The pharmaceutical and biomedical-devices industry preferentially targets the most senior medical researchers — especially ones who publish prolifically — in its efforts to influence what information is collected, whether it’s published, and how it will be portrayed (as in spun). Or so concludes a new National Institutes of Health-funded study that surveyed researchers from the 33 U.S. universities receiving the most funding for biomedical and clinical studies.
The number of Instances of attempts to exert undue influence was not huge. Only 9 out some 230 researchers whose trials were funded by industry, for instance, were asked by their sponsor to withhold research results from publication. However, 30 of these investigators were asked to delay publication of their findings and 15 asked to keep their research and its findings secret. Perhaps most troubling: Nine percent of the polled researchers reported first-hand knowledge of compromises to the integrity of research that threatened patients’ well-being.
The majority of studies that test medicines and other therapies in people are financed by industry. They are primarily carried out, however, at academic research institutions.
To probe for potential conflicts of interest in the relationships that develop between academic researchers and the companies that finance them, Patricia Tereskerz of the University of Virginia’s Center for Biomedical Ethics and Humanities and her colleagues sent questionnaires in the summer of 2005 to 1,479 medical- and nursing school researchers. All had been authors on published research studies within the previous five years.
Of the 703 who responded, 66 percent reported they had received industry support. This ranged from research grants and support for staff or students, to honoraria, equipment and personal gifts valued at more than $100. Industry support favored researchers who were male, had worked at their institutions for at least 11 years and held higher academic rank, Tereskerz’s team reports in the most recent issue of Accountability in Research.
The survey participants also answered a number of questions about potential conflicts of interest or questionable practices at their institution; respondents were restricted to describing only things about which they had “first-hand knowledge.”
The survey showed that the more key an investigator considered industry financing was to his or her ability to conduct research, the more likely industry had approached that investigator with requests to withhold research data, delay the publication of findings, keep a research project secret or present published research more favorably than a researcher had wanted to. Some 28 percent of the investigators who considered industry help especially crucial reported being asked or of knowing someone who had been asked to do at least one of these things.
Survey participants were not asked whether they or their colleagues had complied with sponsors’ requests, because there was a concern that this might deter people from answering honestly — even though questionnaires were anonymous. Indeed, Tereskerz’s team explained, the goal was not to identify who acted inappropriately, but to identify who was being targeted, how often, and with requests to do what. In other words, to understand how pervasive a vulnerability to bias was.
For many years, studies have pointed to attempts by industry (many of them successful) to exert bias. Sometimes industry-funded studies were quashed when results weren’t trending in the direction a company liked. Companies have prohibited the publication of data that didn’t reflect well on the sponsors’ interests. Researchers sometimes have been asked to spin their findings in a way the sponsor likes — such as downplaying negative findings, overplaying marginally positive findings, or leaving out data that raise questions.
In that sense, the Tereskerz team doesn’t break radically new ground, observes Eric Campbell of Harvard Medical School and Massachusetts General Hospital’s institute for health policy. For instance, he notes, David Blumenthal’s work in the early 1980s probed industry funding and trade secrecy. A decade later, Campbell says, he and Blumenthal “looked at the relationship between industry funding and data withholding in science.”
And just two years ago, this team reported finding that nationally, 94 percent of the more than 3,150 physicians that they surveyed “reported some type of relationship with the pharmaceutical industry.” That study had focused on clinicians, and there most of the industry ties involved gifts of meals, free drug samples, payments for consulting or lectures, or compensation for enrolling participants in trials.
Over all, Campbell says “the linkages between industry funding and science — and whether research shows a pro-industry bias — has been well documented.” That said, there were several findings of the study by Tereskerz’s group that “are new,” Campbell says.
For instance, “the finding that scientists were asked to present results in a way which reflect favorably on the sponsors’ drug or product is interesting. It’s 15 percent of those who report that research funding is important to them.” One surprise, he says, “is that it’s not much higher than that.” Then again, he notes, the number of participants responding to any particular aspect of the questionnaire (which depended on what they did or knew first hand) tended to be small, Campbell notes. So the percentage values, he says, have to be considered rather “unstable” or squishy.
But what he considers particularly troubling was the study’s finding that “many institutions do not require disclosure to research subjects of investigator financial relationships with industry.” Of the researchers surveyed, 139 had received industry support for research and/or publication of data and had published data from human trials in the past five years. A poll of this group quantified the share whose schools or hospitals did NOT require that they tell participants in a trial of substantial industry ties, such as:
— holding equity interest in the company sponsoring the research: 30 percent
— having served as a consultant or board member to the company: 36 percent
— having key members of the research team paid for by the sponsoring company: 42 percent
— if equipment used had been donated by the sponsor: 51 percent
— or if key personnel had either a royalty agreement with the sponsor or were part of a joint commercial venture with the company: 40 percent.
Where there was no academic requirement to disclose these ties to the sponsor, how often had a researcher done so voluntarily? Depending on the category of industry support, only 10 to 30 percent answered “always.” Some 10 to 24 percent, depending on the nature of their support, answered “never.”
In theory, Campbell says, if an institution’s ties to industry aren’t fully disclosed, “then potentially a lot of subjects could be harmed, because they have been not consented properly.” That is, they haven’t been apprised of potential conflicts of interest that would have the potential to influence a researcher’s practices and interpretation of data.
“Investigators may assume that [such failures to disclose] don’t hurt subjects,” he says, but incomplete disclosure “is by definition a harm.” It doesn’t allow participants to gauge how dependent researchers may be to their funders. And there’s a name for that, Campbell notes: resource dependency.
The idea, perhaps a century old, he explains, is that “individuals or organizations can control others depending on the importance of the resource that they provide.” If a company can transfer its financial support to another research group, Campbell says, it then holds a “lot of leverage over the individual or organization [it finances].” And he notes that because industry tends to favor senior investigators — ones whose reputations will reflect well on a company’s product — “it’s not surprising at all that these [researchers] are more likely to feel the leverage.”
Overall, he says, “there are some intriguing data in this study. But what we don’t know is the extent to which these [attempts to influence researchers] happen from other funding sources as well.” For instance, some patient-advocacy groups work to gain support for funding to tackle specific diseases — from cancer and heart disease to Alzheimer’s, diabetes, and asthma. They could pressure the investigators whose work they fund “to present results suggesting that a new test is needed,” Campbell says. “Because then you can get the funding from the federal government to develop it.”
Bottom line, Campbell argues: “I don’t think that we should assume that industry is the only organization that may have the will and ability to [exert undue influence] on research.”
And now that ample evidence has emerged pointing to how research integrity can be compromised, he says, “I’d like to know what impact these breaches on research integrity have on the cost and quality of care that patients receive. For instance, if results are presented in an overly favorable way, does that lead to overuse of a medication?” And if so, he asks, what costs does that have — either in heightened healthcare costs or in patient harm?