Evaluation of conflict of interest in economic analyses of new drugs used in oncology

Context Recent studies have found that when investigators have financial re lationships with pharmaceutical or product manufacturers, they are less lik ely to criticize the safety or efficacy of these agents. The effects of hea lth economics research on pharmaceutical company revenue make drug investig ations potentially vulnerable to this bias. Objective To determine whether there is an association between pharmaceutic al industry sponsorship and economic assessment of oncology drugs. Design MEDLINE and HealthSTAR databases (1988-1998) were searched for origi nal English-language research articles of cost or cost-effectiveness analys es of 6 oncology drugs in 3 new drug categories (hematopoietic colony-stimu lating factors, serotonin antagonist antiemetics, and taxanes), yielding 44 eligible articles. Two investigators independently abstracted each article based on specific criteria. Main Outcome Measure Relationships between funding source and (1) qualitati ve cost assessment (favorable, neutral, or unfavorable) and (2) qualitative conclusions that overstated quantitative results. Results Pharmaceutical company-sponsored studies were less likely than nonp rofit-sponsored studies to report unfavorable qualitative conclusions (1/20 [5%] vs 9/24 [38%]; P = .04), whereas overstatements of quantitative resul ts were not significantly different in pharmaceutical company-sponsored (6/ 20 [30%]) vs nonprofit-sponsored (3/24 [13%]) studies (P = .26). Conclusions Although we did riot identify bias in individual studies, these findings indicate that pharmaceutical company sponsorship of economic anal yses is associated with reduced likelihood of reporting unfavorable results .