September 11, 2020

Why phase 3 trials can take so long

"Making it all the way through a phase 3 trial is not exactly a piece of cake," writes NCSE Executive Director Ann Reid in response to the recent news that a high-profile COVID-19 vaccine trial had to be paused.

Administering a vaccine

Check out our entire series explaining the science involved in the coronavirus pandemic. Sign up to receive our coronavirus update each week.

I love board games. So when my kids were little, I wanted to get them started as young as possible. Reader, this was a mistake. The problem wasn’t that my kids didn’t enjoy the early board games I found—Chutes and Ladders™ and Candy Land™—the problem was that they loved them very, very much. Sadly, they are possibly the most stultifying adult activity ever. At least watching paint dry eventually comes to an end. Not true, as I recall, of a game of Candy Land.

Do you remember the rules of Candy Land? They’re pretty simple. Draw a card. Go where it tells you, either to the next square matching the card’s color, or to a specific place on the board—the Gumdrop Mountains, say, or the Ice Cream Floats. Those specific place cards can send you forward or backwards (one of the main reasons the game is basically interminable). There are a couple of squares that allow you to take a shortcut, and at least in the version of the game current at the time my kids were little, there were several “stick” squares that you couldn’t leave until you drew a card of a specific color. Whoever reaches the end first wins. That’s it. Perfect for a four-year-old. And, I’d argue, quite a good metaphor for a vaccine trial.

(As an exercise in clear, concise writing, you might challenge your students to write rules for Candy Land—or even an everyday activity such as making a hard-boiled egg—in the fewest possible words. Classmates can review each other’s efforts to make sure the rules are accurate and complete. This skill will prove useful for future scientists who will have to describe their work in the Materials and Methods sections of research papers. Note: “1) Begin playing 2) Throw board in air and quit” is not an acceptable description, however relatable.)

I was reminded of Candy Land by reports yesterday that one of the most advanced phase 3 coronavirus vaccine trials, by AstraZeneca, had been temporarily paused. This is the clinical trial equivalent of landing on a “stick” square during a game of Candy Land. You simply have to stop the trial until you can be sure that the adverse health event wasn’t caused by the vaccine.

Remember, while we’re desperate for a vaccine to stop the spread of coronavirus, vaccines are given to healthy people. Therefore, the vaccine must be dramatically less likely to cause harm than the disease it is designed to prevent. This is one of the reasons that phase 3 trials need to include so many participants; you need a lot of people not only to determine whether the vaccine prevents infection but also to improve your chances of detecting rare side effects.

The vaccine must be dramatically less likely to cause harm than the disease it is designed to prevent.

Determining whether the adverse health event is related to the vaccine is not always easy. An immediate acute allergic reaction or high fever would be suggestive. A fever the following week would perhaps be less suspicious. A case of appendicitis a month after receiving the vaccine would probably not require pausing the trial. After all, among 30,000 people—even though they’re all healthy enough to qualify for trials—some bad health events are certain to occur. The question is whether there is a plausible connection between the symptoms and the vaccine. Serious illness requires a hard stop and an in-depth effort to find out what happened.

In the case of the AstraZeneca vaccine, one of the U.K. volunteers reportedly came down with a condition called transverse myelitis. This is rare—there are only about 1,400 cases per year diagnosed in the U.S, or around one case per 250,000 people. Having transverse myelitis show up in a group of just a few thousand vaccine recipients is unexpected enough that it’s important to stop and try to determine whether the vaccine could have been responsible, especially because viral infection is one possible cause of transverse myelitis and the AstraZeneca vaccine contains a live, attenuated virus. The researchers will be trying to answer lots of questions and carefully monitoring other participants for similar symptoms to convince themselves¬—and oversight boards—that the vaccine was not responsible before resuming the trial.

What exactly does it mean to pause the trial? The AstraZeneca trial is eventually planned to include at least 30,000 people in the U.S., but recruitment only began here on August 31, 2020, so it isn’t clear whether any U.S. participants have even received the vaccine yet. Researchers will not be administering the vaccine to new participants anywhere until they can be sure that the vaccine wasn’t responsible for that case of transverse myelitis. And that means no progress towards determining if the vaccine prevents infection.

There are two other vaccine candidates in phase 3 clinical trials in the U.S., and another six candidates outside of the U.S., so the AstraZeneca vaccine is not the only one racing through Candy Land toward Home Sweet Home. So far, the others have avoided “stick” squares and other setbacks. But the AstraZeneca experience is a reminder that making it all the way through a phase 3 trial is not exactly a piece of cake. That’s why many experts are skeptical that a vaccine will be approved before the end of the year, much less before election day. That would require a perfect Candy Land game, and every parent knows that hardly ever happens.

Bonus: I was very amused to find this research study that definitively found, using math, that Candy Land is, indeed, the worst offender when it comes to potentially interminable games with your toddler. The paper’s conclusion (below) really brightened my day, something to be very grateful for when the sky outside my window is orange from forest fire smoke and after six months of sheltering in place, I still can’t visit my children and grandson (note, I have forgiven the former for all those games of Candy Land and would happily pay the penance of playing it again if I could do so with the latter). To author Barry Wise, I say thank you for the giggles from one parent-nerd to another:

This article demonstrates how Monte Carlo simulation can be used to solve a real-world, every day problem: Of these three games, which one will provide entertainment for my four-year-old yet let me retain my sanity? If your child is inflexible regarding changing the rules, choose Cootie or Chutes and Ladders which have similar average game lengths. Of the two, Chutes and Ladders is probably the more interesting because of the possibility of moving both forward and backward. If your child insists on Candy Land, consider changing the rules as suggested above. An alternative strategy, of course, is simply to let your child cheat. This not only shortens the games, but has the additional incentive that it usually causes the child to win and puts them in a better mood (though it certainly doesn't teach much about ethics). On the bright side, in the few weeks it has taken to complete this study, we have progressed from board games to card games, specifically Uno, which are much more interesting for adults and children. Perhaps there is a God after all.

NCSE Executive Director Ann Reid

Ann Reid is a former Executive Director of NCSE.

reid@ncse.ngo