i agree that gleaning evidence of risks like ccf from a study like the one adwatch refer to is always difficult, because you may not have a hypothesis about what side effects will be caused before starting to collect data, so you do run into multiple comparison problems, so adverse outcome stuff may well be more hypothesis generating/worry-forming than confirming. i’m not sure the criticism of that one point trashes the entirety of the adwatch project (which is also well received in academic quarters etc) but i do agree it is something to think about.

AdWatch claim that there was a statistically significant difference between treatments in coronary failure. The source paper gives the p-value as being 0.05. That’s only statistically significant at the 5% level if the 0.05 was rounded up, rather than down (which just goes to make clear that hypothesis testing is non-scientific). Many people would argue that the test should be adjusted for multiple comparisons, anyway. But in any event, that kind of p-value does not represent strong evidence of a difference by any stretch of the imagination.

It gets worse, though. The source paper makes clear that an independent cardiologist was employed to go through all the data relating to serious adverse events, including CHF, and his/her expert opinion differed in some cases from what was reported by the sites (most of which will have expert cardiologists). The p-value relating to the independent cardiologists classification was 0.26.

The AdWatch people have clearly misrepresented what was reported in the source paper. This isn’t the only place in which they’ve done it either. They criticized the ad for reporting the glucose results rather than glycated hemoglobin. But the clinical trial titrated patients to achieve the then current guideline targets which were for glucose. The authors of the source paper make clear in more than one place. So the glycated hemoglobin data represent what happens to gh if you titrate glucose to target and is not indicative of what would happen if the patients had been titrated to gh targets. Again, the AdWatch people are deliberately misrepresenting the source paper.

There are yet more flaws in the AdWatch criticism of the GSK ad, but if you had any interest in spotting bullshit, you would have spotted them already. Criticising the pharma industry for bad science, and using extremely misleading and deceptive non-science to do so doesn’t seem very consistent.

I go out of my way to find doctors that do not allow drug reps in their office. (nofreelunch.org) If a drug rep does come in and especially if he’s seen before any patients still in the waiting room, then I’ll leave.

- classic ‘Astroturf marketing’

is this where Basant Puri pops up again??

In such circumstances, it is essential to have a Secretary of State for Health who is prepared to stand up to this kind of pressure so that the NICE system of approval is not undermined. What a good job we have Patr… er… oops!

The example that always strikes me as a key one for this issue is the various drugs (mostly enhancers of acetylcholine neurotransmission) touted by the Pharmas for Alzheimer’s. NICE ruled (as I remember it) that there was no convincing evidence that the drugs worked. Much of the “pro” evidence the Pharmas cited related to various surrogate outcomes (see above) , and pretty much no benefit could be shown in standard morbidity and mortality measures. These slight hints, NICE said, weren’t enough to justify the NHS spending money on the drugs.

NICE’s decision was followed by loud protests from patient groups, carer groups, Alzheimer’s charities, and from a fair number of neurologists and geriatricians who argued this was taking off them the only vaguely hopeful drug they could offer the patients and their families.

This sort of scenario is where the murky relationship between the Pharma people and some doctors is at its most contentious. If many leading clinicians had run studies on the drugs for the Pharmas and had other close ties – including financial – to the companies, then there would always be the feeling they might be rather more disposed to be sympathetic to the company line. Of course, distentangling this from their stated genuine belief that they had seen the drugs help their patients was rather tricky. Symbiosis?

This sort of thing of course, is exactly why NICE and the Drugs and Therapeutics Bulletin are so important, the idea being that they deliver analysis of drug effectiveness and value for money while being wholly independent of the drug companies’ embrace.

Though he did once give out some lovely felt tips with a strawberry smell. And the latest pens for some statin or other are very good too. I also particularly liked a magnetic paper clip holder with a synthetic anus in the top. One side felt healthy, the other side was what a possibly cancerous prostate felt like. Marvellous.

His take on what he does? Well, doctors can’t possibly be expected to read up on every drug out there. They’re simply there to offer them the information they need for intelligent prescribing.

nhsblogdoc.blogspot.com/2006/10/feeding-at-trough.html

on the subject of advertisements aimed at doctors. He has plenty of things to say about the practise of ads to doctors. They may be better educated than the general public of course, but the fact is that, just like advertising to the general public, the pharmaceutical companies would not spend money on ads to doctors if it did not change behaviour.

But I agree, over-prescribing is a more natural explaination.

On doctors investing in drug companies: it shouldn’t matter too much under normal circumstances. A doctor who invests heavily ($100,000) in a billion-dollar company (i.e. small for a drug company) owns a ten-thousandth of them. If they send $100,000 in extra business to that company through over-prescribing, their share of that is just $10 (which is further reduced by the cost of producing the drugs.) (The argument changes if it is a million-dollar company instead.) Kick-backs in the guise of drug-rep gifts would be a much more efficient method of corrupting doctors.

one thing i find that is quite corrosive is when drug reps turn parts of teams against each other. a fairly common situation is where the doctors don’t want anything to do with drug reps, and neither does anyone really, but the drug reps were paying for the whole team to go out for dinner together fairly regularly (low nhs wages for all staff remember) and the whole situation causes resentment.

i was shocked when I went to the states for the first time on business and spent some time in my hotel room flicking channels. With the help of the TV I could easily mistake my jet lag for several other serious conditions.

At least the Rozerem web site has quite good coverage of achieving healthy sleep through routine and avoiding stimulants at night. The sleep assessment tool reported my sleeping habits as normal and made no suggestion of using the product based on the results.

It is unfortunate when the doctor is seen simply as the key master to the medicine cabinet.
What I find particularly distressing about drug adverts is that they list symptoms. I would be far happier if they stated “if your doctor has diagnosed you with X, then product A is the best treatment because…”.

There is not enough information about the normal thresholds of human emotion and physiology. “tense nervous headaches? That’ll be the JDs and cokes last night”. Or “Feeling depressed? Try getting up early and going for a run for a month. Then we can talk about the key to the medicine cabinet”.
Or “it’s good to grieve”

Kim – New scientist (31st march) has an Editorial on the rising demand for untested cancer drugs. It points out the correlation between demand and media exposure under the umbrella of “cancer” rather than trail results for specific cancers.

I followed the link from “postgraduate bullshit” in the original post. It leads to something called AdWatch that criticises a GSK ad.

The AdWatch article contains a statement that there was “a statistically significantly higher risk of cardiac failure” in the active treatment group. Supposing the sample size to be 2180 per group, I get a p-value of approximately 0.01. That’s hardly strong evidence for such a sample size. Moreover, we are not told how many tests were performed in order to find a p-value

www.prwatch.org/node/5899

- another term I think is relevant is the creation of ‘astroturf’ coalitions or ‘astroturf marketing’. That is the creation of ‘fake grassroots’ organisations (geddit) and media campaigns to promote a product or point of view:

www.prwatch.org/prwissues/2003Q1/astroturf.html

Bootboy: What’s this unknown third factor? Flying Spaghetti Monster? Aliens? Shakras? To me anyway, it’s blindlingly obvious that it’s a causative relationship

i gave some suggestions about the unknown third factor above already. to expand (because it’s an intersting question):

doctors who are already dodgy might be more likely to see drug reps more, whereas whiter than white evidence based medicine advocates might refuse to ever see them. those doctors might have been a bit duff before they even saw the drug rep.

single handed GPs (although there are fewer and fewer of them) are sometimes a bit professionally isolated and lonely, and might see drug reps just for a bit of pastoral care.

overstretched doctors who dont have time to do proper reading might see drug reps as a surrogate for proper continuing education.

there are lots of possible confounding variables, and it would be extremely unwise to dismiss them just because – and i share your prejudice – we don’t like the look of drug reps much.

the real answer to what causes the association will probably be: a bit of everything.