Larry Husten has a typically excellent blog post up entitled “What Role Should Coca-Cola Play In Obesity Research?” An excerpt:
The question is prompted by a recent article in the Journal of the American College of Cardiology. The “state-of-the-art” paper reviews the relationship of obesity and cardiovascular disease and presents the case that a decline in physical activity is the primary cause of the obesity epidemic. The article downplays the role of calories and diet and does not include the words “sugar,” “soda,” or “beverage.” Three of the five authors of the paper report financial relationships with Coca Cola.
Husten requested and obtained comments from the lead author of the paper, Carl Lavie, in response to a series of (Husten’s) sharp questions. Dr. Lavie summarizes his views:
1) Hopefully the science speaks for itself, regardless of the sponsor.
2) Science itself is improved by data and facts.
3) Not to be too harsh, but my opinion (based on my interpretaion of the science) is not for sale and cannot be bought. I am sure my colleagues feel the same way.
4) All researchers have bias, as do I, but if you try to eliminate every researcher/scientist/clinician with bias, there will be hardly any left.
5) Therefore, the best way to proceed is to focus on well-designed and well-executed studies.
Dr. Lavie also went on to note that, for example, he deemed any restriction on funding of scientific research from the tobacco industry permissible so long as it was disclosed, and also that unconscious or implicit bias is only a concern when one is working directly for a company. A “minor honorarium would hardly taint” one’s views, according to Dr. Lavie.
Here’s what I have difficulty understanding, as Husten goes on to point out: there exists a substantial evidence base regarding motivated bias and financial conflicts of interest in medicine and science that undermines at least some of the points made by Dr. Lavie. Science, and especially epidemiologic data, never speaks for itself; it is always produced by actual people, is highly uncertain and ambiguous, and inexorably requires interpretation on which reasonable and informed people often disagree. Moreover, people’s scientific opinions and clinical practices are indeed affected by relationships with relevant actors, including financial relationships. There is overwhelming evidence of this (see the work of George Loewenstein for a long-running and robust example of this evidence).
It is simply not evidence-based to suggest the contrary. So what is surprising to me in my experience teaching ethics & COIs is how frequently people who (correctly) insist on the significance of following rigorous evidence in terms of clinical practice seem to offer opinions on the effects of COIs that IMO do not sufficiently reflect what the best evidence on motivated bias actually shows.
When I teach conflicts of interest, which I do fairly consistently, I always announce to the learners (at any level — undergraduates, medical students, residents/trainees, and even faculty in grand rounds, etc.) that my approach is as follows:
Reasonable people of good conscience can disagree on what, if anything, ought to be done about COIs in medicine, science, and public health. However, there is much less room for reasonable people to disagree on what the best evidence actually shows regarding the effects of motivated bias on human behavior. Most people’s behavior is so affected. Now, maybe any given individual will be one of those who are not so affected.
But the odds are ever not in your favor.
Moreover, the problems with COIs in medicine and science have to be understood on an iterative basis. The concern is not necessarily that in any given case, any given individual will exhibit what Andrew Stark terms “behavior of partiality.” The concern is rather that, over the long run of cases, people who have significant (and especially) financial entanglements are at much greater risk of making subtle, often unconscious decisions that may reflect the wishes and preferences of those with whom the moral agent is entangled in ways that are not in the best interests either of individual patients or of population health.
So, the learning goal here is to practice what might be called “evidence-based bioethics.” I try to get the learners on the same page in terms of what the social science literature suggests in terms of the effects of motivated bias (subject to all of the usual qualifications when discussing an evidence base, i.e., limitations, disagreements, trends, quality, etc.). Getting to that baseline is the most important portion of the learning, and typically sets up a nice, informed, and open discussion regarding what, if anything, ought to be done in the face of that evidence.
But we have to acknowledge what that evidence shows — or at least indicate an informed basis for rejecting it.