Imagine you are not feeling well. You have had a cough and mild fever for about a week now, and it is not getting better. So, you go see your doctor who, after a careful exam, diagnoses bronchitis and writes a prescription for an antibiotic. You take the medicine as directed and within two days begin to feel much better.
Why Did You Get Better?
Let's take a closer look. There is more than one explanation for your prompt recovery:
The most obvious explanation is that you had a bacterial infection. The antibiotics killed the bacteria, and your symptoms improved. This sounds perfectly reasonable, but it is actually unlikely. Viruses, not bacteria, cause the vast majority of coughs and fevers. And antibiotics have no effect on viruses.
Regression to the Mean
Based on the idea of "regression to the mean," you would have recovered in the next day or two anyway. This concept says that all livings things continuously cycle through a range of physiologic states, but always within reach of a central or average "set point."
When we move away from our set point, various mechanisms kick in to nudge us back. As long as we do not swing out too far, we remain in good health. Even if we do venture out of bounds and get sick, most of us eventually return to our average state. The odds of recovery, therefore, were always in your favor. And in this case, it simply coincided with your treatment.
It was not the antibiotics per se that made you better. Rather, it was the symbolic significance of your doctor's visit that changed the course of your illness. Your doctor made a difference, but not through the specific actions of an antibiotic. This is known as the placebo effect.
Researching the Placebo Effect
Almost 50 years ago, Henry Beecher, the first scientist to quantify this effect, reported that 35% of subjects seemed to benefit from a placebo alone. While this number is often disputed, Dr. Beecher's work contributed to a revolution in the design of clinical studies. The government now requires that all new drugs undergo placebo-controlled trials to determine their safety and effectiveness. Roughly half of a study's participants receive the active medicine, while the other half receives a placebo (often a sugar pill). In virtually all such studies, a significant number of participants show a response, whether they receive the active medicine or the placebo. To legally market the drug, however, significantly more subjects must respond to it than to the placebo.
Few medical researchers dispute the importance of a placebo group in their studies. The controversy arises when the placebo effect is used to explain patients' improvements in clinical practice. Critics of Dr. Beecher's research argue that many factors other than the placebo effect could explain his observations. And, an article in the New England Journal of Medicine (NEJM) carefully reviewed much of the published research specifically addressing the placebo effect. The authors concluded that under most circumstances, placebos are no more effective than no treatment at all. While this study does not disprove the placebo effect, it strongly suggests that researchers have not found it yet. This is not surprising, due to the elusive nature of both the placebo and its effects.
Defining the Placebo
The classic placebo is an inert, "dummy" pill. Placebos, nonetheless, come in many forms. As long as the patient attaches some symbolic significance to a medical intervention, any intervention can be considered a placebo. Though they are rarely considered placebos, the healing relationships between patients and their providers often have profound symbolic significance. None of the research reviewed in the NEJM article took the nature of relationships into account. This complexity makes it difficult to isolate and define what is placebo and what is not.
Describing the Effect
Researches define the effects of an intervention as either objective or subjective, and either binary or continuous. Objective effects are those that can easily be measured, such as blood pressure, while subjective effects depend entirely on a person's description (eg, level of stress). Binary effects have just two possible values (eg, alive or dead), whereas continuous ones may take many values, such as weight.
The NEJM study authors found that, on average, placebos only had an affect on subjective continuous effects, like intensity of pain. The authors also noted that few researchers ever attempt to measure quality of life. This is not surprising since it is very difficult to quantify outcomes, like sense of well-being, that matter most to patients.
The debate over the placebo effect continues. What the research tells us is that our original idea of the placebo effect, as Dr. Beecher described it, was overly simplistic. Patients clearly attach enormous significance to the words, actions, and prescriptions of their healers. Rather than disproving its existence, we are more likely to discover that the true power of the placebo lies deep within the most meaningful healing relationships.