What is a placebo?

I've seen so many attempts at defining a placebo. The most common is that it is "an inert substance", a "sugar pill". This definition has become popular because historically most discussion surrounding the placebo effect pertained to the use of "placebo" groups in drug trials. In this case, subjects were randomly assigned to receive either an experimental drug with some pharmacological agent(s) being tested or to receive a placebo, a pill that looked, felt, and smelled the same as the active one but which was minus the pharmacological agent(s). This methodology reflected the recognition that people will improve on some outcome for a variety of reasons having nothing to do with the drug being tested.

Say Pfizer is testing a new analgesic for arthritis pain. They set up a clinical trial and assign half to receive their new drug and half to receive an inert counterpart, a "placebo". Participants keep diaries of their pain for a week before taking their pills (so a baseline can be established), then take begin taking their assigned pills. After a month, all the data are collected and analyzed and it is found that indeed those in the active drug group improved, they felt less pain under the influence of the pill compared to their baseline period. But so did the "placebo" group! How can this be? How can people suffering from real pain experience relief from nothing?

The answer is that any or all of several phenomena can occur concomitant with the administration of a "placebo" to cause improvement.

1. Spontaneous remission. This means that any condition can improve on its own, without any intervention. There can be many reasons for this: during the period weather conditions may have become more favorable, lessening pain for all. All chronic medical conditions, including pain, have periods in which symptoms wax and wane.

2. Regression to the mean. Clinical trials of drugs (and other treatments/therapies) tend to seek participants whose pain is especially bad, longstanding, etc. Additionally, people tend to seek out and agree to participating in clinical trials when pain is at its worst. But pain is not always at its worst. Pain varies around some average (i.e., the "mean"). Sometimes it's worse than average; other times it's better than average. Beginning a clinical trial when pain is worse than average, means that it is likely that pain will begin to slide back toward the average (i.e., improve) for no other reason than the natural ups and downs of pain.

3. Detection ambiguity. On a scale of 0 (no pain at all) to 10 (worst pain imaginable), please rate your pain. Sound simple and straightforward, right? Well, not necessarily. Pain is highly subjective. Diabetes, cardiovascular disease, and cancer all have established objective indices that can objectively quantify their state. For example, a simple pin-prick blood test can produce a index of blood glucose, which is a reliable indicator of the presence of diabetes and of how well it is being managed. But pain has no such measure; it is inherently a subjective phenomenon and so we must rely mainly on people's reports of their pain. This is all well and fine but there is ambiguity in how we feel.

Now, imagine that you are led to believe that you have been given a powerful painkiller and then asked to rate your pain on that 0 to 10 scale. Believing you received an analgesic may prime your attention toward the detection of evidence of reduced pain. Pain naturally waxes and wanes. Before participation in the drug trial, you might have interpreted the waning of pain for what it is -- a valley in an unending series of hills and valleys. But because you are concomitantly taking what what you believe is an analgesic there is a greater likelihood that you will interpret the waning of pain as evidence that that the drug is working.

Also, detecting pain in the body is a bit like detecting the temperature outside. Numerous factors can influence your perception of temperature: humidity, wind, how sunny it is, your own mood and anxiety level, etc. We might say that the influence of all these factors leads to ambiguity in detecting the temperature. An implication of this ambiguity is that your estimate of temperature can be manipulated by getting you to focus on some of these factors more than others, which is essentially what drug trials can do.

4. Placebo response. The placebo response is that










So clinical trials of drugs must show that a drug produces improvements that are over and above