Box 22 Asks If The Work Is Enhanced. An Administrator Ticks It.

BETHESDA, Maryland — Box 22 is one line long and it has two options.

It asks whether the proposed work "would be reasonably anticipated to enhance the transmissibility or virulence of a potential pandemic pathogen." Yes, or no.

A "yes" is not a refusal. It routes the application out of the ordinary queue and into a departmental review that has now absorbed four years of hearings, two competing frameworks and a great deal of shouting.

This newspaper did not ask whether that review works. We asked something narrower: who physically ticks the box.

What we asked for

We wrote to seventeen institutions and asked for the 40 most recent applications at each that named a pathogen, together with the submission system's audit log for each form — the user account that keyed the answer, the timestamp, and whether the principal investigator's account ever opened the completed form.

Nine answered. Seven released the logs under records law. Two released them after a telephone call, on the view that a submission audit log is administrative metadata rather than research material.

That is 360 forms.

Who ticks it

In six of the nine institutions, Box 22 is completed by a grants administrator.

The administrator works from the abstract. This is not a lapse in the workflow; it is the workflow. At those six institutions the abstract is the only part of the proposal the compliance screen shows on the same page as the box. The research plan is four clicks away, and the logs show those clicks are rarely made.

In three of those six, the investigator's account never opened the completed form before submission. In 118 of the 360, a certification about pandemic pathogens went to a federal agency under a scientist's name without that scientist having seen the answer.

Every answer was right

Not one of the 360 was answered wrongly.

We paid two microbiologists who work with select agents, and who do not know each other, to answer Box 22 themselves from the 360 abstracts with the institutions and investigators stripped out. They agreed with the submitted answer 360 times out of 360. They disagreed with each other twice, on two proposals neither would have ticked yes.

Every submitted answer is "no", and almost none of them is close to the line. The 360 include catheter-infection surveillance, four nutrition trials, a vaccine cold chain in Malawi, and a great deal of work on soil bacteria.

The two who found out from us

Two of the nine institutions told us they had believed their investigators completed the box personally, and learned otherwise when they went to answer our request.

One has since added a required acknowledgement: the form cannot be submitted until the investigator's account has opened the completed compliance page. It took, their research office told us, an afternoon.

The other is still deciding.

What this is, and what it is not

It is not concealment. Nothing here was hidden, and we obtained all of it by writing letters.

It is not incompetence either. The administrators are transcribing an answer the scientist has already given, in the one document a scientist has every reason to make exact.

What we can say is this. The single control on which four years of argument has turned is, most of the time, operated by somebody whose job description does not mention pathogens, from a summary written by somebody else, on a screen that shows nothing else.

That may be entirely adequate. In 360 attempts it was. We found no case near the line, which means we have not seen the workflow tested, and the desk would rather say so than pretend that 360 easy questions answered correctly tells us how a hard one would go.