Placebos and Autonomy

A new study finds that about half of American doctors prescribe placebos for their patients. In order for the placebo effect to work, the patient must believe that the prescribed treatment is generally effective, something the doctor knows to be false. Hence the doctor is misleading the patient, albeit for the patient's own good. This is a clear insult to patient autonomy; the doctor is not being a partner in health with the patient when seeing if the placebo effect -- a very real phenomenon which does lead to the patient's own body curing the ailment -- will work in the case, rather the doctor is putting him/herself in a paternalistic position.

The doctor is lying to the patient for the patient's own good, yet for the doctor/patient relationship to be successful surely requires a significant degree of trust, trust undermined by this approach. Placebos in research are one thing where the participant not only knows of the possibility, but knows that it is a necessary structural element. With one's own personal physician, though, one does have the expectation of both honesty and good faith attempts to improve your health. Is there something wrong with placebos in this case? Is it morally permissible for a doctor to put health promotion above honesty in a patient/doctor relationship? If the placebo doesn't work, would the patient be right in being angry or should s/he be glad that his/her doctor was pursuing every avenue of treatment?