External funding has been found to be associated with publication independently of the statistical significance of the results (Dickersin 1997). Funding by government agencies was significantly associated with publication in three cohorts of proposals submitted to ethics committees (Easterbrook 1991, Dickersin 1992, Stern 1997) whereas pharmaceutical industry sponsored studies were less likely to be published in two studies (Easterbrook 1991, Dickersin 1992). Indeed, a large proportion of clinical trials submitted by drug companies to licensing authorities remain unpublished (Hemminki 1980, Bardy 1998).
In a systematic review, Lexchin et al. identified 30 studies published between 1966 and 2002 that examined whether funding of drug studies by the pharmaceutical industry was associated with outcomes that are favourable to the funder. They found that research funded by drug companies was less likely to be published than research funded by other sources, and that studies sponsored by pharmaceutical companies were more likely to have outcomes favouring the sponsor than were studies with other sponsors (Lexchin 2003). Other studies have since examined these associations and have found similar results (Bhandari 2004, Heres 2006). Heres et al., in a study of head-to-head comparisons of antipsychotics, found that the overall outcome of the trials favoured the drug manufactured by the industry sponsor in 90% of studies considered, and that some similar studies reported opposing conclusions, each supporting the product of the study sponsor (Heres 2006).
The implication is that the pharmaceutical industry tends to discourage the publication of negative studies that it has funded. For example, a manuscript reporting on a trial comparing the bioequivalence of generic and brand levothyroxine products, which had failed to produce the results desired by the sponsor of the study, Boots Pharmaceuticals, was withdrawn because Boots took legal action against the university and the investigators. The actions of Boots, recounted in detail by one of the editors of JAMA, Drummond Rennie (Rennie 1997), meant that publication of the paper (Dong 1997) was delayed by about seven years. In a national survey of life-science faculty members in the United States, 20% reported that they had experienced delays of more than six months in publication of their work and reasons for not publishing included “to delay the dissemination of undesired results” (Blumenthal 1997). Delays in publication were associated with involvement in commercialization and academic-industry research relationship, as well as with male sex and higher academic rank of the investigator (Blumenthal 1997).