While Public Health zealots trumpet their claim to the moral high ground over statins, it’s still worth asking: “are they right?”
Here’s a strange piece of news. This week, two of Britain’s top doctors (Knights of the Realm, no less) recommended in the strongest terms that tens of thousands of their hitherto well patients deliberately make themselves ill. They don’t know precisely who these victims of unnecessary sickness will be. But they know from their statistics that, if the millions of people targeted with prescriptions for statins decide to comply, the treatment will cause a certain percentage of them to develop muscle disease and possibly diabetes.
On the face of it, the exhortations of Sir Rory Collins and Sir Magdi Yacoub don’t seem very chivalrous at all. But, they and their supporters insist, this slightly bizarre enthusiasm for iatrogenic illness is grounded in reason not malice. They think that the ace in their hand is the fact that the overall amount of good that will come from mass medication outweighs the overall amount of harm. Their principle calculation is that, among the millions of statin poppers in Britain, you will find a small number of people who would have had a heart attack or stroke – and now they won’t (as well as the aforementioned unfortunates with the muscle pathology etc.). Some of the lucky ‘coffin dodgers’ (that’s not a technical public health term, you understand) could theoretically be the same people who are now suffering from aching and atrophying muscles. But probably not – because the vast majority of people taking the pills won’t either be saved from circulatory disease or suffer the other stuff. Essentially, they will only be taking the pills to make up the numbers – but they will never know that.
The main arguments this week have centred on the numbers and the balance between them. What is the true number (and percentage) of people who are ‘saved’? And what are the same stats about those that get so-called “side effects”? (a term which cleverly belittles these often painful and sometimes life-changing symptoms). It’s this technical argument about the magnitudes of harm and benefit that has led to some slightly unseemly statistical shroud-waving from Collins, who was widely reported as opining that the confusion caused by a recent BMJ article was killing people.
But I don’t think the exact figures are really the main argument here. The more time and effort that is expended on the technical argument, the more the attention of the general public is diverted from the medical, ethical and social issues. In spite of the self-righteous and vaguely totalitarian braying of the public health zealots, I for one would like to retain a space for debate and disagreement about the whole “lets treat the whole population” idea. So, in the quest for some slightly less toxic information than that found in the smoking ruins of the BMJ, I went to the respected “Numbers Needed to Treat” website which is known as NNT. There I found a sober reflection on some of the data, refreshingly free from the clamour and hyperbole so beloved by those medical knights. The figures relating to statins taken by people with no prior heart disease make fascinating reading. Here they are:-
Benefits (first expressed as % and then as “numbers needed to treat”)
- 98% saw no benefit
- 1.6% were helped by preventing a heart attack
- 0.4% were helped by preventing a stroke
- 1 in 60 were helped (preventing heart attack)
- 1 in 268 were helped (preventing stroke)
Harms (first expressed as % and then as “numbers needed to treat”)
- 2% were harmed by developing diabetes
- 10% were harmed by muscle damage
- 1 in 50 were harmed (develop diabetes)
- 1 in 10 were harmed (muscle damage)
Playing ‘Public Health Top Trumps’ – Image ‘Fear Factor 100’, William Warby, Flickr
I think this is beginning to look like a ghastly playground game of Public Health Top Trumps. Sir Magdi and Sir Rory say that their 2% cardio-vascular benefit beat my 12% muscle and endocrine pathology. I say “no it doesn’t – my muscle and endocrine beat your cardio-vascular”. They say that causing harm is acceptable as long as some good comes from it. I say “no it isn’t”. Causing muscle pain and diabetes in otherwise healthy people is disgraceful. And all the more so when no one knows who is going to win and who is going to lose. Under those circumstances it just becomes a macabre game of chance and a frankly bizarre way to organise the wellbeing of the population.
I also think that, if each individual patient had these numbers truthfully and clearly put in front of them many of them would say “no thanks”. Because they would see that the most likely outcome is the pointless taking of medicines, every day, for years (there will be no benefit, and no harm). They would also see that the second most likely outcome is pain and illness. And finally they would see that the least likely outcome is their personal avoidance of a heart attack or stroke – a fate they weren’t THAT likely to experience in the first place, to be honest.
“To be honest”. That simple phrase gets us to the nub of the issue really, doesn’t it?
Because when a real patient is in front of a real doctor they are usually told about risk reduction, not NNT. So they have it implied (not definitely stated of course, because that would be untruthful) that the pills are likely to benefit them personally. This is a kind of communication that I once called “worthy dishonesty” – a murky ethical area which later became the subject of an interesting article in the Journal of Medical Ethics.
4 Responses
Ewen Speed on May 23, 2014
Hi Charlie, thanks for the post. I’m wondering if there is any link between the relative treatment costs of acute episodes of stroke and heart attack, versus treatment costs for chronic, and often undiagnosed, diabetes and muscle wastage?
Catherine Will & Kate Weiner on May 27, 2014
Thanks for a really clear and thought-provoking piece on this latest statin spat Charlie. It’s great to debate the whole concept of prophylactic medicines for healthy people, and ask questions about whether people understand what it would mean for them to take a statin. We think however there are some other things worth considering which we thought we might add as postscripts to your piece:
Though you ask us to move away from the ‘technical’ argument about the ‘true numbers’ we would be uncomfortable about treating the Number Needed To Treat or indeed Number Needed To Harm as definitive fact. One of the interesting things about the area is how political the collection and interpretation of evidence has become, as acknowledged in the BMJ debacle. It may well be that the NNH actually underestimates the rates of so-called side effects because of the selection of trial participants and design of trials. But there are also important questions about the lack of evidence exploring why some people do benefit and (many) others do not (and why we’ve got stuck with the epidemiological focus on risk assessment as a rather clumsy way of deciding who is offered the drugs).
The second issue is the gap between who is offered statins and who takes them. As you know we’ve done some work on narratives of statin use and written a fair bit about resistance to medication and non-use as a result. We think fairly large numbers of people will be more or less explicit with doctors and refuse these drugs, and others will engage in quieter dissent by not cashing repeat prescriptions, taking drug holidays etc. While the BMJ debate is about the correct rate of side effects, and whether 17% is the appropriate figure, other studies suggest non-compliance (to use the medical term) may be as high as 50%. These factors may well reduce the incidence of the harms that worry you. There is also evidence that reasonably large numbers of GPs don’t particularly push statins with their reluctant patients.
This doesn’t make us unconcerned about the issue, but for us some of the ethics and politics needs playing out with reference to clinical practice, not policy. You say relatively little about clinical practice other than to suggest that GPs should (but aren’t) talking to their patients about NNT, and may instead rely on discussing absolutely risk reduction and other abstractions. We suspect from our research that they often don’t even do that, or are even cognisant of the population based rationales, and that prescribing in practice may be less about knowledge and more about clinical routines. In addition to bolstering efforts to make side effects visible and part of the debate nationally, we might therefore want to work on that in the clinic? The clinician’s duty should surely be to take any harms suffered by individuals seriously and act to reduce or reverse them quickly. One of the things that worries us from our data is that there seems little space for people to raise concerns about side-effects and pill-taking in conversation with their doctors. If we wish to question the ‘common good’ perspective then encouraging a shift in clinical conversations with individual patients might be just as important as entering a debate about the population goods and bads?
Tom Marshall on May 27, 2014
The numerically small probability of benefit applies equally to drugs to treat high blood pressure, but statins attract more debate. Any idea why?
I teach on a course about hypertension management for GPs. Some of the case studies include management of high blood pressure in women in their twenties. Perhaps one in 2000 such women will benefit from 10 years of treatment. But the “right” answer for the clinicians is to try drug x then drug y and so on. Calculating the probability that a patient will suffer from cardiovascular disease in the next ten years is a sort of optional extra in the management of high blood pressure yet without this we cannot even arrive at an estimate of reduced risk, never mind translate it into numbers of patients who will take treatment for no benefit. So communicating risks and benefits is marginalised from the real clinical agenda of normalising a physiological parameter: blood pressure.
Scott Robertson on Feb 21, 2016
now only if a law firm would take on these companies for those of us that have nearly died. it will be a year before I’m back to normal 🙁