[ed. note: Michael Jorrin, who I like to call “Doc Gumshoe” (he’s a longtime medical writer, not a doctor) writes about health and medicine topics for us from time to time. His words and thoughts are his own.]
The word “placebo” is the first person singular, future tense, of the Latin verb “placere,” to please or be pleasing. It means “I shall be pleasing.” It was a familiar word in antiquity, with no medical meaning, as the first word of a dirge for the dead, Placebo Domino in regione vivorum – “I shall be pleasing to the Lord in the lands of the living.”
From that definitely respectable point of departure, the word took on a somewhat negative tone. To sing a placebo meant to be a flatterer, a sycophant, a servile and perhaps deceptive sort of fellow, the kind we would now call a con man or a hustler.
The first citation in my Oxford English Dictionary that relates in any way to the current medical use dates from 1811 and comes from Hooper’s Medical Dictionary: “Placebo – an epithet given to any medicine adapted more to please than benefit the patient.” And then this one, from the New York State Journal of Medicine in 1946: “You cannot write a prescription without the element of a placebo. The fact that it is signed by a doctor, that it has to be taken to a drug store to be made up, that it has, perhaps, a bad taste, all of those are placebo elements in a prescription.”
Now we’re getting to it. The OED’s definition is “a substance or procedure which a patient accepts as a medicine or therapy but which actually has no specific therapeutic effect for his condition, or is prescribed in the belief that it has no such activity.” In other words, according to that definition, the physician knows that the pill is a dummy, but the patient thinks it’s real. When the physician gives the patient a placebo, the physician is engaging in deception, even though the physician’s intent is to benefit the patient in some way.
And from the patient’s belief that he or she has been given a bona fide, effective medication or treatment, flows some kind of benefit, by some mysterious process that has been called the placebo effect.
That, at least, is one way of defining the placebo effect, as a deception by the physician and unwitting acceptance of that deception by the patient. But there is another way of describing the placebo effect. Some workers in the field now believe that patients derive a benefit from the experience of being treated in a therapeutic setting, whether or not the specific treatment has a clinical benefit. No deception is intended by the clinician, and no acceptance of deception needs to take place on the patient’s part. We’ll discuss that version of the placebo effect later on.
A warning to citizens of Gumshoeland who hope that despite Doc Gumshoe’s repeated disclaimers that he does not recommend stocks, nonetheless there will be a tiny hint of a pharmaceutical or other company that might somehow cash in by cashing in on the placebo effect. Be it known that there will be no hints, veiled or otherwise. This is not to say that at some point a way to make serious money legitimately from the placebo effect will not be found, and when (and if!) such a way is found, Doc Gumshoe hopes that he will be able to spot it and let you all in on it.
So why should you be interested in the placebo effect? Two reasons. One, because it is difficult if not impossible to assess the efficacy of any treatment or intervention without understanding the placebo effect. And, two, because at some point in our lives, it’s almost certain to affect us personally, and it’s kind of nice to know what’s going on, right?
A bit of background: what put Doc Gumshoe on this topic?
From the first moment that I began to put my toes in the swirling waters of medical writing, I was aware of placebos and the placebo effect. And it has become increasingly clear that the two terms have somewhat different meanings. That is, the placebo effect, as it is discussed in clinical studies of drugs and various types of interventions, is not always caused by the kinds of substances that were traditionally used as placebos. There’s a lot of confusion out there.
It is certainly the case that physicians have, whether knowingly or not, prescribed a lot of placebos. It used to be fairly common that doctors gave some patients “sugar pills” (whether they were really sugar pills or not, they were meant to have no clinical activity). These might be patients whom the doctors suspected of being hypochondriacs – malades imaginaires – or patients who had vague symptoms and felt generally lousy, but for whom doctors had no specific treatment at hand. So the treatment consisted of a bit of soothing talk and a pill of some kind, in the hopes that the patient would indeed feel better. And, lo and behold, a good deal of the time the patient did feel better. In many cases, this was because the malady resolved spontaneously. But sometimes, it was thought that when patients thought they had received a bona fide treatment, they actually responded with a real improvement in their symptoms.
And it is also the case that there were, “in the old days,” a lot of remedies that didn’t really have any clinical efficacy, but definitely had an effect on patients’ mood. Geritol was a hugely popular tonic back in the mid-20th century. It was marketed to the elderly, and it was supposed to combat feelings of fatigue. It actually did have some active ingredients – iron, and Vitamin B, and as such might have delivered some benefit to people with anemia (although iron supplements also lead to a condition called hemochromatosis, which can be dangerous). But the ingredient in Geritol that made patients feel good was alcohol. Geritol was about 12% alcohol, about the same as a robust cabernet sauvignon.
Another such “health tonic” was Peruna, a sweet gingery-tasting concoction, which was about 18% alcohol. Were Geritol and Peruna placebos? I would contend that they were. Some people might have been taking these (and other) concoctions for the alcoholic content alone, but many genuinely took them as tonics, to perk themselves up and feel better. And lots of people swore by them, convinced that there was something in them that helped with whatever symptoms they were experiencing.
And then there’s the case of my mother, in whom a drug that I think was intended as a placebo did not work. She was then in her late 80s, and she began to experience severe pains in her lower back. So she went to her regular physician, and came away with a prescription. After more than a week, she told me that she was feeling no better. I asked her what she was taking, and she told me it was Elavil. Elavil is amitriptyline, a very common antidepressant. When I explained to her what she had been prescribed, she was decidedly vexed with her doctor, and took him to task for trying to fob here off with what amounted to a placebo.
Although not in my mother’s case, it’s entirely possible that in some persons, an antidepressant might have helped a patient with aches and pains. It could help a person (not my mother!) relax, and thus perhaps ease muscle tensions. And there are likely to be other mind-body connections that might make placebos effective.
I might add here that antidepressants are a standard treatment for fibromyalgia, a diagnosis that groups a number of aches and pains that are sometimes misdiagnosed as rheumatoid arthritis, as well as other symptoms. There are some health professionals who shake their heads at treating that complex of symptoms with an antidepressant, suggesting that for want of a treatment that more directly addresses the underlying disease process, what physicians are doing, when they put patients with fibromyalgia on antidepressants, is prescribing placebos. (By the way, my mother certainly did not have fibromyalgia, which can be a puzzling condition.)
Can a diagnosis in itself be a placebo?
Here’s another little story for your entertainment: a friend of my wife’s had persistent headaches. He went to a neurologist and was informed that what he had was idiopathic occipital neuralgia. Supposedly, no immediate treatment was indicated. And he felt better after being given that “diagnosis.” But when my wife told me what this chap’s diagnosis was, I was baffled. What the neurologist had told him was precisely nothing – namely that he had persistent headaches (neuralgias) of unknown cause (idiopathic) in the back of his head (occipital). If the neurologist had used those plain English words, my wife’s friend would have felt no better. But the medical terminology evidently made him feel better, because he thought that he was in no imminent danger. This evidently had a placebo effect. (It later turned out that he had a brain tumor, not malignant but space-occupying, and it was successfully removed.)
And consider this. The December 12, 2011 issue of The New Yorker had an interesting article by Michael Specter entitled “The Power of Nothing,” with the subtitle “Could studying the placebo effect change the way we think about medicine?” It was serious and well-written, but in my view, deeply wrong.
Specter tells about an episode of worrisome chest pain that he experienced. Even though he has no cardiac risk factors, he thought he might possibly be having a heart attack and went to his doctor, who examined him and told him he was fine. His pain “suddenly disappeared.” He asks himself, “could words really banish a pain I had struggled with for hours? After I got home, I realized that I had been given a placebo.”
Specter’s doctor made sure that Specter was not having a heart attack, and he did nothing more than relieve Specter’s anxiety about the possibility that he might be having a heart attack, and when his anxiety was diminished, the pain went away. If Specter had indeed been having a heart attack, and his doctor persuaded him that he was okay, the pain might possibly have diminished a bit. But the reassurance would certainly not have stopped an MI in progress. He might have keeled over dead.
What Specter got was a form of cognitive therapy, in addition to a genuine diagnosis. And there’s no doubt that cognitive therapy works – it’s not a placebo, and no reputable clinician thinks so.
So, what’s the underlying difference between what happened with my wife’s friend and Michael Specter? In the case of my wife’s friend, what helped him feel better was nothing more than the aura of the physician – the healer’s touch, the magical incantation of medical terminology. Specter, on the other hand, had a bona fide medical examination which turned up no evidence of a myocardial infarction. So he could legitimately stop worrying.
However, it can be difficult and confusing to try to sort out those two potentially benign effects.
The role of the placebo effect in clinical trials
In a great many clinical trials, the new intervention – whether a drug or a procedure – is compared with a placebo, at least initially. (Further along, the new intervention may be compared with an established intervention, but in the early stages the placebo-controlled trial is the standard.) The people running clinical trials go to extraordinary lengths to make it difficult-to-impossible to tell which intervention is the real thing and which is the “placebo.” Not only must the patient or trial subject not know, but none of the health professionals carrying out the trial know. Presumably, the “placebo effect” can have an effect not only on the patient, but also on the investigators – if they believe that the patient is getting the “real” intervention, it may color their interpretation of the patient’s condition. So if the investigational drug is a little capsule, blue on one end and pink on the other end, so is the placebo. That’s easy. It’s not so easy when what’s being tested is a device. If it’s an electronic head-band that’s supposed to alleviate depression and if the device buzzes when it’s switched on, so must the “placebo,” or dummy device. Surgical interventions are not usually subject to placebo-controlled clinical trials, but just about every other kind of intervention is evaluated this way.
And what do they find? Not surprisingly, some patients in the placebo group actually get better. The question that has vexed a great many physicians, and others, for a good many years, is “why?” It’s fairly easy, in many kinds of illnesses, to discount the placebo effect in most cases. For example, if we’re testing an antibiotic against a certain class of infections, the data (and experience) tells us that those infections do resolve spontaneously some of the time. The antennae of suspicion rise when we know, say, that those infections tend to resolve in 8 to 10 days, but many patients in the placebo group report improvement in their symptoms in 3 to 5 days. However, trials in infectious diseases assess not only how the patient is feeling – clinical outcomes – but also bacteriologic outcomes, i. e., whether the drug is reducing the population of the pathogen that caused the infection. If a placebo were able to affect the bacteriologic outcome, the medical community would really sit up and take notice. This doesn’t happen. Placebos do not kill germs, but they do seem to affect symptoms
The placebo response in clinical trials: what does it mean?
When subjects in clinical trials demonstrate a suspiciously high placebo response, investigators quite reasonably look for any possible cause other than the placebo effect. For example, could the patient’s participation in a clinical trial by itself have influence the outcome? The patient is seeing health professionals, is aware that he or she is in a clinical trial, and may be behaving in ways that have health implications.
As we would expect, there is a great deal of variation in the responses of placebo-treated patients in trials in different medical conditions. The more the supposed efficacy of the intervention depends on patient self-report, the higher the response in the placebo cohort tends to be. Pain medication trials frequently report fairly high responses in patients treated with placebo, partly, of course, because some pain episodes resolve spontaneously, but perhaps partly because of the placebo effect: that is, some patients think they have been given the active medication, and convince themselves that they ought to be feeling some relief from their pain, and this – somehow – results in actual, real pain relief. Or perhaps, as we’ll see later, patients benefit somehow from just being in a therapeutic setting.
One of the conditions in which exceedingly high responses are reported in the placebo-treated cohort in clinical trials is major depressive disorder (MDD). The efficacy of antidepressants is usually evaluated according to one or more of several so-called “validated instruments,” such as the Hamilton Depression Scale (HAM-D) – “validated,” because patients’ scores on such tests tend to agree with the assessments of trained clinicians. An antidepressant is considered to be effective if it improves the scores of a patient taking that antidepressant by a previously set measure. It is not at all unusual for as many as 25% of patients in the placebo group to improve their scores by the target amount. And in some cases, the active drug improves their scores by only 15% or even 10% more than the placebo. Even such a small improvement on the depression scale can be enough to have a drug approved by the FDA for treatment of major depressive disorder.
Is the improvement of patients taking the dummy pills in such a clinical trial evidence of a placebo response – or are there other likely explanations?
Well, there are certainly other likely explanations, and some of you have doubtless thought of them already. First, MDD is a condition that waxes and wanes, sometimes with no particular external reason. Second, even when there is what clinicians call a “neurophysiologic” cause – meaning something going on with the brain chemistry and wiring – depression also has obvious external causes, i.e., life goes on and things get better, or worse, as the case may be. Third – and this is where the placebo response issue comes into play – even though the patients in these clinical trials are specifically not receiving talking therapy or cognitive therapy, they are in a therapeutic setting, speaking with health professionals, taking self-assessment tests. And some of them think that they are getting the active drug, and they perceive that their depression is actually lifting.
Is this a placebo effect, or are they just kidding themselves? Or, to put it another way, is the placebo effect nothing more than people just kidding themselves, or is something real going on?
So, back to the original question…
This brings us back to where we started: is the placebo effect real, and if so, can it be put to work in legitimate patient care. Well, a number of highly-credentialed clinicians lean towards answering that question in the positive.
A professor at the Harvard Medical School named Ted Kaptchuk, along with several colleagues at several hospitals affiliated with the medical school (HMS) created the Program in Placebo Studies and Therapeutic Encounter (PiPS), which is headquartered at Beth Israel Deaconess Medical Center. It’s a multidisciplinary program, searching for explanations of how the placebo effect works – not if, mind you, but how – and drawing on the locally available expertise from many fields, not only clinical medicine, but psychology, neurology, biology, anthropology, and whatever else may seem relevant. From Kaptchuk’s perspective, if the placebo effect can be channeled and applied to patient care where appropriate, it would be foolish not to try to employ it. That would be like having a useful tool in the toolkit and not putting it to work.
Kaptchuk has curious credentials for a professor at HMS. He had been an acupuncturist, with a degree in Chinese medicine from an institute in Macao. He is evidently pretty smart, since soon after joining the Harvard faculty, he was winning grants from the NIH and publishing in top medical journals. What got him started on studying placebos is that fairly often, when patients came to him for acupuncture treatment, they started feeling better before the treatment started. So he designed a clinical trial, in collaboration with a gastroenterologist, in patients with irritable bowel syndrome (IBD).
In this trial none of the patients actually received acupuncture treatment. The patients were divided into three groups. Group one was told that they were on a waiting list for treatment. Group two received sham acupuncture treatment – that is, the needles made contact with skin at the appropriate points, but did not penetrate. And group three received the exact same treatment, except that it was accompanied by a whole lot of touchy-feely attention. No surprise, group three experienced by far the greatest relief from their IBD symptoms.
What this meant was that the placebo effect, at least in this case, didn’t arise only from the clinician’s deception and the patients’ acceptance of that deception, but from the patients’ experience of a therapeutic setting and the sympathetic behavior of the therapists.
This trial was followed by another study in IBD patients, where two groups were compared. One group received no treatment whatever. The second group was told in advance that they would receive a placebo – they were actually given inert pills in bottle labelled “placebo.” And even knowing that the treatment they were receiving was a sham, the second group reported twice as much symptom relief as the group that received no treatment.
So what is it that makes placebos work, when they do work? The potential answers come from brain chemistry.
A quick look at the other side of the placebo effect
In some clinical trials, the incidence of adverse effects in patients taking placebos is considerable. Some researchers call this the “nocebo effect,” nocebo being Latin for “I will harm you.” Nobody really believes that a totally inert pill by itself is capable of causing the severe headaches, nausea, dizziness, or other symptoms observed in clinical trial subjects taking placebos. However, their acceptance of the fact that they may be taking an active drug and that active drugs are known to have adverse effects, may somehow trigger physiologic changes resulting in those same adverse effects.
Neuroscientists at the University of Turin, led by Fabrizio Benedetti, have studied this effect, and come to the conclusion that certain specific neural pathways are affected by these verbal cues. Says Benedetti, “What we have learned is that therapeutic rituals move a lot of molecules in the brain, and that these molecules are the very same as those activated by the drugs we give in routine clinical practice.” The brain regions identified by this team include the hypothalamic–pituitary–adrenal (HPA) axis and also the hippocampus, which is involved with memory, but also with anxiety. The adverse effects that were presumably triggered by this awareness were nullified by the anxiolytic diazepam (Valium), suggesting that anxiety played a major role in these effects. Diazepam, by the way, had no effect on any underlying pain experienced by these patients – only on specific pain arising from the “nocebo effect” – that is, (for example), the headache would go away, but not the pain in the arthritic knee.
How might the placebo effect work?
Researchers have also found that other neurotransmitters are active in the placebo effect. When patients are given agents that block the transmission of endorphins, which are active in pain relief, the placebo effect ceases. Thus, it can be surmised that placebos trigger the release of endorphins. And placebos have also been shown to increase the levels of dopamine in the brain. When placebos work, they do so by activating physiologic responses that have well-understood effects.
Kaptchuk and associates did a study of the placebo effect in migraines that demonstrated the very large effect of how the pill – whether real drug or placebo – was labelled. Migraineurs were given either an active antimigraine drug (rizatriptan, Maxalt) or a dummy pill. All the pills were labelled in one of three ways – as placebos, as either placebos or Maxalt, or as Maxalt. In each case, the pills labelled as Maxalt worked the best. Real Maxalt labelled as Maxalt worked better than real Maxalt labelled as placebo, and better than real Maxalt labelled as placebos or Maxalt. And the same relationship held for the dummy pills – a dummy pill labelled as Maxalt worked better than a dummy pill labelled as placebo. Furthermore, the dummy pills correctly labelled as placebo were more effective than no treatment at all. Judging the overall results of the study, the investigators concluded that the placebo effect accounted for about 50% of the effectiveness of the treatment.
In another very carefully conducted study, Kaptchuk and associates evaluated four different interventions in asthma patients: active treatment with albuterol inhalers, placebo treatment with fake inhalers, sham acupuncture treatment, and no treatment at all. When patients’ lung function was assessed, the only intervention that worked was treatment with albuterol inhalers. But when patients were asked to assess their own response to treatment, they reported that the two fake treatments worked as well as the bona fide albuterol inhalers.
This begins to zero in on where the placebo effect might be valuable – as well as where it might actually be dangerous. As in the general case of infectious diseases, where the placebo effect might relieve symptoms, but not attack the cause, something similar apparently happened with the asthma patients. An asthma attack is scary. The patients think that any minute they won’t be able to breathe at all. The more they try to force air into their lungs, the tighter their airways get, and the less lung capacity they have. Easing the stress, relaxing a bit, can help a good deal. And it’s conceivable that the placebo effect helps to trigger the sympathetic nervous system response that leads to the release of norepinephrine, a neurotransmitter that promotes bronchodilation.
At the same time, asthma can be fatal. An asthma attack can require medical intervention. A placebo could well fool a patient into thinking that he/she is getting better and doesn’t need to get help – until it’s too late.
Getting back to the case of Michael Specter, who thought he was having a heart attack, went to a doctor, had tests, was assured that he was not having a heart attack, and felt better. If the doctor had merely assured him that he was not having a heart attack, without doing the tests, that would have been a placebo. But it might have put Specter at serious risk.
And in what types of conditions might it be useful?
For a start, placebo investigators are clear that placebos will not cure cancer or kill viruses or remove plaque from clogged arteries or clear away beta-amyloid from the brains of people with Alzheimer’s disease. When we hear or read of miraculous cures of people with serious, life-threatening diseases, we can be pretty sure that it wasn’t placebos that were at work.
At the same time, the evidence that they relieve a range of symptoms is pretty strong. Mostly, these would be termed “subjective” symptoms. However, those subjective symptoms can have a major impact on people’s lives. Depression, for example, is one of the leading causes of missed work in the United States, is linked with all manner of life-altering comorbidities, and can lead to suicide in some individuals. And no one should underestimate the consequences of chronic pain.
Pharmaceutical companies are exceedingly interested in the placebo effect – not because they would like to harness it, but because they would like – if possible – to minimize it in clinical trials. One of the findings of the Harvard Program in Placebo Studies is that particular variants in a gene linked to dopamine release affects individuals’ responses to placebos, such that some individuals hardly respond to placebos at all, while others are highly responsive. Pharm companies would like to recruit those persons who do not respond to placebos (e.g., my mother) so as to eliminate that source of doubt in clinical trials. If they were able to do that, clinical trials might be able to enroll fewer patients, and cost a whole lot less money.
A tentative conclusion: it’s not primarily about deception. It’s mostly about the patient’s total experience in the therapeutic setting.
It’s not magic, and it’s not quackery. Doc Gumshoe, your sworn skeptic, trained by Miss Truesdell to put the little grey cells to good use, thinks that the therapeutic experience needs to have more to it than the physician entering information on the computer and coming up with a treatment plan as recommended by an algorithm. The physician needs to look away from the computer screen and become a healer. When patients are involved in that kind of experience, the experience itself elicits genuine physiologic responses that result in real relief of symptoms. And if the principal consequences of a medical condition are symptoms which by themselves result in no further harm, this kind of placebo effect might be quite helpful.
* * * * * * *
Many thanks for all the comments! If you think I’ve gone over to the side of the quacks, don’t hesitate to shoot me down. As for future pieces, I’ve been pondering doing one on depression, which I mentioned a couple of times in this piece. Other possibilities: fibromyalgia, psoriasis, skin cancer. Please let me know. Best to all, Michael Jorrin (aka Doc Gumshoe)