Guardian has this slightly odd view on ONS data
www.theguardian.com/world/2020/jun/18/covid-19-excess-death-rates-more-than-twice-uk-average-for-19-english-councils
Odd why? Because it's for W1-23, whereas covid-19 started in W12 approximately, so W1-11 is kind of irrelevant, and because we'd expect more deaths from covid-19 in older LAs, and fewer in younger ones. Also by adding in W1-W11, which might have been a mild winter with fewer deaths than average, we'd show no excess deaths for the year when in fact there were fewer deaths in January-March and then excess in April/May.
Also it claims
[Excess deaths] is considered the gold standard for measuring the full impact of Covid-19 crisis because it accounts not just for people who have died directly as a result of coronavirus. For example, it includes the deaths of those who may not have sought or received treatment due to lockdown measures, deaths caused by a lack of basic care or nutrition, and undiagnosed coronavirus deaths.
However it doesn't really, because the registrations won't have happened yet for some indirect covid-19 causes such as murder, suicidide, etc. A better way would be to compare week-by-week occurrences by ICD-10 code
icd.who.int/browse10/2019/en
and to exclude those causes from previous years that would be registered late.
In particular
V01-X59 Accident
X60-X84 Intentional self-harm
X85-Y09 Assault
Y10-Y34 Event of undetermined intent
Y35-Y36 Legal intervention and operations of war
would be removed from previous years death counts, and compared to this years without those causes, by week of death. We don't really care about car accident deaths registered in June that occured in November for example.
It's not clear what's meant by this:
Excess deaths are calculated by taking the total number of people whose deaths were registered in 2020 and comparing it with the average number of deaths in the five years prior.
Presumably for the same weeks.
It makes this rather bald claim:
"There is a strong correlation between high excess death rates and age. For example, the rural South Lakeland area’s excess death rate stands at 39%, almost twice the national average, and the area is one of the oldest in the UK when measured by median age."
However median age is not particularly relevant - the relevant figure would be the population over 80. And in general if South Lakeland is older than average then the existing death rate would have reflected that.
We expect deaths to include some level of deaths due to illness, which is largely due to age, and some due to premature death from suicide/murder/accidents. Therefore it doesn't obviously follow that age would mean high excess death rates, as while the premature death registrations would not rise in covid-19, the death rate is likely to rise in proportion to previous years - a very young area would have a lower death rate to start with so any rise would be a bigger %.
It's not clear if they are just stating things as fact here without evidence, but it seems that way. I would look more towards a particularly large number of care homes relative to population, and the size of the over 80 population, not just 'the median age' is higher.