Legislation That Made Government-, But Not Industry Funded Research Easier to Challenge

A new article in the American Journal of Public Health showed that industry's priorities may adversely shape legislation affecting how science, including biomedical and health care science gets done. (Requires subscription, full citation is: Baba A, Cook DM, McGarity TO, Bero LA. Legislating "sound science": the role of the tobacco industry. Am J Pub Health 2005; 95: S20-S27.)
Using documents revealed by litigation, the article narrates the history of the tobacco industry's influence the enactment and operation of legislation about government-funded science. The relevant legislation was:

• The 1999 Omnibus Appropriations Act contained a provision that made all the raw data obtained in all government funded research projects available to anyone.
• An amendment added to the 2001 Omnibus Appropriations Act required the Office of Management and Budget to develop government-wide standards for data quality, including "objectivity of presentation" and "objectivity of substance." The latter required that any data disseminated by the government meet quality standards, and that "influential data," which is "scientific, financial, or statistical information that will have or does have a clear and substantial impact on important public policies or private sector decisions" must be reproducibe upon reanalysis by "qualified third parties."

The article described the Philip Morris "'sound science' public relations campaign [designed] to create controversy regarding evidence that environmental toxins [e.g., second-hand smoke] cause disease." The company allegedly developed a complex legislative strategy to address "contested interpretation of underlying data in published studies." However, the company made every effort to "remain in the background" while its strategy was implemented. Once Congress enacted the above legislation, whose provisions about scientific data never were subject to committee hearings or public discussion, Philip Morris engaged a high-powered consulting firm to "work with federal agencies to encourage implementation." The article also alleges that Philip Morris, through this consulting firms, "contributed to the language of the model data quality guidelines."
Since the enactment of the legislation, "corporate interests have initiated the vast majority of data quality act challenges." Furthermore, some have charged that new "peer review" guidelines proposed by the Office of Management and Budget "politicizes science by giving the White House unregulated power to expedite or delay the release of scientific information."
The importance of all this was outlined by author Lisa Bero in an accompanying news release. Of particular importance is that the legislation and accompanying guidelines and rules make it very easy for outside parties, including large organizations with specific financial interests, to challenge government-sponsored research. Privately sponsored research, including, obviously, research sponsored by these same parties, however, is not affected by the legislation and not open to such challenges. As Bero put it, "if we are going to have laws that force greater scrutiny on research related to public health, they should apply to industry-funded research at least as much as to government-funded research." As an example, she noted that most of the recent controversy about the US Food and Drug Administration (FDA) focused on "bad government science." But there should be equal focus on "bad industry science." She summed it up thus,

What is really ironic is that the data quality law applies only to government-sponsored research (such as NIH research), but not industry-funded research. So, industry-funded research does not have to adhere to the standards. This is particularly relevant with all the transgressions we've seen lately related to the quality or failure to publish industry science. The public health community cannot use the data quality law to challenge industry science.

In my humble opinion, her concerns make sense.