Systematic Bias in Industry-sponsored Cost-effectiveness Studies
Industry-sponsored studies on a new drug or health technology are more likely to be found ‘cost-effective’ than independent studies, across a range of diseases, according to findings from a study published in The BMJ.
In a linked editorial, experts make a call for better reporting of results, more transparency, open-source cost-effectiveness models, and more independent studies, to reduce decision makers’ reliance on potentially biased cost-effectiveness analyses.
A cost-effectiveness analysis (CEA) provided the manufacturer is required by some countries to weigh up a product’s costs and effects.
This cost analysis evidence can be used to set the price for a drug or health technology or decide whether insurance policies will cover them. New drugs covered by insurance plans can be much more profitable than those not covered, which could lead to bias in CEAs funded by the drug and technology manufacturing industry.
While previous studies have consistently shown sponsorship bias in CEAs, most studies were limited to specific diseases, and are out of date. To fill in the gaps, Feng Xie and Ting Zhou from McMaster University, Canada, analysed data from all eligible CEAs published between 1976 and March 2021.
They selected CEAs that reported an incremental cost-effectiveness ratio (ICER) using quality-adjusted life years or QALYs – a ‘value for money’ metric of years lived in good health.
The authors used data from the Tufts Cost-Effectiveness Analysis Registry. In total, 8192 CEAs were included in the study, of which nearly 30% were sponsored by industry.
The study defined CEA industry sponsorship as an analysis funded by drug, medical device, or biotechnology companies, either wholly or in part.
The results show that the industry-sponsored CEAs were significantly more likely to conclude that the new medicine or health technology was cost-effective than those not sponsored by industry.
For example, industry-sponsored studies were more likely to report the intervention being studied as cost-effective below the commonly used threshold of $50 000 per QALY gained than non-industry sponsored studies.
Among 5877 CEAs that reported the intervention was more effective but more expensive than the comparator, the ICERs from industry sponsored studies were one third (33%) lower than those from non-industry sponsored studies.
While only having the registry information to work with was a limitation, the authors said their analysis provides a basis for comparison with previous investigations.
As such, they suggested that “sponsorship bias in CEAs is significant, systemic, and present across a range of diseases and study designs.”
In lower and middle-income countries, industry bias can increase drug prices, where fewer resources mean decision-makers often need to rely on published, rather than independent CEAs.
In a linked editorial, Adam Raymakers at Cancer Control Research, Canada, and Aaron Kesselheim at Brigham and Women’s Hospital, USA, argue that decision-makers “should exercise caution when using published cost-effectiveness analysis in coverage decisions.”
They say finding solutions to tackle bias is more important than ever, and make the case for open-source analysis models, increased transparency, and increased funding for independent analyses, to help minimise reliance on industry-sponsored cost analyses.
Source: The BMJ
