Depression, loneliness, and lower social activity as partial mediators of the association between visual impairment and cognitive decline

ObjectivesSensory impairment is a hypothesized risk factor for cognitive decline; however, the psychosocial pathways are not well understood. We evaluated whether the association between visual impairment (VI) and cognitive decline was partially mediated via depressive symptoms, loneliness, or social activity.MethodsWe used data from 2601 older adults enrolled in the Memory and Aging Project in 1997 and the Minority Aging Research Study in 2004 with neuropsychological tests across five domains measured annually for up to 16 years. VI was assessed with the Rosenbaum Pocket Vision Screener. Depressive symptoms, loneliness, and social activity were self-reported using validated scales. We used structural equation models to estimate the associations of VI with baseline and change in cognitive function, directly and indirectly through each mediator (depressive symptoms, loneliness, and social activity). We evaluated mediation via "psychological distress" using a latent variable combining depressive symptoms and loneliness.ResultsThe association between VI and global cognitive decline was mediated via lower social activity (indirect effect) [95% confidence interval (CI)] of linear slope: -0.025 (-0.048, -0.011), via loneliness (-0.011 [95% CI: -0.028, -0.002]), and via psychological distress (-0.017 [95% CI: -0.042, -0.003]). We did not find sufficient evidence for mediation via depressive symptoms alone.ConclusionsThe harmful effect of VI on cognitive decline may be partially mediated through loneliness and lower social activity. We evaluated whether psychosocial factors mediate the association between visual impairment and cognitive decline. With data from the Memory and Aging Project and the Minority Aging Research Study, we used structural equation models to estimate the total, direct, and indirect associations through each mediator (depressive symptoms, loneliness, lower social activity, and a latent variable, psychological distress, combining depressive symptoms and loneliness). We found evidence for partial mediation via social activity and loneliness.