Journal of the American Medical Association
Barnes DE, Bero LA.
OBJECTIVE: To determine whether the conclusions of the review articles on the health effects of secondhand smoke are associated with the quality of the article, the affiliations of their authors or other characteristics of the article.
ORIGIN OF DATA: Review articles published from 1980 to 1995 were identified by electronic searches of MEDLINE and EMBASE and by a database of passive smoking symposia.
ITEM SELECTION: An article has been included if its stated or implied purpose was to examine scientific evidence that secondhand smoke is associated with one or more health outcomes. Articles were excluded if they did not specifically focus on the health effects of secondhand smoke or if they had not been written in English.
DATA EXTRACTION: The evaluation of the quality of the articles was assessed by 2 independent evaluators who were trained, followed by a written protocol, revealed no conflicts of interest and were blinded by all the study hypotheses and by the identification characteristics of the articles. The conclusions on the article were categorized by the 2 evaluators and one of the authors. The authors' affiliation was classified as or not affiliated with the tobacco industry, based on whether the authors were known to have received funding from or participated in activities sponsored by the tobacco industry. Other features of the article were classified by one of the authors using predefined criteria.
SUMMARY OF DATA: A total of 106 reviews were identified. Overall, 37% (39/106) of the reviews concluded that secondhand smoke is not harmful to health; 74% (29/39) of these were written by authors with affiliations in the tobacco sector. In multiple logistic regression analyzes that control the quality of the article, the peer review status, the topic of the article and the year of publication, the only factor associated with the conclusion that secondhand smoke is not harmful was if an author was affiliated with the tobacco industry (odds ratio, 88.4; 95% confidence interval, 16,4-476,5, P <0,001).
CONCLUSIONS: The conclusions of the review articles are strongly associated with the affiliations of their authors. Review article authors should disclose potential conflicts of financial interest, and review article readers should consider author affiliations when deciding how to judge an article's conclusions.