Showing posts with label statisitcs. Show all posts
Showing posts with label statisitcs. Show all posts

Saturday, 6 October 2012

And I thought it was something else that made you blind...

****

Not coffee:

Drinking more than three cups of coffee a day may increase the risk of vision loss and blindness, according to American research.

Even moderate amounts of the drink make developing the devastating eye condition glaucoma more likely. 

Funnily enough they don't tell you what the increased risk is - maybe it's not all that much? You know featuring the words "not statistically significant":

Compared with participants whose cumalatively updated total caffeine consumption was < 125 mg/day, participants who consumed > 500 mg/day had a trend toward increased risk of EG/EGS that was not statistically significant (RR = 1.43; 95% confidence interval (CI), 0.98-2.08); P trend = 0.06).  

Carry on drinking the coffee folks - it's something else that's making you blind!

...

Friday, 7 September 2012

Lies, damned lies and public health

****

It beats me why those promoting public health campaigns feel the need to (how do I put this) make stuff up.

I understand that some people have a nannying, controlling, interfering world view. They're brought up to believe it OK to go around telling other people that their choices are wrong. It's a world view I hate but why do its adherents feel the need to tell massive porkies all the time?

Don't believe me? Here are a couple of recent examples:

The rest of the story is that the European Respiratory Society is lying to the public. The Society claims that electronic cigarettes were developed by the tobacco industry. This is false. The tobacco industry played no role in the creation of electronic cigarettes and for the past four years, has had no involvement at all in the electronic cigarette industry. It was only recently, when Lorillard acquired Blu cigs, that the tobacco industry entered the electronic cigarette market. The claim that electronic cigarettes were developed by the tobacco industry is an outright lie.

And from the same set of public health campaigners and even worse lie:

The claim that electronic cigarettes cause cancer is without scientific evidence. While it is true that trace levels of tobacco-specific nitrosamines have been detected in electronic cigarettes, similar trace levels are also present in nicotine gum and nicotine patches. If electronic cigarettes can be said to cause cancer, then so can nicotine replacement therapy.

There is simply no scientific evidence that NRT causes cancer, because it is unclear that the trace levels of tobacco-specific nitrosamines in these products have any clinical significance, are capable of inducing cancer, or that there has ever been a human case of cancer caused by NRT.

Similarly, there is simply no scientific evidence that electronic cigarettes cause cancer, because it is unclear that the trace levels of tobacco-specific nitrosamines in these products have any clinical significance, are capable of inducing cancer, or that there has ever been a human case of cancer caused by electronic cigarettes.

So you've never heard of the European Respiratory Society. How about the BBC?

Programme makers asked Sheffield University to model the effects of a 50p per unit minimum price, which would push the price of the cheapest bottle of vodka from about £9 to £13.

Statisticians estimated the effect would be 50,000 fewer alcohol-related deaths in England among over 65s, over the course of 10 years.

The current affairs programme, to be broadcast on BBC One at 7.30pm on Monday, examines problem drinking among Britain’s ever-growing population of older people. 

Let's be clear about these statistics. They are utter rubbish. Complete and absolute tosh. Yet the BBC proposes to broadcast them without challenge. Here are the facts on alcohol related deaths (or a close as we can get using the entirely questionable approach preferred by public health people).  The figure - allowing for estimates of homicides, suicides and drownings - is around 13,000 per year. And declining.

The BBC wants us to believe that introducing a 50p per unit minimum price will result in 5,000 fewer of these deaths. That's a 40% drop. More to the point these statistics mean that nearly every single death of someone aged over 65 currently listed as 'alcohol-related' is eliminated. This is complete nonsense as a moments perusal of the actual statistics would show.

I think you get the picture. On everything - from smoking in cars to the effect of salt consumption the public health lobby simply makes stuff up. And repeats it time and time again until everyone assumes it is correct. Policy affecting millions of people and their choices is being made on the basic of dodgy science, misrepresentation of statistics and, it seems, outright lies.

....

Friday, 27 May 2011

Lies, damned lies and alcohol statistics - the truth about those hospital admissions figures

I was rather bothered yesterday by the revelation - leapt on by, it seems, every second media outlet with excitement - that "alcohol-related hospital admissions" had doubled since 2002.

The number of alcohol-related hospital admissions in England has exceeded one million in a year for the first time, a report has found.

Figures compiled by the NHS Information Centre for the year 2009-2010 revealed 1,057,000 hospital visits in relation to alcohol, a 12% increase on the previous year and more than double the amount recorded in 2002-2003. 

I was especially bothered since alcohol consumption has fallen year on year since 2002 and this degree of increase in alcohol-related ill-health simply doesn't tally with that fact.

I now know why thanks to Chris Snowdon and Nigel Hawkes that there's a pretty straightforward explanation:

These figures use a new methodology reflecting a substantial change in the way the impact of alcohol on hospital admissions is calculated. Previously the calculation counted only admissions for reasons specifically related to alcohol. The new calculation, for which the methodology is described in the report, includes a proportion of the admissions for reasons that are not always related to alcohol, but can be in some instances (such as accidental injury).

So these admission figures are not actual people really admitted to hospital for a drink-related reason but an approximation based on the estimation of a proportion of a given condition - say being hit on the head with a hammer - that, in aggregate can be attributed to drink.

And, as Chris Snowdon points out, where we really do know the admission is drink-related, the figures are falling:

Alcohol-related deaths – that is, those caused by conditions directly linked to alcohol – fell from 6,768 in 2008 to 6,584 in 2009. Much of the fall was attributable to a fall of nearly 250 in deaths from alcoholic liver disease.

Now that's not what they were saying on the news was it now?  This is twice in two days where the nannying fussbuckets have either misunderstood or misrepresented statistics about alcohol and health - I'm beginning to believe it's being done deliberately.

....