Here is the challenge: how do you take a treatment or set of treatments that clearly do not work, and in fact defy basic sciences like physiology and chemistry, and argue that they are worthwhile. The National Institutes of Health (NIH) was essentially given this challenge when the Office of Alternative Medicine was forced upon them in 1991. This office has since morphed into a center, with its current name, the National Center for Complementary and Integrative Health (NCCIH).

The NCCIH recently put out a document that is, in my opinion, the pinnacle of their strategy for promoting worthless treatments (whether they think that is what they are doing or not). It reads like a blueprint for how to spin a political narrative out of negative medical studies.

Focus of Subjective Symptoms

The paper is, “Evidence-Based Evaluation of Complementary Health Approaches for Pain Management in the United States.” Pain is a favorite target for CAM because it is a subjective symptom and there are known neurological mechanisms by which the perception of pain can be easily manipulated.

The fact that pain perception can be easily modified makes sense. The purpose of pain is to grab our attention and change our behavior, to attend to an injury and not make it worse by, say, walking on a broken limb. However, there are times when we need to ignore the pain when other survival needs become a higher priority, like running from that lion.

Pain, therefore, has a high placebo effect potential. In fact, pretty much anything you do will distract people from their pain to some degree. When studying pain it is important to control for all non-specific effects (such as improved mood, distraction, expectation, conditioning, and relaxation) and separate them from any specific effects of the intervention you are studying.

However, if it is your goal to pretend that a worthless intervention with no specific effect is actually valuable, pain is the perfect target. Just surround the worthless intervention with non-specific benefits from the ritual of treatment and present the resulting placebo effects as the treatment “working.”

Medical scientists may complain about non-blinding, poor controls, and nonspecific placebo effects. This is a minor inconvenience for your narrative, however. All you have to do is flip medical science on its head and declare that placebo effects are real and valuable also. As long as the subjects report lower pain, the intervention works.

Misinterpret the Scientific Evidence

Medical science is complex, and therefore there are many opportunities to misinterpret the evidence to support whatever your narrative might be. Cherry picking low grade evidence is one time-tested strategy. Another is hunting for statistical significance without considering things like effect size.

Let’s take a look at my favorite example, acupuncture. The definition of acupuncture is inserting needles at acupuncture points, and may also include manipulating the needles to produce a “de qi” sensation. That is the specific intervention, whatever mechanism you think may be in play.

By that definition, acupuncture clearly does not work. After several thousand studies, research has failed to demonstrate that acupuncture works for anything. The best designed studies consistently show that it does not matter where or even if you stick the needles. Reported benefits are due entirely to the placebo ritual that surrounds acupuncture.

Meanwhile, here is how the NCCIH report summarizes the evidence for acupuncture for low back pain:

We found 4 RCTs (total participants, 1092) that assessed the clinical benefit of acupuncture for treatment of low back pain (LBP) (age range, 28-60 years; most participants were white) and used primary study outcomes of self-report of pain intensity (numeric rating scale or visual analog scale [VAS]) and/or functional disability (Roland-Morris Disability Questionnaire, Oswestry Disability Index [ODI], or Disability Rating Index). Cherkin et al reported modest improvement in pain intensity and function compared with usual care. In pregnant women using auricular acupuncture, Wang et al found a significant reduction in pain intensity and improved functional status compared with no treatment. Comparison of verum to sham acupuncture had mixed results, with 2 RCTS finding no significant difference an finding a slight but significant difference. No significant adverse events were reported.

Studies looking at an intervention for a subjective symptom like pain that are unblinded are essentially worthless. The only reason to do such a study is to make sure the intervention does not have some horrible side effect, and to help design blinded studies that will actually test for efficacy.

Acupuncture proponents, however, seem to have specialized in doing worthless unblinded or poorly blinded studies, flooding the literature with useless noise that they then use to argue that acupuncture works. That is what the NCCIH reviewers do here – comparing acupuncture to no treatment for a pain outcome tells us literally nothing.

There are three studies that are properly blinded. Two of them are dead negative, and the third shows a slight difference only (too small for patients to actually notice). Should the FDA approve a drug based on similar results? In general would you want your physician to use treatments for which this was the evidence?

That is a negative review, consistent with the conclusion that acupuncture does not work. However, the NCCIH paper mixes these clearly negative results with worthless unblinded studies and weaves that into a positive outcome.

Use Fuzzy Definitions

From its inception, “alternative” medicine (which has been rebranded “complementary” and now “integrative” medicine) has been a fuzzy category. I would argue that, in practice, the only feature that truly defines this category is a lack of efficacy.

The common joke, of course, is: what do you call alternative medicine that works? Medicine.

But CAM is the brand, and you need to promote the brand, and the brand cannot only consist of useful interventions based on pure superstition. So, proponents hit upon the idea of including science-based treatments that work. Include those treatments under the CAM umbrella and use them to promote the brand, and the nonsense will ride in behind.

The NCCIH continues this marketing strategy by including all non-pharmacological approaches in the CAM category. Right from the beginning they write:

Medications may provide only partial relief from this chronic pain and can be associated with unwanted effects. As a result, many individuals turn to complementary health approaches as part of their pain management strategy.

They explicitly divide treatment strategies into pharmacological and non-pharmacological interventions, with the latter deemed CAM. This, of course, is not true. Diet and exercise, relaxation, various forms of physical therapy, and distraction are all science-based interventions. I prescribe all of the above on a regular basis. These interventions have plausible physical mechanisms and clinical scientific support. There is nothing alternative about them.

They have become, however, what we like to call the Trojan Horse of CAM. They have been co-opted by CAM proponents and rebranded.

The real CAM modalities are ones that arose outside of scientific mainstream medicine, or that have been rejected by mainstream medicine, or simply have not been studied and remain speculative or experimental but are being prematurely promoted as treatments.

The rebranding of anything non-pharmacological (and also non-surgical) as CAM also serves the function of advancing the notion that mainstream medicine is all about drugs and surgery and does not use other interventions. Again, this is not true, but that’s the CAM narrative.

Inflate its popularity

The broadening of the category of CAM also allows for an inflation of the apparent popularity of CAM, which is then used to justify things like government funding of a dedicated center at the NIH. They write:

National surveys going back more than 25 years have consistently found that these complementary approaches are used by about 30% to 40% of the US public in a given year, although use of a given approach may wax and wane over time.

Let’s take a closer look at those statistics. You only get to the 30-40% figure by including things like exercise and taking vitamins. The vast majority of use is from physical interventions like massage and manipulation. Such interventions range from squarely science-based to barely alternative because they make claims that go beyond the evidence.

Interventions that are clearly not science-based or mainstream, because they are based essentially on magic or pseudoscience, represent a tiny minority.

Use of acupuncture (1.1%), homeopathic treatment (1.7%) naturopathy (0.2%), and energy healing (0.5%) was miniscule.

There is no particular reason to lump all these various treatments into one giant category, except as a marketing strategy. Further, the category is falsely inflated by adding things that are clearly not “alternative.” These hide the truly tiny numbers for interventions that are more reasonably defined as alternative. (Although the word “alternative” is itself misleading, replacing the prior term that was used, “fraudulent.”)

Conclusion

In the end the review concluded that the following interventions had more positive evidence than negative:

Acupuncture and yoga for back pain
Acupuncture and tai chi for osteoarthritis of the knee
Massage therapy for neck pain—with adequate doses and for short-term benefit
Relaxation techniques for severe headaches and migraine.

For the other interventions they concluded the evidence is negative. But even for these four categories, the evidence is actually consistent with the conclusion that the interventions are not effective.

For massage therapy for neck pain the data is all unblinded. Even then, the results were short term, with differences disappearing by week 14. Even though the data is weak, having a short term benefit from massage for any muscle pain is plausible and a reasonable intervention, as long as it is gentle and safe (no deep massage of the neck) and as long as no claims are made for specific therapeutic effects not supported by the evidence.

The evidence is similar with relaxation for headaches:

Blanchard et al randomized patients with headache to biofeedback with relaxation training, biofeedback plus cognitive therapy, sham meditation, or a headache monitoring control condition. All of the treatment groups including the sham meditation group had improvements in the headache index score in comparison with the monitoring control group.

Again – a non-specific benefit from doing something vs nothing, but the details of the something don’t seem to matter. And again, I have no problem with helping patients to relax to help them cope with their chronic pain, just leave out the hand-waving magical explanations.

Similarly, Tai Chi is exercise. It does not become alternative because it has an exotic name.

It is unfortunate that the NIH through the NCCIH is now in the position of promoting the CAM brand. They are trying to justify their existence by buying into the standard CAM narrative and strategies – using a false category, inflating its popularity, focusing on subjective symptoms, and relying upon unblinded data.

In the end, despite the attempt at promotion, this review was essentially negative. What it really shows is the extremely limited potential of these interventions. At best some of them are useful for short term subjective relief, but we are basically talking about relaxation, massage, and exercise. At worst, they are straight-up superstition-based quackery (acupuncture) trying to hide in the herd.

Note: David Gorski also covered this topic at SBM.