Bunkers, cellulose, get thee behind me Internet

Ben Goldacre of Bad Science fame is frittering away his time on Twitter instead of doing the decent thing, i.e. posting on his blog. SHAME, Ben, SHAME.

With the unsurprising result that the hardcore fan ends up following Dr Goldacre on Twitter.

Thereby discovering, for example, a link to a fabulous post on Churchill, bunkers and the chemical composition of wood, of which this is a sample:

Wood is made principally of cellulose, which is the same stuff from which cotton clothes and paper are constructed. Cellulose is difficult to break down, as the individual molecular strands are tightly packed together by hydrogen bonding, making a near-crystalline material that is very impermeable to water, and even more impermeable to digestive enzymes. Most herbivorous animals subcontract out the work of breaking-down cellulose to the bacteria and fungi that live in their guts.

Cellulose consists of thousands of glucose molecules chained together.

Although cellulose is difficult to break down, the other main component of wood, lignin, makes cellulose look positively fragile. Plants make lignin by secreting phenolic alcohols into their cell walls, and then semi-randomly polymerising these alcohols together using free-radicals. The mechanisms of lignin synthesis and its global structure are still areas of active research (or furious argument, depending on your point of view). From the plant’s perspective, lignin is a marvellous glue: it creates a substance that cannot be broken down by conventional enzymes, as you’d need hundreds of them, one for each of the many kinds of linkage found in the lignin.

(Note that it would be impossible to quote this splendid post, with diagram of cellulose, on Twitter. SHAME, Ben, SHAME.)

The rest, anyway,  here.  (Yes, I know the font has changed. And if I went into HTML I could fix this. But I am catching a plane at 7 am, so sloth prevails.)

I have now sublet my apartment in Berlin and am going back to the States for a while. I will be spending 2 months in Vermont in a place with no Internet access, with the faint frail hope of finishing a book. After that, who knows? As Bialystok says in The Producers, in the months to come you will hear little of me.  But I leave you with a link to an excellent blog, and those who have hitherto sneered at Twitter may like to follow the feckless Dr Goldacre, @bengoldacre.

split or steal

HT Ben Goldacre, amazing display of psychology and game theory on a game show, here.

bad science

You write – “First of all it is so good to hear you are doing well!”
As you have made no effort to find out the truth for yourself, and as I am the only source of information that is reliable, it is time you were apprised of the facts.

Perhaps, you may then even contemplate apologising.

...Fact 

The MRI scan taken in April 2003 was read accurately by the surgeon Cameron Platell, and the consulting GP.
It was read inaccurately by you.
The lymphs you informed me were simply overworked were already cancerous. The ovary you informed me was swollen due to another cyst was also cancerous. At this point, my uterus and second ovary were healthy and unaffected. I presume you remember - you examined the scans yourself.

Fact

I have had to have a total hysterectomy. You know this because I informed you during one of the conversations we had while I was in hospital, although your card makes no mention of it.
I am now going through early onset menopause.
I can never bear children.
You are responsible for this outcome.
Considering the fact that my original treatment with you was aimed at getting me a pregnancy, and taking into account your assurances that I would fall pregnant and have a healthy child after my cancer was cured, this is all the more grievous and shocking. 

Penelope Dingle's letters to Francine Scrayen at Australian Story, HT Ben Goldacre.

Coroner's report here.

Singh & the BCA

But most damnable is that this case should have taken place in the arena of medicine, where reasonable criticism of each others practises should never be stifled, for one very simple reason: it’s possible, in medicine, to do enormous harm, even when you set out with the best of intentions.

Anti-arrhythmic drugs provide a chllling example. For a few years in the 1980s these were prescribed to everyone who had a heart attack. It made absolute sense in theory: people who’ve had heart attacks often get abnormal heart rhythms, these irregular rhythms can often finish you off, but anti-arrhythmic drugs prevent them. Why not just give these anti-arrhythmic drugs to everyone who has had a heart attack, on the off chance? They were safe and effective when given to people who had abnormal rhythms, after all. But when prescribed preventively, to everyone, after a heart attack, they turned out to increase your risk of dying, and because so many people had them – because so many people have heart attacks – the deaths were on a biblical scale, killing as many Americans as died in the whole of the Vietnam war, before anyone had a chance to notice that something was wrong.

Ben Goldacre on Simon Singh's victory in the libel suit by the British Chiropractic Association, the rest here.

the wrong kind of data

Ben Goldacre's Bad Science, on the results of a 1995 WHO report on cocaine use.

"Health problems from the use of legal substances, particularly alcohol and tobacco, are greater than health problems from cocaine use," they said. "Cocaine-related problems are widely perceived to be more common and more severe for intensive, high-dosage users and very rare and much less severe for occasional, low-dosage users."

The full report – which has never been published – was extremely critical of most US policies. It suggested that supply reduction and law enforcement strategies have failed, and that options such as decriminalisation might be explored, flagging up such programmes in Australia, Bolivia, Canada and Colombia. "Approaches which over-emphasise punitive drug control measures may actually contribute to the development of heath-related problems," it said, before committing heresy by recommending research into the adverse consequences of prohibition, and discussing "harm reduction" strategies.

...

At the point where mild cocaine use was described in positive tones the Americans presumably blew some kind of outrage fuse. This report was never published because the US representative to the WHO threatened to withdraw US funding for all its research projects and interventions unless the organisation "dissociated itself from the study" and cancelled publication. According to the WHO this document does not exist, (although you can read a leaked copy at www.tdpf.org.uk/WHOleaked.pdf).

fudging

But how can you tell if the research literature on a given subject has been rigged? It’s a tricky problem, because you’re chasing evidence for the existence of trials you cannot see. One option is to use mathematical tools, and something called a funnel plot, one of the cleverest ideas of the last century. It’s so clever that you might need to concentrate for the next bit.

Let’s imagine that there are 30 trials on a given drug. Some are big, and more accurate. Some are small and less accurate, with more random noise. You’d expect that the big, accurate trials should all cluster together around the true finding, all giving similar results for the efficacy of a drug. Meanwhile the smaller, rubbish trials - because they are less accurate measures of the drugs efficacy - will be scattered about randomly, some showing the treatment to be better than the good big trials indicate, some showing that it is worse.

You could then plot all your trials on a graph, one dot for each trial. On the x-axis, left to right, is “how good the drug was shown to be by this trial” and on the y-axis, “how methodologically sound and large the trial was”. If there is no publication bias, you should get a triangle shape: at the top of your graph, you will see all your good-quality, accurate trials, clustered together around the true answer. At the bottom of the graph, you will see a broad smear of results, the poor quality trials showing random variation.

But if there is publication bias, you will see a distorted triangle: the small, poor-quality trials at the bottom will be smeared over to the right, because small trials with unwelcome results are much more likely to be overlooked, and dumped in desk drawers, than huge multicentre collaborative studies involving dozens of academics and tens of thousands of participants, which are almost definitely going to get published. If you get a distorted triangle, you know there are some interesting negative trials missing.

Ben Goldacre in the Guardian's Bad Science column, the rest here.