Dr. Andrew Weil is a rock star in the “complementary and alternative medicine” (CAM) and “integrative medicine” (IM) movement. Indeed, it can be persuasively argued that he is one of its founders, at least a founder of the its most modern iteration, and I am hard-pressed to think of anyone who did more in the early days of the CAM/IM movement, back before it ever managed to achieve a modicum of unearned respectability, to popularize CAM. In fact, no physician that I can think of has over the course of his lifetime done more to promote the rise of quackademic medicine than Dr. Weil. The only forces greater than Dr. Weil in promoting the infiltration of pseudoscience into academic medicine have been the Bravewell Collaborative and the National Center for Complementary and Alternative Medicine (NCCAM). Before there was Dr. Mehmet Oz, Dr. Dean Ornish, Dr. Mark Hyman, or any of the other promoters of IM, there was Dr. Weil.
And why not? Dr. Weil looks like an aging 1960s rock star, and, operating from his redoubt at the University of Arizona, is quite charismatic. For all the world he has the appearance of a kindly, benevolent Arizona desert Santa Claus, an ex-hippie turned respectable dispensing advice about “natural” medicines, writing books, and making himself ubiquitous on television and radio whenever the topic of alternative medicine comes up. Before Dr. Oz told Steve Novella that “Western” science and medicine can’t study woo like acupuncture, Dr. Weil was there, paving the way for such arguments, previously considered ludicrous, to achieve a patina of respectability.
In fact, he’s still at it, doing it far better and far more subtly than the ham-handed Dr. Oz. Unfortunately, it’s the same anti-science message and the same appeal to other ways of knowing built upon tearing down straw men versions of evidence-based medicine (EBM) with gusto. This was brought home last week when Dr. Weil co-authored an opinion piece with Drs. Scott Shannon and Bonnie J. Kaplan for the journal Alternative and Complementary Therapies entitled Safety and Patient Preferences, Not Just Effectiveness, Should Guide Medical Treatment Decisions, an article that was noted at the blog Booster Shots in a credulous, fawning post entitled Dr. Weil says there’s a better approach to evaluating clinical drug trials. In contast, Steve Novella put it far more succinctly (and accurately) in the title of his post: Andrew Weil Attacks EBM. That’s exactly what Weil and company did in this article.
While Steve is absolutely correct, I also see it more as Dr. Weil demonstrating once again that, upstarts like Dr. Oz aside, he is still the master of CAM/IM apologia, much as, even though both were Sith Lords, Emperor Palpatine remained master over Darth Vader until just before the end. You’ll see why in terms of the arguments, both subtle and not-so-subtle, that Dr. Weil and his acolytes make. Moreover, even though his disciple Shannon is granted the coveted first author position, the arguments presented leave little doubt that it’s Weil who’s driving the bus.
Efficacy, effectiveness, and safety, oh, my!
Shannon et al begin by emphasizing that there is a difference between “efficacy” and “effectiveness,” and from the very beginning take pains to separate their definitions of the two, pointing out that they will not use the terms interchangeably. Specifically, they refer to the definition of “efficacy” as “demonstration of benefit under ideal conditions, typically in randomized controlled trials” and “effectiveness” as “demonstration of benefit in real-life conditions.” These are relatively new concepts, comparatively speaking. Even so, they are nothing “alternative,” being instead firmly within the bailiwick of EBM. This doesn’t stop Shannon et al from presenting these concepts as though EBM and science-based medicine (SBM) don’t consider questions of whether treatments work under real-life conditions. As you will see, it is also the opening volley in a rather complex set of fallacious arguments designed to support the watering down of scientific rigor in order to allow CAM/IM modalities that can’t stand up under close scientific investigations to be labeled as “effective” (or at least acceptable).
First of all, let’s see a bit more what Weil and company are up to when they compare “effectiveness” with “efficacy.” Mostly, it’s meant to attack the reliance of EBM randomized clinical trials (RCTs) in order to set the stage for proposing something else more to their liking:
The value of efficacy lies mainly in its ability to indicate potential for effectiveness accurately. Sadly, in the drive to emphasize the importance of delineating clearly sound measures of clinical effectiveness, modern medicine has come to equate RCTs as the final arbitrators of clinical decision making. As discussed below, RCTs are but one tool to sort out these complex questions. In integrative medicine (IM), particularly, with its emphasis on patient variables and practitioner participation, evaluation of efficacy is not sufficient.
I agree that one major value of efficacy in RCTs is that it indicates the potential for effectiveness in the real world. I also even partially agree that EBM overemphasizes RCTs as the final arbiters of clinical decision-making, although, I note, not for the same reason as Weil and his acolytes. We at SBM frequently make the argument that plausibility based on science should be a much more important consideration in evaluating treatments than it is in EBM. Indeed, this is a major reason that the term science-based medicine, as opposed to evidence-based medicine, was coined in the first place. EBM emphasizes RCTs without considerations of prior plausibility, relegating basic science considerations that conclusively demonstrate CAM/IM modalities like homeopathy to be so implausible as to be not worth studying in humans to its lowest rung in its scale of evidence. So, while we agree somewhat that EBM overemphasizes RCTs, we most definitely do not agree with Dr. Weil about how it does so. Whereas Dr. Weil seems to think that RCTs are overemphasized because they are too standardized and do not adequately take into account real-world conditions, he most definitely would not like SBM’s plea to EBM to consider scientific plausibility, because doing so would relegate modalities like energy healing, homeopathy, and acupuncture to the hopelessly implausible.
Dr. Weil also does a very clever thing with his argument. While he correctly points out that evaluation of efficacy is not sufficient to determine whether a treatment is effective, as Steve Novella pointed out, Dr. Weil conveniently neglects to acknowledge that it is at the very least necessary. If RCTs indicate that a therapy is not detectably more efficacious than a placebo, which is the case for the vast majority of CAM/IM modalities, then there is no point in doing trials in “real world” conditions because there is no reason to expect the to be effective. Basically, RCTs serve as a screening test to identify promising therapies that are likely to be effective in real world use. In essence, Dr. Weil is echoing the same two-pronged attack that CAM/IM advocates make against RCTs, the first being the claim that “Western science” can’t study his woo, which is the same fallacious argument that Dr. Oz made regarding acupuncture. The second prong of the attack is to point to what Weil calls “alternative sources of valid information” as demonstrating that his woo works when RCTs do not support its efficacy. Although Dr. Weil is careful never to state this explicitly or even use the term “pragmatic trial,” his emphasis on “effectiveness” over “efficacy” is in essence a plea for relying much more heavily on observational trials, in particular pragmatic trials.
But what are “pragmatic” trials? They are trials that seek to determine if a treatment is effective outside the confines of an RCT, in other words, out there in the “real world.” RCTs, because they seek to determine efficacy, need to control for as many potentially confounding factors as possible, which makes them inherently artificial to one extent or another. Once efficacy is established under controlled conditions, it is sometimes then useful to determine whether this efficacious treatment is effective under usual circumstances in the “real world,” where patient compliance might be poor, either due to side effects, cost, or difficulty in adhering to the therapy; where complicated protocols might be more problematic to follow compared to academic medical centers; and where use of the treatment will inevitably be expanded to patient populations not represented in the RCTs used to approve the drug.
Pragmatic trials can go one of two ways. In the case of treatments demonstrated efficacious in RCTs, most commonly pragmatic trials demonstrate a lower level of effectiveness than the efficacy measured in RCTs might indicate. As Harriet Hall discussed a couple of years ago, a classic example of a treatment that was shown to be efficacious in RCTs but potentially more dangerous in pragmatic studies of actual use in community hospitals was the clot-busting drug t-PA, which was effective for ischemic strokes in RCTs but resulted in a higher death rate. For efficacious interventions, the “real world” is almost always less hospitable than the “ideal world” of RCTs. There is one case, however, where the “real world” can make a treatment seem more effective than it is by any objective measure, and that’s for treatments that are essentially placebos. Outside of the rigorous, carefully controlled world of RCTs, placebo medicine can seem to have effectiveness, which is exactly why proponents of pseudoscientific medicine love pragmatic trials. The populations aren’t as well-defined; often there is no placebo control; and they are almost always unblinded. There is a good reason why, after well-designed RCTs have demonstrated acupuncture to be no more efficacious than placebo or sham acupuncture, acupuncturists have returned to pragmatic studies, as, I will note in a shameless plug, Harriet Hall will discuss tomorrow. The hospitableness of pragmatic trials to placebo medicine is also behind much of the attack on RCTs by advocates of pseudoscience, and Shannon et al certainly engage in such attacks with gusto:
RCTs have dominated decision making about efficacy in health care for almost 50 years. Many researchers have explored the difficulty of subjecting IM treatment approaches to RCTs. There are some characteristics of IM interventions that make RCTs particularly difficult to carry out, and perhaps even less relevant, than for conventional allopathic medicine. As Fønnebø pointed out, the gap between published studies of integrative approaches on the one hand, and the clinical reports by practitioners on the other hand, may partially result from the fact that placebo-controlled RCTs are designed to evaluate pharmaceutical interventions.
Or, more likely, they result from the biases and lack of adequate controls for placebo effects inherent in “clinical reports” by practitioners. Also note the Oz-like argument that RCTs can’t evaluate Weil’s favored woo.
Granted, uncontrolled clinical reports from practitioners can, if carefully documented, represent one form of pragmatic trials; more often they are at best preliminary data and at worst nothing more than anecdotes. To support their attack, Shannon et al cite a prominent New England Journal of Medicine study, namely Concato et al, which found that well-designed observational studies often produce the same results as RCTs, meaning that placing observational studies lower on the rung of the EBM hierarchy might not be justified. Unfortunately, the authors failed to note that Concato et al only looked at studies of clinical questions with objective outcomes, including the Bacille Calmette-Guérin vaccine and tuberculosis, mammography and mortality from breast cancer, cholesterol levels and death due to trauma, treatment of hypertension and stroke, and treatment of hypertension and coronary heart disease. Moreover, Concato et al have been criticized for cherry picking their examples. In contrast, most CAM modalities are designed to treat symptoms with a major subjective component. Also going against Weil’s argument that, because RCTs are designed for pharmaceutical treatments, they might not be so good evaluating non-pharmaceutical treatments is a more recent study comparing effects in RCTs versus observational studies in digestive surgery, which found that a quarter of observational studies gave different results than randomized trials and that between-study heterogeneity was more common in observational studies. If this study holds up, it would appear that RCTs work just fine (and better than observational studies) for surgical interventions.
The merits and flaws of Concato et al aside, it is a false dichotomy that we must choose either controlled RCTs or or less rigorously controlled observational studies. EBM and SBM encompass them all; they simply disagree on how much relative weight to give each type of evidence. Fortunately, Weil doesn’t make the argument that we must favor one or the other. Unfortunately, he’s more subtle than that. What he does instead is to argue for decreased influence of RCTs by arguing that they should be only one part of the picture, an argument he makes by bringing up the Bradford-Hill criteria.
Trying to shrink the importance of the RCT
The heart of the argument in this paper is that RCTs should be only one (presumably of many) factors that determine whether medicine considers and intervention to be effective or not. As the title suggests, safety and patient preferences play a prominent role in Weil’s argument, but first Weil yokes poor Sir Austin Bradford-Hill, FRS, a pioneering early to mid-20th century epidemiologist best known for his work linking smoking to lung cancer and pioneering the RCT, into his service. Bradford-Hill proposed a set of criteria for drawing causality about disease etiology, but his nine criteria are sometimes used in considering the effectiveness of treatments.
Shannon et al set the stage:
As he [Bradford-Hill] pointed out, not all criteria are appropriate for all issues being analyzed, but he listed nine in total from which appropriate ones should be selected for any given situation:
- Strength—referring to the robustness of the association between the causative agent and the outcome
- Consistency—meaning being able to obtain similar results across different research sites and methodologies (i.e., replication)
- Specificity—by which Bradford-Hill meant one disease having one specific outcome, which may not be relevant to complex disorders (e.g., psychiatric problems)
- Temporality—referring to the commonsense notion that the cause always precedes the outcome
- Biologic gradient—which is best described as a dose– response curve: increased treatment would presumably result in a proportionate increase in the effect (again, not relevant in all disorders)
- Plausibility—referring to whether the results are biologically sound
- Coherence—which refers to the agreement of a study’s findings with what is already known (hence, not relevant in situations of truly novel interventions)
- Experiment—the situation in which randomly introducing the causative agent results in the outcome
- Analogy—which is the idea that a similar cause results in a similar outcome.
It is particularly interesting to note that the Bradford-Hill criteria, specified by the individual who influenced the methodology we now accept for RCTs massively, lists experiment with randomization methods as only one of nine criteria for establishing causality.
See, you nasty, reductionist scientists? RCTs are only one criteria! They’re not the be-all and end-all of clinical evidence! I can’t help but point out that there is an implicit straw man here in that embedded in many of the Bradford-Hill criteria are the same sorts of arguments we make at SBM, particularly criteria numbers 1, 2, 5, 6, and 7. In particular, we at SBM like criteria numbers 6 and 7. Key to the very concept of SBM is that interventions should have biological plausibility. They should also be congruent with what is known about the disease or, if not congruent, the evidence supporting an intervention must be sufficiently compelling that it justifies overthrowing the existing paradigm (for example, in the case of the discovery that H. pylori causes peptic ulcers). I also can’t help but point out that most CAM/IM treatments still fail most of Bradford-Hill’s criteria. In particular, CAM/IM treatments almost always run afoul of all of Bradford-Hill’s criteria other than #4, and, quite frankly, sometimes the wackier ones even seems as though they could run afoul of #4 as well.
Perhaps realizing this, Weil says nothing more about Bernard-Hill criteria other than to mention them again briefly in his conclusion. Instead, he and his merry trio of woo-apologists move on, Gish gallop-like, to other deficiencies in the “Western, reductionistic” model.
Use and abuse of systems biology and “holistic” methods
Unable to abuse poor Bradford-Hill anymore, Weil moves on to list “unique” features of CAM/IM research that—surprise!—turn out to be not-so-unique and not nearly as difficult to take into account in SBM (or even in EBM) as he implies:
For instance, the healing relationship of a doctor and patient is generally excluded or “controlled for” in conventional RCTs, whereas some researchers would argue that unconditional positive regard forms the underpinnings of the healing relationship between two people.
The enhanced focus on the healing relationship is thus another factor delineating IM from conventional health care models. A second example is the concept of individualized care, which is rarely included in RCTs (perhaps the MTA study in childhood ADHD is a notable exception). The notion that each patient is unique and quite different permeates IM.
I like the admission that most CAM/IM is placebo medicine, though. Oh, you didn’t see that? The emphasis on the “healing relationship” between practitioner and patient tells you that what Weil is talking about is placebo medicine. After all, that relationship is very important to placebo effects.
Be that as it may, I’ve discussed the so-called “individualization” of CAM treatments before, as well as the difference between the “personalized medicine” when practitioners of SBM use the term versus when CAM/IM advocates use the term. For the full discussion click on the links; the CliffsNotes version follows. Basically, the notion that each patient is unique is a notion that is recognized in EBM and SBM. It’s also utterly facile and obvious, given that no two people are alike. What EBM and SBM try to do is to classify and stratify patients based on science-based characteristics that can be objectively related to disease severity, etiology, and response to therapy. As genomic medicine, systems biology, proteomics, and metabolomics become more sophisticated, it has become possible to make finer and finer distinctions between patients based on biology. The hope is that ultimately knowing these fine distinctions will allow us to “personalize” therapies to individual patients, thus ushering in an era of truly personalized medicine.
In contrast, in CAM/IM the idea of “individualization” and “personalized” medicine most frequently boils down to making it up as the practitioner goes along and doing whatever the practitioner feels like, all without a basis in sound science and evidence. In fact, Weil appeals to just such “personalization” or “individualization” based on magic and fairy dust as being a problem with applying RCTs to CAM/IM:
However, classic RCT research design requires patients to be broken out into groups with a similar diagnosis, which impairs the ability to evaluate an individualized treatment system, such as classical homeopathy, Traditional Chinese Medicine, or Ayurveda. In each of these systems, the patient must be individualized into a quite unique pattern that does not lend itself to a more broad disease generalization such as that found in conventional allopathic medicine. Curiously, the cutting edge of modern medicine anticipates that customized and individualized care looms as a result of advances in single nucleotide polymorphisms (SNPs) and the ability to create a specific genetic fingerprint for each individual.
Imagine that! Using groups that are based on diagnosis and actual disease biology, rather than based on prescientific vitalistic thinking! The nerve of those reductionistic “Western” physicians!
Here’s a hint: Genome variations, as demonstrated by SNPs, are not an example of the sorts of classifications one finds in homeopathy, traditional Chinese medicine, or Ayurveda, although I am impressed at the attempt to liken CAM classifications to “cutting edge” genomic variations and science-based classification systems. Sadly, though, leaving aside the utter ridiculousness of the analogy, Weil appears to be behind the times when it comes to personalized medicine and genomics. Ask any systems biologist or geneticist. SNPs aren’t even really “cutting edge” anymore. Maybe they were ten years ago (possibly even five years ago), but other methods and markers are rapidly supplanting SNPs, in particular direct sequencing of relevant parts of the genome, a technique made practical by next generation sequencing technologies and the concomitant exponential plummeting of the cost of such analyses. Also, as Harriet Hall has pointed out, the claim to “personalization” is a sham; many CAM/IM therapies ultimately appeal to one of many examples of the “One True Cause of Disease” fallacy.
Speaking of shams, so is Dr. Weil’s claim of being “holistic”:
A third methodological issue that distinguishes IM from the environment of pharmaceutical RCTs involves systems thinking. With its roots in holistic, natural, and aboriginal medicine, IM has always embraced a more systems-based orientation to patient care than conventional care. It should come as no surprise that a narrow modality for evaluating treatment effectiveness would become increasingly limiting to IM research. The movement toward increasingly narrow scientific evaluations may create an artificial and arbitrary view of human health, medicine, and treatment effectiveness. Fritjof Capra, PhD, the well-known physicist, indicated that the great surprise of twentieth century science was that complex systems cannot be understood by analysis. Ecology and epigenetics are examples of the strong movement toward systems thinking in modern biology.
This is, of course, utter nonsense. “Holistic,” “natural,” and aboriginal medicine relied upon a prescientific, not a “holistic,” understanding of biology and patient care. Appealing to “systems” thinking based on prescientific or outright religious systems, such as homeopathic provings, traditional Chinese medicine, Ayurveda, and “energy healing,” is meaningless when the systems being invoked are without a grounding in science—or even reality. The rise of systems theory in biology and elsewhere has nothing to do with the sort of false “holistic” thinking invoked by Weil and his ilk. Rather, it has to do with the increasing ability to analyze more than one aspect of a complex system at once. For example, in my field (cancer), thirty years ago it was only possible to analyze one gene or perhaps handful of genes at a time. Then, beginning in the late 1990s, the development of cDNA microarray chips made it possible to analyze hundreds, then thousands of genes simultaneously—then every gene in the genome. As technology and computing power increased and scientists and mathematicians developed techniques to analyze ever larger datasets, it became possible to take the data from these sorts of experiments and begin to understand the complex networks inherent in the expression of the thousands of genes that are produced in human cells. Today, it is becoming increasingly possible to integrate massive quantities of data from various sources, including genomics, proteomics, and metabolomics and begin to understand the vastly complex networks that they form, how they resist perturbation, and how they can be restored when they are perturbed.
That is real holism. Not homeopathy. Not traditional Chinese medicine. Not Ayurveda. What Weil represents as “holism” is in reality a series of pretenders to “holistic” understanding that substitute non-evidence-based prescientific belief systems for science, gussying them up in “science-y”-sounding language that co-opts new science the way CAM/IM co-opts science-based modalities like diet and exercise as being somehow “alternative.”
The rest of the rest
Those who have actually read Dr. Weil’s article will note that I have not yet mentioned one point that Weil hammered on relentlessly, apparently thinking himself Maxwell bringing his silver hammer down upon the heads of those who think scientifically, the better to knock the propensity towards SBM (or even EBM) out of them. That issue is safety. Weil goes on and on about patient safety, even going so far as to propose a ranking system for patient safety. This part of his article is, quite frankly, an insult to physicians’ intelligence. Physicians practicing EBM already weigh efficacy/effectiveness versus safety; all Weil is doing is to rename and rebrand something that EBM does already. To a small extent he has a point that EBM does not yet have a as rigorous “hierarchy” of risk or harm to evaluate treatments as it does levels of evidence for efficacy, but even if we did have such a rigorous hierarchy, risk-benefit estimates still require weighing benefit versus risk. If the expected benefit is zero, or near zero, then even a tiny risk can quickly become unacceptable. To take one example, in the case of acupuncture, where the benefit is not distinguishable from that of placebo, then even a tiny risk of, say, a pneumothorax from a needle stuck too deep becomes very problematic. In other words, Weil’s classification system is unlikely to make CAM seem as attractive as he seems to think it will, although no doubt his intent is to make that evil “reductionistic” allopathic medicine seem more dangerous by comparison.
In addition, Weil makes a big show of complaining about commercial influence, in essence using a weaker form of the “pharma shill” gambit to dismiss pharma-funded RCTs whose conclusions he does not like. This serves, more than anything else, as a ploy to imply guilt by association, given that virtually all pharmaceuticals are manufactured for profit. As we have written about here on numerous occasions, commercial influence is a problem that is recognized in EBM (and SBM). It is increasingly appreciated as an issue, and practitioners of EBM do consider funding sources when considering clinical trial results. Indeed, one can’t help but wonder why Dr. Weil’s concern about pharma influence is so conveniently limited. Does he complain about the influence of supplement manufacturers (which, by the way, are increasingly owned by various pharmaceutical companies and wield a great deal of clout in Congress)? If he does, I haven’t seen evidence of it. Steve was right; this part is just window-dressing, as is Weil’s proposal for a different five-level hierarchy of evidence that not-so-subtly tilts the playing field so that some observational studies and other studies that can’t be considered RCTs could be considered Level One evidence.
What’s truly ironic is that Weil would never admit that there are few, if any, CAM modalities that can meet the criteria of level one evidence even under his own proposed system! In fact, most of the very best-supported CAM modalities would fall either into level four (“Weak indicators of uncertain value: poorly designed studies without strong support from the Hill criteria; small observational studies”) to level five (“Very weak indicators of efficacy or effectiveness: expert opinion of effectiveness; case series; multiple anecdotal reports”). That’s why it’s tempting simply to grant Weil his new system and then challenge him to show more than a pitiful handful of CAM modalities that have evidence to support them stronger than level four.
Finally, two of the more pernicious proposals in Weil’s paper are related. First, he repeats his explicit argument for favoring placebo medicine:
Given the history and philosophical preferences of allopathic medicine, it should come as no surprise that the factors defining the healing response become minimized or ignored in current practice. This failing must be remedied, as these factors account for a huge component of how humans heal and recover. The healing relationship has taken a much larger role in IM as practitioners in nearly all CAM modalities place a much higher emphasis on it. The importance of these issues can be demonstrated most clearly in psychiatry research, where the placebo response plays a huge role accounting for as much as 40%–90%+ of the total response.26
Significant placebo and expectancy responses inhabit other areas of medical practice, such as dealing with pain and even life expectancy in patients with terminal cancer. Clearly, patient factors must compose a significant part of all treatment selection processes. Ideally, every treatment should be matched to the individual’s belief system to reach the highest level of response possible.
Aside from stating, “CAM is placebo medicine, and we should use placebo medicine when it fits in with patient beliefs,” a more explicit admission that what is being proposed in this opinion piece is placebo medicine I cannot imagine. Worse, it’s based on misinformation, the most egregious of which is Weil’s claim that placebo and expectancy effects can increase life expectancy in patients with terminal cancer. Would that it were true, but unfortunately it’s not, either in late stage cancer or early stage cancer. Nor is there any evidence that it has an effect on recurrence. This is where Weil’s second, related thrust occurs. Using this assumption that placebo medicine “works” better if its tenets match the patient’s belief system, he proposes turning the entire concept of informed consent on its head with respect to medical decision-making (MDM):
The ideal process in MDM would remodel the process of informed consent. The actively engaged patient would be offered a quick overview of appropriate treatments (both CAM and conventional) with an unbiased reflection of both safety and effectiveness. The patient would then declare a preference for one over the other(s). Once a practitioner ascertains the basic worldview of a patient (natural versus conventional; safety versus effectiveness) many simple elements of MDM would flow quickly in the future.
This is a process that to which I’ve referred in the context of two other discussions, vaccines and supplements. In essence, what Dr. Weil is proposing is an ideology-based version of misinformed consent. Although Dr. Weil would not doubt strenuously object to my comparison, when you boil it all down, what he is doing, although more subtle, is not materially different than what anti-vaccine groups do when they grossly exaggerate the risks of vaccination and downplay its benefits or what “health freedom” groups do when they exaggerate the benefits (and the scientific evidence for such benefits) of supplements. What these groups each do is to give patients a skewed view of the risks and benefits of an intervention, intentionally making it more likely that patients will choose the “alternative” option. Weil is doing the same thing by trying to confuse the issue of efficacy/effectiveness by throwing in the issue of “patient belief” and placebo medicine. As is the case with anti-vaccine activists and supplement hawkers, the purpose is, having failed to win on “conventional science,” to find a way to tilt the playing field back towards the “alternative” using other means.
The proposals in Shannon et al are nothing original, although I must grudgingly tip my hat to Weil and his acolytes for having repackaged old ideas in a very attractive, slick new package. In essence, they boil down to a massive appeal to other ways of knowing, hence the attack on RCTs in which they point out deficiencies that those of us advocating SBM and EBM have recognized for decades; the emphasis on effectiveness over efficacy that ignores the fact that it’s necessary to have good evidence of efficacy before even considering effectiveness in the “real world”; the tarting up of safety considerations as though EBM doesn’t take safety into account; and a direct appeal to placebo medicine via the methodology of misinformed consent. All of these strategies are clearly designed to try to give CAM/IM a leg up on science that it can’t achieve through science alone. Unfortunately, although such strategies are transparent to SBM bloggers (and, I daresay, many of our readers), to others they are not so obvious, even to proponents of EBM. I also can’t help but think: If Dr. Weil detests EBM so much, he’s really not going to like SBM at all. Not one bit.
Perhaps the most ironic part of all of this is that many of us at SBM actually agree with the contention that EBM relies on RCTs too much. Many of us would even agree with several of the Bradford-Hill criteria as considerations that are worthy of more emphasis in EBM. Where we differ is in how to do this. Most importantly, we emphasize much more strongly scientific/biological plausibility. Prior probability appears to be anathema to Weil, given that he mentions it as part of the Bradford-Hill criteria and then completely ignores it thereafter. No doubt, he knows how scientifically implausible many CAM/IM modalities are. Instead, Dr. Weil Gish gallops off to emphasize, as he does in the title of his article, patient preferences and safety, in essence using these issues to bolster an explicit plea for placebo medicine. Rather than accepting that rigorous scientific examination of his favorite CAM/IM modalities fails to find any benefit over placebo, Dr. Weil cleverly embraces placebo medicine and argues for lower scientific standards to permit pseudoscientific and unscientific medicine to appear to be co-equal with EBM.
Unfortunately, such are the arguments that have been the wedge used to insert quackademic medicine into medical academia. Even more unfortunately, they are working.
Shannon, S., Weil, A., & Kaplan, B. (2011). Medical Decision Making in Integrative Medicine: Safety, Efficacy, and Patient Preference Alternative and Complementary Therapies, 17 (2), 84-91 DOI: 10.1089/act.2011.17210