Key points: A popular "hazard difference" estimator for averted outcomes is biased, but an unbiased one can be calculated from status-based aggregated data (knowing y/n vaccinated for those with and without the outcome). For avertible outcomes, this doesn't work; need to know dates of vaccination.

There is surely some basic biology benefit to gain. But the 100s of millions spent were justified with pandemic prevention/response as part of it, and the vaccine part of that doesn't hold water.

The fact that WIV1 and HKU1 were used by Moderna doesn't mean that they were central to the science; I'm no patent expert but I guess patent applicants want to cover as many applications (eg viruses) as possible for commercial reasons. Unclear to me if using them fundamentally changed conclusions.

On the substance of your points, I am not sure about the "not dangerous because the worst viruses are already at the interface" -- depends if ecological encounter opportunities are rate-limiting, which I think we often don't know.

one argument supporting a type of research may be weak while another is stronger, or research may be both dangerous and useful (or safe and not useful) and taking points one at a time helps to keep the questions separate.

Thanks @acritschristoph.bsky.social for these thoughtful comments. Apart from space and time, one reason we didn't address the other arguments was that I think it's important to consider arguments separately to reduce pressure to form "sides" -- ...

There may be other justifications for the expense and risk incurred by such prospecting efforts, but vaccine development is not a strong one. Pathogen identification at the human interface (in spillover cases) and efforts to avoid or remediate human drivers of spillover risk may be more efficient.

For Marburg, a reservoir is known, but serious efforts at countermeasures have been spurred by human outbreaks.

We conclude that virus prospecting is neither necessary, sufficient, nor particularly feasible as a driver for medical countermeasure (at least vaccine) development.

Last, we take a look at detailed timelines of human and animal isolation of filoviruses (ebolavirus spp. and Marburg) and note that in ebolaviruses, animal reservoirs have not been confirmed but countermeasure development has followed large human outbreaks.

Then we identify the large number of targets for which vaccines are arguably more urgent than for any virus known only in non-human hosts: the 26 families of viruses known to infect humans. Among these, there are no vaccines approved for even a single member of 10 of the 26 families.

Then we note that most major virus outbreaks of the 21st century have been caused by viruses first discovered in humans. Next, we show that recent priority pathogen lists for countermeasure development have contained known human pathogens but shown little relationship to virus discovery in wildlife.

First we note that historically, finding a novel virus in animals before the first known human outbreak did not in any case spur countermeasure development. Poignant examples are Zika (1947 in animals, 1952 in humans, vaccines only much later) and monkeypox (1958-1970, vaccine R&D only recently).

In it we asked if prospecting for novel viruses in wildlife is justified, as often asserted, as a means to speeding the development of countermeasures against emerging infections. We find little evidence that it is.

Good to get them <65 in advance of Medicare denying payment for progressives

There is a link for non-NIH individuals to sign in support. Already several Nobelists and others have