Jumping on the winter deaths bandwagon
Several papers this morning bemoaned the latest figures for excess winter deaths, but it’s not clear why.
“Big freeze and higher bills to put more lives at risk” said the Daily Mirror, which reported the figure for excess deaths in 2010-11 (25,700) without disclosing it was lower than last year’s (25,810) and considerably lower than in 2008-09 (36,450).“Pensioners facing winter catastrophe” was the headline in Edinburgh Evening News.
Michelle Mitchell of Age UK was quoted in the Daily Express as calling the figure “a disgrace", and “the tip of an iceberg of illness, misery and anxiety which grips Britain every winter”.
She’s entitled to hold that view, but in fact there is nothing very striking about the latest figures. They are in line with statistics going back to 2000-01 and show no evidence of a very strong link to the severity of the winter (see table).
This doesn’t discourage Jenny Saunders, Chief Executive of National Energy Action, claiming that “Excess deaths show fuel poverty is killing the vulnerable. The Government must act.”
Yet the table also shows that there is no obvious correlation between deaths and the Department of Energy and Climate Change’s index of fuel poverty. While fuel poverty affecting households that include vulnerable people such as children or the elderly has tripled since 2004-05, excess deaths have shown no comparable change.
It’s plain that deaths do increase in the winter months. Countries with mild climates such as the UK tend to show greater seasonal variation than those with tougher climates such as Scandinavia, a paradox that some researchers have explained by suggesting that those used to extreme cold are better protected against it, particularly when outside. There is no strong link to socioeconomic status, at least in the UK. The vast majority of deaths are in over-65s.
The trend since the 1970s has been downwards, which some have attributed to increased car ownership, which protects against the cold when out of doors. But much about excess winter deaths remains puzzling.
Given these uncertainties, while it may be understandable for charities to blame the deaths on high fuel prices and even “the cuts”, there isn’t a lot of evidence to substantiate their claims.
In the case of NEA, the misrepresentations go beyond mere spin. The press release that appeared on its website began: “Figures released today by the Office for National Statistics show a massive 38 per cent increase in deaths across England and Wales during the arctic cold snap experienced last December, the coldest in over 100 years”.
The ONS never mentioned a 38 per cent increase in deaths. Nor does it recognise the figure of “almost 4,000 more registered deaths than the five year average” for the first week of January 2011 quoted by NEA. The Mirror reported the excess at 3,500, not almost 4,000, but neither is right.
The statistical bulletin containing the figures says that between 26 November 2010 and 8 January 2011, deaths had exceeded the five-year average by about 75 deaths a day. By my count, that’s 44 days times 75 deaths, or 3,300 – but over the whole period, certainly not in a single week.
The ONS statistician responsible says that it would be possible to extract exact figures occurring in a particular week, for a price. But the number of excess deaths in that first week in January would be less than 1,000.
The press release on NEA's website today (November 24) no longer claims the 38 per cent increase. It has evidently been changed in response to ONS complaints. However, echoes of the original claim can still be found on the Internet, as on the Energy Efficiency News website, which reports: “NEA says the cold spell in December precipitated a 38% increase in deaths across England and Wales”. Not any more it doesn’t.
But NEA still claims that there were 4,000 excess deaths in the first week of January. The ONS denies it. How about modifying that claim, too?
Javier (not verified) wrote,
Thu, 24/11/2011 - 13:35
"But much about excess winter deaths remains puzzling".
What about the flu?
Congratulations for your great blog.
Eleanor Bradford (not verified) wrote,
Thu, 24/11/2011 - 14:08
It is interesting that the Registrar General for Scotland released equivalent figures for winter deaths in Scotland one month before the figures for England. The way the information was presented was clear and left journalists in no doubt about the long term trend and significance. This was the copy we ran:
"Despite unusually cold weather, there were fewer deaths last winter than during the previous 20 winters, according to statistics released by the Registrar General. The majority of deaths during the coldest months are from heart disease, stroke and pneumonia, with very few caused by flu or hypothermia. Overall, the number of deaths during wintertime has been falling for the last sixty years - whatever the weather."
Which leads me to wonder who is most at fault in the reporting of the English figures: the journalists or the statisticians?
Joanthan (not verified) wrote,
Thu, 24/11/2011 - 16:11
I've read NEA's article and it doesn't claim excess deaths, it refers to winter deaths and it does correlate with the ONS release and their chart that there was an increase in deaths during the cold snap, even if this wasn't the figure originally stated, which the ONS does say is 5% higher than average. This clearly shows there is a correlation between cold weather and deaths. As for the mirror they have read the information wrong, as the release states over the same period, not just that week.
Although the EWD were lower this year if you look at the regional results some of them are higher, the North East by 26%. Is it coincidence the North East suffered the worst of the weather?
Any one person dying as a result of the cold is wrong. 25,00 is a disgrace.
John (not verified) wrote,
Thu, 24/11/2011 - 18:13
In your article, you have included the paragraphs:
"The statistical bulletin containing the figures says that between 26 November 2010 and 8 January 2011, deaths had exceeded the five-year average by about 75 deaths a day. By my count, that’s 44 days times 75 deaths, or 3,300 – but over the whole period, certainly not in a single week.
The ONS statistician responsible says that it would be possible to extract exact figures occurring in a particular week, for a price. But the number of excess deaths in that first week in January would be less than 1,000."
The last paragraph is not clear. Data for particular weeks is published by ONS - but not data for particular days within a week. The use of the term "excess deaths" elsewhere in the article refers to deaths compared to the average summer deaths - i.e. Aug-Nov and April-July. (As it happens, the number of excess winter deaths in the first week of January is around 3,500.) Your reference to "less than 1,000" I assume is comparing the number of deaths in the first week of January 2011 with the 5 year average for the same period.