Identifying major depressive disorder among US adults living alone using stacked ensemble machine learning algorithms

Area of Science:

• Computational Psychiatry
• Machine Learning in Healthcare
• Public Health Research

Background:

• Adults living alone face a higher risk of Major Depressive Disorder (MDD).
• Existing MDD prediction models do not specifically cater to individuals residing alone.
• There is a need for tailored prediction tools for this demographic.

Purpose of the Study:

• To develop and validate a machine learning model for predicting MDD in adults living alone.
• To investigate the relationship between personal health data and MDD risk in this population.
• To identify key predictors of MDD among adults living alone.

Main Methods:

• Utilized data from the US National Health and Nutrition Examination Survey (NHANES) from 2007-2018.
• Developed a Stacked Ensemble Machine Learning (SEML) model using demographic, lifestyle, and health data.
• Employed SHapley Additive exPlanations (SHAP) to interpret model predictions and identify risk factors.

Main Results:

• The SEML model achieved a robust performance with an Area Under the Curve (AUC) of 0.85.
• Identified significant risk factors for MDD including sleep disorders, medication use, and specific health conditions.
• Found protective factors such as higher age and certain dietary habits (e.g., reduced added sugar intake).

Conclusions:

• Successfully developed a novel SEML-based predictive model for MDD in adults living alone.
• The model enhances the identification of individuals at risk for MDD within this population.
• Provides valuable insights into the complex interplay of personal health data and MDD.