“Strong Medicine”: Ted Kaptchuk and the Powerful Placebo

At the beginning of the first edition of The Web that has no Weaver, published in 1983, author Ted Kaptchuk portended his eventual academic interest in the placebo:

A story is told in China about a peasant who had worked as a maintenance man in a newly established Western missionary hospital. When he retired to his remote home village, he took with him some hypodermic needles and lots of antibiotics. He put up a shingle, and whenever someone came to him with a fever, he injected the patient with the wonder drugs. A remarkable percentage of these people got well, despite the fact that this practitioner of Western medicine knew next to nothing about what he was doing. In the West today, much of what passes for Chinese medicine is not very different from the so-called Western medicine practiced by this Chinese peasant. Out of a complex medical system, only the bare essentials of acupuncture technique have reached the West. Patients often get well from such treatment because acupuncture, like Western antibiotics, is strong medicine.

Other than to wonder if Kaptchuk had watched too many cowboy ‘n’ Native American movies as a kid, when I first read that passage I barely blinked. Although the Chinese peasant may have occasionally treated someone infected with a bacterium susceptible to his antibiotic, most people will get well no matter what you do, because most illnesses are self-limited. Most people feel better even sooner if they think that someone with special expertise is taking care of them. If you want to call those phenomena the “placebo effect,” in the colloquial sense of the term, fine. That, I supposed, was what Kaptchuk meant by “strong medicine.”

Turns out I was mistaken. Let’s briefly follow Kaptchuk’s career path after 1983. In the 2000 edition of The Web, he wrote:

In the almost 20 years since the first edition, I have continued to learn and study. I have treated many patients and worked in many hospitals. I have come into contact with the many other health care systems. For the last ten years I have had a full-time academic appointment at Harvard Medical School (HMS). For four years I worked as a series consultant for a nine-hour British Broadcasting Company (BBC) television series on health care and I was sent to visit various healers on three continents. Currently, I am serving a term on the National Advisory Council of the National Center for Complementary and Alternative Medicine (NCCAM) at the National Institutes of Health (NIH) where I have the opportunity to contribute to the evolving reconfiguration of America’s pluralistic medical environment.

Kaptchuk used his newfound academic stature to celebrate all sorts of practices in that “pluralistic environment” that had, only a few years before, been widely recognized as quackery. We’ve seen his views on “Chinese medicine”; here, he and co-author (and boss) David Eisenberg wax eloquent on chiropractic:

Treatment by a chiropractor can generate a sense of understanding and meaning, an experience of comfort, an expectation of change, and a feeling of empowerment. Chiropractic’s combination of vitalist “innate intelligence” and simple mechanical explanation can give rich vocabulary for just those illnesses conventional medicine remains poorly equipped to address.

Chiropractors never have to put a patient’s pain in the category of the “mind.” They never fail to find a problem. By rooting pain in a clear physical cause, chiropractic validates the patient’s experience.

Chiropractic is in no sense passive; it is, from the start, engaged. Except when contraindicated (as in patients with neoplastic disease and those with extreme osteoporosis), some form of therapy is almost always indicated. For most symptoms, there is a suitable manipulation or a designated mode of redress.

From the first encounter on, chiropractors generate different expectations from conventional physicians. Because conventional practitioners assume that back pain, in the absence of systemic signs, is likely to be self-limited, it is not unusual for a patient to wait weeks for an appointment with a specialist or for a radiographic diagnostic assessment. Because a chiropractor believes that back pain is both explicable and amenable to treatment, a patient can usually obtain an appointment within 24 hours of a telephone call. The message of empathy, urgency, comprehension, and support conveyed by such a rapid response is reassuring and provides a heightened sense of care and compassion.

Chiropractic’s ultimate lesson may be to reinforce the principle that the patient-physician relationship is fundamentally about words and deeds of connection and compassion. Chiropractic has managed to embody this message in the gift of the hands.

Did the authors realize that they were writing a perfect description of quackery? Apparently not: in the fashion of duplicity-by-euphemism that is standard for quackademia, they refer to the overwhelming evidence that chiropractic is a pseudoscience as “contradictions and tensions [that] exist…between chiropractic and mainstream medicine.”

At some point during the nearly 20 years between the first and second Web, Kaptchuk seems to have begun to wonder if acupuncture and other offerings of ‘various healers’ act via “enhanced placebo effects.” Well, duh. But wait: for Kaptchuk, “placebo” doesn’t mean what it means to me or Mark Crislip or even to Peter Moran or Fabrizio Benedetti. For Kaptchuk, as we’ve already seen, placebo seems to mean something similar to what it means to Daniel Moerman. Placebo is “strong medicine”—maybe even stronger than “Western” medicine. When a trial of “EDTA Chelation Therapy for Coronary Artery Disease” was proposed at a 2000 meeting of the NCCAM Advisory Council,

Dr. Kaptchuk noted that they could find the equivalent of the placebo effect at work here, but both could be more beneficial than conventional therapies…The vote was 11 in favor, none opposed, and two abstentions, so the concept passed for funding.

Thus was spawned the most dangerous, unethical, and expensive boondoggle yet wrought by that August Body, paid for by you and me (Dr. Briggs, are you reading this? Are you, Dr. Nabel?).

Placebo Power and “Power Relationships”: a Post-Modern Deconstruction of the RCT

Kaptchuk’s interest in the placebo seems to have arisen from his displeasure with modern medicine’s ‘privileged’ status, which had presented a barrier to establishing himself as a ‘doctor’ back in the USA. It must have been particularly rankling to take a back seat to the despised, reductionist paradigm, with its insistence on evidence from randomized, often placebo-controlled trials (RCTs). In the 2001 HMS conference to which I alluded at the end of the previous part of this series, Kaptchuk referred to the aforementioned “well, duh” attitude as an “accusation of placebo therapy,” and to “CAM” as a “domain for methodological doubt.” He based this on his suggestion that various “CAM” treatments may have “enhanced placebo effects,” effects that are even stronger than specific biomedical treatments. “Is a placebo effect,” he asked, “only a placebo?” He explains this further in an appendix to the 2000 edition of The Web:

Until about 1955, both East Asian medicine and biomedicine shared similar explicit standards for determining the acceptability of a medical intervention: legitimacy was determined by beneficial outcome. While physicians in both traditions also spoke of medicine based on ‘proven’ or ‘recognized’ principles, and inevitably were influenced by cultural assumptions, acceptable therapy was ultimately expected to deliver relief if not cure.¹

In the years after World War II, biomedicine underwent a dramatic shift. Major reforms were undertaken in medical research that sought to free therapeutic evaluations from human judgment based on clinical experience and impressions…The apparatus of the double-blind randomized controlled trial (RCT) gradually established itself as the ‘gold standard’ for determining legitimate therapy.

A major shift had occurred. A medical intervention was now scientifically justifiable only if it was superior to a placebo: method became more important than outcome. Superiority to placebo replaced ability to confirm health benefits as judgment criteria in medicine. Ideally, an acceptable treatment was now a relative outcome that could be isolated, disguised, and compared to the entire matrix of effects embodied in an identical healing ritual lacking this single ingredient.* Healing was no longer an absolute outcome that comprised multiple interactive dimensions. Biomedicine reconceptualized legitimate healing as “a cause and effect relationship between a specific agent or treatment and a specific biological result” [citation to an article in Medical Anthropology]. Therapeutics that imitated the laboratory and depended on an isolatable, precise, and single mechanism were privileged. For biomedicine, the masked RCT significantly realigned the power relationships between ‘art’ and ‘science’ in medicine, as it was itself a product of this transformation.

*Theoretically, it was now possible for a ‘proven’ drug to have a smaller effect size on a particular ailment than a ‘debunked’ therapy (i.e., an intervention that equals its dummy control).

Footnote 1:

The explicit notion that medicine was acceptable because of positive outcomes was significantly supplemented by the often-unacknowledged process whereby legitimate therapy was determined by prestige, cultural associations, and sociopolitical power relationships. For example, biomedicine was adopted by Asian and African countries most often not because of medical outcomes but because of the prestige associated with the other developments of science. James Nelson Riley provides many examples where ‘the efficacious therapeutic techniques that Western medicine has gathered from modern science have come much more recently than the zeal for promoting science within medicine and for exporting somewhat scientific medicine to other cultures…An ideological commitment to science antedates pragmatic benefits from science.”…For a discussion on how ‘efficacy’ can be seen as a linguistic tool to control medical knowledge, see Elizabeth Hsu, ‘The Polyglot Practitioner [etc.]‘…

In a 1998 article titled “Powerful Placebo: the dark side of the randomised controlled trial,” Kaptchuk characterizes the RCT as “self-authenticating”:

In a self-authenticating manner, the double-blind RCT became the instrument to prove its own self-created value system. This shift from emphasizing outcomes to the purity of the means directly parallels developments in medical ethics where ‘informed consent’ replaced ‘beneficence’ as the pinnacle of the value system.

(Hmmm? He offers no support for the last assertion). In that article he also asserts, with justification, that the premise for the placebo-controlled trial, as advanced by Beecher in 1955,

…took for granted that the active drug response results partly from a placebo effect and that the placebo effect buried in the active arm is identical to the placebo effect of the dummy treatment. The placebo was a single and stable ‘power’ that behaved in a consistent manner.

Kaptchuk, with reason, disagrees with that assumption. With less reason he calls for research to “disentangle the ‘non-specific’ and ‘art-of-medicine’ aspects of healing and therapeutic evaluation.”

In a 2001 article titled “The double-blind, randomized, placebo-controlled trial: gold standard or golden calf?” Kaptchuk again characterizes the RCT as “authenticating itself.” He quotes the philosopher Ian Hacking:

The truth is what we find out in such and such a way. We recognize it as truth because of how we find it out. And how do we know that the method is good? Because it gets at the truth.

In that article Kaptchuk tries to make what might seem, at first glance, a good case for the RCT not being all it’s cracked up to be. He begins with a mildly surprising statement:

Until very recently, there was a widespread perception that the absence of the usual components of the masked RCT will “exaggerate estimates of treatment effects“…It was generally believed that identical treatments “are much less likely to be judged efficacious in double-blind, randomized trials than in uncontrolled case series or unblinded, ‘open’ comparisons with contemporaneous or historical series of patients.”

Yes, but there is still that widespread perception. This “discrepancy argument,” as Kaptchuk calls it, is the basis for the RCT’s sitting atop the Evidence-Based Medicine (EBM) evidence hierarchy, and if it weren’t true it would certainly throw the whole project into disarray. Kaptchuk appears to do exactly that by citing several reviews (here, here, here, here, and here, for example) that “make a compelling case that poor methodology could either overestimate or underestimate treatment effects” (emphasis added). He asserts that such findings have led to a “modified challenge to the discrepancy argument,” such that one study’s authors have argued that “research design should not be considered a rigid hierarchy.” He mentions that the editorial accompanying that and a similar study “has cast doubt on the validity of these conclusions,” without revealing the basis of that doubt; according to Kaptchuk, this shows that “The discrepancy debate has intensified.”

Reading Kaptchuk’s essay you might think, if you were unaware of the crescendo of EBM-chatter in the medical literature over the previous 10 years, that he had discovered a heretofore barely acknowledged, critical, even subversive (as he might put it) Truth about the nature of ‘evidence’ in modern medicine; a truth that would, among other things, liberate “CAM” from the ‘microfascism’ of EBM. Just at that point, however, Kaptchuk adds an afterthought:

In terms of blinding, recent comparative assessments remains [sic] consistent with the older evidence. Three studies showed that double-blind RCTs yielded significantly smaller treatment results than trials that were not double-blind. Also, three studies showed that successful concealment of randomization (compared with inadequate concealment of randomization) produces smaller treatment outcomes. Proper masking seems to create distinct outcomes; the discrepancy argument is intact in this domain.

In this domain? IN THIS DOMAIN? Excuse the yelling, but what “domain” was he talking about in the first place? In other words, after spending about 1000 words trying to convince us that “the usual components of the masked RCT” don’t do what we think they do, he off-handedly mentions that, er, the RCTs that he’d been talking about didn’t have those usual components; meaning, like, they weren’t masked. Which was exactly the point that the mysterious “accompanying editorial” had made.

The discrepancy debate has fizzled. And so, to the satisfaction of most observers, has the “self-authenticating” gambit.

Never mind: in that and other essays, Kaptchuk goes to great lengths to argue that masking, “investigator self-selection,” randomization, informed consent, and other standard research methods introduce “possible systematic errors intrinsic in even an ideal RCT” (in a term reminiscent of another current pseudoscience that closely resembles Quackademia, he dubs this process “irreducible uncertainty”). In this he is undoubtedly correct to an extent, but not to the extent that RCTs must be judged no more valid measures of efficacy than are less rigorous tests claiming larger treatment outcomes. Kaptchuk might protest that he does not explicitly make that claim, but I would respond that he certainly implies it, over and over again.

Kaptchuk is in the company of other influential academic “CAM” advocates, such as Andrew Weil, David Katz, Wayne Jonas, and Mehmet Oz, who’ve been making similar, special pleadings for years.

“The Need to Act a Little More ‘Scientific’ “

We’ve previously heard this from Kaptchuk:

When people became interested in alternative medicines, they asked me to help out at Harvard Medical School. I realized that in order to survive there, one had to become a scientist. So I became a scientist.

I won’t take Kaptchuk too much to task for the obvious faux pas in that statement, even though it is entirely consistent with everything that you’ve read above; he has cultivated a pseudo-bumbling, Columbo-like persona for the purposes of some public presentations, and he came of age during the 1960s, when it was fashionable to be utterly serious while pretending not to be (think: Bob Dylan). So it’s quite possible that he pursued science not merely to survive at Harvard Medical School, but because he hoped to learn something about nature—a ‘domain’ about which, as demonstrated by the first edition of The Web, he’d been quite naive. But did Kaptchuk become a scientist? He seems to have learned a fair amount about clinical trial design and probably a fair amount about frequentist statistical inference. Does that make someone a “scientist” these days? What about nature?

I’m doubtful. In addition to his difficulty in distinguishing science from cultural ‘narrative,’ Kaptchuk stumbles in other fundamental ways. In a 2003 article titled “Effect of interpretive bias on clinical research,” he offers this list:

Definitions of interpretation biases

Confirmation bias—evaluating evidence that supports one’s preconceptions differently from evidence that challenges these convictions

Rescue bias—discounting data by finding selective faults in the experiment

Auxiliary hypothesis bias—introducing ad hoc modifications to imply that an unanticipated finding would have been otherwise had the experimental conditions been different

Mechanism bias—being less sceptical when underlying science furnishes credibility for the data

“Time will tell” bias—the phenomenon that different scientists need different amounts of confirmatory evidence

Orientation bias—the possibility that the hypothesis itself introduces prejudices and errors and becomes a determinate of experimental outcome

Some of his discussion is perfectly reasonable; some is further conflation of science and cultural narrative; some is confusing and even contradictory; some is selective in a way that suggests an agenda:

Good science inevitably embodies a tension between the empiricism of concrete data and the rationalism of deeply held convictions. Unbiased interpretation of data is as important as performing rigorous experiments. This evaluative process is never totally objective or completely independent of scientists’ convictions or theoretical apparatus…

Well, which is it? How can unbiased interpretation be so important if it never happens?

Science demands a critical attitude, but it is difficult to know whether you have allowed for too much or too little scepticism. Also, where is the demarcation between the background necessary for making judgments (such as theoretical commitments and previous knowledge) and the scientific goal of being objective and free of preconceptions?…

Interpretation is never completely independent of a scientist’s beliefs, preconceptions, or theoretical commitments.

Again, which is it? Why propose a goal of being free of preconceptions if scientists always have preconceptions?

Evidence is more easily accepted when supported by accepted scientific mechanisms. This understandable tendency to be less sceptical when underlying science furnishes credibility can give rise to mechanism bias. Often, such scientific plausibility underlies and overlaps the other biases I’ve described. Many examples exist where with hindsight it is clear that plausibility caused systematic misinterpretation of evidence. For example, the early negative evidence for hormone replacement therapy would have undoubtedly been judged less cautiously if a biological rationale had not already created a strong expectation that oestrogens would benefit the cardiovascular system. Similarly, the rationale for antiarrhythmic drugs for myocardial infarction was so imbedded that each of three antiarrhythmic drugs had to be proved harmful individually before each trial could be terminated. And the link between Helicobacter pylori and peptic ulcer was rejected initially because the stomach was considered to be too acidic to support bacterial growth…

Let me explain. Kaptchuk is correct that unbiased interpretation can never happen, so it makes no sense for him to propose otherwise on the same page. Biased interpretation is inevitable, moreover, not merely for the mundane or mischievous reasons that Kaptchuk implies, but because it makes no sense to attempt “unbiased interpretation.” That was shown by Thomas Bayes more than 200 years ago.

Kaptchuk seems to know something about this: in the Comments section he writes, without further explanation, “the arguments presented are obviously compatible with a subjectivist or bayesian framework that formally incorporates previous beliefs in calculations of probability.” If he understands what Bayes demonstrated—that attempting to interpret the results of an experiment without regard to previous knowledge is logically incoherent—he withholds it from his readers, whom he has nevertheless presumed to be largely naive.

My view: if he understood the point when he wrote his treatise, he was dishonest in not explaining it and in not citing at least one pertinent article, such as Steven Goodman’s (which I’m willing to bet he had read). If he didn’t understand the point he should have withheld his paper.

Kaptchuk’s portrayal of interpretive bias being based on “beliefs, preconceptions, or theoretical commitments” implies that such bias is more about cultural assumptions and “power relationships” than about science. That, I believe, is his agenda: to suggest that such ‘commitments’ are likely to ‘privilege’ the status quo and to distract from the truth, thus casting enough doubt on the entire project to give at least a tentative free pass to, well, almost anything.

The reality of science is somewhat different: While it is undoubtedly true that an individual scientist’s biases often reflect personal wishes (not so much cultural assumptions), many of the biases of scientists are not only inevitable, but desirable. Kaptchuk’s discussion of “mechanism bias” is notable for its selection of examples that serve only to condemn such bias. A different sort of example would demonstrate an important truth about science, a truth that Kaptchuk either doesn’t understand or would prefer to conceal from his readers.

Such an example would be a ‘positive’ trial of a claim that, if deemed valid, would require discarding a firmly established principle of nature, such as the 2nd law of thermodynamics—a principle that is not merely a preconception or a “theoretical commitment,” but a fact based on data vastly more rigorous and voluminous than the entire body of biomedical literature. In such a case, not to invoke plausibility would (and has, many times) “cause systematic misinterpretation of evidence.”

The converse of what Kaptchuk calls “this understandable tendency to be less sceptical when underlying science furnishes credibility” is “this understandable tendency to be more sceptical when underlying science furnishes no credibility,” which I would call “good science.” Does Kaptchuk even have a sense of what to look for in a hypothesis? Simplicity, conservatism, fruitfulness, and scope, for example?

Kaptchuk seems to believe his own hype: “method” is more important than science itself. I don’t think that he appreciates, or at least acknowledges, some of the fundamental aspects of science. I also suspect that he isn’t forthcoming about some of the things that he does know, if they don’t suit his agenda.

To be fair, I’ll report that Kaptchuk specifically denies, in the final paragraph of his “interpretive bias” paper, one of the accusations that I’ve made against him:

I do not mean to reduce science to a naive relativism or argue that all claims to knowledge are to be judged equally valid because of potential subjectivity in science. Recognition of an interpretative process does not contradict the fact that the pressure of additional unambiguous evidence acts as a self regulating mechanism that eventually corrects systematic error.

I can’t help but be reminded of another of Kaptchuk’s articles, quoted here a couple of years ago. Its abstract:

The advent of scientific research on complementary and alternative medicine (CAM) has contributed to the current state of flux regarding the distinction between biomedicine and CAM. CAM research scientists play a unique role in reconfiguring this boundary by virtue of their training in biomedical sciences on the one hand and knowledge of CAM on the other. This study uses qualitative interviews to explore how CAM researchers perceive and negotiate challenges inherent in their work. Our analysis considers eight NIH-funded CAM researchers’: (1) personal engagement with CAM, (2) social reactions towards perceived suspiciousness of research colleagues and (3) strategic methodological efforts to counteract perceived biases encountered during the peer review process. In response to peer suspicion, interviews showed CAM researchers adjusting their self-presentation style, highlighting their proximity to science, and carefully ‘self-censoring’ or reframing their unconventional beliefs. Because of what was experienced as peer reviewer bias, interviews showed CAM researchers making conciliatory efforts to adopt heightened methodological stringency. As CAM researchers navigate a broadening of biomedicine’s boundaries, while still needing to maintain the identity and research methods of a biomedical scientist, this article explores the constant pressure on CAM researchers to appear and act a little more ‘scientific’.

Placebo Research: a Critical Positive Impact on Health Care?

What about Kaptchuk having mostly (not entirely) abandoned his advocacy of the purported specific effects of highly implausible medical claims, in the last decade or so, in order to pursue his interest in their non-specific effects? Well, perhaps that has separated him from the rest of the quackpack, to an extent, and for that, some observers feel he deserves credit (Peter?). On the other hand he is no Fabrizio Benedetti, who can call quackery “quackery,” who insists that real, neurobiological placebo effects can’t be studied in clinical trials, not only because of regression to the mean and spontaneous remission, but because of biases introduced by both investigators and subjects (as was likely the case in Kaptchuk’s albuterol vs placebo for asthma trial that triggered this series), and who recognizes that any potential clinical applications beyond the obvious “caring for the patient” are apt to be elusive:

I use the placebo response as a model to understand how our brain works. I am not sure that in the future it will have a clinical application. This is a very important point—a translational research: can we use placebo in routine clinical practice? Well, sometimes it works; but that’s not the important problem. The important thing right now is to understand how our brain works. And I would say the placebo response is a fascinating phenomenon, because it is a sort of melting pot of concepts, of ideas for neuroscience. So, if you use a placebo response, you can understand a lot of brain functions, like anxiety, like social learning, classical conditioning, reward mechanisms, and so forth. So, the clinical application is—I think in English you say it is a ‘different kettle of fish.’

Kaptchuk, on the other hand, in keeping with his belief that the placebo is ‘strong medicine,’ argues that

Learning how to enhance medical outcomes via placebo research could have a critical positive impact on health care and scientific knowledge.

It takes only a bit of perspective to realize that learning how to enhance medical outcomes via placebo research, questionable in itself, is unlikely to have more than a minor impact on health care. That perspective consists of the state of health anywhere in the world prior to the advent of scientific medicine and public health. Kaptchuk, who cut his teeth on “Traditional Chinese Medicine,” might review, in William H. McNeill’s Plagues and Peoples, the appendix titled “Epidemics in China” (part of it can be found here). Does Kaptchuk really believe, or expect the occasional sophisticated reader of The Web to believe, that “biomedicine was adopted by Asian and African countries most often not because of medical outcomes but because of the prestige associated with the other developments of science”?

On Contradiction

For a preview of the final posting in this series, consider these statements:

In clinical practice a sham medical procedure would be fraudulent and deplorable; in research such activities can be legitimate, and this outlying research practice underscores the important ethical differences between clinical trials and medical care.

—Miller and Kaptchuk, “Sham procedures and the ethics of clinical trials”

…both traditional and sham acupuncture were shown to provide clinically significant improvement in back pain and function, as compared with either no treatment or usual medical care, suggesting that the benefit from acupuncture probably derives from the placebo effect. Although this alternative standard for evidence-based validation deserves critical scrutiny and serious debate, we believe that it reflects a more patient-centered perspective for symptomatic treatment. Patients who have continuing pain are interested in pain relief; they are unlikely to care whether this derives from the inherent ability of a treatment to modify pathophysiology or its propensity to promote a significant placebo response.

—Miller and Kaptchuk, letter to the New England Journal of Medicine following the report discussed here, here, and here.

<!--<!--