Industry Sponsorship of “Cost Effectiveness Analyses” Produces Biased Results

New evidence shows how pharma affects “Cost-Effectiveness Analyses” to sell the public on their drugs.


In a new article published in BMJ, Feng Xie and Ting Zhou evaluate the link between industry sponsorship bias and favorable cost-effectiveness analyses (CEA) for industry products. CEAs are essential for the pharmaceutical industry because they inform price negotiations for industry products and allow them to be covered by health insurance.

The current research finds that industry-sponsored CEAs were more likely to reach conclusions favorable to the industry than their non-industry-sponsored counterparts. The authors conclude that sponsorship bias in CEAs is a significant problem and suggest CEAs be conducted by independent bodies to allow payers to negotiate lower prices and improve health insurance policy making.

The authors write:

“Sponsorship bias in CEAs is significant, systemic, and present across a range of diseases and study designs. Use of CEAs conducted by independent bodies could provide payers with more ability to negotiate lower prices. This impartiality is especially important for countries that rely on published CEAs to inform policy making for insurance coverage because of limited capacity for independent economic analysis.”

Sponsorship bias is an issue many authors have investigated in the past. Research has shown how bias in funding sources can influence clinical trials and create misleading results. This level of bias likely undermines the credibility of many popular treatments.

Industry money has likely led to bias in the treatment of depression, with industry-sponsored trials consistently finding that drugs are superior to therapy. Research has also found financial conflicts of interest in a “massive” number of antidepressant meta-analyses. With most medical research in the United States being industry-funded, sponsorship bias in healthcare is likely harming patients.

Researchers have found significant financial ties between DSM-IV panel members and the pharmaceutical industry, with 56% of panel members accepting payments from one or more pharmaceutical companies. Publication bias, in which research unfavorable to industry interests is not selected for publication, is common in clinical trials for schizophrenia and bipolar disorder. This research also frequently misreports results, interpreting them inaccurately in ways favorable to industry products.

In addition to sponsorship bias in research, the industry also sponsors numerous patient advocacy organizations (PAOs). With 67% of PAOs accepting payments from pharmaceutical companies and 11.9% receiving more than half their funding from the industry, even the voices supposedly speaking for service users are likely biased toward industry products. One researcher remarked that industry funding of these groups meant “the patient voice might speak with a ‘pharma accent’ when involved in policy discussions.

Research has found that increased transparency and mandatory disclosure of financial ties to the industry by researchers do not prevent bias. Researchers have also argued that the level of corruption present in psychiatric science “is incompatible with a human rights approach to mental health.

The present research begins by explaining the importance of CEAs to the pharmaceutical industry. The data resulting from CEAs is used to produce healthcare insurance coverage policies and inform price negotiations for industry products. With favorable CEA data, the industry can more easily get its products covered by health insurance providers at higher prices, which leads to significantly increased profits for that company.

Previous research has shown that industry-funded CEAs produce more favorable results for the sponsors. However, the most recent analysis of bias in CEA studies was completed more than 15 years ago. Therefore, the present research seeks to provide a more current understanding of sponsorship bias in CEAs.

The current study utilized the Tufts Cost-Effectiveness Analysis Registry as its data source. According to the authors, this registry is one of the most comprehensive databases for CEA research. They selected CEAs that used incremental cost-effectiveness ratio (ICER) as an outcome measure to allow for study comparisons. The authors marked CEAs as industry-sponsored if it was funded, at least in part, by industry. The authors included 8192 CEAs in the present study. 2347 (29.7%) of the CEAs they investigated were sponsored by industry.

Industry-sponsored CEAs were more than twice as likely as their independent counterparts to conclude that industry products were cost-effective. CEAs for drug-related interventions (using a drug to treat a problem) showed significant sponsorship bias. Industry-sponsored CEAs for a number of issues, including mental disorders, were more likely to be deemed cost-effective at the lowest cost threshold examined in the present research ($50,000).

Industry-sponsored CEAs conducted outside North America also found more interventions to be cost-effective at the $50,000 threshold. Industry-sponsored CEAs that found an intervention to be more effective and expensive than currently available treatments reported ICERs 33% lower than their non-industry-sponsored counterparts.

The authors interpret these findings to mean sponsorship bias is significant and systematic in CEAs. CEAs on drugs were particularly problematic, with drug CEAs representing nearly 75% of industry-sponsored studies compared to just over 33% of their non-industry-sponsored counterparts.

The authors also note that 70% of published ICERs were below $50,000 (the lowest threshold in the present research), and only 20% were above $100,000, indicating a high likelihood of publication bias against studies with results unfavorable to industry interests. If this bias is present, as the authors suspect, the problem with sponsorship bias in CEAs is likely worse than the current research indicates.

The authors conclude that given the amount of sponsorship bias within CEAs, and their importance in determining health insurance coverage policy-making and price points for industry products, the use of independent, non-industry-sponsored CEAs is imperative.



Xie, F., & Zhou, T. (2022). Industry sponsorship bias in cost-effectiveness analysis: Registry-based analysis. BMJ. (Link)


  1. The idea that these drugs are effective is based on a scam, the idea that there is some legitimate effect that they should have, that there is some kind of an illness that they should be treating.

    So to be participating in the research project, even to be reviewing the data is immoral.


    Report comment