Sequential versus synchronized sampling for respiratory syncytial virus wastewater early warning: a city-wide network evaluation

Wastewater-based epidemiology (WBE) offers a non-invasive platform for monitoring respiratory syncytial virus (RSV) at the community level, yet conventional once-weekly synchronized sampling at wastewater treatment plants (WWTPs) is constrained by temporal sparsity and the fragile environmental persistence of this enveloped RNA virus. We evaluated two sampling strategies using 73 weeks of empirical surveillance data from five WWTPs in Nanchang, China, collected from October 2024 to March 2026 and encompassing two distinct RSV epidemic waves. Strategy A employed simultaneous fixed-day (Tuesday) sampling across all five plants, yielding a weekly spatial snapshot with Δt = 7 days, whereas Strategy B deployed sequential Monday-to-Friday single-site sampling, creating a rolling detection window with system-level temporal resolution of Δt = 1 day. Viral RNA was concentrated by polyethylene glycol precipitation and quantified by RT-qPCR targeting the conserved N gene. Epidemic onset was determined by cumulative sum (CUSUM) control charts, with lead time defined as the temporal gap between wastewater-derived and clinical-derived thresholds. Empirically, Strategy B achieved a mean clinical lead time of 26.4 days (median 28.0 days, range 14-42 days) versus 19.4 days for Strategy A (median 21.0 days, range 7-35 days), corresponding to an average gain of 7.0 days (95% CI: 4.2-9.8 days, p < 0.001). During Wave 1, both strategies showed similar CUSUM threshold-crossing times, whereas during Wave 2, Strategy B crossed the threshold one week earlier than Strategy A and captured an incipient concentration surge 2-3 weeks ahead of the clinical case upswing. To interpret these empirical findings, analytical derivations and Monte Carlo simulations (n = 10,000) were used to characterize detection-delay distributions and detection opportunities under the two sampling schedules. Theoretical analysis showed that Strategy B reduced the expected sampling-related detection delay from 3.5 days to approximately 1.0 day and increased the 48-hour detection probability from 0.028 to 0.28 at low prevalence (C ≈ 0.5 × LOD), while the system-level weekly detection probability remained identical between strategies. These findings indicate that rolling sampling can improve the temporal coverage of RSV wastewater surveillance and may enhance early-warning performance, although its advantage varies across epidemic phases.