Loneliness and social isolation causal association with health-related lifestyle risk in older adults: a systematic review and meta-analysis protocol

Background: The health impacts of loneliness and social isolation among older adults are widely acknowledged. Despite this, there is no consensus on the possible causal nature of this relationship, which could undermine effectiveness of interventions. One body of thought is that loneliness and social isolation affect health-related behaviours to indirectly damage health. However, there has not been any systematic assessment of the association between loneliness and social isolation and health-related behaviours which considers the possible impact from confounding factors and the causal direction of this association. Methods/design: The research will comprise a systematic review and meta-analysis to address the evidence gap. EMBASE, MEDLINE, PSYCINFO, CINAHL, SocIndex, Scopus and Web of Science will be systematically searched for quantitative observational studies considering an association between loneliness/social isolation and key health-related behaviours in older adults. Two reviewers will independently check the study titles and abstracts for eligibility. Included studies will be critically appraised using Newcastle-Ottawa Scale by the lead author and checked by the second reviewer. Discrepancies in eligibility or quality assessment will be resolved via discussion or referral to a third reviewer. Results will be synthesised and reported in accordance with the Centre for Reviews and Dissemination (CRD) guidelines. This will be in the form of a descriptive summary, risk of bias assessment together with a meta-analysis and sub-group analyses (for covariate adjusted results) where sufficient heterogeneity of results is established. Finally, any associations identified will be analysed using the Bradford-Hill criteria to explore causal relationships which, if they exist, will be reported by means of a computed causations score. Discussion: This review aims to assess the extent and causal nature of associations between loneliness/social isolation and health-related behaviours among older adults. This data will provide a comprehensive overview of the quality of the evidence base to inform stakeholders in tackling the growing public health challenges arising from loneliness/social isolation in ageing populations. Systematic review registration: PROSPERO CRD42017020845