Group sequential designs for clinical trials with bivariate endpoints.

Area of Science:

• Clinical Trial Design and Monitoring
• Biostatistics
• Medical Research Methodology

Background:

• Clinical trials often use multiple endpoints, necessitating robust methods for decision-making.
• Ensuring scientific integrity requires balancing clinically relevant tradeoffs between endpoints.

Purpose of the Study:

• To present a novel framework for designing and monitoring clinical trials with bivariate outcomes.
• To establish a flexible approach that accommodates various bivariate designs and allows continuous tradeoffs between endpoints.

Main Methods:

• Utilized a rectangular hyperbola to define a bivariate null curve, dividing outcome space into benefit and lack of benefit regions.
• Developed a distance function in bivariate space to enable distinct decisions across four quadrants.
• Derived the distribution of a statistic based on the distance function for trial design and group sequential monitoring.
• Employed a recursive form of the bivariate group sequential density to control operating characteristics.

Main Results:

• The proposed bivariate designs meet or exceed standard requirements for statistical size and power.
• Demonstrated the framework's flexibility in capturing previous bivariate designs and allowing continuous endpoint tradeoffs.
• The methods provide a basis for robust trial design and monitoring in bivariate outcome spaces.

Conclusions:

• The presented framework offers a versatile and statistically sound approach for clinical trials with multiple endpoints.
• The methods are broadly applicable across diverse clinical settings and trial designs, enhancing scientific rigor.
• Successfully illustrated the framework's utility in a real-world pediatric clinical trial (Kids-DOTT).