Disease surveillance is getting local, fast and cheap. Reading it is still hard

What changed

Four preprints posted this week show disease surveillance moving in the same direction: closer to where people actually are, faster to run, and cheaper per sample. The most public test was wastewater surveillance in Atlanta during the 2026 FIFA World Cup. Researchers monitored pathogen trends in the city's wastewater and found no statistically significant differences in target pathogen detection between the pre-tournament and tournament periods. A mass gathering that many expected to move the pathogen needle did not, at least in the wastewater. m:…7459

There was one exception, and it is the instructive part. SARS-CoV-2 detection in wastewater increased significantly during and after the World Cup, with a rate ratio of 3.21, but the observed increase coincided with regional and national increases, suggesting ongoing community burden rather than event-specific transmission. The signal was real and the attribution was not. A surveillance system that can detect a rise but cannot say whether the event caused it is still useful, but only if the people reading it have the regional data in front of them too. m:…7459

The same lesson runs through a four-year programmatic analysis of tuberculosis drug-resistance surveillance in Northern Morocco. Routine Xpert MTB/XDR testing identified rifampicin resistance in 39.9 percent of interpretable episodes, and 71.7 percent of resistant episodes with complete profiles exhibited resistance to two or more drug classes. Molecular surveillance using Xpert MTB/XDR enables systematic characterisation of multiclass resistance patterns in routine diagnostic networks, the authors report, which turns a test that used to be a research project into part of the ordinary diagnostic plumbing. m:…8463

Nine years of respiratory seasons

A retrospective cohort study from Lyon gives the long view that the Atlanta snapshot lacks. Winter peaks of influenza and RSV detections re-emerged after 2022 following pandemic-related disruptions, which means the seasonal pattern the pandemic broke has now re-established itself. That matters for planning: hospitals can once again expect a winter peak, and the question is how high it runs rather than whether it comes. m:…7458

The Lyon data also sharpens who bears the risk. Patients aged 65 years and over had longer hospitalisations, a median of 10 versus 8 days, and twice the in-hospital mortality rate of younger adults, 14 percent versus 7 percent. RSV and human metapneumovirus positivity rates were highest in patients aged 80 and over, while influenza, rhinovirus and SARS-CoV-2 peaked in younger adults under 65. The viruses do not distribute evenly by age, and a surveillance report that averages across ages will hide the group that needs the beds. m:…7458

Cheaper typing, honest limits

The week's fourth paper is a methods advance for tuberculosis tracking. HiSpoT, a one-tube dual-barcode PCR method, enables multiplexing of 96 tuberculosis complex samples in a single tube for sequencing-based spoligotyping. For outbreak tracking in places where sequencing budgets are thin, handling 96 samples at once is the difference between typing every case and typing a sample of them. m:…3406

The authors are also candid about the trade-off. HiSpoT shows 70 percent full concordance with whole-genome sequencing for strain typing, and discrepancies are primarily attributable to single-spacer insertion-element disruptions of PCR primer sites. Seventy percent is not a replacement for whole-genome sequencing; it is a triage tool that tells a programme which samples deserve the full treatment. Used that way, it extends surveillance further than a perfect but unaffordable method ever could. m:…3406