The term “informed consent” implies that patients need to be aware of any possible benefits, risks and complications before agreeing to commence with a treatment. Most patients don’t have a medical or scientific background, so it is not practical for them to be fully informed of every aspect of the treatment; however, with respect to preventative treatments, it is important that patients are aware of just how likely they are to benefit.
Preventative medication is available for a wide range of conditions, most notably cardiovascular diseases, which may be prevented with beta-blockers, statins, aspirin, ACE inhibitors, diuretics and so on.
Let’s take statins – or ‘HMG CoA reductase inhibitors’ – which lower blood cholesterol levels by inhibiting cholesterol production in the liver and increasing the liver’s ability to remove cholesterol from the blood:
Statins are widely prescribed*1. Presently, NICE advises statins be prescribed for patients either with pre-existing cardiovascular disease (CVD), or with a >20% estimated 10 year risk of developing CVD [3]. Their guidelines also state that:
“The decision whether to initiate statin therapy should be made after an informed discussion between the responsible clinician and the individual about the risks and benefits of statin treatment, and taking into account additional factors such as comorbidities and life expectancy.”
This is where I get uneasy. Yes, statins have been show to be beneficial using a number of different measures, but how the extent of their benefit is presented to the patient (if at all) is what concerns me.
To keep things simple, I’m going to use the statistics presented in the NICE guidelines in the document entitled “Statins for the prevention of cardiovascular events”. (They rather conveniently conducted a meta-analysis of all studies which had data in a “usable form”.) [3] The data shows that there is a significant reduction in deaths from cardiovascular disease in patients taking statins. The relative risk given was 0.79, meaning that the death rate from cardiovascular disease is 21% lower in the statin-treated groups than the control groups.*2
Let’s say then, that a patient presents to a physician with an estimated 20% risk of having a cardiovascular-related death in the next 10 years, and that physician believes it is in the best interests of the patient to prescribe statins. Those statins will reduce the risk of cardiovascular death by 21% (the relative risk reduction), to 15.8%.
This means that there is only a 4.2% chance that the patient will benefit from the drug if he takes it for 10 years (the absolute risk reduction) – 80% of the time he wouldn’t have died of cardiovascular disease anyway, and 15.8% of the time, he will die of cardiovascular disease regardless. That means he will take 3,650 tablets, with almost a 96% chance they will have no effect.
For many patients, the side effects and inconvenience associated with taking a regular medication may not outweigh the 4.2% chance that they will benefit – statins are not free of adverse effects and can be expensive. Yet many patients (and indeed doctors) may not be aware of just how small the chance of benefit is: perhaps if they did know, they’d be less likely to adhere to their medication (if they agreed to commence treatment at all).
Consider each of the following three statements:
● “This drug will reduce your risk of a heart attack by 21%.”
● “There is a 96% chance you will not benefit from taking this drug.”
● If 24 people take this drug, only one will benefit from the treatment
All paint very different pictures, but it is of great importance that all of these (and arguably others) be presented to the patient if they are to make a fully informed decision on a particular treatment. The third statement is based on the Number Needed to Treat, the inverse of absolute risk reduction. It is a figure which describes the number of patients who need to be treated to prevent one additional bad outcome, in this case a cardiovascular death. This is a particularly useful measure when discussing risks with patients, since it does not require an understanding of percentages to effectively convey its meaning.
However, while patients have a right to be fully aware about the possible benefits of any treatment they are on, we cannot ignore the economic argument – while a particular preventative treatment may only have a very small chance of benefiting the individual taking it, on a population-wide scale where large numbers of people are prescribed the drug, this may lead to a significant reduction in morbidity and mortality. In terms of life-years and the cost required to save them, preventative treatments are much more cost effective than so-called “rescue treatments” such as surgery or dialysis. This fact is of particular importance in healthcare systems with finite resources such as the NHS.
For medicine as a whole, it makes sense for people to take long-term preventive medicines that have benefit based in evidence. The overall reduction in mortality is cost effective for healthcare systems, much more so than waiting until a patient needs more expensive interventions such as surgery. However, at all times, doctors must respect the autonomy of a patient and should present the more realistic absolute risk reductions or numbers needed to treat when informing the patient about the efficacy of any treatments, even if this means fewer patients agreeing to take such medications.
~*~
References:
[1] http://www.nice.org.uk/media/207/AA/ImplUptakeReportStatins.pdf
[2] http://www.nao.org.uk/idoc.ashx?docId=ED722C6D-06DB-4AD0-958B-5987AF031175&version=-1
[3] http://www.nice.org.uk/nicemedia/live/11564/33151/33151.pdf
[4] http://www.annals.org/content/153/5/307.abstract
[5] Tony Hope, Medical Ethics: A Very Short Introduction, OUP, 2004
~*~
*1 In the 12 months to March 2007, 41.03 million statin items were prescribed and dispensed in the UK at a cost of £550 million [1], about 0.7% of the NHS’s total budget for the same year.[2] NICE estimates the number of patients taking statins regularly to be around 4.2 million (12 months to March 2007). Another report found 13% of those over the age of 30 had received at least one prescription of statins in the 12 months leading up to October 2006.
*2 I ought to point out here that all-cause death rates are also of crucial importance, since they take into account any side effects of the drug which may also increase mortality – there’s no point having a drug which reduces fatal MIs if it going to increase deaths from other causes. Statins have been shown to reduce the risk of disease from any cause by 17%.
One of the more instructive analysis of a statin trial was written by Merrill Goozner concerning the Jupiter trial utilizing crestor for individuals with increased CRP. Abstracted in this article is the folloiwing information
http://gooznews.com/?p=1291
“…The group on rosuvastatin developed diabetes at a higher rate than the group given a placebo, 3.0 percent versus 2.4 percent, an increase of six-tenths of a percentage point.
Keep the size of that percentage in mind as I now turn to the actual benefits of giving the statin for elevated CRP. While the overall rate of cardiovascular incidents fell from 2.8 percent to 1.6 percent by giving the statin, the number of so-called hard events — heart attacks and strokes, including those that were fatal — fell from 1.7 percent in the placebo group to 0.9 percent in the statin group, a drop of eight-tenths of a percentage point.
In other words, for every person who didn’t get a serious cardiovascular event, three-quarters of a person got diabetes.
… look at the benefits another way — in terms of the number of people who need to be treated to avoid a serious event. In this trial, 120 patients had to be treated for 1.9 years to prevent one serious cardiac event. Remember what rosuvastatin costs? $1,250 a year. That’s $285,000 per event prevented just for the statin pills. The physician visits, CRP tests and lab work add additional thousands more. Can you imagine how many heart attacks and strokes could be prevented if that money were targeted at people who are truly at risk of heart disease (the obese, smokers, hypertensives, diabetics) to help them modify their lifestyles and get treatment for their underlying conditions?
There’s one other curious element in the trial data. In table 4, Ridker and his fellow authors report that the number of “serious adverse events” in both arms of the trial was almost exactly equal: 15.2 percent in the statin arm versus 15.5 percent in the placebo arm. Presumably, all cardiovascular events (2.8 percent and 1.6 percent, respectively) were included in this total.
On the one hand, I’m not surprised that one in seven trial participants suffered a serious health event during the two years of this trial. The median age of this predominantly overweight group was 66, with some as old as 90.
But what were those other serious events? … study is silent on this point… the raw number suggests that at the end of the day, both of these groups fared almost exactly the same. In other words, giving a statin to people with elevated CRP did nothing to improve this population’s overall health.
So there you have it. A possibly unethical trial with marginal results gets trumpeted in the media as showing “wide benefit” (New York Times)…, this off-label use of statins will quickly find its way into clinical practice guidelines and drug compendia. Within a few years, health care payers will be forking over billions more dollars to the statin drug makers in the name of preventing heart disease…”
Anyone interested in the “real numbers” (ie absolute risk and number needed to treat) access Mr. Goozner’s other articles concerning statin use.
Just read that Goozner article – very interesting. As I mentioned in my post, the all-cause death rates in the treatment and placebo groups need to be compared to take into account the side effects of any intervention. e.g. Calcium supplements may reduce the risk of bone fracture, but they are also associated with an increased risk of myocardial infarction (http://www.bmj.com/content/341/bmj.c3691.abstract).
We also shouldn’t forget the burden of the less serious side-effects of treatment. These need to be considered when making judgements on treatments. Not only can they affect quality of life, but a drug with unpleasant side effects isn’t going to show very good patient adherence. It doesn’t matter how good a drug is at reducing the risk of heart attacks, if a patient won’t take it because of side effects, it isn’t going to do that patient very much good.
Nice post Fi (did I get the name right?) – an informed and sophisticated treatment of the tricky issue of preventative medicine. I was struck by the contrast between your analysis and that presented by a homeopath who was advising a patient who had asked about options for malaria prevention prior to a trip to the tropics. This was part of a BBC Newsnight investigation last night which you can see here:
http://www.youtube.com/watch?v=qgHRWB6-k-Q
The relevant bit starts at 1:56. Suffice to say, she does not give a very sophisticated analysis!
Thanks for that. I’d not seen that report (BBC iPlayer refuses to work outside the UK).
And yeah, exactly. “Conventional medicine may have a 70% chance of protection”. What does that mean? Nothing, really. I can only think of the following possible explanations:
1) It gives you a 70% of not getting malaria. So there’s a 30% chance of you getting malaria. Seems unlikely…
2) There’s a 70% chance of it having an effect. (ARR). Also unlikely, given that it would require the risk of getting malaria on your trip without treatment to be at least 70%.
3) It will reduce your risk of getting malaria by 70%. (RRR). Somewhat more likely.
It’s a complete non-statement. It means nothing. Frankly, it’s bullshit. Also, it is rather worrying that a “medical” practitioner is “plucking these figures out of thin air”.
We usually teach the medical students that last one using the example of hypertension medication. It is a useful example because of course the diseases is largely symptomless, but all the drugs have noticeable (though different) unwanted effects. So it is also an example of where “risk communication” would be very important.
That’s a key problem with adherence to any long term medication that doesn’t appear (to the patient at least) to affect the subjective symptoms. If a patient cannot see the benefit that a drug may have to them, but only the unpleasant side effects it causes, their cost-benefit analysis is going be skewed in favour of not complying with treatment.
It should be the duty of the doctor to explain about treatments – why they are being prescribed, what they should do to the body, how likely they are to work etc. – so that patients know why they should persevere with treatment, rather than just blindly following doctor’s orders.
Absolutely on both counts, Fi.
I think the more complex /subtle problem in practise is that it depends on both the doctor’s skill at explaining risk, and the patient’s ability to take in, comprehend and process the info – and also on the doctor’s ability to gauge if the patient has done so.
To give a personal example, my dad (a retired biophysics Prof) has been on a kind of revolving carousel of antihypertensives since his early 50s, so for the better part of three decades (and on statins.too) He never misses a dose, even with a fiddly multi-drug regime. But this reflects partly his trust in his doctor, and partly – I would say – that a statistically increased risk of nasty things has obvious power / reality to someone whose working life was about data and numbers.
My wife, who is an occupational health doctor, deals with a rather different sort of patient group. I suspect that for a lot of her patients, telling people the stats for CHD risk will have less power than telling them:
Oops – that was meant to be a reply to Fi’s reply to mml1946. It seems that if you forget to fill in the CAPTCHA and have to use the browser back button, WP doesn’t remember that you were replying to a comment.
This is a very interesting post. I deal with “informed consent” as far as obtaining and keeping biological samples for research purposes. Here, the consent is for a defined, one-time use: and then the patient has consented to the use of their materials in a fairly narrowly defined framework that still allows for new technological ways of analyzing the material, because there are certain criteria of anonymization and prohibition of re-distribution. And in theory they can always ask for it back if they’re nervous and force us to prove we’ve either given it back or destroyed it.
In reality, though, one sees behind the scenes how anonymized samples become pathognomic with the disease being studied, detached from the patient and the memory that the patient has a say over those samples. It’s not a real problem, I think, but it can lead to abuses.
Engaging a patient’s truly enlightened consent, especially on a continuous basis, does in fact imply that both the patient and the doctor understand the long-term implications. Which can not always be the case on either side, from experience.