Prevalence and correlates of loneliness and social isolation in the oldest old: a systematic review, meta-analysis and meta-regression

Abstract

Purpose Conducting a systematic review, meta-analysis and meta-regression regarding the prevalence and correlates of loneliness and social isolation amongst the community-dwelling and institutionalised oldest old (80 years and over). Methods Three electronic databases (PsycINFO, CINAHL and Medline) were searched, including studies from inception to January 5, 2023. An additional hand search was conducted by checking included studies’ references, and studies that cited included studies. We included observational studies describing the prevalence and (ideally) the correlates of loneliness, or social isolation, amongst individuals aged 80 years and over. Study design, operationalization of loneliness and social isolation, statistical analysis, characteristics of the sample and key findings were extracted. A random-effects meta-analysis was conducted. Results We included 22 studies. The estimated prevalence of severe loneliness was 27.1% (95% CI: 23.7–30.4%). The estimated prevalence of moderate loneliness equalled 32.1% (95% CI: 15.8–48.4%). Moreover, the estimated prevalence of social isolation was 33.6% (95% CI: 28.9–38.2%). There was heterogeneity between the studies. Egger tests suggest the absence of potential publication bias. Meta-regressions showed that the heterogeneity could neither be attributed to the assessment of loneliness nor to the continent where the study was conducted. Conclusion Loneliness and social isolation are important problems in the oldest old. In this age group, studies are required, in particular from regions outside Europe. Additionally, longitudinal studies are required to investigate the determinants of loneliness and social isolation amongst individuals aged 80 years and over. Studies using more sophisticated tools to quantify loneliness and social isolation are required.