The placebo effect is a very powerful psychological phenomenon whereby people think they have improved even though they weren't given any therapeutic agent. The effect can significantly skew the results of medical trials and make it impossible to accurately assess the effects of medications being studied. To overcome this, a placebo is used as well as the medication being studied. The placebo is designed so that it looks and tastes the same as the medication being studied and the patients cannot distinguish between the two. After the study is complete, the medication is determined to have been of benefit if the proportion of patients on the medication who reported improvement was larger than the proportion of patients on the placebo who reported improvement. This is called a 'placebo controlled trial'.
An even better designed medical trial is the type termed 'double blinded, placebo controlled' where neither the patients nor the researchers actively involved in the trial know who is getting the placebo and who is getting the real medication. Often a computer determines who gets what. This is done as the people who administer the medication may influence the patient's responses as hard as they try not to. Only after the trial is over is the 'code broken' and the results analysed.
Medical treatments using large numbers of patients are preferable to smaller numbers. It is also preferable that medical treatments be assessed by a number of different research groups to see if they each get the same results.
The inclusion of healthy people, also sometimes called 'controls', is often desirable so that measurements can be compared to the patients being studied.