Radiological and Nuclear Defense (JPEO-CBRND) WWS pilot study was larger than previous DoD pilots and more representative of US military installations globally. We analyze the cost-effectiveness of WWS within the DoD context, based on the results from the USAF and JPEO-CBRND WWS pilot study. We developed a cost model that includes upfront capital expenditures, operational expenditures, and indirect costs of lost work time. Further, we performed a break-even analysis to explore how traditional swab testing and WWS could be carried out in tandem within the budget of existing swab testing schemes. We concluded that WWS is cost-effective as a complementary passive community-level disease surveillance scheme, within the context of AFBs, and therefore likely would be cost-effective as a DoD-wide global multi-pathogen monitoring system that could be operated in complement to swab-based testing in the event of future disease outbreaks.

Of the 26 AFBs, 17 recorded WWS data during the period from September 2021 to January 2022 (Figure 1), demonstrating that sewage can be safely sampled in a field environment and at a fixed lab. The procedures implemented at sites during the pilot were designed to collect and process wastewater samples once per week. There were 53 data submissions recorded, and those submissions amounted to 45 unique viable sample records. Invalid submissions included duplicate records and samples with PCR reaction issues such as incubation temperature and sample concentration (Figure 2). Three sites were used strictly for an early feasibility pilot stage in which protocols were established. The remaining 14 sites submitted data collected over partially overlapping periods of 4.5 weeks on average. The pilot study identified 25 positive (or presumptive positive) samples and 20 negative samples in total. The 25 positive samples came from 12 of the 14 sites that collected samples systematically. However, the sites did not collect the same number of samples, and 2 sites that detected no positives were also the sites that submitted the fewest total samples. This procedure not only validated the feasibility of implementing WWS at AFBs but also highlighted considerable site-to-site variability in executing systematic sampling procedures. These results provided a case study from which we derive assumptions for the economic cost model.

The cost of SARS-CoV-2 WWS was estimated and compared to the estimated cost of tier 4 COVID-19 sentinel surveillance (asymptomatic testing) across 82 selected AFBs. The four scenarios modeled are described in Table 2. For each scenario, we used the cost model parameters to estimate the total direct and indirect costs of both WWS and tier 4 surveillance. Table 3 shows the total annual costs (in millions of 2021 US dollars) for each scenario at the 82 AFBs. In scenarios 1 and 2, we estimated that the direct costs of the tier 4 sentinel surveillance program would cost approximately US $18 million under baseline COVID-19 monitoring. We estimated that WWS would cost between US $5.4 million with once-weekly testing (scenario 1) to US $7.5 million with twice-weekly testing (scenario 2).

Enhanced surveillance is needed to manage the response during an outbreak, here defined broadly as either local- or national-level transmission that is sufficiently high such that strict control measures are put in place. Therefore, under COVID-19 outbreak monitoring scenarios 3 and 4, we estimated the direct costs of tier 4 sentinel surveillance to be US $24 million. The cost of WWS may also go up depending on policy decision-making; for example, sites could choose to test more frequently or utilize alternative pathogen detection methods. In scenario 3, only tier 4 sentinel surveillance is increased during the outbreak response, so the estimated WWS direct costs remain US $5.4 million. In contrast, scenario 4 assumed that the use of both tier 4 sentinel surveillance and WWS go up during the outbreak response, so the estimated WWS direct costs increased to US $8.2 million.

Tier 4 sentinel surveillance PCR testing requires that USAF staff take time to get tested, and this leads to additional costs. We used our model to estimate the cost of lost work, based on typical staff salary ranges and time required to get tested. We estimated that there would be a US $5.6 million cost for the loss of work associated with tier 4 sentinel surveillance PCR testing in scenarios with baseline testing. During a disease outbreak scenario leading to increased testing, we estimated the cost of lost work to be US $7.5 million. In contrast, WWS does not place any burden on staff not associated directly with the implementation of the surveillance program, and thus, we include zero additional costs during outbreaks for those scenarios.

We found that the cost of WWS was between US $10.5-$18.5 million less expensive annually in direct costs when compared to tier 4 sentinel surveillance and that tier 4 sentinel surveillance has an additional cost of US $5.6-$7.5 million annually due to USAF personnel losing time for testing. If WWS were implemented, there would still be the capacity to carry out a substantial amount of tier 4 sentinel surveillance PCR testing. We quantified the break-even point for combined WWS and PCR testing by calculating the number of PCR swab tests that could be conducted per base per month under the WWS paradigm while breaking even with the higher cost of the original tier 4 testing scheme. The results of our break-even analysis are shown in Table 4.

To estimate the break-even point based on direct costs only, we simply took the direct cost differences and divided them by the direct cost per swab test (US $62.39/test for materials and labor), spread across 82 bases and 12 months. Under COVID-19 baseline monitoring with once-weekly WWS (scenario 1), we estimated that an additional 204 swab tests per AFB per month could be added to the WWS protocol for the same cost as the original tier 4 sentinel surveillance scheme. If WWS was increased to twice weekly (scenario 2), then we estimated an additional 171 swab tests could be performed at the break-even point. When the cost of lost work is incorporated, including assumed lost work for swabs used to reach the break-even point, we estimated that 225 and 200 additional swab tests could be performed in scenarios 1 and 2, respectively. The demand for all forms of surveillance increases during an outbreak, so more swab tests can be performed at the break-even cost point. When the cost of lost work is included, we estimated that 323 and 289 additional swab tests could be performed for outbreak scenarios 3 and 4, respectively. We estimated that more than half of the tier 4 sentinel surveillance program could be maintained across all scenarios, while WWS is implemented in parallel with no additional cost (ie, at the break-even point).

The DoD SARS-CoV-2 WWS surveillance pilot studies demonstrated the feasibility of implementing WWS at military installations. The pilot studies revealed some important technical considerations. For example, although dPCR was extremely sensitive, it required shipping of wastewater from remote sites to a centralized location, potentially limiting its use in large-scale deployment. Portable quantitative PCR had a lower throughput of samples than dPCR but was simple to use at the point of sampling. Our simple cost model considered these lessons.

The pilot studies also provided real-world data on the costs associated with WWS when compared to standard swab-based testing, including material costs and labor requirements. In general, WWS for SARS-CoV-2 may offer several benefits, including earlier detection of outbreaks, lower costs and burden for community-wide coverage compared to diagnostic testing, and detection of viral presence regardless of symptoms. Coupling our analysis with the overall results of the DoD pilot studies suggests that those benefits are likely to hold for other DoD use cases. Specifically, our model suggests that the deployment of WWS to AFBs would be substantially more cost-effective than broad asymptomatic swab testing. Our break-even analysis indicated that without allocating additional funding to surveillance efforts, WWS could be implemented for AFB-level monitoring, and swab testing could be used for more targeted purposes or simply in parallel. It is important to note that swab testing and WWS do not provide the same information. Swab testing can enable individual-level actions, such as quarantining and contract tracing, and higher resolution data. Therefore, trade-offs between the public health benefits of WWS and swab testing must be determined case by case. These findings apply to both baseline COVID-19 monitoring and scenarios where outbreaks are occurring on bases throughout the year.

Our cost model was intentionally simplistic to enhance transparency for decision makers. That simplicity is also a limitation in that there may be unforeseen complexities and costs associated with scaling the WWS program beyond the pilot sites. In addition, our cost data is primarily derived from the USAF WWS pilot program, and facilities associated with other branches of the DoD may require different considerations. Many of our parameter estimates were obtained from SMEs (ie, individual DoD staff and contractors associated with the pilot studies) as opposed to independent reviews of pilot study budgets. Therefore, our cost estimates should not be interpreted as formal financial forecasts.

Another limitation of our analysis is the lack of an uncertainty estimation. Any formal program-level financial forecast would require uncertainty ranges to be estimated in addition to cost estimates. Many of the material costs were obtained directly from individuals with knowledge of the actual costs incurred during the pilot studies for this work. Therefore, our model could be framed as an estimate of what the actual cost would have been if the pilot was carried out at all AFBs rather than a forecast of the costs of a DoD-wide program—though, we believe our work is germane to that topic. Furthermore, systematic uncertainty in labor and materials costs due to changes in supply chain issues and inflation are likely correlated such that a proper uncertainty propagation would require estimating the joint distribution of costs, which is beyond the scope of our efforts. Given the magnitude of the point difference and the consensus in the literature that WWS is less expensive for population-level monitoring—albeit not necessarily cost-effective if implemented poorly [25]—we believe that our results are qualitatively robust to underlying uncertainty in the data and model specification.

In conclusion, we found that the USAF WWS pilot was a cost-effective complement to standard swab-based testing in the tier 4 sentinel surveillance program. We believe that WWS in tandem with swab-based testing is the best approach to maximize available resources. Looking beyond the COVID-19 pandemic, the DoD can be an important partner in global pandemic and all-hazard preparedness efforts. WWS is uniquely well suited to multi-threat biological surveillance, and our results suggest that the adoption of WWS across US military installations would help deliver a more comprehensive early warning system. A fully developed WWS program would complement civilian efforts like the National Wastewater Surveillance System and enable rapidly scalable outbreak monitoring in the event of future disease outbreaks.