# A Statistical Concept for Conditional Marketing Authorisation Based on the Intermediate and Final Outcomes of a Single Confirmatory Randomised Clinical Trial

## TL;DR

This paper proposes a new statistical approach for granting early approval of medicines based on a single clinical trial with intermediate and final outcomes.

## Contribution

The paper introduces a co-primary endpoint method using group sequential designs for conditional marketing authorization.

## Key findings

- The proposed method aligns with EMA guidelines and improves overall trial outcome interpretation.
- The approach has minimal impact on sample size or timing of interim analyses.
- An oncology trial example demonstrates the validity and flexibility of the method.

## Abstract

Conditional marketing authorisation (CMA) is a path to early market access of new medicines addressing an unmet medical need in the European Union (EU), and similar concepts exist in other regulatory regions. For justifying a CMA, the benefit‐risk ratio has to be positive, and the applicant must be able to provide comprehensive data post‐authorisation for converting the CMA to a full marketing authorisation (MA). A recent proposal is to plan a single randomised clinical trial with interim analyses and base the decision for CMA and full MA on an intermediate and a final primary endpoint. To control the study‐wise type‐1‐error (T1E), the dual primary endpoint concept, essentially a Bonferroni‐split of the study‐wise T1E between the intermediate and final endpoints, has been proposed. We argue that the resulting statistical definition of formal study success is not in line with clinical assessment of the overall trial outcome. Consistent with the European Medicines Agency (EMA) guideline on multiplicity issues in clinical trials, the intermediate and final endpoints should be co‐primary, and we propose that each is assessed with a standard group sequential design. We illustrate our proposal using an oncology phase III trial with complete remission (CR) as the intermediate and overall survival (OS) as the final primary endpoint. Based on the example, we demonstrate that our proposal is a valid and flexible alternative with minimal to no impact on costs in terms of sample size or timing of the interim analysis intended for applying for a CMA but it improves the interpretation of the overall trial outcome.