Are you worried about how decisions involving public money are made? You should be.
Last week, the National Audit Office disclosed that the Department of Health spent £424 million on the anti-flu drug Tamiflu. The money was used to purchase nearly 40 million units of the medicine between 2006-7 and 2012-13.
It also emerged that some £74 million of this stockpiled Tamiflu had to be destroyed because of poor record keeping.
Stockpiling drugs is standard practice in dozens of countries and such huge quantities of the drug are necessary to make sure the NHS is ready to face an influenza pandemic. In 2006, at the height of the bird flu scare, the Department of Health stockpiled enough Tamiflu, made by Roche, for a quarter of the UK population. This grew to cover 80% by 2009, when swine flu threatened. We now have an outbreak of H7N9 in China.
But with such large sums spent on preventive medicine, shouldn’t we know more about how well the drug works?
According to the published evidence, and to notable health bodies such as the US Centres for Disease Control and Prevention (CDC) and the World Health Organisation (WHO), Tamiflu prevents complications caused by influenza (such as pneumonia) and in some cases can delay or interrupt the spread of the virus.
We are learning more. Researchers from Oxford University based in South-East Asia have shown that giving double doses of Tamiflu to patients admitted into hospital with severe influenza gives no clinical advantage over a standard dose. This has implications for the amount of Tamiflu governments stockpile.
But the Tamiflu story is a complicated one. In 2006, we included statements on the properties of Tamiflu in the annual Cochrane review into antivirals that has been running since the 1990s, and of which I am the lead author. The information was based on data from single studies and from a combination of 10 studies that had been published in prestigious health journals.
But when producing the 2009 review, a Japanese colleague pointed out that eight of the 10 studies had never actually been publicly published and all of them had been sponsored by Roche, the drug’s manufacturer.
‘Trust but verify’
As the Russian proverb that US president Ronald Reagan picked up goes, “trust but verify”.
Three and half years later we are still unable to verify the properties of Tamiflu. All the first authors of the published studies who were contacted by us or by the media were unable to answer questions on their drug trials. One admitted what we already knew: that someone else had written their trial for them and published on the Lancet.
All of the authors referred us to Roche and gave a variety of reasons for not having the data handy - from moving offices to the frank admission they hadn’t actually seen detailed data.
We then got some 16,000 pages of study data from the European Medicines Agency (EMA) to which we added some 1,200 pages of comments made by the US Food and Drug Administration (FDA) on Tamiflu. From this, we discovered that the US regulator did not accept the evidence put forward on the Tamiflu effects on complications and were also unconvinced that it could prevent the spread of influenza.
We also found that CDC ignored this and made claims about the properties of Tamiflu that went beyond what Roche were allowed to do by the FDA. We then asked Roche for the data as it now appeared clear that this all needed to be independently verified.
In January 2012 we published the fourth update of our review. The review is completely based on regulatory material. We ditched publications as a source of data because a reliance on this can be misleading; 60% of trial data remains unpublished and published studies show discrepancies with their regulatory equivalents.
What had been published showed some discrepancies with what the regulator - the FDA - said, and despite years of correspondence we still haven’t go answers from Roche.
It was pointless asking anyone else for the detailed data as it appeared no-one outside Roche had seen it. The WHO and CDC dismissed the findings of our review and refused to answer our questions.
However, the important point is that neither the WHO nor the CDC had seen the regulatory data we got from the EMA, and in the late 1990s/early 2000s some WHO policy documents had been written by key industry opinion leaders.
We should all see the data
Is all this unbelievable to you? Fortunately, the BMJ, a major biomedical journal, was as worried as we were about the issue of transparency and access to data from clinical trials. And all our correspondence is open and available to read. For the first time you can follow the whole story warts and all. Redactions were kept to the minimum, all the emails are there.
We are still busy updating the review with EMA data, some of which we didn’t have in time to include in the January 2012 version. We agree with the FDA that there is no convincing evidence that Tamiflu reduces complications or slows progression. News that there have been some cases of H7N9 bird flu virus that have shown resistance is worrying. But this is hardly the point anymore.
The Tamiflu story, abundantly documented in all its twists and turns, was a powerful reason behind the formation of a coalition of people who believe that the results of experiments on humans should not be secret. It is now also supported by GSK, another leading drugs company.
Take time to read the documentation and think about how your public money is spent. Unfortunately Tamiflu is not the only example of this kind.
Want to know more facts about the flu? Find out here