Abstract
According to the WHO, 15 million people worldwide suffer a stroke annually. Of these, 5 million die and another 5 million are left permanently disabled. Patients endure the impacts of strokes during their rehabilitation and afterward, placing economical and emotional burdens on family and community. Using data from the Health and Retirement Study (HRS) of the USA, the research performed a large-scale prospective analysis to examine how demographics, socioeconomic factors, cognition, emotion, and physical activity predict functional impairment and mortality. Multiple regression was employed to identify statistically significant variables that predict longitudinal Activities of Daily Life (ADLs). The least absolute shrinkage and selection operator (LASSO) logistic regression, a supervised machine learning approach, was deliberately chosen to obtain the subset of predictors that provide the best possible classification for the dependent variable. The LASSO regression produced a model with a fair mean Area Under the Curve (AUC) of 0.75 in predicting the risk of the patient's mortality. My findings also uncovered the important roles of BMI, mobility, muscle strength, memory, mental status, and socioeconomic status play in the long-term ADLs and survival of patients with stroke.