Rank invariant tests for interval censored data under the grouped continuous model

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Methods

Background:

• Interval censored data presents analytical challenges in survival analysis.
• Existing score tests often rely on proportional hazards assumptions.
• Generalizing Finkelstein's work is crucial for broader applicability.

Purpose of the Study:

• To develop rank invariant score tests for grouped or interval censored data.
• To generalize existing methods for analyzing survival data with censored observations.
• To provide a flexible framework for survival data analysis.

Main Methods:

• Framing the problem as a linear rank test for location shift.
• Utilizing a known error distribution for test construction.
• Developing adjustments for a large number of observation times.

Main Results:

• The proposed rank invariant score tests are effective for grouped/interval censored data.
• The study provides a generalization of Finkelstein's score tests.
• Graphical tests for the location shift model assumption are presented.

Conclusions:

• The developed score tests offer a significant advancement for survival data analysis.
• The methods are applicable to various scenarios involving censored data.
• Alternative interpretations are provided when location shift assumptions are violated.