OBJECTIVES: The National Healthcare Institute of the Netherlands plays a pivotal role in the assessment of the (cost-)effectiveness of novel medicine to safeguard effective and affordable healthcare to the basic benefit package. Assessors of the Institute are often faced with premature efficacy results derived from interim analyses. However, it is unclear how to interpret interim efficacy results for decision making in the pharmacotherapeutic assessment. The objective of this study is to construct a methodological guideline on the interpretation of interim efficacy outcomes for the pharmacotherapeutic assessment of novel medicine.

METHODS: First, a systematic research was conducted to construct a theoretical framework of the available interim analysis techniques and the consequences of early trial termination on the efficacy outcome measure. Then, the assessment process of the Institute with regards to interim analyses was reviewed by studying published assessment reports which included interim efficacy data. Finally, guidelines instructing on the assessment of interim analyses were collected from the EMA and EUnetHTA databases and reviewed.

RESULTS: Five interim analysis techniques were identified that preserved the type I error at nominal levels (5%). Early stopping for benefit at an interim analysis often results in an overestimation of treatment effect. The magnitude of overestimation is greatest in small sized trials and when interim analyses are performed with a small information fraction. The assessment process of the Institute showed that overall the confidence in interim efficacy outcomes is moderate and uncertainty with regards to the outcome estimate is usually textually described. Finally, guidelines of the EMA and EUnetHTA highlight the importance of documenting prespecified interim analyses, to minimize the risk of statistical bias.

CONCLUSIONS: Based on these findings, the Institute constructed a guideline on the interpretation and the grading of evidence derived from interim analyses using the GRADE methodology, thereby assuring a consistent assessment of interim efficacy outcomes.