Read on Twitter
Short thread on UK All Cause Death Date. One issue with looking at official #COVID data is that they are deaths with people who had a positive COVID test within 28 days. This causes issues in ascertaining mortality burden of COVID in an environment of mass testing.
Excess death data helps solve this issue but also problematic as need to account for lockdown deaths.
One way to disentangle this is to look at the gender + age profile of all cause mortality. COVID deaths have a fingerprint. It kills more men than women + age of death is higher.
One way to disentangle this is to look at the gender + age profile of all cause mortality. COVID deaths have a fingerprint. It kills more men than women + age of death is higher.
Looking at all cause deaths by gender 2018-2020 we can see a pattern. More women die in winter (dec-feb) where as more men die in spring/summer (mar-sep). As women live longer than men, most old people are female, which is why more women die in winter flu season.
We can see the March-April 2020 COVID peak, a large spike in % male mortality. Interesting the male peak falls off earlier than all gender excess. This may be the care home breakouts, as there are around 2-3 women in care homes for every male (again because they live longer).
Currently excess deaths are at baseline and the % male is in line with normal for this time of year. If we see spikes in excess deaths, but not % male deaths, this may suggest they are not predominantly COVID deaths.
Looking at deaths by age. The vulnerable group to COVID is mainly over 65. Looking under and over 65s deaths we can see a normal pattern & a COVID fingerprint.
In winter deaths skew older (in line with more female deaths) as the frail are knocked out by flu etc. In summer deaths skew younger. Now we can clearly see the over 65 bump during COVID peak, which matches excess deaths exactly.
But interestingly, now the age profile is back to or slightly younger normal age range. Also notice the large increase in younger deaths right after COVID peak in April. If excess deaths rise, but % over 65 doesn’t, this suggests the deaths aren’t predominantly COVID.
