Placebo Effect and Adherence: A Challenge to Pharma?

As all four of the regular readers of this blog are aware, I am an avid listener to Rick Bukata's and Jerry Hoffman's monthly audio recordings, Primary Care Medical Abstracts (free advertising for them: www.ccme.org). So here I am all innocently driving my car and listening to the CD for their August 2011 issue, when I am shocked to hear the CD shouting out my name. Jerry is asking that I respond to a query about one of the papers he and Rick had been discussing, in my capacity as someone who has a long-standing interest in the placebo effect.

So here is my answer to Jerry. You'll naturally wonder why it's here on this blog which is about ethics and Pharma and not about placebo effect. In the end I'll suggest an important connection.

The paper that started all this is a thoughtful editorial by Wilson (subscription required) about the placebo effect and adherence. There have now been a good number of studies that show that when you do a double-blind trial with a placebo arm, there is quite often (indeed rather consistently) a significant improvement of outcomes among those who take their placebos faithfully, compared to those who are relatively non-adherent to taking their placebos. Wilson does a neat analysis of what we know, and what we don't yet know, about this adherence phenomenon, and suggests linkages to what we are learning about placebo effects and why this phenomenon might be viewed (at least as a working hypothesis) as a variant of placebo effect.

Jerry then raises the question: what's the message here for clinicians? Should we give patients pep talks to try to both encourage and energize them about the treatments we're prescribing (whether drug or nondrug) to try to enhance their expectations of a good outcome, which has been shown to be positively associated with a placebo response? Should this pep talk include advice to be sure to take their pills (or other treatments) faithfully? Or might it be the case that the adherence research shows that what really matters is what's already inside the patient's head, not what we say--that those in the trials that were more adherent were simply that sort of person, and being that sort of person is what matters in terms of triggering a placebo effect--and our pep talk is worthless?

OK, Jerry, here's my answer, followed by my hunch.

My answer, as Wilson's nice review suggests, is: we don't know. No one has yet done the sort of fine-grained study of the more-adherent research subjects, analogous to some of the latest generation of placebo-effect research done in the last decade with brain imaging etc. So the underlying psychological and neurochemical factors that might explain the adherence-placebo effect link are unknown.

Now my hunch. The placebo effect is almost certainly multifactorial. Indeed, Fabrizio Benedetti of Turin titled his excellent 2009 book Placebo Effects (rather than Placebo Effect) to make the argument that continued research will almost certainly reveal multiple underlying mechanisms that may operate in different diseases and different organ systems.

The best available evidence that we have suggests two very general psychological mechanisms for most placebo effects--expectancy and conditioning. Expectancy is basically forward looking--your body is likely to heal itself when you think it will get better in the future. Conditioning is backward--your body is more likely to heal itself when you associate the circumstances you're in now with circumstances in which your body experienced healing previously.

The adherence effect probably partakes of both. Subjects who take their medicines regularly probably anticipate a good outcome with greater confidence. These same people probably got better in the past when they religiously took their pills, and so conditioning can contribute to their getting better this time by reactivating the same neural pathways.

Now what happens if a physician acts enthusiastic and encouraging about the nature of the treatment? This is likely to increase both expectancy and conditioning effects--expectancy for obvious reasons, conditioning because the patient probably associates an emotionally supportive environment with past healing (going as far back as when Mommy kissed your boo-boo and it got better afterwards). So I cannot see how the encouraging physician could detract in any way from the patient's inner tendency to experience an adherence-placebo reaction, and I can see several ways that the former might enhance the latter. So: bring on the pep-talk.

I have recently become interested in the connection between placebo response and medicine viewed as ritual/performance/theater. We scientific types are used to dismissing ritual as meaningless superstition, but the current placebo research indicates the neuroanatomical and neurochemical reasons why ritual can be efficacious in changing bodily function as well as in altering our cognitive and emotional views of the world. Much of medicine, when we think about it, is ritual and/or performance. (Science writer Nicholas Wade once wrote something like, "All medicine is a form of theater.") Rituals include taking one's pills once or several times a day and can readily trigger both expectancy and conditioning responses. Smart physicians who prescribe exercise and other lifestyle changes try whever possible to suggest rituals to patients to increase adherence, in some cases going so far as to write the instructions on a prescription pad, which when in practice I always found especially powerful. All of these measures seem well calculated to increase placebo effects, as well as to make patients healthier by way of the drug or the exercise or whatever.

I promised in the end to bring this back around to Pharma, so here goes. I suggest that you read Wilson's article and look especially at the effect sizes reported for the adherence-placebo effect. Just for example: Mortality difference between adherent and nonadherent placebo group subjects in the Coronary Drug Project (1980): 15% vs. 25%. More recently, adherence effect in mortality in heart failure, based on the SOLV-TT and SOLV-PT trials: hazard ratio 0.52 (Avins 2010). And remember, according to the body of data Wilson reviews, these are not one-time flukes. And remember too we are talking here about people keeling over dead, not some meaningless surrogate endpoint.

Why is this of importance? The drug companies would kill to come up with a new drug that had efficacy numbers this good. So you could spin this in a pro- or anti-industry fashion. On the side of the industry, have a pity--look what they have to overcome to show that a new drug is better than placebo, when the placebo effect alone can be this powerful. But on the other side of the coin, when mere encouragement and positive thinking can have this much beneficial effect on patient outcomes, why in heaven's name would we want to give drugs that have dangerous side effects and that cost an arm and a leg, unless the drugs had been clearly shown to be really superior to cheap and safe encouragement?

Wilson IB. Adherence, placebo effects, amd mortality [editorial]. Journal of General Internal Medicine 25:1270-1272, December 2010.

Avins AL, Pressman A, Ackerson L, et al. Placebo adherence and its association with morbidity and mortality in the studies of left ventricular dysfunction. Journal of General Internal Medicine 25: 1275-1281, December 2010.