The placebo effect

The power of suggestion is said to be such that it can cure physical ills. For example, you might have heard that by getting somebody to take a pill made of nothing but tasteless sugar, and telling them that it is in fact a powerful pain killer, that a mysterious phenomenon called the "placebo effect" can be conjured to somehow cause the sugar pill, or "placebo", to effectively act as the painkiller the deceived person believes it to be.

This idea has considerable medical implications, and yet it is widely misunderstood. I have done my best to come to grips with it here, by attempting to answer the following questions:

Is it true?  What the hell is the placebo effect?  And what are it's implications?

Part I: It's True!...ish

The word 'placebo' has been kicking around for quite some time, but it's meaning has been continually evolving. At the dawn of the 1900s, the consensus of the medical establishment seems to be that the placebo effect did not exist outside of the mind of the patient. However, despite this somewhat dismissive attitude, it was apparently common practice for hospitals and doctors to keep such medically inactive substances as sugar pills and saline (saltwater) injections on hand, probably as a means of comforting those whose pain (or, possibly, hypochondria) could not be treated more effectively (this practice, that of prescribing a placebo to a patient, is called 'placebo medicine').

But around mid-century, placebos were finally seriously investigated; and, surprisingly, evidence emerged to indicate that the placebo effect was not only alive and kicking, but that it kicked like a horse. It was shown that, for whatever reason, statistically significant numbers of patients taking part in "clinical trials" would respond positively to 'sham' treatments, as long as they sincerely believed that there was a chance the treatments they were receiving would be effective. Indeed, simply for knocking them out, giving them a scar, and them waking them up and giving them the impression that they had just received surgery, it was found that doctors could receive get positive feedback from their patients.

(Click here to see Penn & Teller demonstrate the phenomenon.)

For medical science, the implications of this discovery were enormous. As R. Barker Bausell writes in his [excellent] 2007 book, Snake Oil Science, to which this post is greatly indebted:

"this single finding implies that any non-harmful therapy evaluated by a trial ... has the potential to produce positive results due to the placebo effect." (my emphasis)

Suddenly, it was realized that the credibility of clinical studies depended on them finding a way to compensate for this effect. It was realized that to demand any less of them would be to render them virtually meaningless, as it would mean that any positive results they obtained would be tainted by the possibility that they were merely a testament to the power of the placebo. To this end, the practices of including a "placebo control group" and of "double-blinding" clinical trials has become standard (even in trials involving non-human animals); and the placebo effect has become the hurdle that any therapy aspiring for medical legitimacy is required to clear. The reasoning behind this, is that if it can't be shown to work better than a placebo, then we can't be sure that it works at all.

So it would seem that the placebo effect does indeed exist. Or, more accurately, a phenomenon was discovered that is now known as the 'placebo effect'.

Or was it?

Because while it became accepted that the placebo effect was an important consideration for medical science, the possibility remained that it was purely imaginary. Is the sugar pill in my lead-in story bringing about an actual reduction in pain? Or can it merely seem to?

Part II: What The Hell The Placebo Effect Is

Eventually it became clear that the phenomenon being observed in clinical trials was not in fact being caused by a single influence. Because of this, as Steven Novella writes on the blog Science-Based Medicine, the term 'placebo effect' is actually "a misnomer and contributes to confusion, because it is not a single effect but the net result of many possible factors."

As an example of one of these "factors", he mentions "reporting biases" which result from the patient's desires to please the researchers; and the researchers desires to help their patients, and to prove that the intervention they are investigating is indeed effective. Subtle, cumulative, and boring influences like these have nothing to do with the idea that a placebo can heal or reduce pain. Nonetheless, to this day they are often lumped under the heading 'placebo effect', and from there they have long encouraged bold overestimates of the power of suggestion, as we will later see.

However, neither are suggestions entirely impotent: although difficult to separate from its complimentary influences, it was eventually found that suggestion alone could (sometimes) result in real, albeit limited, physiological responses in people. In fact, it turns out that a possible explanation for this already existed. As Bausell notes in Snake Oil Science: 

"There happens to be some indirect evidence... that while the placebo effect may indeed be in our minds, it is not a figment of our imaginations. This evidence come from someone whose name is very recognizable to most of us: Ivan Petrovich Pavlov, whose seminal contribution to science, classical conditioning, is hypothesized to be the primary triggering mechanism for the placebo effect."

"Classical conditioning", of course, is the name that was given to the phenomenon that Pavlov famously demonstrated by getting dogs to salivate by ringing a bell. The reason he was able to do this is because the dogs had been 'conditioned' to associate the bell with food. In the same way, the thinking goes, humans can be conditioned to respond in specific physiological ways to certain types of stimuli. In addition, a couple other distinct suggestion-related influences have also been proposed, namely: "the subject-expectancy effect"; and "suggestion by an authoritative healer".

But there is also good direct evidence that an authentic, suggestion-caused placebo effect indeed exists. This includes the results of clinical trials that have investigated the matter, neuroimaging, and the observation, as Bausell documents, "that a drug called naloxone" can be used to "at least partially block the placebo effect."

This finding, he notes, is important for two reasons, First, it shows that the placebo effect is probably not entirely imagined or due to reporting biases. Second, it suggests that "the body's very own opioid system" is at least partially responsible for it, as naloxone is, Bausell writes, known to be "an opioid antagonist, which means that it is capable of nullifying the analgesic effects of opioid-based drugs such as morphine", or, in this case, of nullifying those substances that the human body naturally produces which are mimicked by morphine and other opiate drugs (neuroimaging also seems to support this idea).

So yes, although it's effect is still often exaggerated, it seems that the power of suggestion really can lead to reduced levels of pain under certain conditions. Just like a dog can salivate upon hearing a bell that has preceded food in the past, it seems that humans can respond to the acts of swallowing a pill, receiving an injection, or participating in a medical experiment, by producing endorphins or other biochemicals that can give rise to something that might properly be known as 'the placebo effect'.

But can it heal??

Part III: The Medical Implications

Here, friends, is the real meat. Or is it "tofurkey"? You decide.

As medical science became obliged to make the placebo group the benchmark by which interventions were judged, so too was it obliged to to turn its back on the practice of placebo medicine (at least, to a large degree). Sugar pills and saline injections began to lose their traditional place in the medical cabinet. Behind this was the rationale that if we can't be sure a given therapy is inherently effective, then it probably should not be prescribed to patients.

In fact, it is exactly on these grounds that "alternative interventions" (interventions that have not, for whatever reason, been demonstrated in controlled conditions to work better than a placebo) face criticism: how can we be sure that alternative treatments work at all, if they haven't yet been shown to work any better than comparable sham treatments do?

Interestingly, the proponents of these alternative medicines often don't bother to deny that their therapies work no better than comparable placebos. Instead, they will argue that it is unfair to say that a therapy "doesn't work" just because it "only" works as well as a placebo. Apparently, they see placebos as having true healing power. And, therefore, they see the mere inducement of the placebo effect as being sufficient to qualify their favorite alternative therapies as "effective medicine".

This alternative view has been conveniently summed up for me in the brief passage on "The Placebo Effect" that appears in the essay Acupuncture And Evidence-Based Medicine: A Philosophical Critique by Dr. Micheal T. Greenwood. This essay was posted to MemeScreen by a valued reader in response to my post on Skepticism, but it is worth reproducing here. Greenwood writes:

"Objective research usually includes a placebo arm to demonstrate that the tested therapy is superior to no intervention. Implicit in such studies is the idea that the placebo effect does not constitute real medicine, and that only medicine that is better than [a] placebo is "real". Yet the placebo group often demonstrates a 30%-35% positive response (indeed, sometimes as high as 70%), which implies that many people somehow heal themselves without medicine. That they might have been deceived into healing themselves is irrelevant. An irrational dismissal of the potential of self-healing occurs when such healing is rejected because no overt outer action was taken. Such a position is not only subjective, but also ethically questionable because it robs patients of their personal power." 

Greenwood's attitude is evidently shared by a scientist who put it even more succinctly in a recent issue of The Wall Street Journal ("I don't see any disconnect between how acupuncture works and how a placebo works. The body knows how to heal itself. That's what a placebo does, too."); and by the makers of the "Q-Ray Ionizing Bracelet", who unsuccessfully defended their $200 'miracle' product in court on the grounds that it it had been shown to induce a placebo effect.

It seems to me that there are a number of major problems with the argument that placebos can 'heal'. And the most imposing of these problems, is that there seems to be little to no evidence that placebos can heal. In fact, Greenwood's numbers don't imply 'healing' at all, because as we have seen, the "positive response" that placebo groups "often demonstrate" in clinical trials are often due to phenomenon as mundane as reporting biases. Moreover, it turns out that while suggestive influences can (sometimes) serve to affect (limited) desirable physiological responses; their effectiveness quickly disappears when we look beyond this. And although it seems that they can help kill pain, or make one feel 'buzzed' as if on caffeine, or even induce a slightly drunken stupor, there is no reason to believe that a placebo can make a broken leg heal faster; help cure cancer or appendicitis; fight infections; or indeed, to do pretty much anything of therapeutic merit.

In part, this is because real painkillers can provoke placebo effects too. And in part it is because, as Novella writes, "any potential placebo benefit worth having can be fully realized with science-based interventions." Some might point out here that at least placebo medicine cannot have any possible negative side effects (which is true since, strictly speaking, it doesn't have any effects), but that is only because an acute waste of one's time and money, for some reason, does not seem to count.

Which brings us to the final point I will raise here: we live in a dog-eat-dog capitalist society. And it has long been home to people who are more than willing, for whatever reason, to sell useless goods or services on the basis that they are actually medically effective. If placebo medicine came to be considered effective, that would mean that anything might be legitimately sold as being "medicine" - as long as the customer thinks it is. While some, no doubt, see few problems with this prospect, I imagine that it is nonetheless hard to deny that it would see a lot of desperate or gullible people put at even greater risk of being taken advantage of than they are already.

* * *

Don't allow the existence of the 'placebo effect' to be portrayed to you as evidence of a mysterious, internal self-healing force that needs to be unlocked and 'promoted' by alternative medicine. Our bodies our incredible things - mine in particular - and they are indeed remarkable in their ability to heal themselves without assistance; with ineffective assistance; and even with counterproductive 'assistance'. And the people and things around us can indeed have real and beneficial effects that may not be entirely understood by science.

But at this point, it seems fair to me to say that, to the extent one cannot fill a dog's stomach with food by merely ringing a bell, placebo medicine cannot heal.