Statistic of the day

Browsing in the forest of viral data yesterday, I came across the graphic from ONS (reference 1) snapped below.

The first impression was good. Nice clear presentation, a graphic not trying to do too much and with a reasonable supply of supporting information in the way of notes.

But then I started to get more picky. We are told when they stopped counting the deaths but not when they started. And given that all the deaths were registered in 2020, the wording of note 3 is not that clever.

And before we can use the figures, don't we need an analysis of population by age and sex, an analysis of deaths at large by age and sex, an analysis of infections by age and sex? Do the tall second, third and fourth blue columns really reflect an increased vulnerability of the male, with the tall fifth yellow column (right) just reflecting the fact that there are a lot more women than men in the top age band? Is it not a question of vulnerability at all, rather one of risk and exposure. With BH suggesting that males are much more prone to risky behaviour, perhaps in this case resuming attendance at crowded public houses. Which all goes to illustrate the problems that statisticians face when offering statistics to the world.

In the meantime, maybe we can make some use of the raw fact that more men than women are dying of the new virus. Perhaps provision for males is not the same as provision for females, and so one does need to know the numbers of each. Do we, for example, still have segregated wards for men and women, wards which these figures suggest would need to be of different sizes? 

PS 1: rather to my surprise Bing knows all about 'U07.1' and 'U07.2' without my needing to supply any further clues, offering a great raft of stuff about them, including: 'U07.1 is a billable/specific ICD-10-CM code that can be used to indicate a diagnosis for reimbursement purposes' and 'the reason you are getting errors in billing this code is that because 'U07.2 Covid-19, virus not identified' has not been imported into ICD-10-CM ... coming up as invalid. It is actually a valid code'. The medical insurance industry in the US is clearly both very highly organised and very visible to Internet search engines.

PS 2: Cortana having another bad week. Snaps from my telephone have still not reached my laptop from Tuesday. Not a problem here as the snap was sourced locally.

Reference 1: https://www.ons.gov.uk/.