Loneliness and global cognitive functioning in racially and ethnically diverse US midlife and older adults.

Introduction: Few studies have examined the association of loneliness and cognitive functioning in the US. We used two common measures of loneliness and examined their association in a large sample of US Black, Latino, and White adults (ages ≥ 50). Methods: We analyzed Wave 3 of the National Social Life, Health, and Aging Project (N = 2,757). We examined loneliness using one item from the CES-D and the Felt Loneliness Measure (NFLM); cognitive functioning was assessed using the Montreal Cognitive Assessment (MoCA) tool, where higher scores indicated better functioning. We used weighted ordinary least squares regressions to examine the effects of loneliness (CES-D loneliness and NFLM in separate models) on MoCA scores. In exploratory analyses, we examined if these relationships varied by race and ethnicity. We adjusted all models for sociodemographic and other salient factors (e.g., chronic disease, depressive symptoms, living alone). Results: Mean age was 63.49 years, 52% were female, and 9% were Black and 6% Latino persons. Approximately 54% endorsed feeling lonely on at least one measure; 31% (CES-D) and 46% (NFLM). The relationship between loneliness measures was positive and significant, X 2 (1, N = 2,757) = 435.493 p < 0.001. However, only 40% of lonely individuals were identified as lonely on both assessments. CES-D loneliness was inversely (߈ = -0.274, p = 0.032) associated with MoCA scores and this association did not vary by race and ethnicity. Greater NFLM loneliness was positively associated (߈ = 0.445, p < 0.001) with higher MoCA scores for Latino participants only. Discussion: Loneliness appears to be an important predictor of cognitive functioning. However, the association of loneliness and cognitive functioning varied when using the CES-D loneliness item or the NFLM. Future work is needed to understand how loneliness and its clinically relevant dimensions (social, emotional, existential, chronicity) relate to global and individual cognitive domains. Research is needed with racially and ethnically diverse midlife and older adults, particularly to understand our counterintuitive finding for Latino participants. Finally, findings also support the need for research on interventions to prevent cognitive decline targeting loneliness.