Corporate manipulation of research: strategies are similar across five industries.

AuthorWhite, Jenny

INTRODUCTION

The release of over forty-three million pages of internal industry documents as part of a number of legal settlements in the 1990s, most notably the Master Settlement Agreement with forty-six state attorneys general in 1998, (1) has revealed the inner workings of the tobacco industry. Recently several other industries have been required to release to the public many of their internal documents as a result of litigation, and in some cases congressional inquiry. In an era when thousands of newly engineered products are being developed and marketed and novel production processes implemented with insufficient regulatory oversight and little understanding of the long-term risks, industries worldwide are greatly expanding their sponsorship of risk research. The release of these documents provides an opportunity to systematically examine the strategies each industry used to manipulate research in ways that would promote their products, or create doubt about the deleterious health effects of their products and manufacturing processes--thereby enhancing their credibility and profits, and shielding them against unwanted regulation or legal liability. Corporate manipulation of research has been described through case studies of individual industries or products. (2) The objective of this study is to systematically categorize these practices across a specific set of industries, using a consistent data source of internal corporate documents.

Using a previously developed framework, (3) we compare the strategies used by the tobacco, pharmaceutical, lead, vinyl chloride, and silicosis-generating industries (mining, foundries, sandblasting, and others) to manipulate research. Strategies included manipulation of the research question to obtain predetermined results; funding and publishing research that supports industry interests; suppressing unfavorable research; distorting the public discourse about research; changing or setting scientific standards to serve corporate interests; and disseminating favorable research directly to decision makers and the public, bypassing the normal channels of scientific discourse. We hypothesize that the five industries used similar strategies to manipulate research.

  1. Methods

    The tobacco documents were made available through settlement of major lawsuits against the industry in the 1990s, culminating in the Master Settlement Agreement. The Legacy Tobacco Documents Library at the University of California, San Francisco (UCSF) has over eleven million of these documents on its website. (4) The pharmaceutical industry documents were also obtained through litigation and drawn from UCSF's Drug Industry Document Archive (DIDA). (5) Hard copies of the lead, vinyl chloride, and silica industry documents were provided by Dr. David Rosner, author of Deadly Dust, on the silicosis crisis, and Deceit and Denial, on the lead and vinyl chloride industries. The lead and vinyl chloride documents were originally obtained through discovery proceedings in lawsuits against these industries.

    The first author undertook the initial review of the document archives. She reviewed over a thousand documents in the DIDA, searching on "research" and terms such as "product defense" and "publication strategy," then extending the search using document numbers, individual names, and other terms found in the documents. She reviewed all 103 documents provided by Rosner for the lead and silicosis-related industries. For vinyl chloride, the second author reviewed extensive notes prepared by Rosner and Markowitz for their book Deceit and Denial on a thousand available documents, and selected thirty-nine of these that were relevant to research practices. The original documents were then obtained at the Chemical Industry Archives website. (6) These examples were supplemented with additional searching on keywords ("publish," "OSHA," names of researchers, etc.) in the archives. As numerous articles have already been published on tobacco industry manipulation of science, the authors relied on documents that have been cited in previous peer-reviewed publications. (7) After review and discussion by the two authors, each document was coded using previously developed categories of research manipulation,8 as follows: (1) fund research that supports the industry's interests, (2) publish in scientific literature research that supports the industry's interests, (3) suppress industry-sponsored research in cases where the results do not support the industry's interests, (4)distort public discourse on research that does not support the industry's interests, (5) set scientific standards that favor the industry's interests, and (6) disseminate favorable research directly to decisionmakers and the public.

    We identified 241 episodes of research manipulation in the approximately 2250 documents reviewed and entered them into a database. We then selected exemplars from each industry for each category based on the following: 1) the example is unambiguous as an indicator of research manipulation; 2) it is representative of a number of statements in our dataset; 3) it concerns projects that were implemented, rather than just discussed in the planning stage and then dropped; and 4) it reflects an important development in the history of the industry's promotion or defense of its activities.

  2. Analysis

    We present exemplars of the six manipulative techniques that reflect common practices and perspectives of the time for each industry. We include historical, political, or scientific contexts for examples as needed. We calculated the total number of incidences we found in the databases for each type of strategy and each industry. We then tested for independence of distribution of strategy between the tobacco industry, which had the highest frequency of incidences, and the other industries. (We were not able to compare all industries simultaneously because the expected frequencies were too small in some cells.) We used exact two-sided tests of independence of contingency table frequencies. The overall significance level for all four tests was 5%, so we used a significance level of 0.05/4 = 0.0125 for each of the four tests.

  3. Results

    Table 1 shows the numerical occurrences of each strategy by industry. The distribution of strategies used by tobacco was different from each of the other industries. The null hypothesis of independence was rejected for all four tests with an overall significance level of 5%. An exploratory analysis of the source of this lack of independence looked at the largest differences between frequencies expected (assuming independence) and the observed frequencies for each strategy, using the chi-square contribution. The largest differences in strategy frequency between the five industries are explained at the bottom of Table 1. (The P-VALUE for each test is very small for each comparison, far less than the significance level of 0.0125 for each test.)

    1. MANIPULATION STRATEGIES

  4. Fund Research that Supports the Industry's Interests

    Each of the five industries sponsored research (both internal and external) intended to produce findings that would support their commercial interests. Internal documents indicate that much of this research was designed specifically to refute independent research. A lack of disclosure, or even deliberate secrecy, about research funding from industry is also a characteristic of these funding arrangements. The federal government has promulgated numerous policies and standards regarding conflicts of interest, to which private entities in contract with the government must adhere. (9) The nongovernmental American Academy for the Advancement of Science (AAAS) has also published guidelines on conflict of interest. (10) These various standards can serve as guides for research practices in the private sector, but this is not always the case.

    1. Tobacco

      In 1976, Brown & Williamson's General Counsel stated explicitly to company executives that the purpose of the tobacco industry's investment in tobacco and health research was to raise doubts about tobacco's harmfulness:

      During the past 20 years the industry has committed more than $50 million to scientific research related to tobacco and health. The significant expenditures on the question of smoking and health have allowed the industry to take a respectable stand along the following lines--"After millions of dollars and over twenty years of research, the question about smoking and health is still open." (11) 2. Pharmaceutical

      STEPS (Study of Neurontin: Titration to Effectiveness and Profile of Safety) was an uncontrolled, open-label dosing study in which physicians were recruited to prescribe Neurontin to epilepsy patients. (12) While the purpose stated in the published article was to examine the effectiveness of Neurontin, internal documents state the purposes as to "communicate with ... patients ... the benefits of adding Neurontin" and "teach physicians to titrate Neurontin to clinical effect." (13) The ultimate goal was to increase market share and per patient dosage of Neurontin, and to preempt launch of competitor drugs. (14)

    2. Lead

      In 1935 the head of the Lead Industry Association (LIA) complained to members that:

      We are constantly investigating alleged cases of lead poisoning and endeavoring to correct misstatements about lead poisoning, to calm misapprehension about the toxic properties of the metal.... [In this regard,] [a]t the request of Dr. Aub of the Harvard Medical School ... we are contributing $10,000 to Harvard College to continue the lead investigations.... (15) Aub's research generally minimized the harm of lead. (16)

    3. Vinyl Chloride

      In 1967 the Manufacturing Chemists' Association (MCA) agreed to fund a study "hopefully expected to ... confirm that the condition [a degenerative bone disease called acroosteolysis] is purely an occupational disease and in no way affects the general public using polyvinyl chloride products." (17) In...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT