This is a key point. We REALLY need to know that we have good surveillance infrastructure in place for novel vaccine technologies. We haven't used mRNA vaccines at the population level before; nor have we used adenovirus vaccines at scale. https://twitter.com/pvot40/status/1335994242574000136

The major issue with vaccine "phase 4" surveillance is mis-attribution of effects via a "post hoc ergo propter hoc" logical fallacy. Just because someone gets an illness after a vaccine doesn't mean they got the illness FROM the vaccine.

My favorite example from my AMOH days...was asked to sign off on a report from an individual who had tripped and chipped a tooth after being vaccinated. This was listed as a vaccine adverse effect.

We can see how much damage spurious linkage of vaccines and medical conditions can do to population health, eg with the spurious association pushed for MMR vax and autism, which has helped fuel a measles resurgence ...

despite no evidence for MMR increasing autism risk (such evidence as there is actually points in the opposite direction).

But new vaccines can have unexpected adverse effects, and even if they are rare they can be important, because we are hoping to vaccinate a very large fraction of the Canadian (global, actually, population).

A one in 100,000 risk may not be picked up in a large RCT of 40,000 or 50,000 people, and even if it does occur the sample size would be too small to link it to vaccination.

That's why you need "phase 4" studies: studies of vaccine safety once it's going into arms.

Those need to (a) be able to compare rates of illness in vaccinated individuals to some baseline (and it's a lot of work to generate those baseline estimates!);

and (b) need information systems that can record and integrate post-vaccine events across time and space...in a country the size of Canada, you need to know whether you have a weird event that shows up in Newfoundland this week and Revelstoke in 2 weeks, and which is unexpected..

relative to baseline incidence.

To say that Canada has not done well with national electronic data systems to date would be a bit of an understatement.

Pandemics are a particularly important moment for phase 4 trials, because (a) we are giving vaccine to a ton of people at all ages over a short period of time (so increase in absolute risk) and (b) because by definition we have a new vaccine for a new bug.

This can go badly, as in 1976 with swine flu vax and Guillain Barre.

https://wwwnc.cdc.gov/eid/article/12/1/05-1007_article

Even in 2009, with a pretty safe vaccine, we had a surge in narcolepsy with some vaccine types.

https://www.who.int/vaccine_safety/committee/topics/influenza/pandemic/h1n1_safety_assessing/narcolepsy_statement/en/#:~:text=During%202009%2D2010%20they%20found,vaccinated%20in%20this%20age%20group.

Note that the per-dose risk here is < 1 in 10,000 but the RELATIVE risk was 9 x higher. Again, this is really hard to see in real time if you haven't invested in surveillance, and it's something you need to know about.

I am very enthusiastic about vaccines. I am excited to roll up my sleeve for a jab. I think they're the doorway out of this f&**$ing pandemic.

But in all the excitement it's easy to forget about phase 4, which is critical, not only for specific vaccines (and there will be...

...dozens by the time this is over) but for the vaccine enterprise as a whole....unexpected, nasty, rare adverse effects can undermine confidence in vaccination if they're not caught and dealt with expeditiously.

For more on phase 4, here's some great stuff from a brilliant colleague, Dr. Grace Lee from Stanford: https://jamanetwork.com/journals/jama/fullarticle/2772137