Improved Functional Causal Likelihood-Based Causal Discovery Method for Diabetes Risk Factors.

Area of Science:

• Endocrinology and Metabolism
• Biostatistics
• Computational Biology

Background:

• Diabetes mellitus presents a global health challenge, necessitating effective prevention and treatment strategies.
• Existing research on diabetes risk factors primarily relies on correlation, overlooking crucial causal insights.
• Understanding causality is vital for developing targeted and effective diabetes prevention interventions.

Purpose of the Study:

• To develop and validate a novel causal discovery method for identifying diabetes risk factors.
• To address limitations of existing methods, such as redundant and false causal edges.
• To establish a foundation for future causality-focused research in diabetes.

Main Methods:

• Development of an improved functional causal likelihood (IFCL) model incorporating an adjustment threshold.
• Design of an IFCL-based causal discovery algorithm.
• Simulation experiments using datasets of varying sample sizes (768 and 2000).

Main Results:

• The IFCL-based algorithm successfully identified causal relationships among diabetes risk factors with fewer redundant and false edges.
• A larger dataset (n=2000) yielded a more informative causal structure compared to a smaller dataset (n=768).
• Identified causal links include: insulin→plasma glucose concentration, plasma glucose concentration→body mass index (BMI), triceps skin fold thickness→BMI and age, diastolic blood pressure→BMI, and number of times pregnant→age.

Conclusions:

• The developed algorithm effectively discovers causal relationships among diabetes risk factors.
• The IFCL model enhances the accuracy and reliability of causal structure identification.
• This methodology offers a valuable reference for future causality studies in diabetes research and prevention.