Placebos: What Do You Believe?

By Ray McBeth, PhD

In medicine, treatments are typically tested in what is called a randomized control trial. In the trial, a proposed new treatment is tested against “care as usual” in a carefully controlled manner wherein patients are randomly assigned to either receive the new treatment or care as usual. When randomly sorting patients into treatment groups, within the control group (“care as usual”) it is preferred that the patients be of similar age, gender, race, severity of condition and the like to those in the treatment group so that those variables are controlled and only the treatment effects the outcome. If the new treatment is shown, statistically, to have better outcomes than “care as usual” then, in theory, the new treatment becomes the new standard of care. It is important to note that “care as usual” suggests that it uses the most up-to-date understanding of what works best for treating the condition under study.

In drug trials, however, the process becomes even more complicated. Human testing occurs late in the process once the proposed drug is proven to be safe and potentially effective. In the later part of clinical testing that is required for a new drug to be approved for sale in the United States, the drug is studied to see if it is successful in bringing about the intended result. Historically during this step, the new drug was tested against a placebo (e.g., sugar pill, saline injection). The test is done in what is called a “double-blind” manner, meaning that neither the testers nor the patients know who is receiving the new drug and who is receiving the placebo. This is done to avoid any subjective bias that might be inadvertently passed on to the patients, which could effect the outcomes. An even more rigorous model also includes a control group (e.g., a group that receives no care, such as someone who had been wait listed). Again, if the new drug is proven to safely offer better results than the placebo, then it may be approved by the Food & Drug Administration (FDA) for sale. Today, due to ethical considerations, most new drugs are not tested against a placebo, but instead are tested against the drug(s) that are currently approved to treat that condition; not unlike the “care as usual” approach.

What is frequently not understood by the general public is that the placebo often works. Typically, the group that is given no care shows some improvement (simply as a product of time), the placebo group shows even more improvement, and the treatment group shows significant improvement. But in many cases, the treatment is no better than the placebo. How can that be? The developers of the drug (or other treatment) have spent a lot of time and money on the development of the treatment. They would not have done that if they didn’t think it would be successful.

So why do placebos work?

Some have suggested that it is in large part due to confirmation bias. When we expect something to happen and it does, then it is confirmed. When we are involved in a drug trial and are given a medication, we expect that it will work (even though we know it might be a placebo). The medication is given in a professional research setting by a caring treatment team so of course we believe it is going to work. And it does. The ritual of receiving and taking the treatment leads to enhanced expectations which, when met, become reinforced. This is not to suggest that we can just “believe” and be healthy, but that attitudes and expectations play a role in health that is only just beginning to be explored.

There is also a negative perspective in this phenomena call the “nocebo effect.” In one study, two alternatives to reducing severe arm pain were administered: pain pills and acupuncture. In both treatments, people began to call in complaining of terrible side effects, the very side effects that the patients had been warned about. But most of the other patients reported real relief, with those receiving acupuncture feeling better than those receiving the pain pills. What is especially interesting about this study is that neither of the “treatments” were real. The pain pills were cornstarch and the acupuncture needles were retractable and never pierced the skin. One of the conclusions that the researchers reached was that the perceptions of patients matter and that the way providers frame those perceptions can have a significant impact on the outcomes.

This ability of the brain to “fool” us does not only occur in healthcare. For example, in one study, participants were shown three bottles of wine with the prices of each clearly marked. One was priced very low (about $4.00 per bottle), another was medium priced (about $14.00 per bottle), the final one was somewhat higher priced (about $22.00 per bottle). In fact, all three bottles contained the same wine (the medium priced one). While in an MRI scanner, participants were given a small amount of each to drink and asked to rate each one. As you might expect, the most expensive one was rated as tasting the best. More importantly, the MRI scanner showed that when drinking the “more expensive” wine, the parts of the brain used to evaluate expectations were more active than when drinking the less expensive ones. The brain linked the price to a greater expectation of reward and was, therefore, more active. As far as the brain was concerned, it did taste better. We see similar outcomes from many other marketing activities for products including automobiles and other luxury goods. Expectations matter.

But, as you might imagine there are many ethical considerations regarding placebos. Should providers prescribe placebos without telling their patients because they believe that it is the best course of action (and it may well be)?

But is that ethical? Is that lying to patients?

One physician once told me that if he (as a patient) was prescribed a placebo without his knowledge and then found out, that he would “fire” his provider and maybe even report him for malpractice. There are also more serious concerns. In one study, asthma patients were assigned to three groups: one with an albuterol inhaler, one with a placebo inhaler, and one with no treatment. After three treatments, all three groups self-reported an improvement in their symptoms with the placebo inhaler group reporting nearly as much improvement as the actual inhaler group. However, when the patients’ airflow was measured, only the albuterol group actually demonstrated improved airflow. Thinking you are better and actually being better may not be the same.

So what if the patient knows it’s a placebo? What then are the ethical implications?

In one study of migraine sufferers, one group was given a migraine drug labeled with the drug’s name, one group was given a pill labeled placebo, and the third group took nothing. The placebo group was nearly 50% as effective as group taking the actual drug.

Not all placebo treatments provide the same effects. For example, placebo injections are more effective than placebo pills and two placebo pills even more effective than one. Even the color of the placebo pill can make a difference. While placebos will not shrink tumors or cure viruses, they can have an impact on conditions like pain, depression, anxiety, coughs, and irritable bowel syndrome, among others. Placebos have no serious side effects (except in situations like the clinical trial discussed earlier) and cannot be overdosed. Researchers are also beginning to discover who exactly can benefit most from placebos.

All of this leads us to a few, core questions.

First: If placebos work, why are they not used more often? 

There are many reasons, typically starting with the belief on the part of many providers that they are not “real” medicines. Nonetheless, many of the researchers cited above argue that since they do work they should be used. To do otherwise is to deny patients’ treatments that are known to be successful, which, they suggest, might also be unethical.

The second question is: If placebos are to be used, what issues need to be addressed?

Historically, placebos have been used in medical research in such a way that neither the provider nor the patient knew which was the real treatment and which was the placebo. While this may be acceptable in clinical trials, it is not appropriate in professional practice. The American Medical Association has published a statement on ethical use of placebos  that provides some guidance on when they can be used. The statement asserts that placebos may only be used for valid medical reasons with the patient’s cooperation and consent and that placebos should not be used just to “mollify a difficult patient.” The provider must also believe that a placebo is the appropriate treatment and clearly communicate that belief to the patient. This is no less true for “real” treatments.

What about over-the-counter medications? While they have active ingredients, those ingredients alone may not account for all of the relief gained. I was told, after a recent surgery, that extra strength Tylenol should be adequate to control my pain post discharge and it seemed to so. However, I couldn’t help but wonder how much of the relief was the medicine itself and how much was the surgeon’s assurance that it would be adequate. I also wonder if when my primary care physician recommends an over the counter remedy for a cough or a muscle ache, how much of that advice is purely medical. It may well be that it is the process that really works (i.e., my PCPs assurance that this will help), not only what is actually in the pill or ointment.

The use of placebos is part of an incredibly complex interaction between provider and patient that, among other things, includes the attitudes and beliefs of both parties, the setting of the interaction, and a myriad of other psychosocial factors. Nonetheless, many believe that placebos (or placebo-like interactions) will play a significant role in medicine in the future.

What do you believe?