The morality of magic kisses: Ethics and placebo in physiotherapy

When my daughter hurts herself, her placebo of choice is a “magic kiss”. This therapeutic intervention must be applied with care specific to the area of injury. Anecdotal evidence suggests that it is very effective. I use placebo freely at home but is it right to do this in the clinic? In a recent post I suggested that we can be more confident that acupuncture is essentially a placebo treatment. Still, patients feel benefit, so is it still OK to use it? I have often heard clinicians declare “I don’t care if it is a placebo if it works”.

Franklin Miller from the Department of Bioethics at the NIH considers these kinds of questions for a living. He has published a number of papers around this topic including this great review with his colleague Luana Colloca on the ethics of placebo in clinical practice. I was lucky enough to have a correspondence with him that had me thinking about this more deeply.

The big tension around offering a treatment that we know to be just a placebo lies with the competing interests of being committed to reducing suffering and the need to be transparent and offer informed consent. Offering a known placebo usually requires an act of deception that in its way, undermines the relationship between clinician and patient.  Of course if you could offer a placebo, be honest about what you were doing and still have a powerful effect there would be no problem.

Thinking about the ethics of all this in the therapies I see something of a division. We have many passive modalities that may well work by placebo (acupuncture, maybe manipulation, electrotherapy etc: let’s call these “magic kisses”).  Any placebo effect of these therapies in part rests on the effects of expectation, belief in the treatment and possibly a re-evaluation by the patient of their symptoms. Then we have therapeutic approaches like advice, pain education, CBT, lifestyle adaptation, which again seek to push patients to re-evaluate their symptoms and alter their behavioural responses (let’s call these “rational hugs”). Even with a partially shared mechanism of effect there seems to me to be a clear difference – rational hugs do not require a belief in a mystery magic ingredient, and aim to empower the patient by offering a solution that they might control. Even if we found that the effect sizes were similar between the 2 types, there are no fibs being pushed in the second.

Two other points occurred to me. The first lies around the issue of patient choice. Why do patients choose acupuncture or manipulation or ultrasound etc? I guess they do because each treatment has passionate advocates who promote them, advertise them, spread the word about them, often with a willing media tagging along like an enthusiastic labrador. So the treatments help because patients expect them to, but patients only expect them to because the culture that delivers those treatments has propagated that belief! It’s a fabulous business model (it all costs) but I smell a conflict of interest. Patient choice is difficult in a world where good information is so elusive.

The second point lies in the possible long-term harms of offering passive treatments of limited efficacy. These are much harder to measure than the short-term benefits.  Treatments such as manual therapy that are based on models of pathology (that might not be identifiable or relevant) might lead to increased catastrophising and negative beliefs about what should be relatively benign conditions. Recently a number of big studies have suggested that therapies for whiplash associated disorder are a risk factor for poor outcome or delayed recovery.  This isn’t the clear support for my theory of the superiority of rational hugs over magic kisses as it at first appears. Not only passive therapies like chiropractic but also therapies including advice, pain education and clinical guideline-based management (see here and here) show this trend. Maybe offering any treatment increases vigilance and attention to symptoms and thus perpetuates symptoms in somatic syndromes?  I wonder whether any short-term benefits of placebo might be offset by these dangers. It seems that any therapy may have the capacity do harm, even where it is not immediately obvious.

Given the generally small clinical effects of placebo, the potential for undermining trust in the therapeutic relationship and a sinking feeling that it’s a bit demeaning for all involved, I think that if we know a treatment is essentially a placebo we might give it the elbow.  We should celebrate the fact that the placebo effect is still there offering a helping hand when we deliver effective treatment with care. So let’s hear it for rational hugs.

PS: For a nice little chat about placebo from Ben Goldacre click here.

ResearchBlogging.org
Original Article

Miller FG, & Colloca L (2009). The legitimacy of placebo treatments in clinical practice: evidence and ethics. The American journal of bioethics : AJOB, 9 (12), 39-47 PMID: 20013499

Other References

Cassidy, J., Carroll, L., Côté, P., & Frank, J. (2007). Does Multidisciplinary Rehabilitation Benefit Whiplash Recovery? Spine, 32 (1), 126-131 DOI: 10.1097/01.brs.0000249526.76788.e8

Côté P, Hogg-Johnson S, Cassidy JD, Carroll L, Frank JW, & Bombardier C (2007). Early aggressive care and delayed recovery from whiplash: isolated finding or reproducible result? Arthritis and rheumatism, 57 (5), 861-8 PMID: 17530688

Hróbjartsson A, & Gøtzsche PC (2010). Placebo interventions for all clinical conditions. Cochrane database of systematic reviews (Online) (1) PMID: 20091554

Pape E, Hagen KB, Brox JI, Natvig B, & Schirmer H (2009). Early multidisciplinary evaluation and advice was ineffective for whiplash-associated disorders. European journal of pain (London, England), 13 (10), 1068-75 PMID: 19181548

All blog posts should be attributed to their author, not to BodyInMind. That is, BodyInMind wants authors to say what they really think, not what they think BodyInMind thinks they should think. Think about that!

Comments

  1. Neil O'Connell says:

    Hi Barbara,

    I am afraid I don’t recognise your interpretation of the overall achievements of EBM. Massive steps forward in cancer survival rates, improved outcomes and disease control in Rheumatology, the fantastic record on vaccination all tell a different tale. Of course some therapies do not do so well – laminectomy is a good example and is being performed less in light of the evidence. In terms of CABG you seem to imply that CABG was compared to nothing when in the stuies that I am aware of it was compared to percutaneous coronary intervention – not a failure of evidence based medicine but a great example of progress in medicine and the development of equally effective but less invasive procedures. But I would be happy to see the reference source for your assertion. The “little evidence for medicine” argument is something of a fallacy . The oddest part of it is that this erroneous assertion then often leads to the beautiful idea that “therefore any old nonsense will do”. But that is a path that leads us backwards.

    It is worth remembering that placebo effects are generally small, short lived and, beyond subjective outcomes, do not seem to have a strong impact on clinical results (see the Cochrane reference). Helping people to manage their symptoms better may well work via similar pathways as straightforaward placebos but it is not lying to them, selling them a story or a bottle of snake oil. As such I am comfortable with “rational hugs” but not prepared to make pretend with somebody’s symptoms.

  2. Anonymous says:

    Depends on your definition of ethics. Deontology? Utilitarianism? Other?

    I’m guessing too that there’s some confusion over codes of practice and ethics proper. Just because the APTA says something is ethical (or not) doesn’t make it so.

  3. Steve Kamper says:

    I think this discussion illustrates to some extent why ‘the placebo’ is so unhelpful a concept, this is also evidenced in part by the fact that attempts at definition often end in confusion and nonsense (the effect of an inert intervention… etc). In my opinion its beyond time to stop thinking it as some kind of box full of magical stuff (non-specific stuff) that we deliver on it’s own or next to another treatment. If we think treatment effects (of any treatment) are augmented by positive expectations, lets explore how and why this works, if we think conditioning has an impact on how a treatment works, then lets look at this, if there is some other theory lets go there. I do understand the role of sham interventions in searching for the mechanical/physiological effects of treatment, but even to do this we need to deconstruct placebos so we understand what it is we are and aren’t controlling for. I realise that this may sound a little utopian but once we sort out what is in the magical box we can implement what is effective and ditch the rest and the ethical issues fade.

    Luke Parkitny Reply:

    Steve:

    You are quite correct! This is becoming a little like the dualist body vs mind debate. We often try to see the placebo component as a discrete malignant entity that could, with a magic scalpel, be excised from the healthy active tissue beneath. In fact, in reality, the relationship is probably more one where the two are inextricably linked and not so discrete. Having said that, for the purposes of research and investigation, all the earlier comments hold. In interpreting research for clinical practice and indeed in engaging in clinical practice, however, it is worth remembering that the relationship is probably more complex than we sometimes think.

    Neil O'Connell Reply:

    Fair points Steve,

    Like Luke I value the placebo control in research as it tests reasonable predictions about any proposed treatment modality. I would maintain that when those precitions are found false we should dump the treatment. I think this goal is somewhat distinct from investigating the effects of care. Understanding how conditioning and expectation work is key tom undersatanding the cloinical interaction but while the overall concept of “placebo” is perhaps unhelpful and covers a variety of aspects and mechanisms, the idea of placebo controls still gets my vote.

    SnippetPhysTher Reply:

    I think my hang up is I don’t understand how a patient’s perceptions can be yanked out of the equation. My experience is more clinical versus research oriented, so I have a better handle on what happens in the clinic. Okay, reality, maybe I don’t know what the heck goes on during the provision of physical therapy.

    I know in my experience that sometimes, I think, administratively we sometimes shoot ourselves in the foot. In big hospital systems, a physical therapist is just a physical therapist and there doesn’t necessarily seem to be an attitude that each therapist has strengths and weaknesses. The attitude is *all* the therapists are great. But now that I’ve been out on my own, and I have a name and an identity so to speak, for lack of a better word, something “magical” does happen when a patient requests me as a therapist by name and/or knows something about me. If it’s a physician who tells a little something about me – that’s like double “magic.” If the patient searched the net and found me and chose me, “magic” again. Honestly, no, I haven’t measured anything – so maybe I’m wrong in my thinking, but in those situations, it’s easier for me – almost like there is already a burst of trust in me.

    From a purely clinical perspective, if I can understand the non-specific stuff that would be helpful – along with learning of possibilities of changing the non-specific stuff to be more suitable and receptive to physical therapy intervention. It’s almost like during my initial evaluations, I snag a lot of “physical” data, but probably a bigger more relevant chunk of information is missing – a person’s receptiveness to physical therapy… readiness for self-efficacy… belief in being able to successfully manage the situation.

  4. Nice one Neil.

    The one thing I would add is that any misconception by the community (and *groan* the key stakeholders – can you tell I was recently working on a funding application?) of the viability of a treatment may also lead to other – more indirect – ethical and practical problems. For instance, if we accept that a treatment works because of some intrinsic quality it holds (i.e. active component) rather than purely due to the person’s response to the treatment ritual (i.e. placebo) we may in fact not only be changing their “world-view” but also providing foundation for misguided research. If we do not separate the placebo from the active component, the treatments may become absorbed by guidelines and receive community support, funding and ultimately research time that would otherwise be spent on investigating treatments that have a greater positive effect. For instance, we may continue to spend time working out how to better use the blue pill for anxiety – playing with dosages and comparing it to other blue pills, when in fact we should be working on finding a different drug to go inside the pill.
    Although the ethics of placebo are important to consider from the patient’s perspective, they are just as important for the clinician and researcher. We all work from specific paradigms after all and are all influenced by those beliefs.

    Neil O'Connell Reply:

    Couldn’t agree more Luke, we wrote about a good example of this is the (you guessed it) acupuncture literature : http://ehp.sagepub.com/content/32/4/393.abstract

  5. If neurophysiological changes occur, even with something you seem to think is placebo, is it really placebo?

    Neil O'Connell Reply:

    Yep I would say it is.

    Taken to its logical conclusion that position means that as long as the unsuspecting patient buys into it any old nonsense will do. So, in extremis, I could wave my magic cabbage (its all in the green aura) over the affected part whilst chanting to maximise the “healing energy”.

    My argument only applies to treatments that have reached that point in the evidence where the most plausible explanation is that they don’t have an “active ingredient” apart from placebo/ non-specific effects. At that point it becomes ethically problematic.

    What if the quality evidence strongly suggests that a treatment is just a placebo (e.g. acupuncture) but the clinician still believes strongly that it is more than that? I would strongly argue that the problem still stands. After all, willfully ignoring the evidence doesn’t make it go away.

    Just because science has started to unpick some of the physiological processes underlying placebo effects it doesn’t make them viable as an approach to clinical practice. Also placebo effects are unreliable, generally small and shortlived if the Cochrane review is anything to go by. If our role is to educate and empower patients then selling a story (even a good one) seems contrary to that goal.

    But of course thats all in my humble opinion! Thanks , as always, for the comments…

    SnippetPhysTher Reply:

    I’d like to continue the conversation because I am intrigued not just with hard core quantitative research, but also qualitative research.

    Please do not take my questioning to assume I firmly believe it is ethical to push placebo… but the brain – the brain is quite interesting. Since you seem to agree that neurophysiological changes can occur with placebo, what if sometimes that response is actually helpful for creating some change that alters the brain enough to get it out of a rut and lead to maybe an easier path for long term change, especially if tag-teamed with evidence informed intervention choices?

    Neil O'Connell Reply:

    Don’t worry I don’t think you’re a placebo peddler!
    The point you make is interesting but the thing is that if I go looking for it, any interaction I have with you might be likely to alter your neurophysiology – it would have to for you to experience it. That the brain changes in response to placebo treatments is interesting but were I to image the brains of a bunch of people whilst shouting the word “sausages!” at the top of my voice we might also measure a change in the brain. Of course changes that correlate with pain relief perhaps seem more important and desirable but they don’t take away from the fact that when we deliver a known placebo we are basically deceiving a patient. It seems paternalistic, disempowering for the patient and ultimately, given the generally tiny and rather unreliable effect sizes not worth the compromise.
    Brain imaging studies do raise an ethical question – if we could pin the placebo effect down to specific mechanisms then how is it different to an active treatment? Well if we consider all the non-specific effects of care under the placebo umbrella then it is different precisely because it is the non-specific effects of interacting with somebody who is perceived as being there to help. Surely we could achieve kind of an effect that without lying or telling a fairy story. Of course the flip side is that if we argue that brain imaging findings legitimise placebo as a treatment approach if there were no such findings would the approach cease to be legitimate?
    Ultimately using placebo as a control for researching treatments is the best way we have of assessing whether our hypotheses about that treatment are correct. It seems a bizarre state of affairs that if after we test those ideas and find them false we would want to continue with the treatment anyway. Surely then we could go look for something genuinely efficacious? When a patient steps through the door they (I reckon) want a treatment that does what it says on the tin. We have to test the claims on the tin and where they are false throw the tin away!

    Barbara Drummond Reply:

    But these days, evidence based medicine is turning up very little evidence for medicine. The 10 year outcome after a CABG shows survival rates do not differ from those who did not have a CABG, pain relief of people having a laminectomy versus having conservative treatement is no different after 6 months; etc. etc. etc. (Of course, emergency room treatments and public health medicine are not included in my generalization; there are some things that actually have been statistically proven to work.) Perhaps, the future of medicine may be to evolve and empower the brains of patients to the point where they can generate the placebo effect for themselves…..

  6. This question often comes up when I consider whether to completely debunk homeopathy and other alternative therapies, and eliminate them from say my family, or to encourage them to the extent that they serve as effective placebos for certain ailments.