Barnes and Bero report that while most reviews of passive smoking conclude that it is harmful to human health, when the author has an affiliation with the tobacoo industry, the review is more likely to reach the opposite conclusion.

What did they do? Barnes and Bero analyzed 106 reviews of the health effects of passive smoking published from 1980 to 1995 for which it was possible to determine whether one or more authors had received funding from or participated in activities sponsored by the tobacco industry. Articles were identified through electronic literature searches and a database of symposia related to the topic.

Each review article was evaluated by 2 independent assessors with no conflict of interest in the outcome and without knowledge of the identity or affiliation of the review authors. The assessors categorized the article as concluding either that passive smoking is or is not harmful. They also assessed a series of other parameters related to the quality of the review article, its peer review staus, year of publication, etc. These variables were all then used in a multivariate analysis to determine which article characteristics were most closely associated with the conclusion that passive smoking is not harmful to health.

What did they find? Of the 106 review articles, 29 were written by tobacco-affiliated authors (see Table, below). Of these 29, 94% concluded that passive smoke is not dangerous. Only 2 written by tobacco-affiliated scientists found that passive smoking is harmful. In contrast, of the 75 articles written by independent scientists, 87% found that passive smoke is harmful.

The pattern in this table is statistically highly significant (P< .001). When an article is written by a tobacco scientist, it is highly unlikely that the conclusions will be unfavorable to the industry.

The multiple regression analysis yielded a similar result. "The odds that a review article with tobacco industry-affiliated authors would conclude that passive smoking is not harmful were 88.4 times higher than the odds for a review article with non-tobacco affiliated authors, when controlling for article quality, peer review status, article topic, and year of publication (95% CI, 16.4-476.5; P < .001)."

What does this mean? This is one example of a series of analyses showing that when an author has a financial interest in the outcome of a study, the conclusions are likely to be biased in favorof the financial interest. There is every reason to believe that this finding is relevant for studies of the health effects of contamination, also. Indeed an analysis of papers obtained through legal proceedings in tobacco suits has revealed that the Chemical Manufacturers Association (known now as the American Chemistry Council) participated with the tobacco industry in co-funding a organization established explicitly to weaken epidemiological conclusions.

The debate over lead poisoning is a past example from the chemical industry in which industry scientists produced consistently biased findings. A similar case appears to be playing out currently, with multiple independent scientists reporting many low-level effects of bisphenol A, while industry labs report they cannot replicate the studies.