Researchers have warned of the risk of sponsorship bias in health economic studies in the field of cancer.
Industry sponsored studies were twice as likely to be cost minimisation analyses, which compare equally effective treatments on the basis of costs. The industry studies were 2.5 times less likely to be cost effectiveness analyses, say the authors of a study in the British Journal of Cancer ( 2003:89: 1405-8).
"Industry sponsored studies were 1.9 times more likely to have positive qualitative conclusions about costs than studies sponsored by non-profit organisations. This relationship can be attributed to a sponsorship bias," says the report.
"In order to counter the risk that the type of sponsorship might influence health economic analyses, it is necessary to improve the quality of such studies," say the authors. "Furthermore, full disclosure of all financial interests involved in health economic studies should be provided routinely."
In the study, the authors examined a database of health economic studies in the field of oncology to investigate whether any relation could be established between the type of sponsorship and the type of economic analysis, the health technology assessed, the sensitivity analysis performed, publication status, and qualitative conclusions about costs.
In all, 150 sponsored studies were included in the study. The sponsor of 71% of the studies was a non-profit organisation, with the remaining 29% being supported by an industry sponsor in whole or part.