4. Research and publication
34. Conflicts of interest in research funding
Michael Nurok and Carl C. Hug, Jr.
An academic physician has a long history of receiving industry support for clinical research on a medication that she had helped to develop. The university sold the patent rights to a local biotechnology company. The company is small but rapidly growing, and the physician knows most of the founding members. She and her husband are friendly with the Chief Executive Officer, and their children attend the same school along with children of other company members.
To date, clinical reports have supported the new drug as superior to the much less costly and generically available alternative. But data analysis from her recent clinical trial yields no statistical difference between the two products. She discusses her findings with members of the sponsoring biotechnology company, and they suggest that she allow one of their statisticians to check the results.
The company’s statistician segregates the study population according to age and finds that there is greater variability in the group less than 40 years of age. He then repeats the standard statistical test after omitting the under 40 group and shows a clear statistical difference in favor of the new drug. Reanalysis of the data for the entire group with a novel statistical test also shows a definite benefit for the new product compared to the generic drug. A statistician at the university tells the physician that the novel statistical test was done correctly, but points out that this is an unusual – but not incorrect – statistical test to apply to this body of data. Because it is very important to prospective investors, the company officers urge that the positive data based only on the novel statistical method be published as soon as possible.
The investigator is an associate professor at the university and, with one additional peer-reviewed publication, feels confident that her promotion to professor will go through. She knows that her chances of publishing a paper on the study are greatly increased if the study findings are positive. In addition, she is in discussion with the biotechnology company about funding a large project to develop another drug. Without this funding she would be required to do more clinical work to support her salary and probably would have to dismiss members of her laboratory because no other future funding is envisaged. The economy is slow and two members of her laboratory are the sole support of their families.
The United States Institute of Medicine defined conflict of interest as:
Circumstances that create a risk that professional judgments or actions regarding a primary interest will be unduly influenced by a secondary interest. 1
The report defines primary interest as promoting and protecting the integrity of research, the quality of medical education, and the welfare of patients. Secondary interests are listed as financial, the pursuit of professional advancement and recognition, and the desire to do favors for friends, family, students, or colleagues.
A number of different sources of potential conflicts of interest can be identified in the introductory case. These can be divided between financial and non-financial conflicts.
There is a risk that the physician’s professional judgments or actions with respect to her duty to promote and protect the integrity of research could be unduly influenced by the following financialmotivations:
• Reliance on the company for partial support of her salary.
• Reliance on the company for financial support of her laboratory.
Sources of non-financial conflict exist with respect to this physician and include:
• Desire for publication of positive findings to facilitate academic promotion.
• Personal and spousal relationships with members of the biotechnology firm and its CEO.
• Children’s social relationships with company members’ children.
• Sense of responsibility towards members of her laboratory in regard to continued employment.
Discussions of financial conflicts of interest tend to be based on certain assumptions. Stossel2 describes these as beliefs that:
(1) Academic–industrial relationships promote research misconduct.
(2) Commercial intrusion leads to subtle or overt bias in the interpretation of research data, limitations of academic freedom, degradation of quality of research, and violation of important research values.
(3) If commercial intrusions even appear to compromise scientific integrity, public trust in and support for research will be eroded.
Stossel also points out that there is little empiric evidence to support these assumptions. Many point to the positive effects on medical innovation that academic relationships with industry have produced in contrast to the often-noted theoretical effects of conflicts of interest. Nevertheless, potential conflicts are apparent in the above scenario and are common in the practice of academic medicine. The most concerning effect of a financially motivated conflict of interest is that it may introduce bias into science and medicine.
The clinician investigator in this case has a number of potent incentives to publish positive findings of the trial using the novel statistical test. She would please the pharmaceutical company and enhance her ability to receive further research funding from the company. She would likely acquire the additional publication needed to be promoted. The social problems that may arise between her family members and their network of friends would be avoided. In addition, if her laboratory is funded, all of her employees would retain their jobs.
Disincentives to using the novel statistical test and publishing the positive findings include the risk to her reputation as an investigator and the increased costs of health care resulting from reliance on her study as support for prescribing the drug in question. More broadly, maintaining the public’s trust in research is critical to ensuring participation in studies that may lead to new medical breakthroughs and also to ensuring faith in science and medicine. Using only the novel statistical test risks jeopardizing this trust.
A number of facts support the concerns that financial conflicts influence the integrity of medical research. Corporations in general, and industry and drug companies in particular, are focused on their responsibility to stockholders and some have acted aggressively to increase profits through various mechanisms such as controlling the use of data concerning their products.3 A 2003 meta-analysis found that industry-associated clinical trials were significantly more likely to report product-favorable results than those in which industry relationships did not exist.4
Contracts between industry and academic institutions have included restrictions on the publication and other use of data without advance permission of the sponsor. Examples of the various contractual provisions related to publications of industry-sponsored trials were published in 2005.5 A minority of institutions allowed sponsors to decide whether a study should or should not be published. Many institutions allowed contractual provisions to permit the sponsor to insert its own statistical analysis, and sometimes even to draft the manuscript.
Conflicts of interest are not just limited to researchers. In a recent study, 94% of physicians reported some sort of relationship with industry. Interestingly, anesthesiologists were less likely to have received samples, reimbursements, or payments than were family practitioners, internists, pediatricians, cardiologists or general surgeons who write more prescriptions for long-term use of drugs.6Psychiatrists have been a major focus of the investigations by US Senator Charles E. Grassley (R-Iowa), a sponsor of The Physician Payment Sunshine Act, 2009 (S.301). Despite evidence to support public concern regarding conflicts of interest, with the exception of isolated cases, little evidence points towards actual harm having been caused by conflicts.
Physicians have multiple loyalties, interests, and obligations that go beyond their professional role. These relationships lead to the potential for competing interests that are non-financial in their nature. Non-financial conflicts – although inherent to research and the practice of medicine – are more difficult to manage. Critics who perceive the focus on financial conflicts to be overzealous, have argued that industry is subject to scrutiny by the FDA and other regulatory bodies whereas academics are not.
Because the potential for nonfinancial conflicts is so ubiquitous, it is a challenging subject to study and there are far fewer data to assess their impact. Ideally, nonfinancial conflicts should be disclosed and acknowledged. However, criteria for doing so have not been established, and further research is needed to determine the magnitude of the problem and its consequent effects.7
Several options are open to the academic clinician in this case.
The clinician may chose to publish using only the novel statistical test and to make the raw data from her study available online. This puts any uncertainty about the findings into the public domain and renders it available for transparent discussion and debate in addition to allowing other statistical analysis based on the raw data. The industry sponsor may not like this approach; however, it would most likely be better received by the company officers than Option 2. In addition, this approach would mitigate the social and professional risks discussed previously. An ethically defensible argument for this approach is that the novel statistical test is not inaccurate, simply unusual.
The clinician may choose to use only the conventional statistical test and attempt to publish the negative findings of her results along with the raw data from her study. This approach may be difficult to achieve if the contract with the sponsor specifies limits on publication without the sponsor’s approval. In addition, this option may antagonize the industry sponsor and expose the researcher to many of the social and professional risks noted above. Another important risk of this approach is that, if the conventional test is providing a false-negative result, use of this test may lead to the product being abandoned when it could provide a clinically important benefit. Placing the raw data in the public domain may mitigate this risk.
The clinician may chose not to publish any results from her study. This is an ethically acceptable stance provided that the clinician is motivated by genuine uncertainty regarding her results in the face of opposing statistical interpretations. This stance is harder to justify if it is motivated by avoiding the social and professional conflicts brought about by her predicament. In addition, this approach means that data from her study are not available in the public domain.
The clinician may choose to apply for additional funding either (1) to expand the number of patients in the under 40 group; or (2) to conduct another, better powered clinical trial in an attempt to answer the question definitively.
The clinician may choose to publish using only data from the over 40 groups which represent the vast majority of people for whom the new drug would be prescribed. This approach is neither scientifically nor ethically appropriate because it is misleading and motivated by conflicting interests. It exposes the clinician to the lowest risk of adverse financial, social and professional consequences of all options at the cost of jeopardizing her professional reputation with respect to the duty to promote and protect the integrity of research.
Mitigating conflicts of interest
The potential for conflicts of interest can never be eliminated, but conflicts of interest can be managed and, in some cases, reduced. Strategies for doing so are most useful when they are matters of institutional and editorial policies that are focused on preventing the adverse effects of conflict of interest and are not unduly burdensome.
The first and most common step in managing conflicts involves disclosure. Following recent scandals in the United States, several institutions are voluntarily publishing information regarding conflicts of interest online.8 Agreement is needed on criteria and thresholds for disclosure. Another mechanism for managing conflicts is monitoring of research by independent bodies, and public disclosure of raw scientific data. This step would help prevent (by functioning as a deterrent) or detect the most concerning effect of conflicts of interest – the introduction of bias. Research on other mechanisms for preventing bias is needed.
Managing financial and other types of conflicts of interest extend beyond the individual investigator to the academic institutions themselves, because the institutions have become dependent on research productivity, scientific, and medical advancements for their financial well-being and enhancement of their reputations.9 The American Bayh-Dole Act of 1980 (pub L. No. 96–517) and the Federal Transfer Act of 1986 (Pub L. No. 99–502) provided strong incentives to investigators and institutions to patent findings of publicly funded research, and led to an explosion of academic–industry interactions highlighting that academia and industry have different as well as common interests (Table 34.1).
Table 34.1. Diverse and common interests of academic anesthesiology and industry
Policies and guidance on conflicts of interest are changing rapidly. The United States the National Institutes of Health10 and the Department of Health and Human Services Office of Human Research Protection offer updated guidelines.11 In the United Kingdom, the General Medical Council offers such guidance.12
• One of the most harmful adverse consequences of conflicts of interest in research is the potential introduction of bias into the science of medicine and the denigration of the integrity of scientific research.
• The potential for conflicts of interest in research can never be eliminated, but can be managed, and in many cases reduced.
• Strategies for reducing conflicts of interest in research are most effective if they involve clear institutional and editorial policies.
• Disclosure is the most common first step in managing conflicts of interest.
• Public disclosure of raw scientific data and independent monitoring of research are additional mechanisms for managing conflicts of interest.
1* Institute of Medicine. (2009). Conflict of Interest in Medical Research, Education, and Practice. Washington, DC: National Academies Press.
2* Stossel, T.P. (2005). Regulating academic-industrial research relationships – solving problems or stifling progress? N Engl J Med, 353, 1060–5.
3* Angell, M. (2004). The Truth about the Drug Companies: How They Deceive Us and What to Do about It. 1st edn. New York: Random House.
4* Bekelman, J.E., Li, Y., and Gross, C.P. (2003). Scope and impact of financial conflicts of interest in biomedical research: a systematic review. JAMA, 289, 454–65.
5* Mello, M.M., Clarridge, B.R., and Studdert, D.M. (2005). Academic medical centers’ standards for clinical-trial agreements with industry. N Engl J Med, 352, 2202–10.
6* Campbell, E.G., Gruen, R.L., Mountford, J., et al. (2007). A national survey of physician–industry relationships. N Engl J Med, 356, 1742–50.
7* The PLoS Medicine Editors. (2008). Making Sense of Non-Financial Competing Interests. PLoS Med 5(9): e199. doi:10.1371/journal.pmed.0050199
8* Steinbrook, R. (2009). Online disclosure of physician–industry relationships. N Engl J Med, 360, 325–7.
9* Johns, M.M.E., Barnes, M., and Florencio, P.S. (2003). Restoring balance to industry–academia relationships in an era of institutional financial conflicts of interest. Promoting research while maintaining trust. JAMA, 289 (6), 741–6.
10 NIH Guidelines. (Accessed 2 April 2009, at http://grants.nih.gov/grants/policy/coi/.)
11 DHHS Guidelines. (Accessed 2 April 2009, at http://www.dhhs.gov/ohrp/special/conflict.html.)
12 GMC Guidelines. (Accessed 2 April 2009, at http://www.gmc-uk.org/guidance/current/library/conflicts_of_interest.asp#1.)
Levinsky, N.G. (2002). Nonfinancial conflicts of interest in research. N Engl J Med, 347, 759–61.