Facing the Next "Geriatric Giant"-A Systematic Literature Review and Meta-Analysis of Interventions Tackling Loneliness and Social Isolation Among Older Adults
Lukas A Duffner,Niels Janssen,Kay Deckers,Sarah Schroyen,Marjolein E de Vugt,Sebastian Köhler,Stéphane Adam,Frans R J Verhey,Marja Y Veenstra
DOI: https://doi.org/10.1016/j.jamda.2024.105110
Abstract:Objectives: Loneliness and social isolation are associated with adverse health outcomes, especially within the older adult population, underlining the need for effective interventions. This systematic review and meta-analysis aims to summarize all available evidence regarding the effectiveness of interventions for loneliness and social isolation, to map out their working mechanisms, and to give implications for policy and practice. Design: Systematic literature review and meta-analysis. Setting and participants: Older adults (≥65 years). Methods: A systematic search was conducted in MEDLINE, PsycINFO, and CINAHL for studies quantitively or qualitatively assessing effects of interventions for loneliness and social isolation in older adults, following predefined selection criteria. Risk of bias as well as small study effects were assessed and, wherever appropriate, information about effect sizes of individual studies pooled using random-effects meta-analyses. Sources for between-study heterogeneity were explored using meta-regression. Results: Of n = 2223 identified articles, n = 67 were eventually included for narrative synthesis. Significant intervention effects were reported for a proportion of studies (55.9% and 50.0% for loneliness and social isolation, respectively) and 57.6% of studies including a follow-up measure (n = 29) reported sustained intervention effects. Meta-analysis of n = 27 studies, representing n = 1756 participants, suggested a medium overall effect of loneliness interventions (d = -0.47; 95% CI, -0.62 to -0.32). Between-study heterogeneity was substantial and could not be explained by differences in study design, year of publication, outcome measures, intervention length, participant demographics, setting, baseline level of loneliness, or geographic location. However, non-technology-based interventions reported larger effect sizes on average (Δd = -0.35; 95% CI, -0.66 to -0.04; P = .029) and were more often significant. Qualitative assessment of potential intervention mechanisms resulted in 3 clusters of effective components: "promoting social contact," "transferring knowledge and skills," and "addressing social cognition". Conclusions and implications: Interventions for loneliness and social isolation can generally be effective, although some unexplained between-study heterogeneity remains. Further research is needed regarding the applicability of interventions across different settings and countries, also considering their cost-effectiveness.