Addressing loneliness in emerging adults in primary care: a pilot feasibility study

Loneliness, a subjective perception of inadequate social relationships, is increasingly prevalent in the U.S., with emerging adults showing higher rates than other age groups. This developmental period involves transitions (moving, education, living arrangements) that can disrupt social networks, and unaddressed loneliness is linked to hypertension, anxiety, depression, substance use, poor sleep, and longer-term mental illness. Primary care, as the first point of contact in the health system, is well positioned to detect and treat loneliness, given frequent visits by lonely patients. Despite evidence supporting cognitive behavioral therapy (CBT) and promising social prescribing (SP) approaches, these interventions have rarely been adapted or tested for emerging adults in U.S. primary care. The study’s objective was to pilot the feasibility, acceptability, and preliminary impact of adapted group-based CBT and SP interventions to reduce loneliness among emerging adults in primary care.

Meta-analyses indicate CBT can reduce loneliness, with most studies in older adults and variable acceptability/adherence across settings. SP has been implemented in European outpatient primary care to strengthen social networks and engagement in purposeful activities, though the most effective components remain unclear. SP has limited study among emerging adults and in the U.S. Interventions for young adults often target at-risk subgroups and address loneliness as a secondary aim, and neither CBT nor SP loneliness interventions for emerging adults have been widely evaluated in primary care. Prior work suggests primary care patients with loneliness frequently attend physician visits, highlighting an opportunity for early intervention. Evidence for CBT shows small-to-medium pooled effects on loneliness across ages, with recent youth-focused trials showing larger effects; SP shows promising results for depression and loneliness in young adults, but effectiveness and active ingredients require further research.

Design: Pre-post study assessing differences in loneliness and mental health outcomes before and after interventions, with assignment to CBT or SP groups. IRB approval: University of Washington (STUDY00018180). Recruitment and sample: Emerging adults aged 18–25, English-speaking, meeting UCLA 3-item loneliness cutoff (≥6), at least one primary care visit (Jul 2022–Oct 2023) in an academically affiliated urban health system (16 primary care practices), and willing to participate virtually. Excluded if currently receiving CBT. As a pilot feasibility trial, not powered for between-group differences. Procedures: Patients identified via EHR, contacted via text to complete an eligibility REDCap survey (UCLA-3 and demographics). Incentives: $15 for screening survey, $25 baseline, $50 post-intervention. Eligible patients provided e-consent and baseline assessments, then were sequentially assigned to CBT or SP based on group slot availability to balance groups. Two CBT and two SP cohorts ran Fall 2023–Winter 2024. Interventions: CBT—5 weekly group sessions (2–5 participants), ~75 minutes via videoconference, delivered by two trained master’s-level psychotherapists. Adapted from CBT protocols for loneliness and youth depression; core components: psychoeducation (behaviors linked to loneliness), values and goal setting (develop behaviors facilitating meaningful connections), behavioral activation and coping (overcome avoidance/anxiety), cognitive reframing (address maladaptive thoughts), relapse prevention (maintenance plans). SP—5 weekly group sessions (4–5 participants), ~60 minutes via videoconference, delivered by two trained facilitators (not licensed psychotherapists). Adapted from UK primary care SP practices for emerging adults; core components: social world mapping, creating connection plans (connect/reconnect/explore/join), measuring success and self-relationship, overcoming barriers, sustaining plans. Assessments: Baseline (pre-assignment) and post-intervention surveys within two weeks of completion, including UCLA Loneliness Scale-20 (version 3), PHQ-9, GAD-7. Engagement and adherence: After each session, facilitators recorded attendance (yes/no) and rated effort during activities and engagement with take-home practice on 1–5 scales; summed across sessions to yield total scores (5–25). Qualitative interviews: All participants offered post-intervention interviews; 10 completed semi-structured interviews (30–45 minutes; $50 incentive) exploring loneliness experiences, barriers/facilitators, suggested adaptations, and behavioral changes. Interviews were recorded, de-identified, transcribed. Analysis: Descriptive statistics for demographics and adherence; paired t-tests conducted separately for CBT and SP to assess pre-post differences; no between-group comparisons. Qualitative content analysis using Dedoose (v9.0.46): codebook development, double-coding for reliability, independent coding with consensus meetings; thematic synthesis.

Sample and flow: From 2,199 EHR-identified patients, 243 (11%) completed screening; 112 (46% of screened) were eligible and interested; 26 consented; 20 completed baseline and were allocated (CBT n=10, SP n=10). Final analytic sample N=15 (CBT n=6, SP n=9); overall attrition 25% (5 excluded: 3 CBT, 1 SP did not start; 1 CBT no post-survey). Demographics: Mean age 22 (SD=2.4); majority female (87%) and identified as women (73%); racially and sexually diverse. Adherence and engagement: 14/15 (93%) completed ≥3 of 5 sessions; 6 (40%) completed all five (3 CBT, 3 SP). Facilitator-rated engagement (out of 25): CBT group activities mean 19.7; CBT take-home practice 19.3; SP group activities 16.6; SP take-home practice 17. Quantitative outcomes (paired t-tests): Both groups showed non-significant pre-post reductions. CBT: UCLA-20 mean decreased 38.0 (14.4) to 24.8 (14.8), difference 13.2, p=0.15, Cohen’s d=−1.09; PHQ-9 decreased 12.7 (5.2) to 8.2 (3.3), difference 4.5, p=0.11, d=−0.10; GAD-7 decreased 14.5 (3.8) to 10.3 (4.5), difference 4.2, p=0.11, d=−1.08. SP: UCLA-20 decreased 27.6 (8.4) to 24.0 (8.5), difference 3.6, p=0.39, d=−0.35; PHQ-9 decreased 12.4 (6.4) to 8.7 (3.5), difference 3.7, p=0.15, d=−0.60; GAD-7 decreased 8.2 (5.2) to 6.6 (2.7), difference 1.6, p=0.41, d=−0.32. Qualitative themes (n=10 interviews; 4 CBT, 6 SP): 1) Experience of loneliness tied to transitions and dynamic factors (mental health, relationships, friends). 2) Changes in self/behavior: normalization via group, increased self-compassion, perspective change, greater confidence and control; CBT-specific gains in emotion exploration, cognitive reframing, relaxation techniques; SP-specific gains in goal-setting, accountability, social network mapping, practicing social engagement. 3) Barriers/facilitators: mixed views on virtual vs. in-person (CBT appreciated accessibility; SP preferred hybrid); facilitators included compassionate, engaging leaders and applicable materials. 4) Suggested adaptations: longer sessions, larger cohorts to mitigate absences, lighter use of worksheets (CBT), better utilization of online communication tools (SP).

Findings indicate feasibility and acceptability of adapted group-based CBT and SP interventions for loneliness among emerging adults in primary care: substantial interest at screening, high completion among enrolled participants, and favorable engagement. Despite non-significant statistics due to small sample, pre-post improvements across loneliness, depression, and anxiety suggest potential clinical relevance, particularly large effect sizes in CBT for loneliness and anxiety and in SP for depression. Recruitment challenges—low conversion from eligibility to enrollment—may reflect reliance on EHR-based outreach lacking relational engagement valued by emerging adults; multi-pronged recruitment (clinician referrals, community outreach) may improve uptake. Delivery mode preferences varied, highlighting the need to optimize virtual, in-person, or hybrid formats to balance accessibility with social fulfillment. Results align with emerging evidence that CBT can benefit youth loneliness/depression and that SP may reduce depression/loneliness, warranting rigorous testing in primary care. Group processes and mechanisms (cognitive/behavioral change in CBT; increased social opportunities in SP) may be key ingredients; tailoring interventions to heterogeneous loneliness experiences in emerging adults could enhance effectiveness.

This pilot demonstrates that adapted CBT and SP group interventions are feasible and acceptable for addressing loneliness in emerging adults within primary care, with signals of improvement across loneliness, depression, and anxiety. The work informs intervention content, delivery, and engagement strategies for this population and setting. Future research should employ adequately powered pragmatic randomized controlled trials with control groups, longitudinal follow-up, and analyses of mediators/moderators and active ingredients, alongside optimized recruitment (multi-pronged strategies) and delivery modes (virtual, in-person, hybrid) to establish effectiveness and scalability in primary care.

Small convenience sample from an urban health system limits generalizability and statistical power. Absence of a control group prevents attribution of changes to the interventions; sequential assignment introduces potential allocation and temporal biases. Attendance variability and engagement differences may have influenced outcomes but were not analyzed due to sample size. Reliance on EHR-based recruitment may have reduced enrollment effectiveness compared to relational approaches.