One of the reasons (just one of many) that observing a vaccine effect in Covid cases in the UK is that cases are falling so quickly due to lockdown. So we risk misinterpreting variable impact of a change in behaviour by age group with a vaccine effect.

Lots of clever people doing great work to try and untangle that. See @danielhowdon for example. But what if we also took the chance to look at a country with stable case rates. Stable now for about 7 weeks, with a vaccine program rolling out (much slower than the UK) within that.

The French vaccination campaign really started on the 4th of January. It has been slow and thus focused quite sharply on the oldest and most vulnerable in society. But it does now produce quite good data, which combines well with the good French testing data.

Two weeks for vaccines to work from the 4th of January is the 18th of January, plus seven days for case rate smoothing, let's index case rates by age on the 25th and have a look at how they've evolved since then. It's only two weeks of data, but let's try.

The good news is that all the numbers line up and the curves go in the same direction as you'd expect them to if the vaccines were as effective as we know they are, both from trials and from Israel. A few more weeks until this meets a good standard of proof, but it looks good.