Latest on excess deaths in UK - May 2023
Yes, the deaths thing again.
Sometimes it helps to step back a bit. Here’s a long, simple view of the number of weekly deaths in England and Wales over time vs their previous 5-year average (4 week rolling average of both to smooth out registration bumps on holidays).
Oversimplifying a huge amount, when people talk about “excess deaths”, they mean the places where the red line goes above the blue one. Excess deaths are given by the area between them.
This read is not perfect for a bunch of reasons. It’s by registration date rather than occurrence rate, and it’s not age-adjusted (which leads to errors with an ageing population). But it displays some top-level points, which survive more careful analysis:
Deaths were roughly predictable pre-COVID, with some variation for good and bad winters (mostly flu).
COVID killed a lot of people in two huge waves in spring 2020 and winter 2020-1.
Since July 2021, there’s been a consistent, significantly higher death rate than previously - and this excess is higher than can be accounted for by those with COVID mentioned on their death certificates.
None of these three points are seriously debated. But the reasons for point 3 very much are, especially as this excess is maintained through a period where COVID is becoming a smaller and smaller factor in overall mortality - at least as measured by death certificate numbers.
We can get a clearer view if we zoom in on 2020 onwards, and use the ONS’s deaths and averages not by registration date, but occurrence date (a modelled number for the most recent weeks to make up for registration date). Then, some of the shenanigans created by things like holidays falling on different days fall away, and the picture is clearer right up to the present.
People are dying at more than the 5-year average across a whole bunch of conditions associated with being generally unwell (notably, excluding cancers, which seem to be roughly normal).
But - and I feel this is news that is both good and important - this unexplained excess seems to have more than halved. That is, it’s <50% lower than the consistently appalling levels that started in July 2021 and continued until the first week of January 2023 (but for the missing flu winter of 2021-2). It’s now ~20-30 excess per day rather than the 80-100 it was stuck at through 2021-22.
We’ve only got a very imperfect read on the 2023 rates so far, and they might still float up again. But I guess I’m sticking my neck out and predicting that they won’t, just because I believe the temporal coincidence of these high excess deaths coming at the same time as the period of overwhelming pressure on the NHS is not a coincidence.
This pressure is most dramatically indicated by looking at ambulance response times, which jacked up to absurd levels in July 2021 and continued until the first week of January 2023.
These response times are still way away from target (and NHS pressure is still high), but they have not spiralled back up into insanity. And so long as they don’t, I think it’s reasonable to predict that England and Wales’ excess deaths will maintain their new, non-zero, but - tentatively - lower level, at about half what they were before.