Disentangling social isolation, loneliness, and later-life cognitive function for older adults in the United States: evidence from causal inference modeling

Objectives
Older adults are at high risk of the negative health impacts of social isolation and loneliness. One of those possible negative health impacts is Alzheimer’s disease, a leading cause of death for adults in the United States and many high-income countries. Taking a life course perspective, we explore whether there is a direct causal effect of social isolation on later-life trajectories of cognitive function, the extent to which any effect of social isolation on cognitive impairment operates indirectly through loneliness, who may be most vulnerable, and the potential efficacy of a statistical intervention for those living alone.

Methods
We use a counterfactual approach, the g-formula, with the U.S. Health and Retirement Study, analyzing data from 30,421 individuals with 137,653 observations across 2004-2018.

Results
We find a consistent pattern of social isolation having a detrimental direct causal effect on cognitive function, with only 6% of this effect operating through loneliness. Reducing social isolation has a protective effect on cognitive function for all subpopulations regardless of gender, race/ethnicity, and educational level, with only minor differences among social categories. Our statistical intervention shows that targeting social isolation in those living alone may be one viable public health strategy for protecting against cognitive decline.

Discussion
Our results suggest that addressing social isolation—and, by extension, its effects on health—requires both a broad understanding of its heterogenous impacts on the general population and a nuanced approach to targeting public health interventions where they can be most effective.