Quantitatively assessing early detection strategies for mitigating COVID-19 and future pandemics

The study addresses how much earlier detection of emerging pathogens via proactive surveillance could mitigate pandemics, focusing on the initial stages of COVID-19 and generalizing to future diseases. Prior work largely evaluated non-pharmaceutical interventions (NPIs) such as mobility restrictions, school closures, and testing strategies, but comparatively little quantified the benefits of earlier detection itself. Earlier detection could enable intervention when outbreaks are smaller, making resource-intensive strategies like test-trace-isolate more feasible and allowing time to scale healthcare capacity. Given major policy proposals to invest in early warning systems (e.g., hospital-based syndromic testing, wastewater surveillance, and air travel monitoring), this work quantitatively examines the benefit such systems might offer, using COVID-19 in Wuhan as a case study and extending to diseases with varied epidemiological parameters.

The paper situates its contribution against studies evaluating NPIs for COVID-19, work optimizing testing frequencies and turnaround times, and investigations into surveillance at hospitals, wastewater systems, and travel hubs during COVID-19. It references methodological advances in wastewater surveillance and multiplex detection technologies (PCR, CRISPR-based diagnostics, metagenomic sequencing). Historical observations that pathogens like Ebola are often first detected in hospitals and that wastewater surveillance has been effective for polio inform expectations about system performance across diseases with differing hospitalization rates and fecal shedding characteristics.

The authors develop a branching process-based model to predict cumulative cases at detection for specified detection systems and disease parameters. Offspring distribution is negative binomial with mean R0 and dispersion parameter. Detection systems modeled include: (i) hospital monitoring (testing patients with severe infectious symptoms for panels of high-priority pathogen families), (ii) community wastewater monitoring (detection thresholds defined in prevalence; converted to cases based on catchment size), and (iii) air travel monitoring (testing symptomatic passengers or airplane wastewater/air on incoming international flights). The model incorporates detection thresholds, per-case detection probabilities, and delays from infection to detectability (e.g., infection-to-hospitalization vs infection-to-fecal shedding). For wastewater sensitivity, the authors derive a distribution using US data linking PCR-positive wastewater samples to contemporaneous local incidence, fitting a log-normal distribution with median approximately 2.5 daily new cases per 100,000, and adjust for under-ascertainment, sewer connectivity, and shedding dynamics. Simulations compare predicted detection times against status quo detection, with validation via US state-level detection timelines early in 2020. A mathematical approximation is derived for mean cases at detection, showing approximate negative binomial behavior and yielding an interpretable formula relating expected cases at detection to detection delay, R0, detection threshold, and per-case detection probability. The model is applied to multiple diseases (COVID-19, mpox, polio, Ebola, pandemic influenza, and HIV/AIDS), exploring performance across epidemiological parameter spaces (R0, serial interval, hospitalization rate, time to hospitalization, fecal shedding probability) and catchment sizes (e.g., 650,000 vs 30,000). Statistical comparisons to actual detection times use one-sided tests to evaluate whether proposed systems detect earlier than status quo.

- COVID-19 in Wuhan: Hospital monitoring would have detected the outbreak modestly earlier than actual detection, with an estimated ∼0.4–0.43 weeks gain; the difference is statistically significant (e.g., p ≈ 1.9e-09 in a one-sided Welch t-test). Wastewater monitoring would not have accelerated detection in Wuhan and on average would have lagged actual detection; model predictions are consistent with Massachusetts wastewater analyses. Air travel monitoring did not accelerate detection in most evaluated scenarios due to the low probability of simultaneously traveling and being infectious and detected. - General disease comparisons: System performance depends on epidemiology. Hospital monitoring tends to outperform wastewater when hospitalization rates are high (e.g., Ebola), whereas wastewater can outperform hospitals for diseases with low hospitalization rates and significant fecal shedding or long/asymptomatic courses (e.g., polio, HIV/AIDS). Wastewater performs better in smaller catchments (e.g., 30,000 persons) due to prevalence-based thresholds translating to fewer absolute cases. - Potential benefits across outbreaks: Early detection systems can detect outbreaks when they are up to 52% smaller (e.g., wastewater for polio) or as much as 110 weeks earlier (e.g., hospital for HIV/AIDS) compared to status quo detection. Relative median detection time rankings among systems are robust across diseases in simulations and the mathematical approximation closely matches simulation outputs. - Mathematical insight: A compact approximation shows expected cases at detection scale with detection delay and R0 and inversely with the fraction of cases entering the detection system, clarifying how system design and disease parameters jointly determine timeliness.

Findings indicate that while early detection systems offer limited gains for COVID-19 (on the order of days), they can substantially accelerate detection for other diseases depending on hospitalization rates, shedding profiles, and delays to detection. However, earlier detection only translates into improved outcomes if it triggers a rapid, coordinated response; factors such as feasibility of interventions, availability of countermeasures, and predefined response plans are critical. Cost-effectiveness and informational value differ by system: wastewater may detect earlier but cannot directly indicate disease severity, whereas hospital detection provides evidence of clinically significant disease. The model is intended for preparedness planning ahead of future pandemics with uncertain parameters and can guide prioritization of surveillance investments. Policy relevance is underscored by ongoing WHO treaty discussions and proposed national investments in early warning systems; the framework can inform decisions alongside cost-effectiveness analyses and pilot implementations.

The study introduces and validates a quantitative framework, including a practical mathematical approximation, to assess early detection systems for emerging pathogens. Applied to COVID-19, the model suggests hospital monitoring would have achieved only marginally earlier detection, with wastewater and air travel providing no advantage in Wuhan-sized settings. Across broader epidemiological landscapes, system performance varies: hospitals excel for high-hospitalization diseases, while wastewater can be superior for low-hospitalization or fecal-shedding diseases and smaller catchments. Early detection could substantially reduce outbreak sizes at detection or bring forward detection by months to years for some pathogens. Future work should integrate cost-effectiveness, operational pilots, and strategies to ensure that earlier detection triggers timely responses. Establishing and calibrating surveillance systems in advance of the next pandemic is essential to realize potential benefits.

- The text indicates reliance on multiple assumptions and parameter estimates (e.g., detection probabilities, delays, wastewater sensitivity distributions, underreporting factors, sewer connectivity), which may vary by setting and pathogen. - Wastewater performance depends on catchment size and infrastructure; generalizing US-derived sensitivity to other contexts may introduce bias. - Air travel monitoring effectiveness is limited by low joint probability of travel while infectious and detection on arrival; modeling may not capture all operational nuances. - Some historical outbreak detection benchmarks are estimated rather than directly observed, introducing uncertainty in comparisons to status quo. - The model focuses on detection timing; it does not directly assess downstream response effectiveness or cost-effectiveness, which are necessary to translate earlier detection into improved health outcomes. - The framework is not intended for real-time use in the early months of a novel pandemic but for advance planning; early parameter uncertainty may affect accuracy.