Date of Award
Campus Access Dissertation
Doctor of Philosophy (PhD)
This study investigated the longitudinal causal relationships among depressive symptoms, cognitive functioning, and daily functional ability in older adults by means of a mediational analysis with latent growth curves. Two major hypotheses derived from the literature were investigated: risk factor hypothesis postulating that late-life depressive symptoms represent a risk factor for progressive decline in cognitive and daily functioning, and comorbidity hypothesis proposing that late-life depressive symptoms accompany and/or result from cognitive and functional decline.
The design of the study was a secondary data analysis from an ongoing longitudinal multicenter study, Alzheimer’s Disease Neuroimaging Initiative (ADNI). The final sample used for the analysis was comprised on 690 participants (age: 73.6±7.02, education: 16.09±2.77). The analyses employed a stepwise procedure: first, first-order latent growth curves were modeled for each variable, then second-order models were analyzed for higher level latent constructs, and finally a full mediational analysis with first- and second-order latent growth curves was performed.
Results indicated a good fit for both the risk factor hypothesis (χ2(df) = 919.79(352), p
The limitations of the current study included limited generalizability to populations different from the ADNI cohort, invariability constraints, and limited number of observed variables for inclusion into latent constructs.
Our findings identified two distinguishable pathways, which deepen the understanding of the nature of longitudinal associations among affective symptoms, cognitive ability, and functional impairment in older adults. Regardless of whether late-life depression precedes or follows the onset of cognitive symptoms, it was found that its effect on the affected individual’s daily functioning was substantial. Implications for clinical practice and future studies were discussed.
Tsoy, Elena, "Late-Life Depression and Cognitive Decline: A Latent Growth Curve Modeling Analysis" (2018). Graduate Doctoral Dissertations. 420.