Three factors are of major importance in the suffering of badly wounded men [during the Second World War]: pain; mental distress; and thirst. Therapy has been almost entirely directed to pain, and this usually limited to the administration of morphine in large dosage.
—Henry Knowles Beecher, American anesthetist and medical ethicist
The American anesthetist Henry Knowles Beecher put his academic career on hold to serve in the US Army during the Second World War. While working on the front line in southern Italy, he reportedly saw something extraordinary. With supplies of pain-killing morphine running out, a nurse injected a wounded soldier with salt water instead of morphine before an operation. The soldier thought it was real morphine, and it seemed to work: he didn’t appear to feel any pain. Beecher is said to have continued the practice and returned to the United States convinced of the power of placebos.
While there are no historical records to confirm this often-repeated account, in 1955 Beecher carried out a systematic review of fifteen previous studies (with more than 1,000 patients) that had groups of patients who received placebos in them. He found that a third of patients who receive placebos get better. He published his findings in a widely cited paper titled “The Powerful Placebo.”
One confusing thing about Beecher’s study is that we don’t know if it was the placebos (like salt water) that made the patients get better or if it was the body’s self-healing ability at work independently of the placebo treatments. After all, many common illnesses like colds and flu clear up no matter what we do. The medical term for something going away on its own is natural history. (Geeky aside: philosophers call the mistaken inference that placebos have cured a condition based on some people getting better after receiving a placebo the post hoc ergo propter hoc fallacy: the “after, therefore because of,” fallacy.)
Many people have written about Beecher, debating how powerful placebos really are. Some say placebos are even more effective than Beecher claimed, but one major study coming out of Denmark claims that placebos have hardly any effect at all. I led a team of researchers to do a systematic review that helped to resolve this controversy. To do this, we compared the size of placebo effects with the size of “real” treatment effects. Our review included 152 trials (containing more than 15,000 patients), and we found that, on average, placebos have almost the same effect sizes as “real” treatments.
The overall effect of a treatment is a combination of both the placebo effect and the drug effect, so we would not always want to replace treatments with placebos. Say that someone ranks their pain at 5 on a scale of 0 to 10 (where 0 is no pain and 10 is the worst pain imaginable). If we gave that person a placebo treatment, their pain might be reduced from 5 to 4. If we then gave that person a drug, their pain would be reduced from 4 to 3. So taking the drug would have a bigger effect than taking the placebo alone, even though the drug effect compared with the placebo effect is the same as the placebo effect compared with nothing. You can still often—but not always (see the antidepressant example below)—get additional benefits from the treatment. That being said, you can also get additional harms from drugs: placebos are generally much safer than “real” treatments. What is more, if we construe placebo effects more broadly to include positive thinking and doctor empathy, we get an even larger placebo effect. This means that there are many common ailments for which placebo treatments are the best and safest option.
As I did my research, I learned that there are many misconceptions about placebos and placebo effects. The biggest one is that placebos are “inert” substances—basically annoying noise that needs to be controlled in a trial, or something that quacks use to sell their snake oil. Placebos are neither inert, nor inactive, nor nonspecific. They are active, and their effects can be as specific as “real” treatments. When the patient expects to get better after taking a placebo painkiller, their body will produce its own morphine (endorphins), which is as active and specific as the morphine that a doctor might inject.
I also learned and have written in my academic work about two other interesting things about placebos that I had not thought about previously:
In fact, sometimes all that is required for a placebo effect is for a doctor to watch someone, which is called a Hawthorne effect.
In a series of experiments between 1924 and 1933, scientists fiddled with the lighting in parts of the Hawthorne works belonging to the Western Electric Company factory near Chicago. When they turned the lights up in one part of the factory, they found that productivity went up. Then they tried turning the lights down: productivity went up again. They kept playing with different light intensities until the lights went so low that workers said they couldn’t see and production unsurprisingly dropped. Since productivity increased whether the lights were turned up, turned down, or remained the same, they reasoned this outcome was not related to the change in lighting. Eventually the researchers worked out that productivity was increasing because the workers knew they were part of an experiment and worked harder. Scientists called the effect of being observed the Hawthorne effect. Something similar can happen to patients in clinical trials.
When a patient is in a trial, they are usually observed carefully by doctors. Knowing that they are being monitored, the patients might do things that are good for their health, like exercising more, eating healthier, drinking less alcohol, or simply expecting to recover. These could all produce positive effects. The contact that patients have with health-care practitioners could also have a positive effect (more on this in Chapter 10). Confirming the likelihood of Hawthorne effects within medical trials, a systematic review found that even patients who are untreated (for example, they are on the waiting list for a trial) improve by 24 percent.
SSRIs are a class of antidepressant drugs that includes Prozac, Zoloft, and Luvox. With around one in ten adults in developed nations taking them to help reduce symptoms of depression, these drugs have made drug companies billions of dollars. According to the serotonin theory, depression is caused by the lack of a brain-signaling chemical called serotonin. SSRI drugs don’t actually produce serotonin, but they prevent the body from absorbing it once it has been produced, so that more serotonin ends up in the body, thereby reducing the symptoms of depression. And many people who take SSRIs report feeling much better. But here is the rub: placebos work almost as well as SSRIs for people with mild or moderate depression.
People with mild or moderate depression have some, but not all, of the symptoms of depression:
The symptoms must persist for at least two weeks, in such a way that they affect the ability to function normally and they cannot be due to some other reason. People with severe depression have all or almost all the symptoms. When Irving Kirsch did a review of forty-seven previous trials comparing SSRIs with placebos, he found that unless a person was severely depressed, the difference between the drugs and the placebos was so small it would be hard for a depressed person to detect. Kirsch’s study (and others like it) have even led some scientists to believe that the “serotonin theory” is mistaken, and that depression isn’t really linked to serotonin.
The main point here is not about serotonin: it is that the trials seem to show that placebos are quite good for treating depression. Does this mean that we could or should give mildly depressed people Tic Tacs instead of drugs? Probably not, because we don’t believe in the power of Tic Tacs as much as we believe in drugs, so Tic Tacs would be unlikely to have the same benefits. However, for many people with mild to moderate depression, a doctor (or patient) familiar with all the aspects of self-healing and placebo effects that I present in this book is likely to have the same antidepressant benefits as a drug, and without the side effects or risk of dependence.
Even when researchers try to keep a trial blind, patients in trials are quite good at detecting whether they have been given the real thing or the placebo. This is because it can be really hard to make placebo treatments that look, taste, and smell exactly like the real drugs. And some studies show that even if researchers try their hardest to keep a trial blinded, more than half the participants can guess whether they are taking the placebo or not. Even if patients can’t tell the difference at the beginning of the trial, they sometimes get side effects and so realize immediately that they are taking the “real” treatment.
For instance, a common side effect of tricyclic antidepressants is dry mouth. Patients in tricyclic antidepressant trials who get a dry mouth might realize they are receiving the “real” intervention and develop a positive expectation about recovery. Meanwhile those who don’t experience dry mouth may realize they are receiving only a placebo, causing them to have negative expectations. I’ll explain in Chapter 8 how negative expectations about a treatment can cause worse outcomes.
To get around this problem, researchers sometimes put chemicals into placebos so that they cause the same side effects as drugs. These active placebos are better at tricking patients into believing that they are getting the real thing. Psychiatrist Joanna Moncrieff and colleagues at University College London compared the effects of active and standard placebos in trials in which they were up against antidepressants. Active placebos had greater effects than normal placebos.
Placebos are not a panacea for everything, but neither is “real” medicine. Yet like “real” medicine, placebos work quite well for some things. Also, placebos are not just pills. They can be injections, sham surgery (more about this in Chapter 10), or funny-tasting drinks (more about this in Chapter 9). We will also see in Chapters 8, 9, and 10 that you don’t need placebo treatments to have placebo effects. Moreover, placebos are less likely to harm patients, and doctors who use placebo effects (either on their own or in addition to effective treatments) will have better outcomes than doctors who don’t. But if placebos work, why can’t you get a prescription for one? Typical answers to this question no longer make sense.
The main reason doctors hesitate to give placebos is that they believe it requires them to lie to their patients. Recent studies suggest this isn’t true. Ted Kaptchuk, professor of medicine at Harvard Medical School, conducted a placebo-controlled trial for people with severe irritable bowel syndrome (IBS). Those for whom normal treatment was not working were randomized either to a waiting list, or to be given pills they knew were placebos. (I will solve the mystery of how these open-label placebos work in Chapter 9.) The placebos were presented to patients as: “placebo pills made of an inert substance, like sugar pills, that have been shown in clinical studies to produce significant improvement in IBS symptoms through mind-body self-healing processes.” Kaptchuk found this type of placebo to have a similar effect as many “real” drugs for IBS (although the existing therapies were not all that effective either).
Linda Buonanno, a woman who took part in the trial, had been suffering from IBS for years and nothing had helped. Some days she was in so much pain she could barely leave the house. The open-label placebo had a great positive benefit. So much so that she said, “I never felt better in my life.” But then at the end of the trial she stopped receiving the placebos and her IBS became worse again. A pharmacist whom she asked for some placebos said he was unable to give her any because it would be unethical.
Kaptchuk’s trial using placebos that people knew to be placebos shows that there is no need for doctors to deceive their patients. We also don’t need to give someone a placebo to have placebo effects. As we saw with the Hawthorne effect, simply paying attention to someone can improve their symptoms, and as we will see in Chapters 8 and 10, positive messages and empathetic care can also induce placebo effects without any inherent deception.
So, placebos are sometimes a good option and we need not lie, and in those cases there are no good ethical objections. It is hardly surprising then that surveys carried out in many different countries show most doctors have used placebos. Instead of worrying about how to ban them, we need to explore ways to expand their use in an ethical way.
A final problem people have with placebos is that they say placebos might work for others, but not for them. They say things like “Placebos might work for naive, gullible, uneducated people, but I am rational and smart and nobody fools me, so placebos won’t work for me.” While it is true that placebos (like “real” treatments) work for some people but not others, we can’t predict who they will work for. So even people who think placebos won’t work for them could be surprised.
In fact, our beliefs about ourselves are often mistaken. If you ask people whether they are better-than-average drivers, most say yes. But as long as we make a nerdy statistical assumption that the distribution is not skewed (which we can), only half of us can be better than average. The other half are worse than average.
Likewise, if you ask people if they can tell the difference between expensive and cheap wine, most say yes, whereas hundreds of studies show that only a small handful of people actually can. With placebos it is similar: it is hard to tell who placebos will work for. Some studies are starting to show that placebos appear to be most effective for those who are more suggestible, those who are more optimistic, and those in whom brain structures associated with pleasure are more sensitive to placebo effects. But the studies are not definitive, so for now we can’t predict exactly who will respond to placebos.
All this is to say that since placebos work, we do not need to lie to patients to use them, and doctors should exploit placebo effects more. We can also exploit placebo effects for ourselves (see the takeaway exercise at the end of this chapter). This brings us to the question of how placebos work.
There are many different types of placebos that can work in different ways, and you don’t need placebo treatments to have placebo effects. So we cannot really say how placebos work without identifying what kind of placebos we are talking about. However, in general, most placebos work in three main ways:
Also, being treated by an empathetic doctor can reduce stress and help us relax, which can have additional health benefits (see Chapters 8 and 9).
I’ve adapted the technique for this exercise from the positive psychology movement. Systematic reviews suggest it is quite good for anxiety-related disorders and depression. The cool thing about positive psychology is that if you have a “mental illness,” practitioners do not focus on it. Instead, they try to focus your attention on personal growth and what it means to have a meaningful, happy life. You will need about thirty minutes to complete this exercise, and you will need a pen and some paper:
Here is a very short version of a “best possible life” story: “I envision spending quality time with my family and having amazing vacations together. The character trait I see in that picture is ‘prudence,’ and that’s just what I’ll need to enhance my financial situation in a way that maintains my work/life balance so that I can spend time with my family. If I had that kind of prudence now I would pay off and cancel all my credit cards so that I never pay high interest rates.”
Some studies suggest that “alternative” practitioners are much better at using placebo effects to their advantage than conventional doctors, but all doctors can (and should) exploit placebo effects. No matter how serious the illness, you can enhance the effectiveness of your interaction by being positive and empathetic. I’ll describe the details of how this is done in Chapters 8, 9, and 10. If you have a patient with a complaint that you are sure does not require any “real” medicine, but they refuse to go away until you give them something, then in addition to being positive (Chapter 8) and empathetic (Chapter 10), you can offer them something harmless such as vitamin C (or some other vitamin, after verifying that it will not harm the patient, perhaps by inducing an allergic reaction).
You don’t need to lie to the patient. You can say something like Kaptchuk told his IBS patients: “These vitamin C tablets are pills that don’t have a specific effect for your problem; however, pills like these have been shown in clinical studies to produce some improvement in [you fill in the blank] symptoms through mind-body, self-healing processes.”