Associations Between Funding Source and Results of Cost Effectiveness Analyses of Drugs Used in Breast Cancer

Abstract

Background: Cost effectiveness studies are increasingly included in the regulatory decisions of many countries and in formulary decisions in the United States. Pharmaceutical company sponsorship of economic analyses of oncology drugs previously has been associated with reduced likelihood of reporting unfavorable results. Demonstrating persistence of this relationship may help enable better interpretation of study results and development of strategies to address potential bias. Methods: Breast cancer was selected for analysis since it is the cancer with the largest number of cost effectiveness studies and the cost of recent drugs has caused questioning of their incremental cost effectiveness. Search of the Tufts Medical Center Cost Effectiveness Analysis Registry resulted in 105 studies published between 1991-2012 that evaluated the cost effectiveness of drugs used to prevent or treat breast cancer. Overall study conclusions regarding cost effectiveness of the investigated drugs were evaluated using three thresholds: $50,000 per quality adjusted life-year (QALY), $100,000 per QALY, and $150,000 per QALY. A logistical regression was performed to determine how study characteristics including funding source were associated with study findings. Results: Overall, 65 studies were funded by industry (62%). Studies with pharmaceutical company funding were more likely than studies with other funding to report favorable cost effectiveness estimates (75.4% vs 40.0%, OR=4.07 CI=1.44-12.26 at the $50,000 threshold; 80.0% vs 57.5%, OR=3.15, CI=1.07-9.84 at the $100,000 threshold; and 87.7% vs 67.5%, OR=3.65, CI=1.02-14.66 at the $150,000 threshold). Conclusions: Industry sponsorship continues to be associated with a higher likelihood of reporting favorable results. These findings suggest that steps are necessary to ensure that the cancer cost effectiveness literature is not biased by sponsorship. Expanding funding sources other than pharmaceutical companies for these studies or pre-registering cost effectiveness studies may help mitigate this potential source of bias.