In clinical trials, a sham treatment is a medical procedure, analogous to a placebo, which is given to a control group of subjects, to enable the effects of the supposedly "active" treatment to be assessed objectively. A sham treatment is not necessarily expected to be ineffective; it may have effects derived from the placebo effect or nocebo effect, but there may also be other 'non-specific' effects of the sham treatment. For example, if a sham treatment involves a physical manipulation, there may be unanticipated physiological consequences of that manipulation which affect the trial outcome. The purpose of a sham treatment group is to identify any specific benefit of one element of a medical treatment above and beyond all benefits that might be attributed to everything else about that treatment.
Sham treatments are also an essential part of the rigorous design of experimental studies in biology, whether on humans or animals, and indeed also on studies of isolated cells in vitro. Studies of the effects of any drug on cells must be controlled for strictly be a sham treatment involving exposing the cells to the drug "vehicle" (the solution it is dissolved in) alone, using a protocol otherwise identical to that used with the drug. Studies in animals of the effects of a drug microinjected into a small defined region of the brain would be expected to be controlled for by a matched sham group, treated identically to the experimental group but with vehicle microinjections. In this case there are expected effects of the sham treatment (effects due to anaesthesia, surgery and to incidental, unavoidable damage to brain tissue, and possibly effects of the particular vehicle).
In a double-blind study of a drug, the clinician gives the patient a pill or injection, without knowing if the pill or fluid contains active drug or only inert ingredients (i.e. a placebo). In the double-blind technique, neither the clinician nor the patient can be influenced by knowing the actual contents given.
However, if the sham treatment involves a physical manipulation in the control arm of a randomized controlled trial, it is impossible for the clinician performing the procedure to be "blinded" to the treatment. For example, in a study of acupuncture, needles are inserted in both the control and experimental branches of the trial. In the control arm, the needle is not inserted in a designated acupuncture point, but in an area of skin considered neutral in acupuncture theory. Sham acupuncture, given the extremely thin needles, is rather benign, as it involves no significant risk or pain.
Sham surgery, however, is in a different category of risk and discomfort, and raises ethical questions. General anesthesia, if required, is definitely not a procedure without risk. Further, for the patient to believe that surgery was performed, there must be at least an incision, which further introduces risk and probably postoperative pain. Even a superficial incision has some risk, but if the research trial is of a surgery that requires opening the chest, skull, or abdomen, it is unlikely that a safety board will permit that risk.