Article in Slate HERE by Darshak Sanghavi, pediatric cardiologist at U. Mass.

Dr. Novella has an interesting discussion of the phenomenon and Dr. Sanghavi’s article HERE.

Interestingly, although they agree that the placebo effect is not real, Novella argues that trials should be placebo controlled so as to “minimize bias” by blinding the doctors as to which patients are getting treatment and which ones are not.

Dr. Novella is exactly correct here. I’m not convinced that the reasons Sanghavi cites for why we include placebo in clinical trials, or the impression he has about how phsycians view it, is true. Placebo controls are not included because the placebo effect can cure disease, and I doubt most doctors think so. It is included, as Novella says, because it can create the impression, in patients and researchers, that it has improved the disease. Accepting that 30-50% of patients in most trials receiving placebo show improvement doesn’t mean we think the placebo, and the power of positive thinking, is making them better, only that the power of psoitive thinking and self delsuion is making them, and unblinded researchers, evaluate them as better. I don’t think it is possible to underestimate the power of subtle bias. A new drug recently entered the veterinary market to control nausea and vomiting. In a clinical trial (unblinded) 75 % of patients (dogs) who recieved it stopped vomiting. But 50% of dogs who recieved the placebo also stopped vomiting. Now I sure don’t believe that taking a sugar pill creates a will to heal in a dog that actually effects disease. Yet we see placebo effects of roughly the same percentage of subjects in vetreinary trials as they do in human trials. Something is biasing the results, and the placebo control is necessary to account for it until, and if, we figure out what.

Interesting, Brennen, that hadn’t occurred to me. Yes, there’s no way that we’d expect to see a placebo effect in animals, since they don’t know they are being treated for anything. They can’t expect to get better by taking a pill, since they wouldn’t know what it was for. And anyway the pill would probably be hidden in their food ...

Has there been much literature about “placebo effects” in animals?

I think it’s sort of like a comment that a psychiatrist speaking on PBS quite a few years ago made.  It was something like this.  “Two years of treatment seems to be able to cure two-thirds of patients with depression.  This appears to be a decent success rate until we realize that two thirds of depressed people who receive no treatment improve significantly after about two years.” 

Quite likely, if the study had included a third sample of dogs with a vomiting problem but which were not given any pills, either the treatment or the placebo, fifty percent of them would have recovered from their vomiting problem.  We must recognize that the animal body, human or other, has a great ability to repair insults or problems without any outside treatment.  This has to be taken into account when any medication is tested.

Occam

Well, I’m not sure how much research there is on the placebo effect specifically, but good studies usually include one. THIS is a link to a journal article on the effect of gold-bead implantation of hip arthritis in dogs. What is interesting about it is that owners’ perception of what treatment their pet is recieving affects their evaluation of the clinical signs. If they guess they’re in the treatment arm rather than the placebo arm, they rate the success higher than if they guess or know they’re not getting the actual treatment. So the effect is likely largely one of perception rather than physiology, at least in animals. But I’d be interested to see if it affects more objective measures since this might imply either 1) a real physiological effect or more likely 2) a degree of extremely subtle and hard to control bias even when measuring something supposedly purley objective. I’ll look into it.

Brennen, I think this is actually an extremely interesting and potentially new angle on the question of the placebo effect ...

I haven’t been able to find much research specifically looking at placebo in non-human subjects, though I suspect it’s out there. I do know placebo controls are ubiquitous in research on veterinary therapeutics and effects seem to be about what one would expect in human subjects. Interesting indeed. The quote below is about the only comment I could find specifically addressing the topic, and it doesn’t say much.

Placebo Effect
Pain Management for the Small Animal Practitioner; Second Edition
William J. Tranquilli, DVM, MS, DACVA; Kurt A. Grimm, DVM, MS, PhD, DACVA, ACVCP; Leigh A. Lamont, DVM, MS, DACVA

 

The term placebo literally translates to “I shall please.” All proposed mechanisms for the placebo effect suggest an interaction of brain states and somatic health processes (Table 1-1).

When objective and subjective pain assessment scores improve with the administration of a placebo, interpretation of the effectiveness of analgesic treatment is difficult.

Expectancy (a probable mechanism of the placebo effect in humans) requires the ability of the patient to comprehend and anticipate a response to a treatment (e.g., faith or hope). The existence of expectancy in animals is questionable. However, expectancy may lead to observer bias when assessing response to treatment.

Table 1-1. Some proposed mechanisms for the placebo effect in animals.

Human Contact—Visual and tactile contact from a human can cause changes in the subject’s physiologic state, response to painful stimuli, and productivity.
Classical Conditioning—Pavlov first demonstrated classical conditioning with morphine in dogs. Classical conditioning usually requires repeated exposures to the conditioned stimulus and therefore may be a factor in chronic pain treatment.
Endogenous Opioids—This mechanism may explain analgesic placebo responses, but not physiologic responses such as immune modulation or tissue healing.