Social connection interventions and depression in young adults: a systematic review and meta-analysis

Mental disorders are a leading cause of disease burden in young people, with three quarters of mental health problems emerging before age 25. Depression has substantial health and social impacts. Loneliness—defined as a discrepancy between desired and actual social relationships—is strongly associated with increased risk of depression and has been rising among young adults, including during the COVID-19 pandemic. Early adulthood involves significant developmental and societal changes (mobility, technology) that may heighten vulnerability to loneliness and depression. Given the link between loneliness and depression and evidence that social capital protects mental health, social connection interventions that increase quantity/quality of social relationships or reduce loneliness may help prevent depression in young adults. Existing intervention literature has focused primarily on children/adolescents, older adults, or clinical samples, revealing a gap for non-clinical young adults. This review aims to (1) identify interventions addressing social connection/loneliness in young adults, and (2) evaluate their effectiveness on depression and social connection/loneliness outcomes.

Prior research documents a robust association between loneliness and depression in young adults, with loneliness predicting outcomes such as antidepressant use. Cross-temporal analyses indicate increases in loneliness since the 1970s. During the COVID-19 pandemic, young adults reported greater loneliness and depression than other age groups, with increases in loneliness accounting for much of the rise in depression in some settings. Social capital is viewed as protective against depression, yet most social connection intervention studies target children/adolescents, older adults, or clinical populations. There is limited evidence for general-population young adults, who face unique developmental challenges and societal changes affecting relationships. The review addresses this evidence gap by synthesizing interventions that explicitly aim to improve social connection or reduce loneliness and assessing their impact on depression and social outcomes.

Design: Systematic review and meta-analysis following PRISMA; protocol preregistered on PROSPERO (CRD42023395595). Data sources: PubMed, Scopus, and PsycInfo searched for English-language studies from 01/01/2000 to 01/01/2023 using terms related to young adults, social interventions/connectedness, depression, and randomized controlled trials. Inclusion criteria: Peer-reviewed quantitative studies (RCTs, quasi-experiments, pre-post with comparison) including young adults aged 18–24 from the general population; interventions explicitly aiming to increase social connections or reduce loneliness (in-person or online); inclusion of a comparison group (no intervention, waitlist, or active comparator); depression measured pre- and post-intervention using validated tools; and at least one validated measure of social connection and/or loneliness. Exclusion criteria: Qualitative studies, reviews, commentaries, editorials, book chapters; clinical patient populations recruited from clinical settings; interventions that exclusively used psychological approaches (CBT, mindfulness, ACT) without explicit social connection focus; social media platforms not explicitly designed to increase social connections/reduce loneliness in this context. Screening and selection: 2,197 records identified; 381 duplicates removed; 1,816 titles/abstracts screened; 59 full texts assessed; 5 studies met criteria; 1 additional study identified via author contact; total included: 6. Screening conducted by one reviewer with 10% full-text double-screened; 100% agreement. Data extraction: Conducted independently in duplicate using a predefined Excel form; discrepancies resolved by a third author. Extracted items included study design, location, sample characteristics, intervention details (delivery modality, duration), measures (depression, social connection/loneliness), outcomes, and funding. Risk of bias: RCTs assessed with Cochrane RoB 2.0; non-randomized studies assessed with ROBINS-I; dual independent assessment with consensus procedures. Synthesis and analysis: Narrative synthesis by delivery modality (in-person vs online) and sample type (higher education students vs other). Random-effects meta-analyses using standardized mean differences (SMD) to accommodate varying measures; heterogeneity assessed with I²; analyses run in STATA/MP 18.0 (metan). Funnel plots visually inspected for publication bias; Egger’s test not performed (<10 studies). Sensitivity analyses excluded studies with serious risk of bias.

Study set: Six studies (3 USA, 1 Canada, 1 Australia, 1 South Korea); sample sizes 23–438; mean ages ~18.7–21.6 years. Five studies recruited higher education students; four were in-person interventions, one online app-based; one study targeted street-involved youth. Four interventions were group-based; durations typically 4–10 weeks (weekly 60–90 minutes), plus one self-directed app and one single-session reading/reflection. Controls: Five waitlist/no intervention, one active comparator. Measures: Depression (PHQ-9, BDI, DASS-21, CES-D); Loneliness measured in four studies; various social connection indices across all. Risk of bias: All RCTs had some concerns (e.g., lack of blinding). Non-RCTs: one moderate, two serious risk of bias (confounding, deviations, missing data, measurement issues). Meta-analysis—Depression: Overall (all studies) SMD = −0.22 (95% CI −0.35 to −0.08; p=0.002); I²=0.0%, p=0.784. Subgroups: In-person higher education interventions SMD = −0.19 (95% CI −0.34 to −0.03; p=0.020), carrying most weight (73.62%); Online app (Nod) SMD = −0.27 (95% CI −0.54 to −0.002; p=0.048); Street-involved youth SMD = −0.64 (95% CI −1.71 to 0.42; p=0.235). Sensitivity (excluding serious risk-of-bias studies): overall SMD = −0.19 (95% CI −0.33 to −0.05; p=0.008); in-person higher education after exclusion SMD = −0.16 (95% CI −0.33 to 0.01; p=0.058). Publication bias: Funnel plot did not suggest bias. Meta-analysis—Loneliness: Pooled across three studies SMD = −0.10 (95% CI −0.24 to 0.05; p=0.188); I²=0.0%, p=0.487; sensitivity not conducted (none rated serious risk of bias). Social connection outcomes (narrative): Group-based interventions showed improvements in belongingness, social functioning, interpersonal relationships, and social support (some effects stronger in minoritized or lower-SES groups). The app-based intervention showed no overall improvements in social connection, though interaction analyses suggested buffering effects for participants with higher baseline loneliness or depression. The street-involved youth program improved social connectedness but did not reduce depression.

Findings indicate that social connection interventions can reduce depressive symptoms among young adults overall, despite mixed effects across individual studies. Diverse social connection outcomes generally improved, though pooled loneliness reductions were not statistically significant. The review underscores the potential of group-based interventions in higher education settings and suggests that individuals with higher baseline loneliness/depression may derive preventive benefits, particularly from app-based approaches. Effects were often stronger among minoritized racial/ethnic groups and lower-SES students, highlighting the importance of targeting at-risk subgroups. The evidence base is limited in scope, with few studies outside high-income university contexts and challenges engaging young adults with complex needs (e.g., street-involved youth). No study assessed whether depression improvements were mediated by changes in social connection/loneliness, leaving mechanisms unclear. These results address the research question by demonstrating preliminary efficacy for reducing depression via social connection-focused approaches, while emphasizing the need to expand interventions beyond higher education and to rigorously test mechanisms, modalities (online vs in-person; individual vs group), and long-term outcomes.

Social connection interventions show potential for improving social connectedness and reducing depressive symptoms in young adults, with small overall effects observed primarily in higher education settings. The current evidence base is limited in size and scope. Future research should recruit larger and more diverse samples, extend follow-up durations, compare delivery modalities (online, in-person, self-directed, group-based), tailor interventions to at-risk subgroups (e.g., minoritized, lower-SES, socially isolated), and develop/assess interventions for young adults in workplaces and those not in education, employment, or training.

Evidence limitations include a small number of studies predominantly from high-income higher education settings, variable risk of bias (including serious risks in some non-randomized studies), inability to blind participants, low engagement in the app-based intervention, and challenges engaging high-risk populations (e.g., street-involved youth). Meta-analysis power was constrained, and heterogeneity estimates may be biased in small meta-analyses. The review excluded grey literature and qualitative studies, potentially missing ongoing or unpublished interventions and limiting insights into acceptability and process mechanisms. Many primary studies had modest sample sizes and limited follow-up.