Sequential tests of promise with discrete time-to-event data

Area of Science:

• Biostatistics
• Clinical Trial Design
• Epidemiology

Background:

• Futility study designs are crucial for efficient clinical trials, especially in pediatric HIV research.
• Early antiretroviral therapy (ART) aims to achieve viral remission in infants with perinatally-acquired HIV.
• Traditional designs may lack efficiency for time-to-event data and small sample sizes.

Purpose of the Study:

• To introduce a novel family of sequential test procedures for futility studies, termed 'tests of promise'.
• To adapt these procedures for time-to-event data, enhancing efficiency in clinical trials.
• To provide methods for calculating operating characteristics and optimal boundaries for these sequential tests.

Main Methods:

• Development of sequential test procedures for futility analysis in time-to-event data.
• Comparison of proposed methods with existing designs like Simon's two-stage design for binary outcomes.
• Calculation of operating characteristics and determination of optimal/near-optimal stopping boundaries.

Main Results:

• The proposed sequential tests of promise demonstrate significant efficiency gains.
• Improvements are noted in early stopping and statistical power compared to traditional methods.
• Optimal boundaries are provided for small to medium sample sizes, suitable for specific trial contexts.

Conclusions:

• Sequential tests of promise offer a more efficient approach for futility studies with time-to-event data.
• These methods are particularly relevant for trials testing early interventions, such as ART for HIV-infected infants.
• The design is also applicable to trials investigating disease cures requiring treatment interruption and small participant numbers.