Title: Predicting Ten-Year Trajectories of Depressive Symptoms from Demographic, Health, Social, Psychological, Behavioural, and Genetic Factors: An Analysis of Middle-Aged and Older Adults in the Health and Retirement Study
Thesis Supervisors: Dr. Eli Puterman
Committee member: Dr. Mark Beauchamp, Dr. Guy Faulkner
Defence Chair: Dr. Bill Sheel
Abstract:
Depressive symptoms in midlife and older adulthood are common, persistent, and associated with cardiovascular diseases, cognitive impairment, and earlier mortality. Previous research has identified risk factors for depressive symptoms in middle and older adulthood, such as increasing age, female sex, race and ethnicity, major life stressors, lack of social connectedness, neuroticism, lower life satisfaction, physical inactivity, alcohol use, and polygenic risk scores, to name a few. However, these multidisciplinary predictors have not been compared to determine their relative contribution to depressive symptoms in middle and older adulthood. Furthermore, previous research has only assessed depressive symptoms and their predictors cross-sectionally or prospectively (i.e., 2 time points), which misconstrues the dynamic temporal nature of depressive symptoms. Therefore, using data from the Health and Retirement Study (N = 13,994), the aims of this study were threefold: 1) estimate trajectories of depressive symptoms in middle- aged and older adults using growth mixture modelling (GMM), 2) identify predictors of trajectory groups in order of importance using random forest (RF) analysis, and 3) compare how trajectory groups differ in relation to the most important predictors using multinomial logistic regression (MLR). GMM indicated that there were five latent classes following distinct trajectories of depressive symptoms over 10 years: “consistently low”, “increasing”, “consistently moderate”, “decreasing”, and “consistently high”. Out of a total of 72 predictors ranging from demographic to genetic factors, the top five predictors of depressive symptoms trajectories were trait anxiety, lower life satisfaction, neuroticism, perceived constraints, and hopelessness. MLR demonstrated that trajectory groups differed significantly in relation to psychological, health, and social factors, which highlights potential avenues for intervention.