Using the NANA toolkit at home to predict older adults' future depression

Area of Science:

• Geriatric medicine
• Computational psychiatry
• Psychological assessment

Background:

• Depression is frequently underdiagnosed in the elderly population.
• The Novel Assessment of Nutrition and Aging (NANA) validation study collected mood data from older adults.
• This study explores the predictive potential of mood data for future depression.

Purpose of the Study:

• To investigate the utility of machine learning in predicting depression status in older adults.
• To identify specific mood indicators that can forecast future depression.
• To develop a tool for early detection of depression in geriatric populations.

Main Methods:

• Machine learning, specifically least absolute shrinkage and selection operator (LASSO) logistic regression, was applied to mood data.
• Six mood measures were averaged and used to predict depression status (Geriatric Depression Scale) 10 weeks later.
• Cross-validation and receiver operating characteristic (ROC) curve analysis were employed to validate the model.

Main Results:

• A predictive model was developed using self-reported sadness and tiredness.
• The model achieved a cross-validated area under the ROC curve of 0.88 (CI: 0.69-0.97).
• Sadness and tiredness emerged as significant predictors of future depression.

Conclusions:

• Self-reported sadness and tiredness are sensitive indicators for predicting future depression in older adults.
• The methodology shows potential for wider studies and clinical deployment.
• This approach can help mitigate the underdiagnosis of depression in the elderly.