Yes yes yes AND: it would also behoove us to move away from an almost-total reporting focus on vaccines, and do more reporting on another very important tool in this effort: testing. See, e.g., the Twitter feed of @michaelmina_lab. The disproportionate coverage of vaccines is another journalism fail/blind spot.

Yes, though (sorry) your math is incorrect. I completely agree with your fundamental point about needing to report on the actual (fairly low) level of risk. However, your math comparing covid deaths to dog attack deaths is not accurate: The dog-attack data is over a **lifetime** (per the link you posted, "Lifetime odds..."). The CDC covid breakthrough data is since May 1, 2021, so about **four months** -- not even a full year yet. Also, I've been tracking the CDC's weekly numbers, and they have, alas, ticked up quite a bit since the Delta variant has taken hold. It's about 235 breakthrough deaths per week taking the difference between the 8/30 and 7/26 CDC data. (About 200/week if you subtract the deaths that probably aren't really covid related.) Assuming the higher number, and if it were to remain constant, 235/week equals 12,200 breakthrough deaths a year -- per about 170 million vaccinated people -- so about a 1 in 14,000 chance of dying from a breakthrough covid case per year. To your point, that's still a pretty low risk. The vaccines are extremely good. (BTW, dog attack deaths are like 30-50 per year in the US. So, not comparable at all, sorry.) That puts it in the ballpark of skin cancer. (Per https://www.worldlifeexpectancy.com/usa-cause-of-death-by-age-and-gender ) The perspective I like to put it in is comparing it to dying in a car crash or from a fall: Which are both about 39,000/year. So if you're willing to take the risk of riding in a car or climbing a ladder, your odds of dying from a breakthrough covid case are about three times less. So, yeah, the media should report the accurate numbers and do a better job describing the risk. But, er, ah, eek, so should you. :)

A couple of things if we're going to dive deep into the numbers. First, I don't know the future course of breakthrough deaths, and neither do you. Delta may persist, or fall back. Second, the denominator (those vaccinated) is also rising, and likely to continue to do so as well, also at an uncertain rate.

In any event, appreciate your agreement on the core argument calling for a greater sense of proportionality. And very much appreciate your engagement with the piece.

For sure. (FYI, as for the denominator, yeah, that's why I calculated deaths per million, to normalize it: 12,200 est. annual deaths at current pace / 170 million vaccinated so far is how I got the "1 in 14,000 chance of dying per year". So that part of the math doesn't change as more people get vaccinated; though the odds themselves may change as we get more data, the virus mutates, etc.) Another stat that needs to be shouted more is that the odds of dying unvaccinated is 25X the odds of dying if you're vaccinated. Since there appear to be very very few harmful side effects (proportionally speaking), it's a mathematical no brainer to get vaccinated. :)

Dave is totally right. You are comparing lifetime statistics to a 6-month statistic. You really should take this article down. if you assume that the data accurately represents a 6-month risk with an average 70 year lifespan, then the arithmetic yields a ~ 1:585 lifetime risk, which is still 5 times less likely than an auto accident, but nowhere near 1:87,000.

Yes yes yes AND: it would also behoove us to move away from an almost-total reporting focus on vaccines, and do more reporting on another very important tool in this effort: testing. See, e.g., the Twitter feed of @michaelmina_lab. The disproportionate coverage of vaccines is another journalism fail/blind spot.

Yes, though (sorry) your math is incorrect. I completely agree with your fundamental point about needing to report on the actual (fairly low) level of risk. However, your math comparing covid deaths to dog attack deaths is not accurate: The dog-attack data is over a **lifetime** (per the link you posted, "Lifetime odds..."). The CDC covid breakthrough data is since May 1, 2021, so about **four months** -- not even a full year yet. Also, I've been tracking the CDC's weekly numbers, and they have, alas, ticked up quite a bit since the Delta variant has taken hold. It's about 235 breakthrough deaths per week taking the difference between the 8/30 and 7/26 CDC data. (About 200/week if you subtract the deaths that probably aren't really covid related.) Assuming the higher number, and if it were to remain constant, 235/week equals 12,200 breakthrough deaths a year -- per about 170 million vaccinated people -- so about a 1 in 14,000 chance of dying from a breakthrough covid case per year. To your point, that's still a pretty low risk. The vaccines are extremely good. (BTW, dog attack deaths are like 30-50 per year in the US. So, not comparable at all, sorry.) That puts it in the ballpark of skin cancer. (Per https://www.worldlifeexpectancy.com/usa-cause-of-death-by-age-and-gender ) The perspective I like to put it in is comparing it to dying in a car crash or from a fall: Which are both about 39,000/year. So if you're willing to take the risk of riding in a car or climbing a ladder, your odds of dying from a breakthrough covid case are about three times less. So, yeah, the media should report the accurate numbers and do a better job describing the risk. But, er, ah, eek, so should you. :)

Thanks for taking the time to write this.

A couple of things if we're going to dive deep into the numbers. First, I don't know the future course of breakthrough deaths, and neither do you. Delta may persist, or fall back. Second, the denominator (those vaccinated) is also rising, and likely to continue to do so as well, also at an uncertain rate.

In any event, appreciate your agreement on the core argument calling for a greater sense of proportionality. And very much appreciate your engagement with the piece.

For sure. (FYI, as for the denominator, yeah, that's why I calculated deaths per million, to normalize it: 12,200 est. annual deaths at current pace / 170 million vaccinated so far is how I got the "1 in 14,000 chance of dying per year". So that part of the math doesn't change as more people get vaccinated; though the odds themselves may change as we get more data, the virus mutates, etc.) Another stat that needs to be shouted more is that the odds of dying unvaccinated is 25X the odds of dying if you're vaccinated. Since there appear to be very very few harmful side effects (proportionally speaking), it's a mathematical no brainer to get vaccinated. :)

Dave is totally right. You are comparing lifetime statistics to a 6-month statistic. You really should take this article down. if you assume that the data accurately represents a 6-month risk with an average 70 year lifespan, then the arithmetic yields a ~ 1:585 lifetime risk, which is still 5 times less likely than an auto accident, but nowhere near 1:87,000.

The dog attack statistic is lifetime risk. It's not comparable.