Testing Equality in Ordinal Data with Repeated Measurements: A Model-Free Approach

Area of Science:

• Biostatistics
• Clinical Trials Methodology
• Ordinal Data Analysis

Background:

• Randomized clinical trials frequently involve repeated measurements of ordinal categorical responses.
• Analyzing such data requires robust statistical methods to accurately assess treatment effects.

Purpose of the Study:

• To propose a model-free approach for measuring relative treatment effects using the generalized odds ratio (GOR).
• To develop procedures for testing treatment effect equality and treatment-by-period interaction.
• To provide interval estimators for the GOR.

Main Methods:

• Utilized a model-free approach based on the generalized odds ratio (GOR).
• Developed statistical procedures for hypothesis testing on treatment effects and interactions.
• Derived interval estimators for quantifying the GOR.

Main Results:

• The proposed methods allow for the analysis of repeated ordinal outcomes without assuming a specific data distribution.
• Procedures for testing treatment equality and treatment-by-period interaction were successfully developed.
• Interval estimators for the GOR were derived for practical application.

Conclusions:

• The generalized odds ratio (GOR) offers a flexible and powerful tool for analyzing repeated ordinal data in clinical trials.
• The developed procedures provide reliable methods for assessing treatment effects and interactions in complex trial designs.
• The approach is illustrated with real-world data from surgical outcomes and insomnia treatment studies.