“Untreated a cold will last a week, but treat a cold and it will only last 7 days.”
There’s a widespread understanding that a placebo effect is one where some kind of “mind over matter” phenomenon takes place. That, somehow, a “believer” of a treatment being given to him, when taking a placebo and thinks he’s getting real medicine, will have his immune system “waken up” to do the work it could not do without the placebo or the belief. This is certainly not the case, as we’ll see below.
The word placebo is derived from the Latin for “I shall please” and in 1811, Quincy’s Lexicon-Medicum defined placebo as “[any medicine] adapted more to please than to benefit the patient”. And that’s exactly what it does. It pleases, without having to benefit the patient. A placebo, for those who might not be aware, is a substance that does not have any active ingredients relating to an illness that the patient who receives it has. Essentially, we can call it, as it is usually called, a sugar-pill (though in reality it can be something other than a pill – it can even be a “fake” surgery). This is not to say that there is no such thing as a placebo effect. Far from it.
A placebo response is defined as “any perceived response to a medical intervention other than a physiological response to an active treatment” (Steven Novella). In fact, there are many different kinds of placebo effects, mostly psychological:
Every patient has a desire to improve or find relief. This goes without saying. All patients want to get healthier and this is why they invest time and money (and hope) for a treatment. So they have an obvious motive to believe that the treatment works.
Some patients tend to desire to please the person giving the treatment, whether a doctor or a researcher. Whether conscious or not, certain people think it a “duty” to satisfy the expectations of their doctor, and either report exaggerated benefits or even convince themselves of them.
Also, the doctors/experimenters desire to be successful and to please the patient. This all-too-human attribute of wanting to be helpful can actively distort the perception even of people who are trained to avoid these errors in observation. Add to that the selfish reasons of treating the ill -which might bring them money, fame or another grant for their research- and you have a very effective recipe for self-delusion. This is why the double-blind method in research is so important, so that the researcher cannot know whether he’s providing a placebo to the patient or a real drug. It’s been observed that when doctors know they’re giving placebos to their patients, instead of real medicine, even when they are not vocalizing this information to their patients, there is a weaker benefit to their pain relief (we’ll come back to the subject of pain later).
Both patients and doctors/researchers desire to justify the risk and expense of the treatment (risk justification). And actually, the more the risk and the greater the expense, the more either of them seem to overestimate the treatment’s benefits or exaggerate the patients’ improvement (the observable benefits rise according to risk and expense). Amazingly, it has been observed that four sugar pills are better than two sugar pills, in healing ulcer (as Goldacre notes on page 67). Also, a pain relief treatment was shown to be more effective when patients were told the treatment cost 2.50 dollars than when they were told it cost 10 cents (Goldacre p.70).
Apart from psychological reasons, errors in inductive reasoning and logical fallacies also play important roles in observing a placebo effect. The confirmation bias (related to emotionally charged issues, belief perseverance and other psychological factors) includes “the notion that people tend to notice and remember events that confirm their beliefs, biases, and desires” (Steven Novella). They will also miss, ignore, or rationalize any disconfirming evidence.
One might make the mistake of attributing an improvement to a treatment when the reasons for the improvement are irrelevant to it. Some illnesses are self-limiting, meaning that they will resolve on their own or have no long-term harmful effect on the patient’s health. So, the observer might think that the reason for the illness’ cure is the treatment he was given when, in reality, a cure would take place no matter what the patient was doing at the time. In other words, if phenomenon A is followed by B, one might think that A caused B while B would happen anyway, irrelevant to A.
Another effect, related to the previous, is the “regression to the mean”. In any system of data that fluctuate over time, any extreme in this fluctuation is likely to be followed by a return to the mean. Therefore, just from pure statistics, any extreme symptoms will be followed by their lapse, near the mean. This can be misattributed to whatever the patient was doing at the time (like taking a placebo or real but ineffective medicine). The observer has little, if any, way of knowing what would happen if the patient was left untreated. What helps to counter this is a large group of “guinea pigs”. This is why anecdotal evidence is not actually evidence. As the saying goes, “the plural of anecdotes is not data”.
The “observer effect” takes place when the observance itself alters the behavior and the outcome of an experiment. When people know they are being observed or treated regularly by a health-care provider they appear more compliant with their treatment (see Hawthorne effect).
There are also the “non-specific or incidental benefits”. When patients disrupt their daily routine to receive treatment, they might avoid their usually hectic and stressful activities they would otherwise be involved in. This might decrease stress levels and force them to relax spending time lying down to rest. Any benefits these factors produce, might be confused with benefits from the treatment itself.
Note that all of the above do not necessitate a belief in the effectiveness of the treatment in order for them to take effect on the patient. However, belief or expectation of benefit may help to enhance said effects.
Also, note that you can experience a placebo effect without being given a placebo, but also being given real medicine, which might turn out to be ineffective in the treatment of the particular illness but produces the appearance of benefit.
For someone to think that a placebo effect is a mind over matter effect, it would mean for him to find great difficulty explaining how a placebo can “work” on animals. How can you convince an animal that you’re giving it a placebo for it to have its immune system activated? However, the aforementioned view of what a placebo effect is (and isn’t) can easily explain how it can be observed in animals and babies, which it is.
Apart from placebo effects, nocebo (“I shall harm”) effects can also be observed. This is when an observer is led to believe there is harm to the patient. It can also be enhanced by expectation of harm and experiments have shown that, if a doctor gives medicine to his patients while telling them that he’s not convinced of its benefit, the patients are less likely to report improvement.
So, are there any actual benefits from placebos? Yes there are, though they are quite limited. Placebos can reduce stress and therefore reduce the risk factor for psychological and mental diseases, heart conditions, depression and other subjective symptoms. They can also decrease pain, which is a purely subjective experience (there is no “pain-o-meter”, as Novella says). We can only know what the subject tells us about how much pain he or she feels. Thirty per cent of people will respond with a significant decrease in their feeling of pain due to placebo interventions. Anything that provides an improved mood, comfort or the expectation of benefit will decrease the perception of pain. They can also decrease muscle tension, which can be the cause of real physiological symptoms. Also, patients with abnormal symptoms but with no signs of concrete medical diagnosis have observed improvement when given a fake “placebo” diagnosis (Goldacre p.75). This probably means they were never ill to begin with.
This is not to say that placebos can cure all sorts of illnesses, as some have attempted to convince us of. “Systematic reviews of clinical trials found no measurable mind over matter or biological placebo effect, no placebo effect for any objective biological measure, so therefore cancer patients do not survive longer due to placebo effects” (Novella). So, for things in which stress is not a risk factor and for other non-subjective symptoms, we should not expect benefit through placebo effects, and there is no evidence to support that there is any.
Placebos are very useful in trials of medicines so that we can separate the placebo effects from the real benefits a medicine might have. This is why anecdotal stories about a treatment working in some cases can, and should be, ignored as they cannot be deemed reliable in attributing the improvement either to the treatment or to a placebo effect. So, whenever someone tells us that they were cured thanks to an “alternative” treatment, for which there is no confirming evidence that it has true beneficiary results, we should take them as seriously as we would if they told us that they were cured after praying to Poseidon.
Placebo effects can actually be used beneficially in a number of ways. Even real and effective medical treatments come with placebo effects. For example, a good therapeutic relationship between doctor and patient has been shown to improve the doctor’s effectiveness, in relation to a doctor who is more formal and doesn’t offer reassurance to his patients. Note that this “ritual” is exactly the method most homeopaths use during their consultations. They take their time with their patients, showing genuine care and interest for their condition and their life in a general “holistic” way; something a medical doctor in a hospital would find impossible to do, if only for lack of available time (or interest, or training, or motivation).
We are all “victims” to the placebo effects, which are not privy to medical conditions. It is wrong to suppose that if you observe improvement due to a placebo there is something wrong with your brain. We are all placebo responders and Maria Konnikova tells us that the smarter you are (or, perhaps, think you are) the more easily you can be tricked.
Bloodletting had been used as treatment for more than 1.5 millennia and was considered effective, because of placebo effects. We now know that people can die from bloodletting (and many surely have died because of it) but, at the time, most thought it was beneficial. We can easily imagine people at the time claiming that they knew someone who was cured thanks to this method while in reality they had confused, for example, a “regression to the mean” effect for treatment.
So, in conclusion, what’s the harm in using a placebo in real life? It’s probably harmless enough when addressing the common cold or a random headache, but it can dissuade us from taking real medicine for more dangerous illnesses, with very serious consequences (try googling ‘death by homeopathy’ and see what comes up). A placebo effect is an illusion, an artifact of observation, not a necessary tool for countering a dangerous condition.
As Novella concludes, “it should be kept in mind that all of the benefits that you can get from placebo effects you will also get from treatments that are science based and actually work.”
Most of this article is based on Steven Novella’s course Medical Myths, Lies, and Half-Truths for The Great Courses. Also, some of his articles here have helped me in the past to get in touch with what placebos are and do.
I also used Ben Goldacre’s Bad Science – in particular the chapter called “The Placebo Effect”.
Richard Dawkins’ contribution to my understanding of biology in general and placebos and homeopathy in particular has constituted a revelation. You can start here.
For more on the confirmation bias, you want to check the wonderful blog You Are Not So Smart here.