Abstract
The current study aimed to evaluate specific mechanisms of interventions to improve loneliness among older adults. EMBASE, MEDLINE, and PsycINFO databases were searched for articles published through June 2024. We selected randomized controlled trials (RCTs) that sought to improve loneliness in older adults, were published in English, and used previously published measures to assess loneliness. We used random-effect models to calculate pooled standardized mean differences (SMDs) for each mechanistic approach, and random-effects meta-regression to assess heterogeneity of studies. Thirty-five RCTs from 31 published studies (total N = 5291 participants) met review criteria, including six distress tolerance interventions, 11 cognitive engagement and restructuring interventions, 12 social behavioral activation interventions, and six acceptance of aging interventions. Study samples included older adults from the community. Meta-analysis revealed significant pooled SMDs for improving loneliness (pooled SMD = -1.11 (95% CI [-2.19, -0.03], p = 0.043), 25 studies). Heterogeneity of studies was high (I(2) = 99.57%). There was a trend for publication bias among cognitive and social behavioral activation interventions, though effect sizes did not change with adjustment. Meta-regression analysis found that distress tolerance interventions, group interventions, interventions with longer and more frequent sessions, and cognitive or acceptance of aging-focused interventions in older adults had larger effect sizes.