The Power of Placebo In Weight-Loss TrialsFriday, April 8, 2016
Placebos are powerful – even when people know they’re on placebo!
This scientific fact has been long established and probably accounts for much of the hype and “success stories” around dietary supplements and a host of “alternative” medicines, which often lack any demonstrable “biological” effect (other than placebo).
Although placebo-controlled trials are generally considered the highest-quality evidence, it is often wrongly assumed that the placebo group is not receiving treatment, when it actually is.
Thus, contrary to a common belief, taking a placebo is a form of “active” treatment, and cannot be equated to “doing nothing” (which is perhaps why even study participants randomised to a “placebo” or “control” group in clinical trials often do so much better than individuals who are not in a trial).
Now a paper by Kevin Fontaine and colleagues, in a paper published in OBESITY, look at the problems around the use of “placebos” in weight-loss trials.
As the authors note,
“It is becoming increasingly clear that placebo responses operate not merely from believing one is ingesting a bioactive compound or undergoing a legitimate medical procedure but also from environmental and psychosocial factors embedded within clinical and research encounters. These encounters may engage several potential conscious and non-conscious mechanisms, including expectations, conditioning, anxiety reduction, interactions with practitioners/research staff (e.g., emotional support, empathy), and so on. Collectively, placebo-related factors may promote improved outcomes even when participants do not receive a bioactive treatment and regardless of whether they are aware of this.”
The authors cite a number of studies that have specifically looked at the power of “placebo” in weight loss trials.
In one study,
…participants consumed identical 380-calorie milk shakes. On one occasion, the label said it contained 140 calories (“sensible shake”), and on another, the label said it contained 620 calories (“indulgent shake”). The “indulgent” condition produced a steeper decline in ghrelin (a satiety hormone) and greater self-reported satiety compared to the “sensible” condition.”
“…all participants [received] a placebo pill and [were] then randomized to conditions in which they were told that they had taken either a placebo or weight loss pill. Those who were told that they had taken a weight loss pill ate more during a taste test and preferred larger quantities of sugary drinks compared to those who were told that they had taken a placebo. The findings suggest that those who believed that they took the weight loss pill may have perceived that the medication gave them greater “leeway” with regard to intake when offered food during the taste test.”
Thus, it is important to remember that,
“….disclosing information about specific aspects of the trial, even during the consent process, may influence participant expectations and potentially affect outcomes.”
(Incidentally, the preceding statement raises the issue of whether or not “full disclosure”, as prescribed by current ethical practice in clinical research, in fact obstructs rather than supports good research – but that is a story for another day.)
It is evident from the examples that the “placebo effect” can make an “active drug” seem less effective than it is, partly because weight-loss studies are virtually never perfectly blinded.
Take for example the case where people randomised to the “placebo” group lose considerable weight due to the “placebo effect”, whereas, the “active drug” group, figuring out that they are on the active treatment, rely more on the drug to do its thing than individuals who are evidently on placebo, thereby losing considerably less weight than if they took the active drug while believing they were on placebo.
Obviously, this poses all kinds of challenges for obesity trials (or for that matter any diet or exercise trial), where it is virtually impossible to conceal from the participant the group they are in.
Take for example a study that is advertised as studying the effect of carb-restriction on weight loss, which naturally will attract people who at least to some extent believe that carb-restriction is indeed the best way to lose weight (which is why they are interested in the study).
As there is no way to blind the diet that participants are on, it is easy to imagine that those randomised to the low-carb diet (who happen to also believe in its power) do better than those on the control diet (who automatically assume that they will lose less weight and likely end up doing exactly that).
Although such a study may easily be touted to demonstrate the superiority of a low-carb diet, its results could also be fully explained by “placebo-related effects”.
It is important for us to always remember the power of “placebos” in the weight-loss literature, where “doing nothing” simply be virtue of being in the study, is never “doing nothing”.
Friday, April 15, 2016
Most instances of “placebo effects” are explained by a statistical phenomenon called “regression toward the mean”. Unless specifically controlled for this phenomeon, it is not meaningful to discuss power of placebo.
Wednesday, April 20, 2016
I loved this sentence in the conclusion of this paper:
Therefore, we suggest that, as testing and imaging techniques become more sophisticated, study designs might begin to include placebo conditions that measure potential genetic, neurobiological, and psychophysiological placebo responses, which might serve as biomarkers of the placebo response.