by Jerome Burne
What exactly counts as evidence in medicine? Everyone is obviously in favour establishing a scientific basis for treatments and against drug companies hiding unfavourable trial results – see this recent report on HealthinsightUK. For self-styled sceptics who gleefully dismiss anyone practising or supporting non drug approaches to health and healing as quacks that’s about all there is to it.
A common sceptic claim runs something like this: “Medicine tries to follow the scientific method…SCAM (ho ho) does not follow the scientific method. It holds to the old folk ways of anecdote, tradition, “sympathetic magic” and so on.” It’s a view which conjures up a sort of idealised medical UKIP world where objective science and evidence rules with an impartial empirical hand.
The biomedical lottery
Unfortunately for the rest of us today’s medical research is not driven by an abstract search for objective truth but by the need to produce a product. Unless there is a drug at the end of promising new line of research it’s unlikely to attract funding from the biggest payers – the drug companies. Instead it will be parked with the comment that it needs more research, the funding for which will be virtually impossible to obtain, however effective it is.
So drug research rather than being a scientific project aimed at preventing or curing disease, uses science as the engine for a giant biomedical lottery that can earn the winners billions, although the tickets are paid for by patients.
Several recent posts by HealthInsightUK contributors illustrate the way that multinational pragmatism rather than unbiased empiricism rules.
For instance, a basic scientific principle is that you abandon a hypothesis when its not supported trials and tests. The failure of the companies and doctors to do this in the case of blockbuster drugs such as statins – used to cut your level of cholesterol to lower your risk of heart disease – has long been tracked by Dr Malcolm Kendrick.
“I have seen evidence that directly refutes this hypothesis again and again and again and… indeed… again. I can produce far more evidence contradicting it, than supporting it,” he posted on his blog earlier this month. “Yet still it stands, untouched (and indeed has mutated into something hugely complicated).
The mutating hypothesis
“Now we have ‘good’ cholesterol and ‘bad’ cholesterol, and ‘light and fluffy’ bad cholesterol and ‘small and dense’ bad cholesterol (which really should be called ‘evil’ cholesterol, I suppose). We have the ratio of good to bad cholesterol, apob-100 levels, particle numbers, sub-fractions of good cholesterol, dyslipidaemia, LDL particle size, or number, or… the list goes on and on.
“I have come to realise that this constant creation of new types of cholesterol, and sub-fractions, and ratios, is all part of the game that is played to protect the cholesterol hypothesis from refutation. How can you refute a hypothesis that can change into any shape it likes? Answer, you can’t.” But most doctors still maintain that all statin use us is well supported by science.
But if medical treatments lived and died according to the quality of their evidence, a drug such as the anti-depressant Cymbalta (duloxetine) would not be the number four on the drug best seller list in the USA.
It’s licenced for depression, anxiety, fibromyalgia, neuropathic pain and incontinence and it may soon get a licence for arthritis and back pain. “This is so-called indication creep,” writes BMJ columnist Dr Des Spence in the current issue. “(It) broadens pharmaceutical companies’ potential for profit, which is certainly good business if not good medicine.”
Not clinically detectable or relevant
This is because chronic pain is a valuable market but one where the bar for evidence of effectiveness is set pretty low, as Dr Spence explains: “Pain research uses myriad soft rating scores that patients report themselves. These are converted into numbers to enable the production of statistically significant outcomes—even if they are not clinically detectable or relevant.”
So duloxetine will have jumped through all the scientific hoops yet, as Spence makes clear, this drug is unlikely to be a valuable addition to a clinician’s armoury. It is only marginally more effective than a placebo (all pain treatments have a strong placebo response), scoring at about a single point more effective on a 1-10 rating scale.
It is also a drug that patients will be taking for years yet those in the trials were only followed up for a few months. “Where are the hard objective data of long term benefit?” he asks.
Not only is the evidence not hard, the balance between benefit and risk is remarkably close. For each person who benefits, according to one of the trials, six or seven people need to take it. However for every 8 people who are treated, one will suffer side effects.
This looks like bad medicine
These include being hard to come off; stories of people’s problems with withdrawal choke the internet. “The rapid rise of any psychoactive drug for multiple medically unexplained symptoms surely looks like bad medicine,” Dr Spence concludes. And yet it is clearly a winner in the biomedical lottery.
So medicine’s scorecard for “trying” to follow a scientific approach is hardly impressive. But what about the “old folk ways” that CAM supposedly clings to? Dieticians are usually held up as being on the scientific side healthy eating while clinical nutritionists are lumped with the other CAM unreliables.
However in a post just before Christmas Dr John Briffa challenged all-too familiar advice on carbohydrates that has recently been reinforced on the NHS Choices website. This restates the long running claim that carbohydrates should make up the largest portion of a healthy balanced diet and promotes bread, pasta, potatoes and rice as foods that can help you lose weight, providing you keep fat intake low.
Dr. Briffa sweeps aside these claims: “There is absolutely no recognition in this article … of the influence of hormones such as insulin and leptin on fat storage, and the differing effects of the major macronutrients (e.g. fat and carbohydrate) … In fact, there is abundant evidence to show that low-carb diets generally satisfy far more effectively than high-carb one.”
Nutritional Flat Earth Society
He concludes with a flourish: “To my mind, dietician Sian Porter and the NHS Choices website have done an appalling job of communicating the facts and summarising the evidence. For me, the article reads like a missive from the nutritional flat-Earth society.”
Could “folklore” clinical nutrition, which has long abandoned such dietician’s homilies as the low fat diet and that calories in should equal calories out be the ones practising in the most scientific way?
The logical flaw in attempting draw a line between scientific medicine (good) and CAM (bad) is in assuming that what is scientific is automatically valuable. Science is value free. It is a tool, a technique, a way of gathering information which can be put to good or bad ends. It can result in both bombs and bariatric surgery.
The crisis facing medicine at the moment is how to deal with the epidemic of metabolic disorders such as obesity, diabetes, Alzheimer’s and cancer. What links them all is lifestyle, so a sensible way to slow or even reverse the rise would be to put some of the resources devoted to winning the biomedical lottery into equally science-based public health programs to develop the most effective forms of exercise or new ways of tailoring diets.
In practice funding for most clinical trials comes from drug companies who are understandably not remotely interested, which leaves drugs holding all the science cards, however faulty or fraudulent. But there is nothing inevitable about this arrangement; it’s ultimately a political decision. Right now the hunt is on for new pills that replicate the benefits of lifestyle change.
Press releases trumpet the discovery of genes that get turned on or off during exercise or when cutting back on calories. Doing the same things with drugs, they say, might make you live longer, as well as treating obesity, diabetes, neurodegenerative disorders and cancer. The very limited success of targeting biochemical switches to lose weight suggests this approach is fraught with problems and prone to nasty side effects.
But the research is done under the umbrella of science and prizes from this lottery will be huge. Meanwhile the number of people with metabolic disorders will continue to rise.