The New England Journal of Medicine published in the July 14 issue, a study of 46 patients with asthma and compared albuterol, sham acupuncture, placebo albuterol, and no intervention. Each patient had one of the four done in random order at four office visits. Measured outcomes were self-report of breathlessness and lung function. The lung function test used was the forced expiratory volume in one second (FEV1) which is a measure of how hard can you blow out for one second and is a traditional asthma outcome.
In brief summary, albuterol improved function, and all three 'treatments' improved the reported symptoms, but the "no treatment" did not increase function or improve symptoms. Rather than concluding that albuterol is the best of the four options for treating asthma, the authors conclude that self-reported symptoms are not a reliable outcome to measure and that placebo effects can be clinically meaningful even when compared to active treatments.
Peter Lipson, blogging for Science Based Medicine, has a much more thorough review of this paper, but one of his conclusions is the same as one of my first reactions to the paper. Namely, the placebo effect is probably largely due to the interaction of being kind and caring for someone, which (miraculously!) is part and parcel of real treatments as well as placebos. Therefore, we can focus our attention on real treatments without having to worry about "losing" the benefit seen with placebos simultaneously making our patients feel better and actually be better at the same time.