How does that approach work, and how has it been used in influenza research?

The test negative design has been used both in influenza work and with other vaccines, including the COVID vaccine most recently.

And the way this works is you take a population, let's say people coming to the clinic for a respiratory infection that meets a case definition, and then you test whether they are positive for, let's say, influenza virus or negative for influenza virus.

You then go and see whether they have been vaccinated or not, and you can use those data to come up with an effectiveness outcome.

Now, the reason you use this design is that if you just went to the general population, not people who have a specific illness, you might get bias entered.

Because let's say the healthier people are more likely to get vaccinated, paradoxically, because they're more mobile, and they can get in to see the doctor more easily and go and get vaccinated.

So this controls for the kind of bias that you have to avoid when you use the observational studies.

Finally, moving forward, how should regulators weigh evidence from each of these types of studies to get an accurate understanding of a product's population-level effects?