February 12th, 2013

Conflict Of Interest Is A Complex Issue


In a recent editorial in the Journal of the American College of Cardiology, Robert Califf discusses a few of the many complex questions related to the conflict of interest issue. Here CardioExchange’s Harlan Krumholz further explores the issue with Califf.

CardioExchange: Do you think that conflict of interest issues are mostly an issue about optics and not substance – or are you concerned that relationships with funders sometimes influence the science in ways that are not favorable to society?

Califf: I believe that financial conflict of interest has real substance, but that focusing exclusively on the medical products industry and failing to consider other sources of conflict of interest is a huge mistake and leads to sensationalism that then engenders reactive rules that add to bureaucracy rather than addressing the real issues. Yes, relationships with funders (medical products industry, NIH, foundations) have a huge influence on science as do professional rivalries, egos and desire to appear in the press (fame).

CardioExchange: One concern is that many trials are not under the control of the principal investigator– do you think that the profession should insist on trials that are not under the control of the funder?

Califf: I do not believe that a human experiment should be “under the control” of any single individual. Every aspect of a clinical trial is a compromise and no one gets everything they want. In fact a despotic, egotistical principal investigator can be just as dangerous as a controlling funding organization. Like our constitution, trials need a balance of power: PI, executive committee, steering committee, investigators, DMC, regulators, funders—all have a role to play. Yes, the PI ultimately is responsible, just as Hillary Clinton and Barack Obama were responsible for the security of our embassies—but there are many players. People understand this, and we should think of trials the same way. Indeed, one of the biggest influences on trials will now be the participants themselves,who are rightly becoming invovled in design, conduct and interpretation of trials.

CardioExchange: Even when the trial is not under the control of the funder, there may be other types of control. Should the PI and investigators always have full control of the data?

Califf: For now, I believe the sponsor and the trialists should both have the data under a set of rules agreed to before the trial starts. The biggest issue here is that most academic institutions are incapable of correctly managing complex trial data or analyzing it, so the PI’s are dependent on analyses from the sponsor or a CRO (clinical research organization) hired by the sponsor. I DO BELIEVE THAT INVESTIGATORS SHOULD ANALYZE THE DATA INDEPENDENTLY WITH ACCESS TO THE FULL DATA SET.

CardioExchange: What is the best path forward with conflict of interest, assuming that disclosure is not enough (since it is hard to interpret)?

Califf: The best path forward is transparency and public access to the data so that many people can analyze and interpret. This will take time. Importantly, my interpretation of the available information is that:

  1. Academic analyses have more major errors and are not as reproducible as industry analyses. The oversight and punishment of bad behavior by FDA is important.
  2. Bias in choosing the question is a much bigger issue than lying about the data.
  3. In industry-sponsored trials, lying about the data is rare. Lying by omission (not looking at aspects of the data) is much more common.

3 Responses to “Conflict Of Interest Is A Complex Issue”

  1. Conflicts of interest are unavoidable as there is an inherent conflict when a “lipidologist” does a lipid study or a cardiologist does a heart study.

    However, there are egregious conflicts which continue to be undisclosed when a researcher with a financial interest in a product is continually published despite questionable value of the research.

    The worst conflicts of interest are in the editorial boards of journals which continue to publish their powerful friends despite the poor quality and undisclosed conflicts within the research.

    A conflict that enriches the financial coffers of the institution that pays one’s salary is no less a conflict than one that enriches the investigator personally.

    It is easy to become jaundiced regarding all medical literature. Those of us on the front lines however have learned which authors to respect and which ones to read only to find the corruption. It is just a matter of time before these conflicts reach the light of day and some of our most esteemed medical institutions are brought to shame.

  2. Judith Andersen, AB, MD says:

    I agree with Rob Califf that there are layers of bias in any clinical trial.I grew up at Duke during the MIRU era, and became inbued with the notion that investigators proposing a trial have an hypothesis in mind and are rarely neutral as to what they expect the outcome to be. NIH/other nonprofit research organizations that support non-industry trials may be less biased as to outcome, but have much vested in a distinction between control and treated populations (few of us are excited by studies that reveal no difference between control and treatment groups) — and the cost of large clinical trials makes the careful review of a study hypothesis and oversight of study design critical. Those folks most vested in the hypotheses and study design are not necessarily the best critics. And industry only complicates the issue of objectivity and impact upon both the treatment of patients and the economics thereof: witness the recent Sidney Gilman disaster. Who could have predicted that a respected teacher and previously relentlessly objective data analyst could have moved so far to the dark side that he provided data revealing the lack of benefit of much-awaited drugs for amelioration of Alzheimer’s disease to industry and its investors prior to their release the medical and scientific community?

    It’s a no-brainer that we can never achieve absolute objectivity. We, who develop hypotheses with regard to disease etiology and possible prevention, are biased by our prior training, with its attending thought processes, those who support us with methodologic and institutional support, are biased by our enthusiasm and our preliminary data, and those industry folks who have an agent that addresses our hypotheses cannot help but support us with their financial input and their drugs. We need adjudicators who are neither research nor clinical experts in the fields in question (but do have expertise in clinical investigation and data analysis in other fields, and both a willingness to criticize bad thinking or performance, and support for expressing those criticisms) to vet our trial designs, data collection and data analysis (and realize that even they, accepting the obligation to criticize the process, may still become wed to it in some powerful way).

    With rare exception, human beings are optimists. Investigators, particularly those whose research hypotheses are at the center of a clinical trial, expect the outcomes they espouse to be supported by the obtained data — and professional pressures may induce them to declare positions less than securely supported by fully analyzed clinical trial data. Industry income and prestige, both professional and personal,all hinge on both vetting and calling a winner. Our habits — of publically rewarding those who call a winner — and discounting a loser– at the time of the initial publication of clinical trial results — run the risk of missing the messages sent by clinical trials — particularly those of the often-delayed sub-group analyses — and may throw the baby out with the bathwater — or possibly treat the baby in a way that will create a distant disaster.

    We need a systematic overhaul of our clinical trial strategies — with particular attention to those that are funded by industry with the singular goals of proving efficacy of an agent in a specific disease state. And including a review of what our psychology colleagues know about hope and optimism — and disappointment and disaffection with less than dramatic results. And a more thorough understanding of the diversity of our subpopulations. Subpopulation analysis data arising from prior clinical trials is rarely treated with the same enthusiasm (and market interest) as the primary data, and valuable lessons lost in that process.

    As a human being, optimist and believer in the clinical trial process, I think we need to slow the juggernaut just a bit. Without restricting creativity concerning new drugs and new devices for identified prevalent clinical disease states, spend a bit more time looking at the roots of disease in our vulnerable populations. We have a huge, profitable and largely unregulated market for dietary supplements and unregulated therapies for minor diseases, rarely controlled for in clinical trials. There are enormous overlaps of dietary and supplement inputs to individual health, with the emphasis of “holistic physicians” on supplements — and virtually no data. The benefits of simple prophylaxis: small dose aspirin, etc — are being undercut by the routine use of non-aspirin NSAIDS for minor pain, with little education as to the interference in aspirin efficacy related to non-aspirin NSAID use. And most of our primary care doctors — and their cardiology consultants — think that these are trivial issues.

  3. Jean-Pierre Usdin, MD says:

    “Forgive us we confess”
    In Richard Lehman’s literature this week, we have a good example of a new generation of conflicts of interest.
    “Rivaroxaban vs Enoxaparin in acute ill patients”
    Totally sponsored by the manufacturer of Rivaroxaban:
    Biased conclusion: tested drug non inferior (close your eyes on the rate of bleeding more important with Rivaroxaban!)
    Cost effectiveness: no discussion.
    And finally total forgive of the searchers because they declare their full interest in the product! “Forgive us we confess”
    So let’s publish in a famous periodic!

    The same way was used by two authors (in the same journal) of a Review Article to conclude that DES are better in, roughly, all situations than BMS. They both received fees from DES manufacturers.

    Furthermore in the same periodic few weeks ago a group of internists, discussing an imaginary study, were asked if they prefer Government sponsored studies or manufacturers sponsored ones.
    Some voices did not understand why more participants preferred government financing!
    Personally I know why. I am more suspicious and continue to be. But conflicts are now frequently masked “in online version”…

    (Special thanks to dr Richard Lehman.)