Generalizing the Finkelstein-Schoenfeld Test to Incorporate Multiple Alternating Thresholds

Area of Science:

• Biostatistics
• Clinical Trials
• Cardiovascular Research

Background:

• Composite endpoints combining fatal and nonfatal events are common in cardiovascular clinical trials.
• The standard Finkelstein-Schoenfeld (FS) test prioritizes fatal events, potentially underutilizing data from significant nonfatal events.
• Existing methods may not fully capture the clinical importance of all events in composite endpoints.

Purpose of the Study:

• To introduce the Finkelstein-Schoenfeld with multiple thresholds (FS-MT) test as a more flexible alternative for analyzing composite endpoints.
• To extend the capabilities of the traditional FS test by incorporating sequential, alternating thresholds.
• To develop a weighted adaptive approach for threshold determination within the FS-MT framework.

Main Methods:

• The proposed FS-MT test extends the FS test by applying multiple thresholds sequentially and alternating across different event types.
• A weighted adaptive approach is developed to optimize threshold selection for the FS-MT test.
• Simulation studies were conducted to evaluate the operating characteristics of FS-MT under various conditions (follow-up time, event correlation, treatment effects).
• A case study using the Digitalis Investigation Group trial data was performed to demonstrate the practical application of FS-MT.

Main Results:

• Simulations indicate that the FS-MT test retains the statistical properties of the FS test.
• The FS-MT test allows for more flexible utilization of information from lower-priority, nonfatal events.
• The method's performance was evaluated across different scenarios, demonstrating its robustness.
• The case study illustrated the practical utility and application of the FS-MT method.

Conclusions:

• The FS-MT test provides a statistically sound and more flexible approach to analyzing composite endpoints in cardiovascular trials compared to the standard FS test.
• This methodology allows for a more comprehensive assessment of treatment effects by better integrating information from both fatal and nonfatal events.
• The developed R package "FSMT" facilitates the implementation of this advanced statistical method in clinical research.