Thread by @EmilyG_DC

Time to talk about Bayes' theorem.

Mathematically, the probability of being vaccinated conditional on being sick [P(A|B)] rises as the vaccination rate [P(A)] rises.

In other words, vaccinated ppl being a bigger share the sick itself doesn't imply vaccines are less effective.

Vaccinated people as a percentage of those sick [P(A|B)] is **not the same thing** as the probability of getting sick while vaccinated [P(B|A)].

Please don't confuse the two!

Here's an illustration (h/t @CDCgov). Suppose vaccines were NOT effective (0%). The % vaxxed among the sick would be = to the % in the population, at any vax rate.

When effective, the % of sick people who are vaxxed grows as the vax level increases.

https://twitter.com/EmilyG_DC/status/1421183780602191880?s=20

So headlines like this, while factually true, are shoddy SciComm and emphasize the wrong message at a time when it is CRUCIAL for our collective health that more people get vaccinated against #COVID19.

More good data viz to keep in mind when we see vaccinated increase as % of cases and hospitalization in the near future.

The virus is changing and the population’s vaccination rate is too.

https://twitter.com/jburnmurdoch/status/1421186304700735494

Lastly, vaccine efficacy is a wholly different concept. It is neither P(vax|sick) nor P(sick|vax).

Efficacy measures the *relative reduction* in risk. By definition it compares rates in a vaccinated population vs an unvaccinated (placebo) population. who.int/news-room/feat…

Bottom line: while delta is more transmissible and cases are rising, vaccines still reduce infection and confer VERY POWERFUL protection against severe disease and death.

Mask up in the short run, and help your community get vaccinated now for the long run.