RESUMO
BACKGROUND: Loneliness and social isolation are increasingly recognised as prevalent among people with mental health problems, and as potential targets for interventions to improve quality of life and outcomes, as well as for preventive strategies. Understanding the relationship between quality and quantity of social relationships and a range of mental health conditions is a helpful step towards development of such interventions. PURPOSE: Our aim was to give an overview of associations between constructs related to social relationships (including loneliness and social isolation) and diagnosed mental conditions and mental health symptoms, as reported in systematic reviews of observational studies. METHODS: For this umbrella review (systematic review of systematic reviews) we searched five databases (PsycINFO, MEDLINE, EMBASE, CINAHL, Web of Science) and relevant online resources (PROSPERO, Campbell Collaboration, Joanna Briggs Institute Evidence Synthesis Journal). We included systematic reviews of studies of associations between constructs related to social relationships and mental health diagnoses or psychiatric symptom severity, in clinical or general population samples. We also included reviews of general population studies investigating the relationship between loneliness and risk of onset of mental health problems. RESULTS: We identified 53 relevant systematic reviews, including them in a narrative synthesis. We found evidence regarding associations between (i) loneliness, social isolation, social support, social network size and composition, and individual-level social capital and (ii) diagnoses of mental health conditions and severity of various mental health symptoms. Depression (including post-natal) and psychosis were most often reported on, with few systematic reviews on eating disorders or post-traumatic stress disorder (PTSD), and only four related to anxiety. Social support was the most commonly included social construct. Our findings were limited by low quality of reviews and their inclusion of mainly cross-sectional evidence. CONCLUSION: Good quality evidence is needed on a wider range of social constructs, on conditions other than depression, and on longitudinal relationships between social constructs and mental health symptoms and conditions.