The New England Journal of Medicine recently published a troubling article on acupuncture, which was ably deconstructed by Dr. Mark Crislip. This incident has reignited a discussion of what, exactly, "placebo" is.
A common argument is that placebo is like any other intervention, something that can be intentionally harnessed for the benefit of patients. This is both true and overly simplistic.
First, we must review what "placebo" is. There are basically two primary uses of the word. The first is technical. In randomized controlled trials, making sense of an intervention requires some sort of control group. If I give Fabulostatin to one group, any changes I see in the group may be due to chance alone or to bias. I can minimize this effect by subjecting a similar group to an identical-appearing sugar pill, one which we call a "placebo". If the only significant difference between the two groups is Fabulostatin or placebo then significant differences between the two groups can be more readily attributed to Fabulostatin.
This doesn't mean that the placebo group will be unchanged. If there is a third arm of the study, one in which no pill was given, we can often measure a difference between this group and the placebo group. This difference is usually called the "placebo effect", that is, the group that is treated with a sugar pill improves despite being given no active drug.The second use of the word is in a therapeutic sense. A doctor may give a patient a pill or treatment they know to be essentially inert, but then give the patient the explicit expectation that it will help them. The patient may indeed feel better. This is a tricky concept, one that is difficult to measure. It is also ethically problematic.
There is little doubt that a visit to a physician is more than just a trip to a drug dispensary. Simply visiting a compassionate doctor can improve subjective measures of health. But can this be harnessed in a more specific way? According to this study, yes. But the truth is rarely this simple.
Most studies that claim to use the placebo effect as a therapy (including the study in NEJM) have a few things in common. They confound the two main meanings of placebo, and they evaluate subjective measures of self-limited conditions. One way to look at low back pain, for example, is that medical science is not very effective. Most interventions, including anti-inflammatory medications, back exercises and physical therapy, are not much more effective than doing nothing. Most low back pain, even severe low back pain, resolves over the course of several weeks, no matter what intervention we use. The use of placebo (whether we call it a sugar pill, or acupuncture, or chiropractic) fails to improve on the natural course of the disease. When we administer "placebo" to these patients, it is tempting to take credit for their improvement, but all we've really done is let nature take its course. Sometimes this is a good thing; if we help steer a patient away from back surgery when they don't need it, we've helped protect our patient from harm. But this is qualitatively different from "doing something".
There are significant ethical problems in using an inert treatment and taking credit for its "success". Medical ethics (at least in North America) generally frown on deception. Deception removes a patient's ability to give informed consent for a treatment and creates a destructive doctor-patient relationship. If I lie about a sugar pill, why should the patient trust me about a blood pressure pill?
Ethical problems aside, placebo isn't a therapy in the same sense as surgery or medication. It is the sum of many otherwise-intangible effects, such as being listened to, touched, and cared for. There is no need to add needles, back-cracking, or sugar pills. It's time to give up the idea that we can ethically prescribe placebo treatments. Prescribing therapies that are "as good as placebo" simply creates deception, cost, and possible harm through the application of physical interventions that cannot improve upon holding a patient's hand and listening to them.