OBJECTIVES: The gold standard to evaluate efficacy of new therapeutics is the Randomized comparative phase III trial. However, it faces challenges due to precision medicine advancements. External control arms (ECAs) offer alternatives, in particular for rare diseases, enabling comparisons without head-to-head trials. Employing ECAs is complex, in part due to varying regulatory guidelines. We aimed to assess their acceptance in oncology for different health technology assessment (HTA) institutions.

METHODS: We reviewed the opinions for reimbursement of oncology drugs between 2021 and 2023 that involved an ECA from England, France and Germany HTA bodies. Automated electronic searches were conducted on their websites to search eligible HTA opinions. The extracted information included available data for the control arm (aggregated data (AgD) vs individual patient data (IPD)), statistical methodology, acceptance and critics related to data limitation, design and methodology.

RESULTS: A total of 148 ECAs were extracted (49 England; 66 France; 33 Germany). The acceptance rate was 41% in England (AgD 13/20, IPD 7/29), 14% in France (AgD 6/30, IPD 3/24, ND 0/12) and 0% in Germany (AgD 0/9, IPD 0/23, ND 0/1). For AgD, the most used method was Matching-Adjusted-Indirect-Comparisons (England 95%, France 93%, Germany 100%). For IPD, the most used methods were unadjusted comparisons (England 59%, France 8%, Germany 30%) and weighted comparisons (England 28%, France 63%, Germany 39%). Major criticisms were related to heterogeneity/bias (England 47%, France 56%, Germany 58%), unclear data (England 43%, France 56%, Germany 64%), statistical methods (England 24%, France 39%, Germany 9%) and sample size (England 33%, France 30%, Germany 9%).

CONCLUSIONS: Our research confirms disparities between countries to assess treatments using ECAs, with key issues in data clarity and methodology, even if more treatments are available for rare cancers. This emphasizes the importance of improving how methods are applied to ensure consistency and reliability.