Nathan Grubaugh's bet on far-UVC: tuberculosis hotspots will show results first

Nathan Grubaugh, an infectious disease epidemiologist, is putting a specific, testable bet on the table: far-UVC installations in tuberculosis hotspots will produce measurable reductions in transmission quickly, and the evidence will arrive well before comparable data from schools or general community settings.

The logic behind the call rests on epidemiological signal density. Tuberculosis transmission in high-burden institutional settings is frequent enough that a meaningful drop, if it occurs, will stand out against baseline rates within a relatively short window. In lower-burden environments like schools or residential communities, transmission events are rarer, attack rates are harder to measure, and any protective effect from a germicidal intervention can take considerably longer to become statistically visible. Grubaugh is not arguing that far-UVC fails in those settings. He is arguing that tuberculosis hotspots are where the field will get its answer first.

That framing gives the prediction a useful structure. It is not a claim that far-UVC eliminates tuberculosis, or that the technology is ready for broad deployment. It is a claim about sequencing: which environments will generate credible evidence soonest, and why. As Grubaugh puts it, those high-burden places will see “pretty noticeable drops in transmission fairly fast and pretty reliably,” while community and school settings offer no such speed of feedback.

The underlying principle that dense transmission environments produce faster, cleaner feedback is not controversial in epidemiology. When baseline incidence is high, even a modest proportional reduction generates enough events to be distinguishable from noise. When baseline incidence is low, a trial needs either a much larger population, a much longer time horizon, or both, before results become interpretable. Grubaugh’s prediction applies that general principle to a specific technology in a specific class of settings. That is what makes it checkable rather than merely plausible.

I think those places will see, for example, pretty noticeable drops in transmission fairly fast and pretty reliably. And I think that kind of thing will be forthcoming relatively quickly compared to, you know, if you put this in a school in a community, we don't really know that fast. Nathan Grubaugh

What the prediction does not specify is a precise timeline. Grubaugh uses the phrase “relatively quickly” and “forthcoming,” but he does not attach a number of months or a minimum effect size threshold. A reader who wants to hold him to the call will need to watch for published results from high-burden clinical and institutional settings and ask whether drops in transmission rates are appearing faster there than in community trials running in parallel. The absence of a hard deadline is a limitation of the call, though it does not dissolve it. The directional claim, that hotspot data will arrive first and will be more reliable, is a falsifiable prediction over any reasonable horizon.

The stakes of the call being right are not small. Tuberculosis remains one of the leading infectious causes of death globally. If far-UVC installations in high-burden settings do produce fast, reliable reductions in transmission, those results would carry significant weight for public health policy, procurement decisions, and the broader question of where germicidal light technology gets prioritized. A clear positive result from a setting where the signal is strong would also do more to advance the field than years of ambiguous data from low-transmission environments where the technology may be equally effective but far harder to evaluate against a noisy background.

The call is also grounded in a practical observation about how scientific evidence actually accumulates. Grubaugh is not simply predicting that far-UVC works. He is predicting something more specific: that the order in which the evidence arrives will be shaped by the epidemiological structure of the settings being studied. That second-order claim, about evidence sequencing rather than effect size alone, is the part worth watching. If community and school data somehow arrives before hotspot data, or if hotspot data turns out to be as ambiguous and slow to interpret as community data, the reasoning behind the prediction will need revisiting even if the underlying technology still shows promise.

The settings Grubaugh is pointing to exist. Installations are being pursued in various jurisdictions, and transmission data in high-burden environments is measurable. Whether the evidence arrives on the timeline he implies, and whether it is as reliable as he expects, is a question the next several years of deployment will answer.