Estimating the effects of hypothetical loneliness interventions on memory function among middle-aged and older adults in the United States

Loneliness is associated with faster memory decline in mid- and late-life, but it remains unclear whether interventions to ameliorate loneliness protect memory function. We examined the impact of sustained and one-time loneliness interventions on memory function among US middle-aged and older adults. Using the nationally representative Health and Retirement Study in 2006-2018, we estimated counterfactual average memory scores over 12 years of follow-up under the following scenarios: [A] a hypothetical intervention eliminating loneliness only at baseline (baseline intervention), [B] a hypothetical intervention eliminating loneliness for 10 years (sustained intervention), and [C] the natural course (no intervention). We used targeted maximum likelihood estimation to account for time-varying confounding. The analytic sample included 10 136 participants (median baseline age 64 years; representing 50 million community-dwelling adults). Over 12 years, estimated mean memory scores declined by 0.58 standardized units (95% CI, 0.56, 0.60) under the natural course; the difference in decline (vs. natural course) was 0.00 standardized units (95% CI, -0.01, 0.01) under the baseline intervention and 0.02 standardized units (95% CI, -0.02, 0.05) under the sustained intervention. Compared to the natural course, we did not find evidence that either the baseline or sustained interventions was associated with better memory function over follow-up.