Plain Language Summary

What is it about?

Surrogate endpoints are measurements used in clinical trials that substitute for outcomes that matter most to patients, such as survival or quality of life. About 60% of new drugs approved in the United States over the past 20 years relied on surrogate endpoints because they allow shorter, smaller trials. However, most surrogate endpoints lack strong evidence that they predict real clinical benefit, creating uncertainty for healthcare decision makers. This gap is particularly problematic for health technology assessment agencies that evaluate whether healthcare systems should pay for new treatments but receive limited guidance on evaluating surrogate endpoints. This report provides a practical framework for validating surrogate endpoints using statistical methods and incorporating them into cost-effectiveness analyses. The report’s main contribution is offering unified recommendations that address 3 evidence levels: biological plausibility, the link between surrogate and patient outcomes, and the relationship between treatment effects on the surrogate and patient outcomes.

How was the research conducted?

The research used a structured expert consensus approach to develop practical recommendations. A multistakeholder task force of statisticians, health economists, regulators, decision makers, academics, and industry representatives convened in 2024. The researchers reviewed existing methods for surrogate endpoint evaluation, examined guidelines from health technology assessment agencies worldwide, and analyzed real-world case studies of drug appraisals. The team synthesized this evidence through multiple rounds of expert feedback. This consensus method was used because it integrated diverse expertise while ensuring recommendations would be practical for all stakeholders involved in healthcare decision making.

What were the results?

The most important finding is that multivariate meta-analytic methods that account for within-study correlations and measurement errors are the best approach for evaluating surrogate endpoints. The task force found that Bayesian methods can improve surrogate evaluation by borrowing information across different treatments or diseases when data are limited. A surprising finding was the large gap between available statistical methods and actual practice: few health technology assessment submissions used rigorous surrogate validation despite appropriate methods being available.

Why are the results important?

These results improve transparency and consistency in healthcare decisions when trials use surrogate endpoints instead of patient-relevant outcomes. The findings could change practice by giving health technology assessment agencies clear evaluation criteria and manufacturers better guidance on statistical methods. Patients, clinicians, health technology assessment agencies, payers, and manufacturers all benefit through more rigorous evaluation of whether treatment effects on surrogate endpoints translate into real clinical benefits. Long-term impacts could include common international standards for surrogate endpoint evaluation and better alignment between drug approval and reimbursement processes.

What are the strengths and weaknesses of this study?

The main strength is the comprehensive, multistakeholder approach that connects statistical validation methods with practical health economic modeling guidance. The primary limitation is that recommendations require sufficient high-quality data from multiple randomized trials, which often does not exist when health technology assessments occur, especially for rare diseases. Future research should develop methods for surrogate evaluation with extremely limited data and establish international consensus on minimum validation standards. 

^{Note: This content was created with assistance from artificial intelligence (AI) and has been reviewed and edited by ISPOR staff. For more information or for inquiries on ISPOR’s AI policy, click here or contact us at info@ispor.org.}