Yesterday, with scant evidence, the CDC recommended, and all 50 states adhered to, a pause in J&J vaccinations. While a committee is meeting today (see https://www.nytimes.com/2021/04/14/health/johnson-johnson-covid-vaccine.html ), it will likely be days or more before they re-start vaccinations (update: it will be at least 10 days, because they are the CDC is delaying a decision).
What's the harm of receiving the vaccine?
To date, 1 person has died of a rare blood clot shortly after receiving the J&J vaccination, out of more than 7 million vaccinated. Since 100,000 people die of blood clots a year (CDC blood clot information), many of them with no symptoms beforehand, this means around 300 people per million die of blood clots per year, or about 6 per million per week. This is about 40 times higher than the 1 in 7 million for the J&J vaccine (and perhaps more than that if they measured the deaths out to more than a week).
But for the time being, let's assume this was not happenstance, and that we cannot prevent this from happening again, so that around 1 in 7 million who receive the J&J vaccine will die from a blood clot.
What's the harm of *not* receiving the vaccine?
Right now, there are around 60,000 confirmed COVID cases a day. Currently, about 1 in 100 confirmed cases results in death (keep in mind the COVID infection fatality rate is a fraction of this, because confirmed cases are only a fraction of actual cases). This means that the 60,000 confirmed cases each day will result (a few weeks down the road) in about 600 deaths.
If we assume around 130MM are immune (from either a vaccination or from having had COVID -- and this is conservative enough to allow for some people getting COVID twice and a small number of vaccinated people getting COVID), we have about 200MM people exposed daily. This means the chances of getting and eventually dying from COVID is 600/200MM per day, which is 3 per million. Another way of looking at it is that for every day for every million unvaccinated (exposed) people, there are 3 additional deaths.
Now, we just need to compute the additional person-days of exposure to figure out the resulting deaths. J&J was distributing about 1.5MM vaccines a week (about 200K a day), so we can compute the people exposed times days delayed for the total additional person-days of exposure due to a "pause" for different durations of the pause:
1 Day pause: 200K person-days (200K people by 1 day)
2 Day: 600K person-days (200K people by 2 days plus another 200K by 1 day for those who would've got their shots on day 2)
3 Day: 1.2MM person-days (K person-days from before plus 450K (150 for each day's delayed people) additional person-days for this extra day.
It turns out there is a simple equation for this kind of series, and it is 200 multiplied by the sum of days delayed, which is simply 200*(d+1)*d/2, where d is the number of days delayed.
So we can compute, that a 1 week delay will mean 200*8*7/2=
-5.6MM person-days, and a 2 week delay will equal
-21MM person-days of exposure.
Now it's easy to compute deaths. At 3 per million person-days, a 1 week delay will cause about 17 deaths, and a 2-week delay will cause about 66 deaths.
What if the vaccine is permanently pulled? I don't think this is a real possibility, but if that happens, if would be the total person-day delay before those people received an alternative vaccination.
Caveats and Assumptions
I'm assuming above that:
1) Delays mean canceled appointments. If the shot is just replaced by another vaccine shot, it may make little difference. This is certainly not true right now, as many appointments were canceled yesterday and today, but some (in NY) were kept and the vaccine was just replaced with a different one. As the delay extends, this will surely happen more, so as we get further out, the deaths will surely be fewer than 1 per person-day of J&J delay.
2) The death rate is unchanging. This is true in the short term (of a few days), but virus rates and death rates are certainly falling in the coming weeks, so the rate of 3 deaths per million will fall over time. This makes little difference in the next few days but an extended delay would mean the estimates above are too high.
3) All the delayed vaccine will be given the day the delay ends. This might be possible if the delay is short, but a longer delay would mean it takes a few days to use the built-up supply, even after the pause ends, thus killing more people.
4) Delays will not become permanent. Some people who were scheduled to get a shot and then had it canceled may decide the vaccine is just too risky, no matter what the ensuing review says. These people will never get vaccinated and this means the delay causes more deaths (and may already have done so).
Conclusion
Pausing the J&J vaccine shots might cause more than a dozen additional dozen COVID deaths the first week of such a pause and roughly 60 or 70 if the pause lasts for two weeks. It would take around 4 weeks of continuing inoculations for another person to die from the vaccine, if indeed the vaccine causes deadly blood clots. So we are looking at around 66 excess deaths from a delay of two weeks versus less than 1 had we kept giving the vaccine.