American Society of Clinical Oncology Annual Meeting abstracts with deemed financial conflicts of interest receive more prominent placement and better peer review scores than research without such conflicts, according to a study in the July 20 issue of ASCO's Journal of Clinical Oncology (2013;31:2678-2684).
“We need to dig down deep and figure out the reasons behind this,” said the first author, Beverly Moy, MD, MPH, Clinical Director of the Breast Oncology Program at Massachusetts General Hospital Cancer Center and Assistant Professor of Medicine at Harvard Medical School.
The study also found an increase in the number of such financial conflicts of interest in the meeting abstracts over a period of several years, although the researchers are limited in their ability to say why, Moy said in an interview.
ASCO CEO Allen S. Lichter, MD, said the increase “reflects the steady shift from federally funded clinical research to industry-funded research: “This trend has been going on for a number of years as federal funding has diminished and as industry has stepped in to take its place.”
In an accompanying editorial (JCO 2013;31:2645-2647), Ralph M. Meyer, MD, Director of Canada's NCIC Clinical Trials Group and Professor of Oncology at Queen's University, Kingston, in Ontario, notes, though, that whatever the cause, documenting such conflicts is useful because it can foster discussions about the relationship between industry and research in the oncology community.
An abstract of the study was reported at last year's ASCO Annual Meeting, but Moy said that no significant data changes had been made between then and the full research paper.
She and her colleagues evaluated 20,718 abstracts presented at the ASCO Annual Meetings in 2006 through 2011, except for 2007, since the information for that year was incomplete in the ASCO database. For the purposes of the study, “financial conflicts of interest” reflected categories in ASCO's 2005 conflict-of- interest policy (it was since updated, earlier this year—see below), which included industry employment, owning stock, consulting, receiving honoraria or research funding, or being an expert witness.
Moy and her coauthors—Angela R. Bradbury, Paul R. Helft, Brian L. Egleston, Moktar Sheikh-Salah, and Jeffrey Peppercorn—assessed financial conflicts of interest and the relationship with meeting placement prominence, defined in descending order of plenary session, clinical science symposium, oral presentation, poster discussion, general posters, and published only. Two individuals assigned peer-review scores of 1 to 5 to the abstracts, with the lower value indicating higher merit. The researchers also assessed the peer-review score.
Overall, 36 percent of the abstracts had at least one author who reported a financial conflict of interest, and the rate increased from 33 to 38 percent from 2006 to 2011, a rate that was considered significant.
Abstracts with conflicts of interest were also found to have been given significantly more meeting prominence—specifically, the odds ratios compared with the rates in general posters were 7.3 for the plenary session, 2.2 for clinical science symposium, 1.9 for oral presentation, and 1.7 for poster discussion.
Moreover, abstracts with financial conflicts of interest had significantly better peer-review scores compared with abstracts without such conflicts. For all abstracts, the peer review score was 2.76 for those with the conflicts compared with 3.01 for abstracts that didn't have them Excluding publish-only abstracts, the peer review score was 2.62 compared with 2.73, respectively.
The study was funded in part by the National Institutes of Health, the National Cancer Institute, and ASCO.
Increasing Relationships with Industry
Potentially, the increase in financial conflicts of interest reported in ASCO abstracts may reflect the fact that oncology researchers need to have more relationships with industry to conduct research, Moy said. For example, cuts in government funding have likely had an impact on oncologists' need to work with industry to conduct clinical trials. In addition, the number of cooperative group trials has declined as government research funding has become scarcer. Additionally, oncology researchers may be more aware of conflict-of-interest policies and the need for self-disclosure, and so the change can be a reflection of the information now being included.
BEVERLY MOY, MD, MPH
Moy said that the researchers had anticipated the results of the study. One reason for the increase may be that as oncologists work with industry more, they increasingly have access to interesting cutting-edge drugs. Additionally, as the study notes, more prominent industry-funded abstract placement may have occurred because of “commercial bias in the conduct, analysis, or reporting of our most prominent cancer research,” with important studies without industry support receiving less attention.
Another possibility is that collaboration with industry may actively encourage relationships with well-known academic researchers, resulting in more prominent data that also happens to be commercially funded, she said.
This study is especially timely because ASCO released its latest revised conflict-of-interest policy (http://bit.ly/1aQTWln) in April of this year, Moy noted. Explaining the differences, Lichter noted that the previous version contained certain prohibitions about industry relationships, but ASCO had “a very liberal exception clause.” The updated policy, he said, “sets some straightforward prohibitions, recognizing that while company-sponsored research is an important part of the landscape of clinical oncology research today, company employees cannot be the first author, last author, or the corresponding author.”
ASCO leadership wants to ensure that clinical researchers not employed by a company occupy the key authorship roles and have participated in the work, and the new policy should give readers and meeting attendees the greatest sense of confidence that research has been fairly evaluated and presented, he said.
Pros and Cons of Working with Industry
“There are certainly many pros and cons of partnering with industry to do oncology research,” Moy said. Gaining access to promising anticancer agents is one benefit, and the pharmaceutical industry also has sound mechanisms in place for conducting clinical trials as well as numerous resources for ensuring that data are well analyzed, she explained.
Also asked for his opinion for this article, Benjamin Djulbegovic, MD, PhD, Distinguished Professor at the University of South Florida & H. Lee Moffitt Cancer Center & Research Institute, said that while the methodology is usually better in industry-sponsored trials compared with non-industry trials, it's not clear whether this is related to study design. Industry may be asking questions that aren't relevant to the public, avoiding, for example, direct head-to-head comparisons.
Moreover, the biomedical research community has to be on guard about potential drawbacks such as publication bias, Moy said. “People are very concerned about publication bias, and studies have documented this bias. If a trial is negative it should still be published.”
The editorial writer, Meyer, noted that he does not subscribe to concerns about bias and scientific rigor: “I think that, in general, private sector research is of high quality. There are very capable people in industry who do good science and good work.”
Djulbegovic said that while the relationship between scientific research and industry cannot be viewed in black-and-white terms, commercially sponsored research does need to be conducted in a way that serves the public.
Industry-sponsored research by itself won't meet all of societal needs, Meyer said. “Industry is generally going to ask questions about the efficacy of its drug, and will need to frame this research in a manner that includes addressing its potential success in the marketplace. Other types of research are needed to address societal issues such as the best therapies for a given medical condition, the roles of non-drug therapies, and health care delivery systems.”
Understanding of Underlying Mechanisms
Djulbegovic said he sees room for improvement with disclosure policies, which he views as “a blunt tool that is not successful in all circumstances.” Instead, researchers need to identify the mechanisms underlying the conflicts and how they work, he said. “We need to identify the different points where conflict of interest impacts research along the continuum of study design, conduct, analysis, and reporting.” Once readers know the role of individual authors in the research process, a better understanding of conflict of interest may emerge.
Still, managing the conflicts by disclosure only and without attempting to describe how they could theoretically affect study results will not achieve the intentions to improve the integrity of the published scientific record, he added. “By explicitly identifying the mechanisms of conflicts of interest, we may be in a position to identify when the studies in which authors reported [these conflicts] are or are not, in fact, at risk of generating biased results.”
Whatever the mechanisms, though, researchers who are going to work closely with industry need to figure out how to best manage these relationships, Moy said. “The partnership needs to be productive and avoid some of the pitfalls that the public sees as harmful.”