Regional variations

Trying to answer my own questions about the fictional swine flu (see yesterday's entry below) has sent me surfing around the Scottish, Welsh and English/UK official swine flu sites. The Welsh seem to produce the clearest and the one written (ironically) as much as possible in plain English - presumably plain Welsh as well. I find the UK/English one the most confusing - not least as to whether graphs refer to England or the UK as a whole. But the Scottish and the UK/English reports both suffer from very confusing multi-coloured graphs that are hard to read. Indeed on occasion it seems nonsensical to even bother publishing them. Where are my friends at Straight Statistics when you need them!

It's also been interesting to see the regional variations in swine flu - there is the "Tayside effect", highlighted by the Courier and Flutrackers.com. And what about this throwaway remark in the latest Welsh report: "As at 23 September, there were 11 clinically-diagnosed patients in Wales hospitals in connection with swine flu, with three in critical care and no reports of underlying health conditions in any patients". That echoes WHO warnings about the riddle of patients with no underlying health problems.

I shall be back, hopefully tomorrow, with the identity of the members of my swine flu family most at risk from swine flu.

Meet the swine flu family

Pictured left - and there's a better version on my Flickr account - is the swine flu family. With the help of my "learn to draw" book and a few 2HB pencils, I've drawn them up for use as a graphic a la the brilliant Hans Rosling (the best use of statistics ever seen). They'll be appearing in various forms on this blog but my first questions are:

* Which one is most likely to get swine flu?


* Which one is most likely to have to go to hospital with swine flu?


* And which (although it is a very rare event) is most likely to die from swine flu?

It's not easy to find the answers in the UK statistics so answers on a postcard ...

Lack of swine flu data is a scandal

My thanks to one Linda Robinson who gave me a link to the www.flutrackers.com website in my quest to try and collate more details of swine flu victims in the UK. I've added the details from that site to my list. I've tried to be very careful but the information is so confused that I can't rule out omissions or duplications. Email me if you have any corrections or additions. It's appalling that the details from official sources are so sparse that I am having to do this. It is very reminiscent of Daniel Defoe's Diary Of A Plague Year (set in 1665) and going round the streets to try and find news of how many people have died! It's difficult to see what this secrecy achieves - except suspicion that the government is hiding the true situation. The public should be given as much information as possible. For instance, most believe the young are worst affected by swine flu but this doesn't seem to be borne out by the ages and genders in my list. It's difficult to pick out the worst offender in this blanket of secrecy but I'm going to have a go anyway: Newham University Hospital Trust. The good work of reporter Susan Smith on the Newham Recorder enabled the paper to discover that someone on their patch had died of swine flu. The Trust grudingly confirmed the fact but the most telling line in Susan's story is "Newham University Hospital NHS Trust will not reveal the gender of the patient." For goodness sake, why not? One can understand the name and address not being given out (although personally I think those details should be given) but what possible reason is there for not saying whether it was a man, woman or what age they were. (One wonders how the public would react if they later found out the adult was worked at a nursery but no parents were told). This nonsense - the lack of data - has got to stop.

List of swine flu deaths in the UK

It's time to start blowing the lid on the scandal that has become swine flu in the UK. And I don't mean a conspiracy that it's all down to deliberate germ warfare by MI5. I mean the dearth of information about the deaths and the flippant remarks everytime someone dies that they had "underlying health problems" (UHPs). As others have pointed out, we have almost no data about those who die from swine flu - whether they are men, women, how old they are, and what UHPs if any that they had. So I thought I'd draw up my own statistics from a trawl of the net. Here's the list of UK deaths that have been made public to date. Please email me (cleaverw@aol.com) with any corrections or additions.

Known Swine Flu Deaths in the UK

14 June 2009 - First death - Jacqui Fleming, 38/39 died giving birth at Carnwadric, south Glasgow (UHPs*)
15 June 2009 - Jack Fleming, son of Jacqui, dies
29 June 2009 - man, 73, Paisley, Scotland (UHPs) - second Scottish death
30 June 2009 - Sameerah Ahmad, 6, Birmingham (UHPs)
30 June 2009 - Girl, 9 (family asked details to be private), Birmingham
3 July 2009 - Man, 19, South London. Serious UHPs (first to die in London)
5 July 2009 - Abdullah Patel, 42, Dewsbury, Yorks (UHPs - lung and kidney problems
6 July 2009 - Asmaa Hussain, Dewsbury, Yorkshire
9 July 2009 - Chloe Buckley, 6/7, West London (septic shock as a result of tonsilitis)
9 July 2009 - Dr Michael Day, Luton, 64 (UHPs - heart disease, high blood pressure, viral pnuemonia)
10 July 2009 - healthy man, Basildon, Essex
15 July 2009 - woman tourist visiting Inverness, Scotland (significant UHPs)
17 July 2009 - pregnat woman, 39, Whipps Cross, London. Baby in intensive care
17 July 2009 - Sky news reprots Death from swine flu of woman who had just given birth and Baby less than six months old (serious UHPs)
17 July 2009 - Woman, Ruptara Miah, 39, at Leytonstone (just given birth)
21 July 2009 - Girl, 15, Glasgow, Scotland (UHPs) - fourth scottish victim
22 July 2009 - no sex, age given. West Midlands
4 August 2009 - Soldier Lee Porter, Bisley, Surrey, 30 (no UHPs)
8th August 2009 - man, 26, Glasgow (fifth scottish death)
9th August 2009 - Madelynne Butcher, 18, Southampton
14th August 2009 - adult dies in Birmingham (UHPs)
21 August 2009 - woman, 55, Wales
22 August 2009 - woman, Northern Ireland
24 August 2009 - woman, 59, Islay, Argyll, Scotland (UHPs) - sixth scottish death
28th August 2009 - man, 52, Edinburgh; "significant UHPs" - seventh in Scotland
5 September 2009 - Boy, 7, Strood, Kent? - later said not to be swine flu

And with the death toll at 70 at the moment, that means people in the UK are dying from swine flu at the rate of 17 a month.

*UHPs - Underlying Health Problems; a glib phrase trotted out by medical professionals. At the moment they seem to range from viral pneumonia to being pregnant

Information please

Can anyone else find UK statistics for how many men have died from swine flu compared with how many women have died from swine flu because I'm damned if I can find out!

Swine flu and statistics - again!

I'm glad to report a website I've discovered called Straight Statistics. Its mission statement says: "We are a campaign established by journalists and statisticians to improve the understanding and use of statistics by government, politicians, companies, advertisers and the mass media. By exposing bad practice and rewarding good, we aim to restore public confidence in statistics." Amen to that! A campaign that's is long overdue and most definitely needed here in the UK.

There are some worrying reports about numbers of swine flu cases in American schools that have already returned from their summer holidays. Most UK schools will be starting to return in the first few days of September.

Latest swine flu figures (end of August)

The latest swine flu figures have been released by the UK government. They show "Pandemic influenza activity continues to decrease across most regions of the UK and in all age groups, though remains at levels higher than expected for this time of year." The total number to have died so far this year in the UK is 65. But as the figures were being issued Scotland suffered its seventh death from swine flu. The 52-year-old not only had Underlying Health Problems but "significant" underlying health problems. As usual, no information is given on what these UHPs are or how many of the rest of the population have them.

Count me out

Journalists, by and large, don't do maths. Ask them what 40 per cent of 40 is and they won't have a clue. Ask them what the difference is between a median average and a mode average and they'll just look blankfaced. It's only when they come to fill out their expenses that they suddenly become boy geniuses about tax and compound interest. And I count myself as among those who are statistics-blind. However, I hopefully can at least ask questions and spot when some table or graph looks or smells a bit dodgy. So I'm going to start an occasional series of swine flu stats presented by journalists. And here's the first, courtesy of the BBC website...

It purports to show swine flu increases in the last week but (and it may be just me being thick) surely without any figures it's meaningless. If eight people died in China one week and 12 the next, that's a 50 per cent increase - worrying but 'only' four more people. But if 600 died one week and 900 the next, that's still a 50 per increase but now it's 300 who died in one week. So the BBC table tells us almost nothing. Or is it me?

It's not just me

I was encouraged to see the Royal Statistical Society questioning the lack of detail in the government swine flu statistics - though I fear this is one press release that is going to pass most journalists by. President David Hand had written to Sir Liam Donaldson with a few questions. You will find his letter online but the key points are:
The dangers of producing weekly reports on swine flu instead of daily ones

• Some vagueness on statistical standards
• Lack of demography in the data
• Lack of other data relating to sex and age of patients

And interestingly a concern that those who die from H1N1 away from hospital might be missed out of the statistics. This line from David Hand's letter particularly interested me:

"Readers of the UK’s weekly H1N1v updates need to know numerators and denominators, not just percentages, and the empirical basis for multipliers used in inferring from GP-consultations for influenza-like-illness (ILI) to headline figures on the estimated number of H1N1v cases. Moreover, the provenance for intervals of uncertainty should be properly explained."

I think that's what I've been saying - but the RSS puts it in their own jargon. It seems to be saying there's a basic lack of statistical data missing. If so, journalists should be highlighting the government's failure to understand the necessity of such information.

But I fear not much is going to come of this. Journalists and maths/statistics just do not mix. Very, very few journalists have the understanding to look at a graph or table of stats and ask pertinent questions. And there's another problem. The Royal Statistical Society can't talk in plain English. They need a press officer who can explain in layman's term precisely what their concerns are with H1N1 statistics. It's no good hiding behind "oh it's all very complicated and can't be simplified". It's the job of press officers and journalists to make sure the public can understand such information.

One bright light on the horizon: The RSS does hand out awards on statistical excellence in journalism which hopefully encourages the hacks to improve their act.