Placebo

Placebos are medications that have no pharmacologic action but somehow exert a beneficial effect. This is actually a multifaceted and complex phenomenon. First, someone might feel better after taking a placebo simply because of the natural course of the disease. For example, if you always take something at the peak of a headache, your headache will get better no matter what you take (since it was at the peak). The effect of your headache getting better could be mistakenly attributed to the medicine rather than how headaches normally progress. Also, medications are favored by a "regression to the mean" effect. That is, when you see a doctor because of chronic low back pain, it's usually because you had an episode of extra severe pain. Well, it's likely that subsequent pain episodes will be average (that is, less painful than the episode that brought you to the doctor). Lastly, the actual placebo effect is an effect based on suggestion and expectation. If you think something will get better, it might be more likely to get better. Indeed, endogenous opiates are released when placebos are administered; your body produces pain-killers simply from the suggestion that you're actually getting pain-killers.

I think this is interesting, especially since placebos are standard in the design of clinical trials. Surprisingly, the very first clinical trial conducted by James Lind in 1747 on citrus fruit and scurvy included a placebo.