Yellow Card: parcelling up risk

This was the first posting on on this blog, all the earlier ones are articles. 

Following up on my feature in the Mail today  around Professor Gotzsche’s excellent new book about why it is a really bad idea to have a mammogram I’ve been picking up a lot of material around the issue of evidence and medicine and just how much of mainstream medicine is lacking  a good evidence base.

Don’t intend to plunge into debate over CAM but it is becoming very clear that there are serious problems with hiding of unfavourable evidence from drug trials and other medical interventions. Gotzsche’s book is a detailed account of his 10 year battle to get his detailed analysis of the failures of  the mammogram program taken seriously. Far from being a life saver he’s very convincing that they are very damaging on a population level.  The way his research has been ignored suggests a major lack of evidence for this 100 million pound program.

But he’s not the only one raising the lack of evidence issue. I’ve also been writing a piece on  another passionate and thorough book Pharmageddon by Dr David Healy that makes a far more wide ranging case for for the systematic lack of evidence for the safety and effectiveness of drugs in particular. That’s due out in next few weeks so I won’t go into detail now.

But one of his themes is the inadequacy of the Yellow Card system and the way it denies personal reports the status of proper scientific evidence. This was something I came across doing the Panorama program on the diabetes drug Avandia last year and which I also wrote about for the Mail

Yellow cards – only anecdotal

If you suffer from a side effect from a drug you and/or your doctor can make a Yellow Card report on it. Summaries of the reports say all over them that this doesn’t prove that the drug caused it, but they are supposed to be one of the sources of information that the UK drugs watchdog the MHRA uses to build up a picture of the safety of the drug.

The story with Avandia was whether it caused heart attacks – evidence that it did had been concealed by the drug company but discovered by an American cardiologist and made  public. However there was another diabetes drug called Actos that worked in much the same way as Avandia but didn’t seem to have the damaging effect on the heart. So what doctors and patients needed to know was was Actos safer and should you switch?

Through all this the  MHRA was saying that the benefits of Avandia outweighed the risks. So it seemed to me that was just the sort of question the  Yellow Card record could throw some light on. Very simply: which one had the more  reports of cardiovascular problems?

Anyone can access summaries of Yellow Card data at the MHRA website so I got records for the two drugs. Given that it was such a current issue and that patient’s lives were at stake I was amazed to find that Avandia showed up as having eight times as many reports of deaths linked to heart problems as Actos did – 25 vs. 3.   A similar pattern showed up with non-fatal side effects linked to the heart 275 compared with 75.

No alarm bells rang

Those might look small numbers but it is generally reckoned that somewhere between 1% and 10% of actual side effects are ever reported.And when the side-effect is one that is also one of the effects of the disease – diabetes certainly raises cardiovascular risk – the  reporting rate is likely to be lower still.

Did that set any alarm bells ringing at the European Medicine Agency, the official body that licensed Avandia for use in the UK? “The agency takes all post-marketing data into account, such as the Yellow Card data from the UK,” said a spokesperson. “But it is only part of all the data we rely on to update our guidelines.”

The European regulators advice wasn’t that helpful says Anthony Barnett endocrinologist and professor of medicine at the University of Birmingham. “It does warn against giving the drug to people with heart problems but anyone who has have diabetes for several years is very likely to have markers for heart disease such as raised plaque and narrowed arteries, so why give it at all? Actos reduces heart disease risk, it brings down triglycerides and LDL cholesterol while Avandia raises both of them.”

Heart attach riisk from the begining

My investigation was in July and the Avandia was withdrawn from the market in September just at the time the Panorama program went out.

But for anyone concerned about evidence and how well patients are protected is is not at all reassuring.  The two drugs has been on the market for a decade by then. There were questions about the possibility of a heart attack link with Avandia from the very beginning. The concealed data showing that Avandia seemed to raise the risk of heart attack was made public in 2007 and all that time the Yellow card data showing a hugely bigger risk from Avandia would have been accumulating in the Yellow Card database. And yet it prompted no action until 2010. Patient reports come far too low down in the “evidence” hierarchy. It is something Dr Healy is campaigning to change. More on that later

But the signals weren’t only coming in from patient’s reports they were coming in from another publicly accessible source – prescribing data. Every year precise details of how many prescriptions have been written are available on the web. So I checked them to see how doctors had been responding to the concerns over Avandia.

Doctors sharper than the watchdog

This showed that they had been far quicker off the mark than the MHRA. By 2009 (just two years after the study finding that Avandia raised the risk of heart disease by 40% prescriptions for the drug  had plunged to half what they had been in 2006. Meanwhile those of the rival drug Actos had soared by almost the same amount. Back in 2006 1.1 million prescriptions were written for Avandia in England and just over 450,000 for Actos. By 2009 the situation had completely reversed. Actos prescriptions were running at 1.2 million and Avandia had dropped  to just over 550.000.

So which group of patients were being treated according to evidence based medicine? Those whose doctors were following MHRA advice (benefit outweighs the risk) and giving them Avandia or those who followed the trial evidence and maybe, who knows, looked at the Yellow Card data and gave Actos? The simple division between real evidence based medicine and the evidence free complementary stuff that fuels all those attacks on CAM practitioners doesn’t look quite so clear cut.