This post is the second in a series examining the claims made in a recent essay that seeks, in part, to refute common criticisms of homeopathy (Milgrom, 2009). I have already examined the empty assertions about evidence for clinically useful specific effects. Now, I would like to move on to examine an attempted refutation of claims that, “Homeopathy is deadly”.
How deadly is homeopathy?
This is entirely irrelevant; no critical discourse that I have come across has made the claim that the remedies themselves are toxic*. As I pointed out in my last post: the problem is not in the pills, but in their uselessness; and the attitudes of some homeopaths. He then moves to the actual concerns of sensible critics:
“The claim arises over concerns that those taking homeopathic remedies might forgo ‘life-saving’ drugs. This is a false perception: many who come to homeopathy do so only after conventional treatments have failed.”
And this is not right either: those who are able to turn to homeopathy after conventional treatment has failed are not going to be suffering from life-threatening illnesses. They are generally people who are suffering from chronic complaints for which modern medicine has no good treatments (such as certain kinds of back pain, stress, medically unexplained fatigue, and modest viral illness – Goldacre, 2007). The danger, such as it is, lies in choosing homeopathy instead of proper medicine for serious illness. That many will be using homeopathy to treat illnesses that are not life-threatening doesn’t mean that all users (or practitioners) of homeopathy are as conservative.
Milgrom’s rejection of any suggestion that homeopathy can harm is disappointing. There are documented cases of people choosing homeopathy, or having it chosen for them, and dying as a direct result. The number of fatalities appears to be low, but denying that there is any problem at all is rash.
For instance, Gloria Thomas died at nine months of age, from sepsis, after her homeopath father ‘treated’ her eczema with homeopathy instead of seeking proper medical aid. A UK GP, Dr Marisa Viegas was eventually struck-off after a she advised a patient with idiopathic dilated cardiomyopathy to take homeopathic treatments instead of the drugs she needed. The patient died as a result.
An example of the dangers of the disregard that some homeopaths show for conventional medicine and evidence is seen in the untimely death of Russell Jenkins, a CAM practitioner. He took the advice of homeopath Susan Finn, who suggested that he treat an electrical burn with Manuka honey. As a result of this improper treatment, he died from gangrene. This is a different form of harm, but no less dangerous.
Add to this the documented incidents of UK homeopaths advocating homeopathy for malaria prophylaxis and the activities of homeopaths in developing countries who believe they can treat AIDS and malaria: there are real risks.
Again, I would not want to over-state the problem; but it is inappropriate for Milgrom to ignore it.
Other homeopathic harms
Not all the harms of homeopathy are directly deadly ones. A risk analysis that only focuses on extreme outcomes is too simplistic: there are real harms that don’t kill. For instance, Goldacre (2007) identified a range of other risks associated with homeopathy. These include medicalisation, “the reinforcement of counterproductive illness behaviours, and promotion of the idea that a pill is an appropriate response to a social problem, or a modest viral illness.”
Also, by knowingly prescribing placebos medical practitioners can undermine the notions of informed consent and patient autonomy.
As Milgrom’s essay shows, homeopaths are apt to denigrate conventional medicine. This attitude can also lead some homeopaths to undermine public-health campaigns, like those promoting vaccination.
Finally, as Milgrom shows, homeopaths have a tendency to misrepresent scientific evidence, undermining the public understanding of both science and medicine.
It is important to recognise that even placebo medicine has a range of risks associated with it.
What are homeopaths for?
Milgrom then moves on to flirt with the placebo effect. After making what we have seen is the unjustifiable assertion that there, “is evidence to support homeopathy is more than a placebo response.” He notes that, “homeopaths like other health practitioners, responsibly encourage expectation of positive outcomes”. This is fair enough. The evidence shows that any benefit that homeopaths deliver through the therapeutic encounter is due to expectation effects (Shang et al, 2005). If homeopaths were open about this then, perhaps, there could be a role for them in a clinical setting.
However, there is a sizable fly in the ointment. As Milgrom says, health practitioners, “responsibly encourage expectation of positive outcomes.” Proper medical practitioners can deliver specific effects through their interventions, along with non-specific expectation effects. They also have the advantage of diagnostic training and don’t disparage other medical disciplines. Given this, who needs homeopaths?
Neither does encouraging positive expectations reduce the risks involved with homeopaths pretending to treat malaria, AIDS or other dangerous diseases. This line of attack has little relevance to the matter at hand.
Wouldn’t that be NICE?
Similarly irrelevant are the author’s claims about Prozac: a particular pharmaceutical being either ineffective or unsafe doesn’t mean that homeopathy is either effective or safe. Anyway, Milgrom’s analysis is problematic in its own right:
“One of the world’s top-selling drugs, the anti-depressant Prozac, was recently shown to be no better than placebo [22]. Yet, with an effect size of only d ~ 0.3 (the National Institute for Health and Clinical Excellence – NICE – recommends d = 0.5 for clinical efficacy), there are no urgent calls for Prozac’s withdrawal through ‘lack of efficacy’.”
His reference [22] is to Kirsch et al. (2008) and it does not say what he claims it says. First, this paper looks at what evidence was available before Prozac was licensed, not the totality of the data. As Ben Goldacre has observed**:
“It is common for quacks and journalists to think that the moment of licensing is some kind of definitive “it works” stamp of approval. It’s not, it’s just the beginning of the story of a drugs’ evidence, usually.”
So this paper does not show what the best evidence is for the efficacy, or otherwise, of Prozac (fluoxetine) for the treatment of depression.
Milgrom has also confused the result for a specific drug, fluoxetine, with a pooled analysis of all the drug groups against their placebo groups (Table 2, Model 3a). The paper says that the drug group:
“[…] does not meet the three-point drug–placebo criterion for clinical significance used by NICE. Represented as the standardized mean difference, d, mean change for drug groups was 1.24 and that for placebo 0.92, both of extremely large magnitude according to conventional standards. Thus, the difference between improvement in the drug groups and improvement in the placebo groups was 0.32, which falls below the 0.50 standardized mean difference criterion that NICE suggested.”
So, this is where Milgrom’s “d ~ 0.3″ comes from. It does not relate specifically to fluoxetine, but rather to a pooled analysis for all the drugs covered in this review. In fact, the mean difference between the drug and placebo groups, “easily attained statistical significance.”
The paper does show that the drugs studied achieved both statistically and clinically significant improvements, compared to placebo, for the most severely depressed. As Figure 3 shows (below), they also exceeded the NICE criterion for these patients (the green bit).
Neither does the essay contain any mention of the weaknesses of this study, or the criticisms that have been levelled at it***.
Milgrom’s argument is further weakened by the fact that NICE have, since 2004, taken the position that:
“Antidepressants are not recommended for the initial treatment of mild depression, because the risk–benefit ratio is poor.”
So, for cases where the evidence does not support the use of drugs like Prozac, NICE recommends they are not used.
Here Milgrom overstates the scope of Kirsch et al. (2008) by implying that it is a definitive assessment of the efficacy of Prozac. He also turns the argument into a simple binary choice: either the drug works or it doesn’t. Reality is more complicated: the effectiveness of this (and other) drugs varies with the severity of the depression. Whilst they may not be justifiable treatments in some cases, they are in others.
By arguing that drugs which don’t meet the NICE criteria should be withdrawn, he is also setting the bar too high for homeopathy. It’s notable that he provides no “d” values for any single homeopathic treatment. Prozac may not be very useful for treating all but the most severe cases of depression, but there is no evidence that homeopathy can help at all. After reviewing the literature Pilkington et al (2005) concluded:
“Evidence for the effectiveness of homeopathy in depression is limited because of a lack of high-quality clinical trials.”
I am sure that it’s possible to argue that anti-depressants are over-prescribed and their benefits are at times over-stated. However, their limitations appear to be appreciated by the medical community and strategies are in place to align their use with the available evidence. There may be legitimate controversy here, but it is clear that drugs like fluoxetine have some benefit, compared to placebo: not the zero benefit Milgrom alleges. Furthermore, Milgrom appears to have misunderstood the paper he has cited: ascribing the results for a pooled analysis of number of drugs to a single drug. He also has missed the guidance from NICE to limit the use of anti-depressants based on their risk-benefit ratio.
This essay is meant to be making the case for homeopathy. In this context the discussion of Prozac is irrelevant. This section of the essay is also meant to be overturning the notion that homeopathy is dangerous; again, a flawed analysis of Kirsch et al. (2008) does not contribute to this objective.
Real medicine has risks … but are they this big?
Then again, neither does his next argument, which claims that:
“Those who denounce homeopathy as ‘deadly’ should consider conventional medicine’s safety record; something recently scrutinised by the UK’s House of Commons Public Accounts Committee [23]. Including fatalities, this committee found that in 2006 alone, at least 2.68 million people were harmed by conventional medical interventions; representing 4.5% of the UK population …”
Milgrom’s reference [23] is this report:
Leigh E: A safer place for patients: Learning to improve patient safety. 51st report of session 2005–06 report, together with formal minutes, oral, and written evidence. House of Commons papers 831, 2005–06, TSO (The Stationery Office). July 6, 2006.
It can be found here. The first thing that stands out is that it was published in early July 2006: so it’s obvious that it cannot provide data for “2006 alone”. This raises some suspicion about the rest of the claims.
As does the fact that I cannot find Milgrom’s figures in this report. What Leigh (2006) examines is how patients can be treated more safely. It looks at the incident reporting systems in the NHS, along with how it can better learn lessons when things go wrong. It does not provide any estimates of the number of people “harmed by conventional medical interventions”. It’s concerned with episodes of unintentional harm: medical accidents. On this specific topic it quotes a previous report which, “estimated that one in ten patients admitted to NHS hospitals are unintentionally harmed”. This is consistant with a recent report on patient safety from the House of Commons Health Committee (Barron, 2009).
So, if we take this rate of harm and apply it to 2006, how close do we get to Milgrom’s figures? For the year 2005/06 the NHS Hospital Episode Statistics count 12,678,628 admission episodes. If one in ten of these admissions resulted in harm, this implies around 1.3 million incidents of harm. The report cites nothing which would allow us to assess the number of people harmed, or include medical accidents that may occur to patients who were not admitted to hospital (under the care of a GP or treated as an out-patient, for example). This report only provides information on harm done to hospital in-patients; and a crude estimate at that.
Terry et al(2005) also points out that injuries due to falls ranks as one of the most common causes of these incidents. To place these events at the feet of conventional medicine, rather than the process of caring for the sick, would seem to be harsh.
Finally, it is, of course, not valid to relate this figure to the UK population as the data refers to incidents not individuals.
Either I’m missing something (always a possibility) or Milgrom’s figures must come from elsewhere. It might be that they are correct and that this is just a mis-citation. In any case, he should make clear where these numbers actually came from.
Ultimately this is another irrelevance. The rate of harm resulting from conventional medicine must be set against the benefits it delivers. We have seen that homeopathy’s magic pills and potions may offer no risk in themselves – because they don’t contain anything – but neither do they provide any benefit.
It must also be recognised that conventional hospitals often care for desperately sick people with complex conditions: under these circumstances mistakes are more likely. However, Milgrom’s simplistic analysis makes no attempt to do this, rendering it useless. Further, it has no relevance to concerns about the safety of the practise of homeopathy. It’s just a bit of tu quoque.
Ultimately, all healthcare providers should be striving to do less harm. The rate of medical accidents within the UK’s NHS – whilst comparable to that in other developed countries – is still too high.
However, safer medical practise should also include stopping homeopaths treating serious medical conditions and the abandonment of medical interventions that incur risk without benefit.
Critics refuted?
So has Milgrom managed to refute the charge that, “Homeopathy is deadly and those who practice it are at best purveyors of a placebo effect”? I don’t think so. This part of the essay is very weak.
It steadfastly ignores the documented incidents of harm caused by homeopaths treating serious diseases. His concept of homeopathic damage is simplistic, limited to deaths which are not acknowledged.
It may be that homeopaths, “responsibly encourage expectation of positive outcomes”. But that is all they have. Real medicine can offer this and more: effective treatments.
Going on the offence, Milgrom offers nothing more than two examples of ill-founded tu quoque. First, nasty critics call homeopathy a placebo, so he says Prozac is nothing more than a placebo. Of course, that’s not really true; Milgrom has mis-read the evidence, not acknowledged its limitations and ignored the effectiveness of anti-depressants under particular circumstances. He also invokes NICE criteria when discussing Prozac, even though it’s clear that no homeopathic intervention could meet this standard, unlike the drug he disparages.
Similarly, critics say homeopathy is dangerous, so Milgrom says conventional medicine is too. However, the source of Milgrom’s figures is obscure – they do not come from the report he cites. Neither are they relevant: the failings of conventional medicine don’t make homeopathy any better. And even with its failings conventional medicine is massively more successful than homeopathy can ever be.
All-in-all this is an empty attempt to justify an empty practise. The scholarship is slap-dash and the arguments are flawed. It is hard to imagine that this essay was subject to any meaningful review.
Next, I’ll look at Milgrom’s attempted refutation of “The claim that homeopathy is unscientific”.
Also in this series
Disclaimer
I am not a doctor. This does not constitute medical advice. If you need that consult a properly qualified and registered medical practitioner.
These are just my opinions, but I try to make sure that what I write is both accurate and fair. If you think that I have got anything wrong please let me know. If you are right I will happily change what I have written.
Notes
**Ben Goldacre made some interesting observations on this paper in a piece published in the Guardian, “A quick fix would stop drug firms bending the truth” (blog version)
***The Pyjamas in Bananas blog has looked at this paper in detail and provides a good reference for this controversy. This topic is clearly more complex than the caricature provided in this essay suggests.
References
Barron K, (chairman). Patient Safety – Health Committee – Sixth Report of Session 2008-09 – Volume I: Report, Together with Formal Minutes. London: The Stationery Office Limited; 2009. Available from: http://www.publications.parliament.uk/pa/cm200809/cmselect/cmhealth/151/151i.pdf.Goldacre B. Benefits and risks of homoeopathy. The Lancet. 2007 November;370(9600):1672–1673. Available from: http://dx.doi.org/10.1016/S0140-6736(07)61706-1.
Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, Moore TJ, Johnson BT. Initial Severity and Antidepressant Benefits: A Meta-Analysis of Data Submitted to the Food and Drug Administration. PLoS Med. 2008 February;5(2):e45+. Available from: http://dx.doi.org/10.1371/journal.pmed.0050045.
Leigh E, (chairman). A safer place for patients: Learning to improve patient safety. 51st report of session 2005/06. Report, together with formal minutes, oral, and written evidence. London: The Stationery Office Limited; 2006. Available from: http://www.publications.parliament.uk/pa/cm200506/cmselect/cmpubacc/831/831.pdf
Milgrom LR. Under Pressure: Homeopathy UK and Its Detractors. Forsch Komplementmed. 2009 September;16(4):256–261. Available from: http://content.karger.com/ProdukteDB/produkte.asp?Aktion=ShowAbstract&ArtikelNr=228916&Ausgabe=248719&ProduktNr=224242
Pilkington K, Kirkwood G, Rampes H, Fisher P, Richardson J. Homeopathy for depression: a systematic review of the research evidence. Homeopathy : the journal of the Faculty of Homeopathy. 2005, July; 94(3):153–163. Available from: http://view.ncbi.nlm.nih.gov/pubmed/16060201.
Shang A, Huwiler-Müntener K, Nartey L, Jüni P, Dörig S, Sterne JA, et al. Are the clinical effects of homoeopathy placebo effects? Comparative study of placebo-controlled trials of homoeopathy and allopathy. Lancet. 2005;366(9487):726–732. Available from: http://dx.doi.org/10.1016/S0140-6736(05)67177-2.
Terry A, Mottram C, Round J, Firman E, Step J, Bourne J. A safer place for patients: learning to improve patient safety. London: National Audit Office; 2005. Available from: http://eprints.whiterose.ac.uk/3427/.
Acknowledgements
dvnutrix for pointing this nonsense out to me.
Part one – original – on this blog
Nessun commento:
Posta un commento