Thursday, May 26, 2005

A Troubling Study of the Contracts Between Medical Schools and Corporate Research Sponsors

The NEJM article about contracts for clinical trials between academic health centers (AHCs) and pharmaceutical companies mentioned in an earlier post deserves further attention. (The study is available here, and its citation is Mello MM, Clarridge BR, Studdert DM. Academic medical centers' standards for clinical-trial agreements with industry. N Engl J Med 2005; 352: 21.)

First, let me summarze the study's design. The authors sent a survey to senior administrators in the offices of sponsored research in 122 medical schools. They received responses from 107, for 91% response rate, excluding four schools that did no clinical research. They asked the administrators questions about provisions they would usually allow in contracts between their medical schools and corporate sponsors. The proportions who would allow some particular research practices follow:
  • The sponsor will own the data produced by the research - 80%
  • The sponsor will store the data and release portions to the investigator - 35%
  • The investigators (whether at your site or any other site) are not permitted to alter the study design after the agreement is executed - 68%
  • The sponsor is permitted to alter the study design after the agreement is executed - 62%
  • The sponsor may prohibit individual site investigators from publishing manuscripts independently of the sponsor or group - 15%
  • The sponsor may include its own statistical analysis in manuscripts - 24%
  • The sponsor will write up the results for publication and the investigators may review the manuscript and suggest revisions - 50%
  • While the trial is going on, the investigators may not discuss research results (including presentations at scientific meetings) with people not involved in the trial - 66%
  • After the trial is over, the investigators may not discuss research results (including presentations at scientific meetings) until the sponsor consents to dissemination - 21%
  • The terms of the clinical-trial agreement are confidential - 62%
  • After the trial is over, the industry sponsor may prohibit investigators from sharing raw research data with third parties - 41%
In summary, this study suggests that some medical schools will acquiesce to contracts that prevent their faculty from having meaningful control over research projects for which they are the ostensible investigators. Some medical schools will allow corporate research sponsors to manipulate the design, analysis, and interpretation of results to increase the likelihood of getting the results the sponsor wants. For example, 50% of medical schools would allow the corporate sponsor to write manuscripts reporting on study results, and only allow faculty to "suggest revisions." Thus, many medical schools acquiesce to contracts that could allow corporate sponsors to ghost-write articles. Furthermore, more than half of the medical schools acquiesce to contracts that are kept confidential, preventing easy discovery of manipulation of the conduct and dissemination of the research by the corporate sponsor.
(To see specific techniques that corporate sponsors may use to increase the likelihood that research produces the results they want, see our earlier post about the article entitled "Medical Journals Are an Extension of the Marketing Arm of Pharmaceutical Companies" by Richard Smith.)
To underline the context, in my experience, to do clinical research, most medical school faculty are strongly pressured by their institutions to obtain external funding to support this work. Those who cannot obtain funding within a few years of unemployment often have to find something else to do other than research. However, although faculty are expected to find sources of funding, it is generally the medical schools or academic medical centers, not the researchers, who negotiate research agreements, and although such agreements may bind faculty members, the money paid under their auspices generally goes directly to the faculty member's employer. Thus, individual researchers are under considerable pressure to find grant and contract funding, and to keep their financial sponsors happy, but must work under whatever contract provisions their employers negotiate.
The New England Journal of Medicine article clearly suggests that some medical schools allow researchers to be bound by agreements that allow the study's sponsor to manipulate study design, analysis, and interpretation to make it more likely that the results will be to the sponsor's liking. Since clinical trials enroll real human beings as subjects with the promise that the data they contribute, and the risks they run, will lead to valid scientific results, allowing such manipulation appears to not only likely to produce biased results, but also to be allow unethical exploitation of human subjects, and thus is in conflict with the medical school's scientific and humanitarian mission.
This is yet another illustration of leaders of powerful health care organizations acting in conflict with physicians' and medical researchers' core values.
One can only hope that publicizing such practices will lead to more ethical research, and less exploitation of human subjects.
By the way, I do not in any sense mean to blame the research administrators who must try to negotiate these contracts. Note that 69% of the administrators "perceived that competition for research funds created pressure on administrators to compromise on the language of the contract." 24% described the pressure as "great." I do blame the higher leadership of the medical schools for fostering a climate in which money is more important than doing valid, ethical research. This raises several important questions:
  • Why don't some medical school leaders enforce clearer ethical standards for clinical research sponsored by industry?
  • Are they so worried about getting the money, and so worried that they will lose contracts if they insist on ethical standards?
  • Are they so steeped in the exteme relativism of the post-modernist university that they don't believe in such standards?
  • Are they so used to violating their own mission to support political causes, as is documented again and again on the FIRE web-site, that they aren't troubled by violating their mission for economic reasons?
Maybe it's time we start trying to address such questions.


Kevin Schulman said...

This study builds on our 2002 work. Multicenter clinical trial research is governed by standard agreements at all sites participating in a study-there really is little variablity within a study, only some varaibility across studies. The final result is that when patients are randomized, the institution has not assuance that the sponsor will publish the overall trial results, not that there will be an independent process to assess whether in fact the trial is publishable. Since the definition of research is the creation of generalizable knowledge, site investigators at academic medical centers are in fact deceiving patients when they recruit them to multicenter clinical trials if they ask them to participate in a research study under current agreements. Even with the 2001 guidelines, the 2002 paper and now this paper, there looks to be little pressure to change practices. Further, we need a single ethical standard for research-otherwise trials will move to non-academic sites in the US or off-shore.

Anonymous said...

To people with experience in the area, these recent revelations will not be a surprise. The new interest is in getting firm information on the scale of it.

Nevertheless, one should be cautions about interpreting some of these things at face value. For example, the 50% of cases where "The sponsor will write up the results for publication and the investigators may review the manuscript and suggest revisions" may seem a bit scandalous, in view of the opportuinity for distortion which it provides.

But one can also consider that a sponsor contracts the investigation to an academic research team because they have the clinical/laboratory/library facilities, and the technical expertise, to perform a good investigation; while the sponsor has expertise in trial design, trial monitoring, regulatory requirements, and statistics. That sponsor may do a thoroughly ethical job.

I'm sure we're all aware of cases where a sponsor would be unwise to keave trial design and statistics to a particular clinical research team!

On the other hand, the sponsor may not do an ethical job (that door is after all open).

So, in interpreting (e.g.) that 50%, one would like to know what proporion of it led to an ethically good outcome, and what proportion led to distortion of the results.

On occasion, individual researchers vlow the whistle. Often, they may be reluctant.

On the context of what the terms of research contracts should be, perhaps they should spell out not only who does what, but also why (e.g. the clinical team do not have experience in trial design).

Ultimately, it would be desirable that all participants should be independent of the sponsor and responsible for the integrity of the operation. But perhaps that is utopian.

ted Harding.

Stan Young said...

There are a lot of bright people paying attention to clinical trials: sponsor, clinical physicians, FDA, editors, referees, and readers. All have interests and different levels of power. The FDA examines raw data sets. It is not easy to imagine that the situation is all that bad. On the other hand, physicians have to evaluate journal articles from non-randomized studies. These are the Wild, Wild West. Hundreds of questions can be asked and a few, most likely false positive, results are reported. No one can check anything as near as I can tell. Ask for study details like the SAS code used for the analysis or the raw data and most likely you will get stonewalled. Most of this research is government funded. The vast majority of epidemiology studies do not replicate and epidemiologists don't appear to care.