Some comments on excess mortality in the second wave, which seems to be hotly contested at the moment.

Headline: there is excess mortality.

ONS data shows us that death numbers have been above the five-year average for 8 consecutive weeks (blue bars taller than red). /1 https://twitter.com/actuarybyday/status/1326098173039964160

Analysis of age-standardised death rates by the Continuous Mortality Investigation (who have been producing objective and authoritative mortality analysis for decades) shows excess deaths year-to-date are increasing again in the second wave (black line on chart is rising). /2 https://twitter.com/covid19actuary/status/1326181697176690690

So, it is clear that deaths are above average. But is that statistically significant?

I can see how people arguing in good faith might disagree on this point. (Which is not to say that I think everyone questioning it is doing so in good faith!) Let’s look at the evidence. /3

Here is @PHE_uk in last weeks Mortality Surveillance report:

“In week 43 2020 in England, no statistically significant excess mortality by week of death above the upper 2 z-score threshold was seen overall or by age group”. /4

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/930814/Weekly_report_mortality_W44.pdf

Meanwhile, here is @PHE_uk in their most recent Flu and COVID surveillance report:

“Significant excess all-cause mortality was observed in week 43 overall, by age group in the 75 to 84 year olds and 85+ year olds...” /5

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/933459/Weekly_Flu_and_COVID-19_report_w45_V2.pdf

(Someone is going to quote either tweet 4 or 5 to reinforce their point aren’t they!) /6

Whilst I’m sure there is a technical reason that the two @PHE_uk reports contradict each other about week 43 mortality, it is somewhat unfortunate that people trying to look at the evidence for themselves get a different answer depending on which report they read. /7

PHE are better placed than me to explain the difference but note that the flu surveillance report (saying excess mortality IS significant) is more recent. The earlier mortality surveillance report was provisional and makes estimations based on typical reporting delay. /8

Let’s go back to the big picture rather than get bogged down debating whether individual weeks are statistically significant. We have had 8 consecutive weeks of above average mortality. If this was just random variation that would be like getting heads from 8 coin tosses. /9

In conclusion then, it is clear that the second wave has tragically resulted in more excess deaths in the UK. COVID deaths have been rising week-on-week and this has not been fully offset by falls in deaths from other causes. How much worse it will get remains to be seen. /end