The purpose of this analysis of health economic studies in the field of oncology was to investigate among sponsored studies whether any relationship could be established between the type of sponsorship and (1) type of economic analysis, (2) health technology assessed, (3) sensitivity analysis performed, (4) publication status, and (5) qualitative conclusions about costs. The Health Economic Evaluations Database (HEED, version 1995(HEED, version -2000 was searched on the basis of oncological ICD-9 codes, sponsorship, and comparative studies. This search yielded a total of 150 eligible articles. Their evaluations were prepared independently by two investigators, on the basis of specific criteria. When evaluators disagreed, a third investigator provided a deciding evaluation. There was no statistically significant relationship between the type of sponsorship and sensitivity analysis performed (P ¼ 0.29) or publication status (P ¼ 0.08). However, we found a significant relationship between the types of sponsorship and of economic analysis (P ¼ 0.004), the health technology assessed (Po0.0001), and qualitative cost assessment (P ¼ 0.002). Studies with industrial sponsorship were 2.56 (99% lower confidence interval (CI) ¼ 1.28) times more likely to involve cost-minimisation analyses, were 0.04 (99% higher CI ¼ 0.39) times less likely to investigate diagnostic screening methods, and were 1.86 (99% lower CI ¼ 1.21) times more likely to reach positive qualitative conclusions about costs than studies supported by nonprofit organisations. In conclusion, our results suggest that there is a greater probability that industry-sponsored economic studies in the field of oncology tend to be costminimisation analyses, to investigate less likely diagnostic screening methods, and to draw positive qualitative conclusions about costs, as compared to studies supported by nonprofit organisations. There has long been discussion as to whether commercial sponsorship of clinical studies produces a conflict of interests (Davidson, 1986;Rochon et al, 1994;Krimsky and Rothenberg, 1998;Smith, 1998;Stelfox et al, 1998;Montaner et al, 2001;Morin et al, 2002). However, only two applied studies deal with this issue for health economic studies conducted in the field of oncology. Friedberg et al (1999) found that, in the case of new drugs developed for oncological use (including haematopoietic growth factors, antiemetics, taxanes), pharmaceutical sponsorship of economic analyses was associated with a low likelihood of reporting unfavourable results. Subsequently, however, the same authors moderated these conclusions about costs in a comparison of industry vs nonprofit-sponsored economic analyses of six novel drugs used in oncology (Knox et al, 2001). After reviewing all the available pharmacoeconomic reports, they established the differences in study reporting -but not in the types of journals in which the studies were published -between pharmaceutical companyand non-profit-sponsored studies. They concluded that these results, and in particular the...