Tell Me the Odds

I have reservations about Scott Gottlieb’s WSJ op-ed, urging a massive effort and a massive rollout of a SARS-CoV-2 vaccine in the U. S.:

The first nation to develop a vaccine for Covid-19 could have an economic advantage as well as a tremendous public-health achievement. Doses will be limited initially as suppliers ramp up, and a country will focus on inoculating most of its own population first. Even with extraordinary international collaboration among multiple companies, it could be years before a vaccine is produced at a scale sufficient to help the entire world. The first country to the finish line will be first to restore its economy and global influence. America risks being second.

China is making rapid progress, with three vaccines entering advanced development. Chinese officials say they could have a vaccine available for widespread use next year. The Europeans are also making progress. While friendly nations will try to share a successful product—to a point—the U.S. can’t rely on vaccines from China or even Europe being available in America quickly. So it’s important to take steps to speed up progress in the U.S., and to prepare to manufacture such a vaccine on a global scale. A more prepared U.S. could inoculate Americans quickly and share the product with others, particularly low-income nations that can’t develop their own vaccines and need protection.

More than 70 companies and research teams are working on a vaccine, but fewer than 20 have the experience and manufacturing scale to pull a product through development. Only five or six operate primarily in the U.S., which means foreign governments might try to make a claim on a vaccine before America can. Each company is taking a slightly different approach, spreading bets and increasing the chance for success. (I serve on the board of one of them, Pfizer Inc. )

To win the race to a vaccine, America needs to engineer a development and regulatory process that is unprecedented in scope and urgency. Testing six or more candidate vaccines at once during a pandemic has never been tried anywhere. But it can be done.

First, the Food and Drug Administration should work with companies to conduct early safety testing while the vaccines are evaluated in laboratory and animal models to assess the full strength of their immunity. This parallel development process will save time and reveal more about the vaccines sooner. Regulators can also allow manufacturers to share common platforms for conducting the necessary studies. The FDA has developed good measures for potency using laboratory tests and animal models for the virus. These platforms should be adopted across industry, which will let regulators get clear answers more quickly.

Next, this effort will require novel approaches to clinical testing that allow us to build a large safety database and get an earlier answer on whether a product is working in people. Given that this vaccine will be deployed for mass inoculation of entire populations, a vaccine will need to be tested in tens of thousands of patients before it is approved for general use. An unsafe product could cause significant harm. The urgency to develop a vaccine quickly is eclipsed only by the need to make sure it is very safe.

Large Covid-19 outbreaks in American cities this fall may be inevitable. Against this grim backdrop, one approach to testing a vaccine is a “stepped wedge cluster.” Under this kind of clinical trial, a vaccine would be administered in the setting of an outbreak. The point is to provide some potential benefit while building a large and rigorous data set to evaluate its safety and effectiveness. This could be done as soon as a vaccine has cleared early safety trials.

The idea is to take a large number of doses and hold a trial in an outbreak city by serially vaccinating big groups of people—perhaps 25,000 at a time—with each cohort spaced two weeks apart, until 100,000 people have been inoculated over about six weeks. To see if the vaccine works, researchers compare the four groups and assess if timing of inoculation had a discernible impact on someone’s likelihood of contracting Covid-19.

The next massive challenge is making enough vaccines. Congress has set aside more than $3.5 billion for this purpose as part of the Cares Act. This allows the government to secure doses in advance of a product’s approval, which is essential for rolling out a vaccine the minute it’s approved. The money will be used to support investments in large-scale manufacturing. Johnson & Johnson recently announced a major collaboration with the Health and Human Services Department’s Biomedical Advanced Research and Development Authority to secure an early supply of vaccines. The government should also give grants to manufacturers with the most promising vaccines to rush the construction of large factories and other facilities.

I think there are some steps missing from his plan. For a massive trial in an “outbreak city”, don’t you need to know the prevalence? We don’t know that and if we stay on our present path we won’t know in three months or six months or a year, either. Otherwise how do you know what you’re measuring?

Or this:

China is making rapid progress, with three vaccines entering advanced development. Chinese officials say they could have a vaccine available for widespread use next year.

At this point how much would you trust a Chinese vaccine? How much would you trust their claims for it? How much would you trust whether what they’re producing is actually the same thing that they tested? Fool me once… I think the only thing we can do at this point is to ignore China.

My final misgiving was stated pretty well by G. K. Chesterton: anything not worth doing is not worth doing well. How confident should we be that, regardless of how great the prospective rewards, an effective vaccine can actually be developed? How do you scale up the production of something that cannot be produced?

We don’t know, for example, if having recovered from the virus conveys resistance. I would be a lot more comfortable if an effective vaccine for any coronavirus had been developed in the past.