Hard cases make bad medicine

I wrote a couple of weeks ago about the Ebola outbreak and the ethics of using drugs untested on humans to combat Ebola. Understandably, the pressure to use untested drugs is always at its highest in an outbreak of disease, as we saw during the swine flu epidemic of 2009.

In that case, governments across the world spent millions on stockpiling and prescribing Tamiflu, an anti-viral drug that had previously been licensed for treatment of flu. Thousands of people took the drug. Tamiflu wasn’t untested – there have been many trials. So it’s all the more shocking that the doctors expected to prescribe the drug and the patients expected to take it don’t actually know how well it works.

I’ve written before about the AllTrials campaign, and the struggle of the Cochrane Collaboration to get drug manufacturer Roche to release full data from their trials of Tamiflu. Even the drug regulators seem not to have had access to all the data. For anyone coming new to the issue, AllTrials have issued a video, presented by Dr Ben Goldacre, which sets out the situation. As he says, Roche have done nothing illegal in withholding the information. So do we need a change in the law, just to ensure researchers, doctors and patients can be sure the drugs they are taking work?

If you haven’t already done so, please do sign up to the AllTrials initiative and support the campaign.

The ethics of Ebola

Protective clothing, 17th-century style

Protective clothing, 17th-century style

On Tuesday I phoned in for the WHO press conference about the ethics of allowing experimental treatments to be used in the current, out-of-control outbreak of Ebola virus in West Africa. (My report for the BMJ is here.)

The situation feels like the plot of a movie. A deadly, infectious disease with no known cure. Healthcare workers, heroic in full-body personal protection suits, struggle to contain the outbreak. And somewhere, in a laboratory, a lone scientist comes up with the formula that might – just might – save the world.

In this case, the WHO admits that usual containment methods have not worked as well as in previous outbreaks. There are a handful of compounds under investigation that show early promise. But most of them have not been used in humans, only in primates. And the compounds have been produced only in tiny amounts, enough for early stage research. We have a few doses of untested new drugs that work in monkeys.

Should they be tried on patients? And if so, which patients? The doctors and healthcare workers risking their lives to save others might seem to have a moral right to priority treatment. But what of the patients’ families, who have often been informal carers? Should children be prioritised over the elderly? It probably didn’t help that the first people reported to have been treated were not Africans, but American missionaries who had been flown back to the US.

Setting aside the morals of distribution, let’s consider why we have drug trials. We need to know if the medicine is safe and effective – whether the benefits outweigh the harms. In an epidemic situation, with a disease of such virulence, this might seem over-cautious. What harm could it do? People are going to die anyway without treatment, right?

Not right. On Monday, as the ethicists met to discuss the situation, there had been 1848 probable or confirmed cases of Ebola, and 1013 deaths. Some people get better. While those figures don’t represent a reliable ratio, doctors have described mortality in the current outbreak to be around 45%. High, but not universal.

What if one of the new treatments unexpectedly made things worse, bringing the mortality rate up to 100%? Or caused susceptibility to another, more widespread killer, such as malaria? Then there’s the potential effect on the virus itself. We know from antibiotic resistance that microbes can adapt to drug treatments. What if one of the new vaccines or treatments actually increased the virulence of the virus, or made it easier to pass on?

These are worst case scenarios. Yet it would not be the first time a new treatment was adopted prematurely, only to make things a whole lot worse. However urgent the situation, we need clinical trials. It might not fit the movie script, but data collection and analysis are the right response to the epidemic. Statisticians should be heroes too.

When prevention is not better than cure

scales of justicePrevention is always better than cure. It sounds like a no-brainer, until you realise disease prevention can contradict that other medical dictum: first do no harm.

Two studies this week demonstrate this principle. Firstly, results of a long-term trial of prostate cancer screening by PSA (prostate specific antigen) blood test. I summarised the study for BMJ news here and you can see the study itself here.

PSA testing saves some lives by diagnosing prostate cancer early enough to treat (about one life for every 780 men invited for screening in this study). The downside is that it over-diagnoses cancers, so many men have treatment for harmless tumours that wouldn’t have grown or caused symptoms in their lifetime. The side-effects of prostate cancer treatment can include sexual dysfunction and incontinence, so the potential for harm is clear.

What interested me was that the authors of the study accept the problem. Despite showing a clear benefit for PSA testing, they don’t call for universal adoption, but say that the medical profession must develop better ways to tell the difference between indolent, low-risk cancers and the aggressive types that kill. In the meantime, they say patients should be given enough information to decide for themselves whether to have a PSA test. It’s refreshing to see researchers giving patients credit for enough intelligence to make their own decisions. (Here’s a link to a patient decision aid on PSA testing that I helped to develop for the NHS).

The second study, which I reported for BMJ’s patient website Best Health here, looked at the effects of taking daily aspirin. The benefits are a reduction in the chances of getting and dying from some types of cancer (mainly colorectal), and of having heart attacks or strokes. The main risk of aspirin is bleeding in the gut or stomach, and stomach ulcers. The researchers crunched numbers from the latest studies on the effects of long-term aspirin use, and concluded that it helps more people than it harms, among those aged 50 to 65. They stopped short of making a recommendation about whether all over-50s should start taking aspirin. You can see the study here.

For the average Joe and Joanna, the question is all about risk. You feel like a healthy person. Do you take a screening test or a drug that might help you live longer and avoid disease in the future, but with a small chance of causing you harm in the here and now? Or do you decide to live for today and worry about the future when it happens?

With preventive medicine making advances all the time, it’s a choice we’ll all have to face more frequently in future.

E-cigarettes: friend or foe?

Anticigarette leagueI’ve spent the past week immersed in the literature about e-cigarettes. The debate is amazingly polarised, from those who fear e-cigarettes will perpetuate the tobacco epidemic to those who think they could save millions of lives.

There is no doubt that smoking tobacco is hugely harmful and e-cigarettes are considerably less toxic to human health. E-cigarettes also seem to be more attractive to smokers than conventional smoking cessation aids such as nicotine replacement therapy patches and gum.

Yet the industry is at present unregulated, meaning a wide variation and uncertainty about manufacturing standards, the amount and strength of nicotine in each e-cigarette capsule and the flavourings used to make them more palatable. Different regulatory authorities are taking different approaches to the problem, meaning that what’s regulated as a tobacco product in one country may be considered a medical device in another, or a food product in a third.

In addition, there’s fear that e-cigarettes will ‘re-normalise’ smoking. I’m sure I wasn’t alone to double-take the first time I saw an e-cigarette advert on the side of a London bus. It brought back memories of the days when cigarettes were freely advertised and the upper desks of buses were a fog of tobacco smoke. In the interim, smoking has become anti-social. Will e-cigarettes make smoking seem glamorous again?

But the key unanswered question is whether they actually help people stop smoking tobacco. Perhaps surprisingly, there has only been one randomised controlled trial in smokers wishing to stop smoking, which compared e-cigarettes both with and without nicotine to NRT patches. The study was inconclusive, with disappointing tobacco quit rates for all three treatment groups.

E-cigarettes are perceived to be a less harmful alternative to smoking, which can help wean people off tobacco. Yet the research base to recommend them simply isn’t there. People can’t make informed decisions about how to stop smoking until big, good-quality studies are complete.

Image of vintage anti-tobacco campaign materials: from Wellcome Images with thanks.

Why smoking is (still) the worst thing you can do for your health

When I saw the headlines claiming that eating a high protein diet was ‘as bad for you as smoking’, I snorted and got on with my breakfast. I know that a headline like that is likely to be based on spin, over-inflation of some observational study results and a slow news day. There were more important things happening in the world.

Then I saw my Twitter and Facebook feeds. People were genuinely concerned, confused and upset by the news stories. People who were trying hard to reach a healthy weight, follow a decent diet and be as healthy as possible. People who have no reason to doubt what they’re told by researchers or health journalists.

I won’t do an analysis here, because NHS Choices’ Behind the Headlines has got it covered in their piece here. It’s a prime example of asking about just one factor (in this case the proportion of protein in people’s diet) and doing a crude analysis of the mortality statistics. It tells you little about the other habits of the people in the study. Did their protein come from lean chicken, or from fried burgers? Did they have a healthy green salad alongside, or a pile of chips? Did they pick up their meal at the drive-through takeaway, or cook it themselves after cycling home from the gym? No idea.

Worse, the results were ‘cherry-picked’. So overall, for all age groups, eating moderate or high amounts of protein was not linked to higher death rates. For people over 65, eating a high protein diet was linked to a lower likelihood of dying. For only one group – those aged 50 to 65, it was linked to a higher likelihood of dying. When the results are all over the place like this, I’m immediately suspicious.

Worst of all, the comparison with smoking – which was made in the press release – was not based on anything in the study. It was a crude comparison between the link to increased death rates in their study, and in other studies looking at smoking. It’s not a valid comparison, because the people in the studies are not the same. And, as NHS Choices point out, ‘We need to eat protein. We do not need to smoke’.

Many, many studies have repeatedly found that smoking causes disease and early death. This one study about high protein diets does not trump those years of patient research. If you do just one thing for your health on No Smoking Day, make the decision to stop smoking. Want to know what will help you stop? See BMJ Best Health’s Stop Smoking topic for more details.


Get every new post delivered to your Inbox.

Join 208 other followers