Classification and regression tree analysis vs. multivariable linear and logistic regression methods as statistical...

Area of Science:

• Medical Statistics
• Biostatistics
• Hematology Research

Background:

• Haemophilia A and B are genetic bleeding disorders due to deficiencies in coagulation factors VIII or IX.
• Multivariable models are crucial for identifying factors associated with binary or continuous outcomes in haemophilia research.
• Traditional regression models can be complex for non-statisticians to implement and interpret.

Purpose of the Study:

• To provide a didactic explanation of Classification and Regression Trees (CART) analysis for haemophilia research.
• To elucidate the application, rationale, and timing for utilizing CART in this specific medical domain.

Main Methods:

• Classification and Regression Trees (CART) is a non-parametric, non-linear method involving sample partitioning.
• Classification Trees (CTs) are applied for categorical outcomes.
• Regression Trees (RTs) are utilized for continuous outcomes.

Main Results:

• CART methodology is gaining traction in the medical field, though its application in haemophilia research is limited.
• Two published examples of CART analysis in haemophilia are presented and explained in detail.
• The study highlights the growing use of CART in medical research.

Conclusions:

• CART analysis offers a more accessible approach to medical decision-making due to its ease of implementation and interpretation.
• The study advocates for the promotion of CART analysis in haemophilia research for analyzing both continuous and categorical outcomes.
• CART provides a valuable tool for researchers and clinicians in the field of haemophilia.