If a patient comes into a clinic, naive to the incredible benefits of Therapeutic Cabbage Rubbing and a physical therapist proceeds to sell them on how truly remarkable cabbage is, informing them of the unique resonance the cabbage creates when rubbed on skin, the science-y neurophysiological effects and stories of the success other people have had with the treatment from a position of assumed knowledge with obvious charisma — That patient might actually leave the clinic feeling a little better and be more likely to seek out therapeutic cabbage rubbing in the future.
Now say they walk into your clinic and as a well informed physical therapist, you know that therapeutic cabbage rubbing has no basis in our current understanding of physiology and lacks clinical evidence to support its efficacy and effectiveness, but this patient REALLY expects that the cabbage will help them out — does this expectation all of a sudden make Therapeutic Cabbage Rubbing a well reasoned treatment option?
When providing treatments that you can reasonably say have no benefit beyond the effect of expectancy (one of the many things included in the nebulous umbrella term “placebo”), you must consider how the patient developed such expectations in the first place. Neil O’Connell summarizes this issue well:
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
This type of reasoning is fatally circular and only perpetuates the use of treatments that can be rightfully discarded. However, this is by no means a suggestion that the literature on expectation, placebo, verbal suggestion and other aspects of clinical interaction should not help inform how we engage with our patients. The lessons learned from the broader literature base in placebo and therapeutic alliance can be vitally important to good clinical practice but, they do not make an ineffective treatment useful.
One recent example that can inform physical therapy practice is the work of Peerdeman et al with their paper “Relieving patients' pain with expectation interventions: A meta-analysis” published in Pain earlier this year. Peerdeman et al took a look at the effects of expectation interventions (that is, verbal suggestion, conditioning and imagery) on acute procedural, experimental and chronic pain. For the sake of discussion, let’s focus on the verbal suggestion aspect which has garnered the most interest on social media and showed the biggest effect size in the analysis. The basic gist of verbal suggestion is that if a patient was receiving an injection of morphine, the patient would experience a greater analgesic benefit if the doctor verbally suggested that the morphine was a powerful pain reliever and would help a good deal with their discomfort, as opposed to someone who receives the same injection without any verbal suggestion or meaningful interaction.
What Peerdeman et al found was that in acute procedural (ex. minutes to hours after venipuncture or surgery) and experimental pain, there were medium to large effect sizes in pain reduction when interventions were paired with verbal suggestion. In patients experiencing chronic pain (the group more representative of those in a physical therapy clinic), such as chronic low back pain or recurrent headache, they found only small effect sizes for verbal suggestion. Medium to large effect sizes sounds exciting, but what does this translate to in terms of pain reduction? The authors found that verbal suggestion had an effect size of g = 0.75 (95% CI 0.50-1.00), or in terms of a 0-10 pain scale it produced a 1.39 reduction in pain with a 95% confidence interval of 0.85-1.93. Other important things to keep in mind is that there was significant heterogeneity in the results of the individual studies analyzed and there was some noted risk of publication bias. When the pooled effect size was analyzed when accounting for the risk of publication bias, it dropped to g = 0.43 (95% CI 0.24-0.62). This should give pause to the initially appealing statement of “medium to large effect sizes in pain reduction.”
So what is the takeaway here? The work by Peerdeman et al is informative in the sense that we should considering incorporating verbal suggestion when providing well reasoned, evidence based treatments. How interventions are framed may be important and clinicians should seek to highlight their benefits to patients. What this bit of research and the broader literature base does not suggest is that ineffective treatments are made clinically useful by leveraging verbal suggestion or that if a patient expects a treatment that is ineffective, that it is now evidence based to provide it under the guise of expectancy. Consider for a moment what that scenario might look like — What is the cost of continuing to perpetuate the use of ineffective treatments? What implications are there to verbally suggesting a treatment is powerful and likely to help given knowledge to the contrary? This type of practice might be considered intentionally deceiving and could possibly undermine the therapeutic relationship for what is likely to be a small, inconsistent effect.
The idea of leveraging placebo has been a popular one in physical therapy as of late, but the profession must keep in mind that placebo effects do not turn an ineffective therapy into an effective one. Instead of glamorizing the power of placebo, the profession should be extremely circumspect of the idea that placebo produces powerful, clinically meaningful effects in isolation. As noted by David Colquhoun, placebo effects in clinical trials are often weak and likely a result of statistical artifacts such as regression to the mean. This is further substantiated by the work of Hróbjartsson & Gøtzsche showing that placebo effects for clinical conditions are generally small, inconsistent and difficult to distinguish from biased reporting.
Physical therapists should leverage the work of Peerdeman and others to maximize the benefit of treatments shown to have meaningful and specific effects. This can include instilling hope, emphasizing positive effects of well reasoned interventions, motivating behavior change and nudging patients towards health. Continuing to unpack how to best leverage these things is essential. Further, the physical therapy profession needs to be comfortable abandoning interventions that do not survive scientific scrutiny, instead of rationalizing their continued use under the false notion that placebo and expectation morphs an ineffective treatment into a useful one. Continuing to repurpose shoddy treatments in this fashion is a disservice to the profession and the patients it seeks to help. Patients deserve to be educated on and provided the best treatments available, not sold a story about the therapeutic effects of cabbage and other rotten products.
Image above is a still from The Truman Show, directed by Peter Weir