Association between the use of surrogate measures in pivotal trials and health technology assessment decisions: a retrospective analysis of NICE and CADTH reviews of cancer drugs

Objective: To assess whether using surrogate versus patient-relevant endpoints in pivotal trials of cancer drugs was associated with health technology assessment recommendations in England (National Institute for Health and Care Excellence [NICE]) and Canada (Canadian Agency for Drugs and Technologies in Health [CADTH]). Methods: Cancer drug approvals from 2012 to 2016 were categorized by demonstrating benefit on overall survival (OS), progression-free survival, disease response, or having no comparator. Approvals were analyzed by benefit category and health technology assessment recommendation. The association between benefit (surrogate vs OS) and recommending a drug was examined using descriptive statistics and linear probability models controlling for unmet need, orphan designation, and cost-effectiveness. Results: Of 42 cancer indications that NICE recommended, 15 (36%) demonstrated OS benefit. Cancer indications with OS benefit were less likely to receive a recommendation from NICE than those without (P = .04). In linear probability models, availability of OS benefit was no longer associated with a recommendation from NICE (P = .32). Cost-effective cancer drugs had a 55.6% (95% CI: 38.9%-72.3%) higher probability of receiving a recommendation from NICE than those that were not. In Canada, 15 of 37 (41%) cancer indications that were recommended showed OS benefit. There was no detectable association between surrogate measures and CADTH recommendations based on descriptive statistics (P = .62) or in linear probability models (P = .73). Conclusion: When cost-effectiveness was considered, pivotal trial endpoints were not associated with NICE recommendations. Pivotal trial endpoints, unmet need, orphan status, and cost-effectiveness did not explain CADTH recommendations.