Certain topics covered by Science-Based Medicine are perennial. They keep coming up again and again, sometimes unchanged from past instance when we’ve written about them, sometimes changed subtly—or not-so-subtly—to match the times, although with the core narrative intact. If there’s one topic covered on SBM over and over and over again since the beginnings, it’s the placebo effect, or, more properly, placebo effects, given that there is not just one placebo effect. In general, going back to the very beginning of this blog in 2008, we have discussed how those promoting medical interventions that have no effect but placebo effects, such as reiki, homeopathy, acupuncture, and any number of alternative medicine staples, use, misuse, and abuse the science of placebo effects in order to promote a narrative that their quackery can be just as effective as real science-based medical interventions. Recently, I’ve seen an article popping up on social media and emailed to me on The Conversation entitled Placebo effect can work as well as real medicine – but your body may need permission to use it. It’s by an associate professor of biology at Tufts University named Phil Starks, and regular readers will no doubt recognize from the title alone what sort of distorted narrative about placebos that the article will promote. However, given my desire to take a mental health break from examining in depth yet more of the depredations and destruction committed against science and public health programs of the federal government on a near-daily basis by the current administration, I thought that discussing this article and placebo effects would be the perfect temporary antidote to the near constant flood of anxiety-producing news coming from the Department of Health and Human Services under Robert F. Kennedy, Jr.
In other words, this week I’m going to party like it’s 2009. (Unfortunately, SBM didn’t exist in 1999, thus necessitating a slight change the allusion.)
The magical, mystical super-powerful, just-as-good as medicine placebo
Before I proceed, I think I’ll simply very briefly summarize my general take on these placebo narratives. It’s probably not surprising that many physicians and scientists find the idea that the mind has the power to heal the body, often referred to as “self-healing” or “mind-body healing,” and an explanation invoking placebo effects to “explain” this “power” so seductive. After all, who wouldn’t want to be able to cure themselves simply by willing it to be so? It’s a concept that, like so many concepts in “complementary and alternative medicine” (CAM), goes far back into ancient times and stretches forward to today in ideas like The Secret, which goes quite a bit beyond the whole idea of “mind-body healing” or healing yourself because you wish it to be so, and declares that you can have virtually anything you want simply by thinking the right thoughts. In fact, to me it appears that the “powerful placebo” is being drafted in the service of supporting what are, at their core, mystical beliefs far more than science, dressing them up in what I like to call “biobabble,” which I define as the medical or biological equivalent of Star Trek technobabble.
In that vein, I note that Prof. Stark’s article doesn’t start very promisingly (as you might imagine). Right off the bat, Prof. Starks wants to convince you that placebo effects are totally biologically real and able to heal with no side effects:
The first time the placebo effect really got under my skin was when I read that roughly one-third of people with irritable bowel syndrome improve on placebo treatments alone. Usually this statistic is presented as a fascinating quirk of medicine. My reaction was anger.
Humanity possesses an extremely effective treatment, with essentially zero side effects – and patients need someone else’s permission to use it.
The placebo effect refers to the improvements in symptoms that patients experience after they’re given an inert treatment like a sugar pill. Driven by expectation, context and social cues rather than pharmacology, the placebo effect is often dismissed as all in the mind. But decades of research have shown it is anything but imaginary.
“No side effects”? Has Prof. Stark never heard of nocebo effects?
I note that the first article linked to is another article in The Conversation entitled In research studies and in real life, placebos have a powerful healing effect on the body and mind. Tellingly, one key passage in this article reads:
But studies have shown that the placebo effect is so strong that many drugs don’t provide more relief than placebo treatments. In those instances, drug developers and researchers sometimes see placebo effects as a nuisance that masks the treatment benefits of the manufactured drug. That sets up an incentive for drug manufacturers to try to do away with placebos so that drugs pass the FDA tests.
Even though Prof. Starks doesn’t (quite) invoke this narrative, which borders on conspiracy theory, he clearly at least tacitly approves of it, given that he linked to an article expressly promoting this idea. Be that as it may, this is as good a point as any to reiterate that placebo effects are most prominent for medical diseases and conditions that have a high degree of subjectivity in symptomatology, such as pain. For objective symptoms, placebos tend not to do so well. I like to point out that the most objective of all is being dead rather than alive and that you will never find a study showing that one-third of patients receiving a placebo live longer. If there is a study that has convincingly demonstrated that a placebo can detectably increase overall survival (OS) in patients with cancer compared to a no treatment group that can’t be attributed to flaws in the clinical trial design, the Hawthorne effect among the healthcare providers taking care of the patients, or observer bias among the researchers, I have yet to find it, and I have looked.
Because I’m a cancer surgeon, another example that I like to use is tumor shrinkage. Chemotherapy, targeted therapies, and radiation, for example, cause reproducible, measurable tumor shrinkage in many, many studies; placebos do not. The main reason for blinding in a clinical trial of a cancer chemotherapy, for instance, would be to prevent observer bias. If, for instance, radiologists knew which group a given patient was in, they might be subtly and unconsciously measure tumor response to be better in the group receiving the experimental drug. The same is true for many other pathological and biological measures of disease.
Again, regular readers will recognize some very common maximalist (and exaggerated) tropes about the “power of placebo.” Trope #1: Placebo effects result in powerful “healing” as effective as actual science-based medicine.
Longtime readers will remember that a favorite narrative of people promoting unscientific placebo medicine is that placebo effects are as powerful as “real” medicine. Prof. Starks even leads with his first section after the introduction entitled Medicine works, even when it isn’t medicine. Let’s see how many tropes we recognize:
The placebo effect is so reliable that researchers must account for it in nearly every clinical trial.
Yes, and no. Again, in clinical trials looking at “harder” endpoints, such as death, accounting for placebo effects is much less critical than in clinical trials that look at endpoints that are strongly affected by emotions, perception, and the like. I note that the link included goes to a page on the website of the National Center for Complementary and Integrative Health (NCCIH), formerly the National Center for Complementary and Alternative Medicine (NCCAM), dedicated to placebos. NCCIH, as longtime readers will know, has been the primary federal source funding the rise of quackademic medicine and the promotion of quackery in medicine, primarily academic but also in non-academic settings. What, you might ask, does that webpage feature? Why, it’s a link to a video interview with Harvard placebo researcher Ted Kaptchuk, a name so ubiquitous in the promotion of placebo medicine and one that’s such a blast from the past as a topic on this blog that he has his own tag. Kaptchuk is, as some might remember, arguably the foremost promoter of the idea of “placebos without deception.” Does Prof. Stark buy into the concept of placebos without deception? Of course he does! We’ll get into that later, but I can’t resist a quick spoiler: Every “open label placebo” study suggesting the possibility of placebo effects without deception, without in effect hiding from patients in a clinical trial which group they are being randomized to, has involved deception elsewhere in the study design.
Next up from Prof. Starks:
When testing a new drug, scientists compare its effects to what patients experience on a placebo treatment like sugar pills, saline injections or sham surgery. If the drug doesn’t outperform the placebo, it rarely reaches the public. Placebo responses are common and powerful enough to rival active treatments.
Even surgery isn’t immune to the placebo effect. In several well-documented studies of knee procedures, patients who received sham operations – incisions without the full surgical repair – improved almost as much as those who received the real procedure.
Clearly something real is happening inside the body. But the strangest part of the placebo effect is not that it works. It’s what makes it work.
No one argues that something “real” isn’t happening in the body due to placebo effects. After all, the brain is part of the body, and even the conclusion that placebo effects are more due to perception rather than any real biological change would be a “real” effect in the brain. The problem is that the word “real” is doing some serious heavy lifting. When Prof. Starks uses the word “real,” he clearly wants you to think that placebos are causing “real” therapeutic effects, just like real science-based medical interventions. It’s a very convenient redefinition that all promoters of placebo medicine use at one point or another.
It is, of course, true that placebo responses can be common enough that they appear to rival active treatments in placebo-controlled randomized controlled clinical trials (RCTs). That doesn’t mean that they actually are “powerful enough” to rival active treatments. What it means is that, under the conditions of the given placebo-controlled RCT, the experimental treatment being tested doesn’t work, because it doesn’t produce improvements in disease symptomatology or progression beyond what the untreated patients who are part of the clinical trial experience. Indeed, acupuncturists love to take advantage of this to try to claim that acupuncture “works” when they design trials with a no treatment/normal treatment group in addition to the sham acupuncture and “true” acupuncture groups. (Usually, the sham acupuncture group involves either inserting needles in the “incorrect” locations not on acupuncture meridians and/or using special retractable needles that appear to be inserted but actually do not break the skin.) When the “true” acupuncture group doesn’t experience any greater relief of symptoms than the sham acupuncture group, often the researchers will try to argue that the proper comparator should be the no-treatment group, just the same way that the article cited by Prof. Starks claims that drug companies whose drugs fail placebo-controlled RCTs want to do.
In any event, as much as evidence-based medicine (EBM) guru Peter Gøtzsche has taken a heel turn in recent years, he is the researcher who produced two of the most compelling reviews failing to find a clinically relevant therapeutic effect attributable to placebo in any RCT, the 2001 New England Journal of Medicine article Is the Placebo Powerless? — An Analysis of Clinical Trials Comparing Placebo with No Treatment and the 2010 Cochrane systematic review and meta-analysis, Placebo interventions for all clinical conditions.
The former article concluded:
We found little evidence in general that placebos had powerful clinical effects. Although placebos had no significant effects on objective or binary outcomes, they had possible small benefits in studies with continuous subjective outcomes and for the treatment of pain. Outside the setting of clinical trials, there is no justification for the use of placebos.
And the latter concluded:
We did not find that placebo interventions have important clinical effects in general. However, in certain settings placebo interventions can influence patient-reported outcomes, especially pain and nausea, though it is difficult to distinguish patient-reported effects of placebo from biased reporting. The effect on pain varied, even among trials with low risk of bias, from negligible to clinically important. Variations in the effect of placebo were partly explained by variations in how trials were conducted and how patients were informed.
Which is what we at SBM have been saying all along, and we have seen no compelling studies or evidence to persuade us otherwise or that we need to change our minds. We keep looking though, and remain open to the possibility. I wonder if Prof. Starks is open to the possibility that the two reviews above come to accurate conclusions.
Meanwhile, another classic paper from 1997 by Kienle and Kiene, The Powerful Placebo Effect: Fact or Fiction?, reported a reanalysis of classic placebo literature, noting that reanalysis of one of the seminal papers on placebo effects, one that has been highly influential, Henry K. Beecher’s 1955 paper The Powerful Placebo, had called into question the rather—shall we say?—expansive interpretation interpretations of the data analyzed in the that paper, which is still widely cited. For their part, Kienle and Kiene reanalyzed other papers cited as being evidence for therapeutic placebo effects and found that misattribution of improvement in many of these papers could be explained by issues unrelated to placebo administration, such as:
- Spontaneous recovery. “Spontaneous improvement was a major factor in Beecher’s misinterpretation of 10 of the 15 trials. This error is wide-spread in the placebo literature.“
- Fluctuation of symptoms/regression to the mean. “Spontaneous improvement of diseases and the spontaneous fluctuation of symptoms are special forms of regression to the mean, i.e., the tendency of extreme values to move closer to the average on repeated measurement. In their interesting article, “How much of the placebo `effect’ is really statistical regression?” McDonald et al. [30] have argued, that “most improvements attributed to the placebo effect are actually instances of statistical regression.” My addition: Recall how, if a treatment is given when symptoms are at their worst, when those symptoms inevitably regress to the mean, it is often attributed to whatever intervention was given before the regression, whether that intervention had anything to do with the improvement or not.
- Additional treatment.
- Conditional switching of treatments. “When the patients felt well, they received a placebo; when they felt worse, they were switched to active treatment, or they were excluded from evaluation until they felt better again.“
- Scaling bias. In three of Beecher’s trials [2, 3, 7] there were false augmentations of placebo effects due to asymmetrical measurement scales [21]. The scales included two or more categories for improvement, and only one or even none for deterioration. Thus the scales tempted patients to falsely give too many positive reports.
There were several other contributors, but I’ll stop there for the moment, as these were the most common issues. Although research methodology has improved since 2010, many of these problems still plague placebo research, which is dominated by researchers who believe in the narrative promoted by Prof. Starks, such as Ted Kaptchuk.
“Placebos without deception”? Not again!
Prof. Starks then proceeds to put a somewhat different twist on very old placebo narratives, specifically that we must somehow give “permission” for placebo effects to work, starting by citing longstanding, well-accepted findings regarding specific factors that modulate placebo effects:
Placebo treatments tend to be more effective when delivered by credible authorities. Pills work better when prescribed by doctors wearing white coats. Expensive pills outperform cheap ones. Injections produce stronger responses than tablets.
All of this is true, but doesn’t imply what Prof. Starks thinks that these findings imply, for example:
Some researchers have even removed the deception from placebo experiments entirely. In open-label placebo studies, patients are directly told they are receiving a placebo; and yet many still report significant improvement.
No, no, no, no, no, no! We have periodically written about these so-called “open-label” placebo trials used to fuel the myth of “placebos without deception” going back to 2010, when I deconstructed the first such study that I had noticed. In my 2010 post, I noted that, although the study did tell subjects whether they were in the placebo group or not, the subjects were told that placebo sugar pills “had been shown to have self-healing properties.” The study protocol itself stated:
The provider clearly explained that the placebo pill was an inactive (i.e., “inert”) substance like a sugar pill that contained no medication and then explained in an approximately fifteen minute a priori script the following “four discussion points:” 1) the placebo effect is powerful, 2) the body can automatically respond to taking placebo pills like Pavlov’s dogs who salivated when they heard a bell, 3) a positive attitude helps but is not necessary, and 4) taking the pills faithfully is critical. Patients were told that half would be assigned to an open-label placebo group and the other half to a no-treatment control group. Our rationale had a positive framing with the aim of optimizing placebo response.
Placebo without deception? Nonsense! The deception was in the script that each subject heard. The same is true of other studies that purport to show “placebo effects without deception.” Another of Kaptchuk’s studies, which tested the drug Maxalt versus placebo pills for their effect on migraine headache symptoms, included in its language about the placebo: “Our second goal is to understand why placebo pills can also make you pain-free.” Not to understand why placebo pills might be able to make you pain-free or could possibly make you pain-free. “Can make you pain-free.” The other study, billed an open-label placebo trial for chronic low back pain, included this language:
After informed consent, all participants were asked if they had heard of the “placebo effect” and explained in an approximately 15-minute a priori script, adopted from an earlier OLP study, the following “4 discussion points”: (1) the placebo effect can be powerful, (2) the body automatically can respond to taking placebo pills like Pavlov dogs who salivated when they heard a bell, (3) a positive attitude can be helpful but is not necessary, and (4) taking the pills faithfully for the 21 days is critical. All participants were also shown a video clip (1 minute 25 seconds) of a television news report, in which participants in an OLP trial of irritable bowel syndrome were interviewed (excerpted from: http://www.nbcnews.com/video/nightly-news/40787382#40787382).
Basically, studies claiming to find “placebo without deception” tend to use variations of the same talking points from the 2010 study that I discussed, often with clips from news reports about happy study participants to prime the patients that placebo effects can be “powerful.”
So blinded by his belief in placebo effects as “powerful,” that Prof. Starks correctly interprets the example of placebo effects in animals and doesn’t seem to realize that what he is saying also applies to humans:
Consider what happens in veterinary medicine. Dogs and cats cannot believe a treatment they’re given will work; they have no concept of receiving medication. Yet when owners and vets believe an animal is being treated, they consistently report improvements in pain and mobility that medical tests do not confirm.
In one study of dogs with osteoarthritis, owners reported improvement roughly 57% of the time for animals receiving only a placebo.
The animals themselves may not have improved. But the humans caring for them perceived they had. The healing signal, it turns out, travels through the humans in the room.
No, it’s not a “healing signal.” In this case, placebo effects in animals are in reality a change in perception in the humans caring for their pets based on expectation bias. Here, Prof. Starks is so close to the correct interpretation of placebo effects. He just can’t generalize how placebo effects in animals are correctly interpreted to what happens in humans. He even correctly points out that observations by homeopaths that cholera patients treated with homeopathy in the 1800s had higher survival rates than those treated with the medicine of the times neglected to note that, when medicine is toxic and ineffective, the way that so much 19th-century medicine was, doing nothing could result in better outcomes, but he couldn’t resist adding:
Death rates were lower not because homeopathy worked but because the placebo effect – combined with not poisoning patients – was more effective than the medicine of the day.
No, placebo effects had nothing to do with it. These observations only indicate that doing nothing (homeopathy, which is water) produced better outcomes in very sick patients than poisoning them with toxic metals (which many 19th century treatments involved) did, no invocation of placebo effects from homeopathy necessary. That didn’t stop him from trying to shoehorn placebo effects into the narrative explaining this discrepancy. Indeed, when you think about it, it should be obvious that this concept of a “health governor” is basically another word for the “wisdom of the body,” a term often invoked by New Age “healers” as what their treatments somehow awaken or activate.
Placebo effects are due to “permission” to activate “self-healing”? Again, nope
As you see, Prof. Starks is very much about common longstanding incorrect narratives about what placebo effects are and what they can do. He does, however, throw in a twist in describing what he believes placebo effects to be in this framework:
The placebo effect is not a trick of the mind. It is a feature of human biology that people have largely surrendered to whoever performs authority most convincingly.
He even portrays himself as the skeptic, critical of “wellness culture”:
If belief can activate biological healing pathways, belief can also be manipulated. Charismatic figures, elaborate medical rituals and expensive treatments may produce real improvement in symptoms even when the underlying treatment is physiologically inert. That is how wellness culture works. It leverages the same social scaffolding of care to trigger the body’s internal pharmacy, regardless of whether the treatment itself does anything.
Amazingly, that first link goes to an article by critic and skeptic of alternative medicine, Jonathan Stea, entitled The multi-trillion-dollar wellness industry is making us sick, and the second link goes to another article in The Conversation, Why do we fall for wellness scams? Our cultural biases and myths are often to blame by Jesse Ruse. Both are strong articles. Neither supports the existence of powerful healing effects from placebo medicine or argue for the use of placebo. Moreover, strictly speaking, Prof. Starks is not entirely wrong when he argues that wellness culture leverages the “same social scaffolding” upon which placebo effects depend. He just leaps to the wrong conclusion:
The placebo effect is often celebrated as proof that the mind can heal the body. But I believe that may not be its most interesting lesson. It also reveals that human physiology evolved to take its cues from other people. Your brain, immune system, and pain response are not isolated machines. They are deeply intertwined with social signals, expectations and trust.
In a world filled with doctors, advertisements, wellness influencers and elaborate medical rituals, that insight is both fascinating and profoundly maddening. People are walking around with one of the most powerful healing systems ever documented locked inside them, and they can reliably access it only when someone in a position of authority gives them permission.
Wait, what? As I said, this is a bit of a new twist on an old argument. The way that this claim has traditionally been framed is that authority figures, such as shamans and priests in the distant past and doctors in white coats right now, are taken more seriously because of their perceived authority, so that their pronouncements that a given treatment can work are more effective in changing the perception of symptoms in a way consistent with placebo effects. Here, Prof. Starks is portraying placebo effects as “one of the most powerful healing systems ever documented” that somehow needs “permission” from authority figures and medical rituals, claiming that it’s all due to our inability to control our stress responses:
You can talk back to the stress response, consciously reappraising the threat – in other words, reframing a looming deadline not as a catastrophe but as a manageable challenge – to help quiet it. But notice what you cannot do: You cannot simply decide to activate your placebo response. You cannot will yourself to release pain-relieving endorphins by believing hard enough in a sugar pill. For that, you still need the ritual, the white coat, the authority figure. You need someone else.
The stress response, misfiring as it is, remains yours. The placebo response has been outsourced: not because it wasn’t always social, but because even now, people still can’t seem to access it on their own.
Did Prof. Starks cite any evidence for this assertion? Of course he didn’t. There isn’t any, not really. It’s his interpretation of the change in perceptual filter that leads patients to perceive improvement in their symptoms, whether or not there really was any improvement, or whether the placebo intervention caused any improvement in the pathophysiology of the disease or condition causing their symptoms. He does cite a paper while claiming:
Some researchers have proposed that placebo responses reflect a kind of biological health governor: a system that regulates when the body invests heavily in recovery. Cues from trusted individuals may be exactly the signal the body waits for before committing resources to recovery. A caregiver’s reassurance, a physician’s authority and the rituals of medicine may tell the body that conditions are finally stable enough to devote energy to healing.
If that interpretation is correct, the placebo effect is not a trick of the mind. It is an ancient biological system responding to social information.
The paper cited in the link in the passage above is remarkable to me, mainly because a highly reputable journal, Current Biology, published such an evidence-light commentary that makes some really amazing conclusions based on little data. (Seriously, this article almost demands a post of its own.) The authors propose that there is a “health governor” in our biology that weighs the costs versus the benefits of activating healing systems:
Given the real benefits, it will usually be adaptive to deploy a particular defense despite the costs. Better be in pain that reduces mobility than risk opening a wound. Better to put up with a fever that helps kill bacterial parasites than risk the infection flaring up out of control. But, equally, given the real costs, there will also be times when deploying the defense would actually be inadvisable. Better not to mount a full-blown immune response when there’s a risk of famine. Better not feel the pain from a sprained ankle when escaping from a predator.
In general, a particular healing measure can be expected to be adaptive only when the anticipated benefits are likely to exceed the anticipated costs. In short, self-healing involves — or ought to involve — a judgment call. It follows that, from early on in the evolution of the vis mediatrix, there must have been selective pressure to develop a secondary control system that can forecast the costs and benefits, and steer the vis accordingly.
Who makes this “judgment”? According to the authors, it’s a system that they postulate and dub the “health governor”:
What this health governor has evolved to do is to perform a kind of economic analysis of what the costs and benefits of self-cure will be, taking account of how dangerous the situation seems to be right now, what can be expected to happen next, what reserves there are in store, and so on. In effect, the health governor acts like a good hospital manager who, with finite resources, has to try to provide a service that maximizes patient satisfaction in the short-term while minimizing long-term risks. Crucially, he needs to be able to make an informed guess about future needs and opportunities, so that he can budget accordingly
The evidence that this “health governor” functions this way—or even exists at all? I didn’t see anything in the commentary that I considered compelling evidence. The paper is largely speculative and concludes that the reason “permission” is needed from societal influences and authority figures for placebo medicine to heal is because we are stuck in the past, evolutionarily speaking, with a health manager that is too conservative in terms of devoting energy and resources to “self-healing”:
The paradox is this. When people recover from illness under the influence of fake treatments, they must of course in reality be healing themselves. But if and when people have the capacity to heal themselves by their own efforts, why do they not simply get on with it? Why ever should they wait for third-party permission — from the shaman or the sugar pill — to heal themselves? How strange that people should be condemned to remain dysfunctionally sick just because — as must still often happen — they have not received permission.
It does indeed seem a puzzle for evolutionary biology. And we contend that the only explanation can be that the health governor remains stuck in the past, with rules for health expenditure that are overcautious for the modern environment.
Again, no solid evidence is presented for this hypothesis. It’s all speculative, such as when the authors suggest an explanation for why we are still “stuck in the past” with respect to self-healing: “A more intriguing one would be that modern humans are descended from a sub-population of Homo sapiens whose conservative approach to health allowed them to survive catastrophes that left their more liberal cousins too exposed [20].” The article referenced says nothing about this. It is a paper on human evolution that examined whole-genome sequences to infer when various human populations started to diverge genetically.
The times might change, but the narrative remains the same
With the rise of the “make America healthy again” (MAHA) movement, there has been a lot more discussion of and tolerance for the sorts of alternative medicine interventions that we used to discuss as being nothing more than placebo medicine. It should therefore not be unexpected that old narratives about placebo effects as being the activation of “powerful self-healing systems” in the human body should resurface with new wrinkles. I’m not saying that Prof. Starks is MAHA. I have no idea, never having heard of him before. What I am saying is that his is a very old narrative about placebos that can be traced back at least to the 1970s, when stories of “acupuncture anesthesia” started reaching the West from Communist China, to great buzz in the media. A more science-based perspective is not that placebo effects represent some sort of activation of an innate and powerful “self-healing” system in the body (whether “permission” needs to be given through authority figures, rituals, and other external influences or not), but rather derive from artifacts from the clinical trial process and changes in perception of their condition by those receiving the placebo.
Eighteen-plus years on since SBM first came into existence, neither I nor any of the others who have contributed to SBM have found compelling evidence to view placebo effects in any other way.