Probably because those rates are from 10 to 60 years, which is quite a broad range. People in their fifties/sixties die like A LOT more than young people and these ages happen to be the most vaccinated compared to the youngest ones. https://plato.stanford.edu/entries/paradox-simpson/

Also, when we are seeing far fewer vaccinated people in hospital in most countries (in my country 70% of the admissions in the hospital are from unvaccinated people), it is not only that the number of vaccinated people in hospital is much lower, you have to keep in mind that the number of vaccinated people is much higher (in my country +80%, almost 90%, are vaccinated) so the effect of the vaccine in death rate is even much higher (before measuring the effect of vaccines we even thought that vaccinated people would be the most prevalent in hospital because the number is vastly superior)

Then why did this trend only emerge in April? The population of young people being vaccinated has increased since then.

There are all kinds of possible explanations for this. If you look at this heatmap of vaccine uptake by age and time (you'll need to select second dose)[1], you'll see that the uptake percentage is fairly uniform across all age groups below 60 e.g. in March, but goes up fast for 55-59 in April, meaning the "10-59 + Second dose" group became older on average during that time. My best guess is that the uniform distribution early on was from healthcare workers; later on, the age-prioritized vaccination rollout changed the distribution.

There's a reverse trend that coincides with an increased uptake in younger groups: The age-specific death rate for "10-59 + Second dose" in June was 2,8 and went down to 2,4 in September.

One thing to keep in mind: When we're talking about the 10-59 age group on their second shot prior to April, that's only a population of 800k with a total of 16 deaths reported between January and April. Confidence intervals for this period are very wide and the upper confidence limit in the ONS data is quite close to what they're reporting e.g. for September.

[1]: https://coronavirus.data.gov.uk/details/vaccinations?areaTyp...

If it's all due to age, then why did the vaccinated group (older) have a lower death rate than the unvaccinated group (younger) prior to April?

It does not seem unlikely to me that the group of health workers that made up most of the vaccinated group prior to April generally has a lower death rate than the unvaccinated group, which would have contained most aged 50+ at the time (vaccine uptake was <6% for the 50-59 group at the end of March).

More importantly, COVID also accounted for more than 30% of deaths in the "10-59 + Unvaccinated" group between January and April (2,761 of 8,665), as this was around the time the third wave peaked. As one would expect, this didn't really affect deaths in the vaccinated group.

I don't think anyone should attempt to draw conclusions from such a small group that is not at all representative of the whole population (in general and within the 10-59 age group).

Probably Simpson's Paradox. A 59 year old is both more likely to be vaccinated and to die from any cause, compared to a 10 year old.

Then why did this trend only emerge in April? The population of young people being vaccinated has increased since then.

Your link also provides the likely answer when you scroll down to the comments:

> One reason not to draw too much of a conclusion is that 10 to 59 is a really, really broad age group, and the underlying death rate at the top of that group will be much higher than at the bottom.

> If the vaccination is also unevenly distributed in that group (and you would imagine the vaccination rate to be much lower among 10 year-olds than 59 year-olds) that would probably provide a sufficient explanation.

Then why did this trend only emerge in April? The population of young people being vaccinated has increased since then.

It seems to me you are really interested in an answer, so maybe you want to workout an answer yourself:

* When did vaccination start for the group 10-59 in the UK?

* How does the makeup of the group "unvaccinated" change over time?

* What is the average distance between the two lines, are there any trends?

Let me know if you have any questions.

I assume you're saying that this is all due to the respective ages of the populations. If that's the case, then why did the vaccinated group (older) have a lower death rate than the unvaccinated before April?

To answer that question you need to dive into the original data set (if you want to see yourself: Search ONS death by vaccination status, in the .xlsx it's sheet 'Table 4').

There is a reason why the person who made the chart started with 19-Mar-2021: There aren't enough deaths before. Even for the 21-Mar data point it's only 5 deaths!

Luckily ONS does provide confidence intervals (column J, K). There is even a marker in the sheet called 'u' to highlight unreliable data due to small numbers. But whoever made the chart was not interested in either (why?).

While we are already in the full data set, note that the death rate for the unvaccinated group is appr. 4.0 (!) in Jan 2021.

There might also be another reason (hypothetical based on anecdotes): The very first to be vaccinated in the 10-59 group were likely nurses and doctors.