Thread by @clearyb1

Some thoughts on risk perception/risk communication in light of this week's @EMA_News statement on the possible link between the AZ #COVID19 vaccine & rare blood clots. A thread 🧵

Low probability, high consequence events are called dread risks. pure.mpg.de/rest/items/ite… There is a risk in avoiding risk. People can make decisions to avoid a rare risk that put them at a greater risk of a more common adverse outcome.

We don't know right now if the observed association between the vaccine and very rare clotting/bleeding issues is causal or if there may be some underlying factor involved. We don't know the baseline risk of these events in the context of the pandemic.

A key message for the potentially hesitant member of the public is the 1st point in the EMA statement- benefits of vaccination outweigh potential risks. The vaccine will likely prevent huge numbers of clots that would otherwise be caused by COVID.

People will have legitimate questions around what the EMA mean by very rare. We don't have solid estimates yet of the exact incidence, but based on evidence to date around 1 in a million would be a reasonable estimate- this may vary as more evidence emerges.

There are good principles of how to communicate evidence to patients dtb.bmj.com/content/57/8/1… Some key elements are: use numbers rather than words, always use a graphic when you can, use absolute rather than relative risk, use a consistent denominator when describing events.

I really like one of their examples where a scale with verbal and numeric descriptions contrasts everyday and clinical risks (examples here from the context of obstetric anaesthesia).

I have used graphics to visualise rare events in the context of medicines use in pregnancy and there is potential to apply this to the data on the possible association between the AZ vaccine and rare haematological outcomes

https://twitter.com/clearyb1/status/1185491619811860480?s=20

It is difficult to apply this though as the reported rates are so rare. As an example, this is what 1 in 100,000 looks like. You'll need to click on the image to see the 1 red square.

We could also think in terms of large crowds that people can visualise- 1 person in 12 All Ireland Final crowds in Croke Park

1 person in 77 fully standing concerts in the Three Arena

1 person in 19 Rugby International crowds in the Aviva Stadium

Or 1 baby in 17 years of births in Ireland at the current birth rate flickr.com/photos/pinksto…

Finally- we could compare to other rare events. Fatalities in car crashes happened at a rate of 30 per million in 2020. Murders happened at a rate of 11 per million in 2019.

To summarise- these adverse effects may have a cause other than the vaccine and they are very rare. I would be happy to take the AZ vaccine. I would like my family to have access to it. If we want to end COVID, we need to use everything at our disposal. #VaccinesSaveLives