Science journalist Sharon Begley wrote a recent piece in The Saturday Evening Post about Placebo Power. The piece, while generally better than the typical popular writing on placebos, still falls into the standard placebo narrative that is ubiquitous in the mainstream media. The article is virtually identical to a dozen other articles I have read on placebo effects in the popular press, and most significantly fails to even question that narrative.
Begley is generally one of the better science journalists, although I have had my disagreements with her – specifically over her attitude toward the relationship between skeptics and the media. She seems to have a distorted and negative view of skeptics and does not think that the media can or should help us in our “debunking crusade.” (The term itself speaks of a fundamental misunderstanding of the modern skeptical movement.)
I have also parted ways with Begley over her view of the relationship between science and medicine. She seems to have a fairly negative view of doctors, fueled in part by her imperfect grasp of medical science. This is the risk with even the best lay science journalists – science is often complex and it is difficult to master the nuances if you are not an expert and steeped in the evidence and the community. Further there is a tendency for people in general (including journalists) to go along with an appealing and available narrative. (For journalists those narratives that are appealing are the ones that make good headlines.) These shortcomings are present throughout her recent article on placebos.
I will start with what the article does well – break down several reasonably established mechanisms for certain placebo effects. These include endorphin release for pain, dopamine release which serves as both a reward signal and also may temporarily reduce certain symptoms of Parkinson’s disease, those caused by a lack of dopamine (although I consider this evidence preliminary), conditioning, and ritual. That is where she does better than most. Still, she follows the standard narrative of proclaiming in breathless terms the surprising power of “the placebo effect” without putting it into a reasonable perspective, or mentioning very high profile data that tells us what placebo effects are not. She writes:
“It is tempting to say that “mere thought” or “mere belief” caused these patients to feel and function better, just as the child’s trust in her mother made her knee feel better and our belief that little white pills will relieve a headache made the calcium tablet do so, even though it contained not a speck of headache-fighting medication. But if doctors and scientists have learned one thing about the placebo response or placebo effect, it is this: There is nothing “mere” about how thoughts, beliefs, and the power of the mind affect the body.”
And there is the standard narrative. I don’t think it is nitpicking to complain about the term “the placebo response or placebo effect” in the singular, which Begley uses throughout her piece. There are many placebo effects, which are different for different conditions. I know it’s probably hopeless at this point to expect general use to change – “the placebo effect” is likely to remain a common vernacular short hand. However, in an article entirely on placebos such distinctions should be made explicitly clear.
The only proper reference of “the placebo response” is the outcome measured in the placebo arm of a clinical trial, where use of the singular is proper because the study is measuring a net combined placebo effect. When we start talking about the placebo phenomenon, however, we have to start talking about placebo effects in the plural. One of the major conceptual problems with the placebo narrative is conflating the placebo effect that is measured in a clinical trial with all the various placebo effects that may make up that measured response.
This is where the placebo narrative almost always breaks down – popular writers like Begley talk about physiological mechanisms like endorphins and dopamine without mentioning all the other effects that are being measured in those clinical trials they are citing as evidence for how powerful those mechanisms are. What is measured in the placebo response includes things like reporting bias, or the desire of subjects to feel better, to please their doctor or the researcher, and to justify their prior decisions (to trust the doctor, take the treatment, enter the study, etc.). Begley fails to distinguish, in other words, between the pain that patients feel and the pain that they report. In studies we never know and cannot measure how much pain patients feel, only how much pain they report. Anything that would affect that reporting will also be lumped into “the placebo response” that is measured in the placebo arm of the study.
This is not quibbling. There is good reason to believe that reporting bias may be the major component of measured placebo effects.
Begley also fails to mention non-specific responses. Anxiety reduction from the belief that one is getting treated can certainly improve one’s mood and outlook on their symptoms (and their reporting of symptoms). Being in a study also involves getting medical attention and is associated with better self-care as well. This is partly the Hawthorne effect – looking to see how patients do is likely to have an effect all by itself, improving compliance and self-care, for example.
I do have to give Begley credit for referring to “the placebo effect on pain” in one instance in her article. This is at least better than most. When discussing specific physiological mechanisms for certain placebo effects, it must be qualified what outcome is being considered. Placebo effects for pain may not be relevant at all to placebo effects for any other condition. Such articles also frequently fail to mention that placebo effects for pain tend to be much higher in magnitude than for other outcomes. In fact – placebo effects are generally restricted to subjective outcomes (ones, probably not by coincidence, that are dependent on reporting and cannot be objectively measured).
In fact Begley did not mention the now famous (or infamous, I guess, depending on your perspective) study by Kaptchuk in which he studied both objective and subjective outcomes in asthma in response to active and placebo interventions. The results – there was a measured placebo response to subjective outcomes, but none for objective outcomes. Asthma is a condition susceptible to things like anxiety and expectation – so it provided a good opportunity to demonstrate objective improvement from placebo interventions, but showed none. This study, while very telling, does not fit the placebo narrative that journalists like to tell, and so is often absent from such articles or misinterpreted when present.
Begley does mention the “placebo effect without deception” study. There are two main problems with citing this study the way Begley (and almost every media report about it) does, as pointed out previously here by David Gorski. The first is that the effect size is small (slight improvement), and may not even be clinically relevant. They were also barely statistically significant, and one outcome (quality of life) was not significant. These are unimpressive outcomes. Second – the placebo without deception wasn’t quite deception free as claimed. Subjects getting open-label placebo were told:
“Additionally, patients were told that “placebo pills, something like sugar pills, have been shown in rigorous clinical testing to produce significant mind-body self-healing processes.””
The bit about “rigorous clinical testing” and “significant mild-body self-healing processes” profoundly overstates and misrepresents the evidence. In other words – it’s deception. It’s just deception with plausible deniability. This hardly settles the issue about the effects of “open label placebos” without deception. But it does fit nicely into the placebo narrative, and Begley dutifully reports the standard presentation of this study without a hint of journalistic skepticism.
Conclusion
I had the pleasure of interviewing Christopher Hitchens a few years ago. He had a dim view of journalists in general, and for me one of his most memorable observations was that journalists generally accept the narratives of the day that are handed to them. Precious few do the hard work of investigative journalism to find something closer to the true story. Based on my experience I have to agree.
Unfortunately the placebo narrative that Begley regurgitates is now well-established in the media and we are likely to read some version of it over and over again. I have become a bit harsher over time on journalists for doing this because there are many articles, by experts, just a Google search away, that point out the various flaws in this narrative. Those skeptics that Begley seems to dismiss have done the hard work for her and other journalists of actually reading the original research, digging down to the salient details, and teasing out the nuances that make all the difference to a proper interpretation of a complex clinical issue.
NOTE: Orac has also weighed in on Begley’s article.