Sleep Quality, Nutrient Intake, and Social Development Index Predict Metabolic Syndrome in the Tlalpan 2020 Cohort: A...

Area of Science:

• Metabolic health research
• Public health
• Data science in medicine

Background:

• Metabolic Syndrome (MetS) prevalence is influenced by complex factors.
• Understanding gender-specific risk factors is crucial for effective intervention.
• Previous studies often lack comprehensive analysis of social and lifestyle determinants.

Purpose of the Study:

• To investigate the relationship between Metabolic Syndrome (MetS), sleep disorders, nutrient intake, and social development factors.
• To identify gender-specific predictors of MetS using advanced machine learning.
• To address data imbalance in a Mexico City cohort.

Main Methods:

• Employed machine learning models (Random Forest, RPART) for MetS prediction.
• Utilized data balancing techniques (SMOTE, ADASYN) to handle an unbalanced dataset.
• Analyzed predictors including body mass index, nutrient intake, sleep quality, and social factors.

Main Results:

• Random Forest achieved a balanced accuracy of approximately 87% in MetS prediction.
• Key predictors for men: BMI, family history of gout.
• Key predictors for women: waist circumference, glucose levels, protein/fructose/cholesterol intake, sleep disturbances, and social factors.

Conclusions:

• MetS predictors significantly differ between genders, necessitating personalized management.
• Diet, sleep quality, and social factors are important, regionally dependent, contributors to MetS.
• Future research should explore the interplay of these factors for targeted MetS prevention and treatment.