Predictors of older adults’ objectively measured social isolation: A systematic review of observational studies

•First systematic review focused solely on objectively evaluated social isolation.•Studies included in this review had moderate to high risks of bias.•Predictors were clustered into four categories.•Building consensus on the appropriate scale and standard cut-off point are necessary. : Preventing social isolation is essential for promoting healthy aging. This study aims to 1) systematically review and synthesize the predictors of objectively measured social isolation from observational studies that have treated isolation as an outcome in community-dwelling older adults, and 2) overview previous studies in this area and reveal an agenda to develop future research. : Peer-reviewed primary studies published in English or Japanese were identified from PubMed, ScienceDirect, PsycINFO, Igaku-Chuo-Zasshi, and CiNii. We followed the PRISMA statement. Risk of bias of included studies was assessed using the Newcastle-Ottawa Scale. : Of the 1,161 studies identified, 10 were included in the review. Most studies were cross-sectional and published in the past 20 years. The risk of bias score ranged from 3 to 6 for the individual studies, indicating a moderate to high risk. The assessment measures for social isolation and proportion of isolated subjects varied greatly across studies. Findings on the predictors of social isolation from each study were clustered into four categories: socio-demographic factors, physical health factors, psychological and cognitive factors, and social and cultural factors. : It was difficult to confirm these predictors conclusively because of the cross-sectional study design and a huge variety of assessment tools for social isolation. This result suggests the need to build consensus on the appropriate scales and standard cut-off points for assessing social isolation. Furthermore, longitudinal studies are urgently required. When considering predictors of social isolation, it is also essential to consider the regional and ethnic background of the samples studied.