Table 2 Study details for the eight included papers from Eccles and Qualter [17] and one from Ma et al. [11].

Conoley and Garber (1985), USA [12]

Randomised control trial

Undergraduate psychology students with moderate depression

57 (0, 0%).

No power calculations are reported.

(n.s., n.s.)

Revised UCLA Loneliness Scale (20-item), Beck Depression Inventory, Causal Dimension Scale (controllability subscale).

Pre, Post, Follow-up (2 wk)

Counselor Rating Form. Post.

Reframing group: interviewer focused on ways to experience loneliness more positively; self-control group: Interviewer encouraged participants to overcome loneliness. Both groups: 2 × 30 min weekly interviews over 2 weeks.

Individual non-tech

Randomisation to one of two interventions (reframing or self-control) or control condition: wait list.

Results indicated that participants in the reframing group experienced a more significant reduction in depression than those in the self-control or control groups. All participants became less lonely over time, but no treatment was more effective than another in reducing loneliness. No differences were found for controllability. No effect sizes are reported.

Kneer et al. (2019), Netherlands [57]

Intervention study with control condition (not randomised)

Adolescent refugees

57 (22., 38.3%)

14: peers (refugees)

16: buddies (peer coaches)

12: control peers

15: control buddies.

No power calculations are reported.

13–18 (14.09, 1.65)

Peers - questionnaire: leisure activities; media usage; school connectedness (3 items); school motivation (8 items); peer norms for school performance (3 items); self-esteem (10 items—1 dismissed); life satisfaction (7 items) emotional support (4 items); social distraction (3 items); affectionate support (3 items) social anxiety (9 items); peer loneliness (16 items: [81]).

Buddies—semi-structured media diaries; weekly.

Pre and Post.

Peer coaching ‘Peer2Peer’ to help refugee adolescents (peers) orient to new culture and build friendships. Buddies received 3 days (4 h) training. Peers met with buddy a minimum of 1 × monthly. Online communication almost daily.

(14 weeks)

Individual, non-tech/ tech

Control condition: peers and buddies received no intervention/ training

Neither buddies nor peers nor. control peers showed any significant differences between pre and post.

Only control buddies showed a higher school connectedness post.

Social media played a central role in keeping in contact and also initiating contact, micro-coordination, and ‘hanging out’.

Larsen et al. (2019), Norway [16]

Randomised control trial

Adolescents enrolled in upper secondary schools

1937 (n.s., n.s.)

670: Intervention 1

809: Intervention 2

458: Control group

(Sample at baseline 2254 (1206, 53.5%)).

Power calculations reported in the study protocol [71] indicated that a sample of 975 students and 49 classes was needed to detect a small effect size of .25. It is not reported how many classes participated but authors report lack of statistical power due to low number of participating schools.

Baseline: 15–19 (16.82, n.s.)

Short form Symptom Check List (5-item); Norway Loneliness Scale (6-item: [82]); Perceived Family Wealth.

Baseline and follow-up (7 mo.).

Whole-school programme ‘Dream School Program’ (DSP): aims to enhance the psychosocial environment to reduce loneliness and mental health problems. (Intervention 1 ‘single tier’);

DSP + Mental Health Support Team (MST) – counsellors and nurses etc target specific students with known mental health problems/other issues (Intervention 2 ‘multi-tier’)

(1 × class and poster—semester 1; 1 × class - semester 2)

Group, non-tech

Randomisation to one of two interventions (DSP or DSP + MST) or control condition: Education as usual

No effect of the intervention on students’ mental health problems and loneliness in either of the intervention groups.

An overall increase in mental health problems and loneliness was found in all groups at follow-up. Compared to girls in the control group, girls in the multi-tier group had a significantly smaller increase in mental health problems. No effect size is reported.

Masia‐ Warner et al. (2005), USA [13]

Randomised control trial

Adolescents with social anxiety disorder (12% at baseline had a subtype with main concerns around performance situations and public speaking in class)

35 (9, 25.7%)

18: intervention (4, 22.2%)

17: control group (5, 29.4%).

No power calculations are reported.

13–17 (14.8, 0.81)

(Intervention (15, 0.59); Control (14.5 (0.94))

Independent rated: Anxiety Disorders Interview Schedule for DSM IV: Parent and Child Versions; Liebowitz Social Anxiety Scale for Children and Adolescents; Social Phobic Disorders Severity and Change Form; Children’s Global Assessment Scale.

Self-rated: Social Phobia and Anxiety Inventory for Children; Social Anxiety Scale for Adolescents; Children’s Depression Inventory; Loneliness Scale (16-item: [83, 84]).

Parent rated: Social Anxiety Scale for Adolescents: Parent Version Pre and post.

Independent raters—pre, post, follow up (9 mo.).

Skills for Academic and Social Success ‘SASS’: Social skills

12 × 40 min weekly group, 2 × 15 min individual, 2 × group booster sessions, 4 × 90 min weekend social events; Parents—2 group meetings (90 min)

(3 months).

Group/ individual, non-tech

Control condition: waitlist

Adolescents in the intervention group demonstrated significantly greater reductions than controls in social anxiety (d = 0.68–2.0 depending on measure), and significantly improved overall functioning (d = 2.3. Post - 67% intervention sample compared to 6% of waitlist sample, no longer met criteria for social phobia. No significant effect for treatment found for self-rated loneliness. Follow up (9 mo.) of intervention involved the sample only (n = 18) - clinical gains maintained.

Mason et al. (2016), USA [45]

Randomised control trial

Adolescents deemed to be at-risk of alcohol and marijuana use presenting at primary care clinics

117 (n.s., n.s.)

57: Intervention

60: Control group

(Sample at baseline 119 (35, 29.41%): 59: Intervention (15, 25.42%); 60: Control group (20, 33.33%))

No power calculations reported.

Baseline: 14–18 (16.4, 1.2)

(Intervention (16.4, 1.23); Control (16.2, 1.35))

Center for Disease Control Youth Risk Behaviour Survey—substance misuse (3 items); Social Stress Scale of the Behavior Assessment System for Children-Second Edition (BASC-2) (2-item: [85])—used to measure loneliness and perceived isolation (“I am lonely” and “People act as if they don’t hear me”, both encoded as 0 = false, 1 = true); Depression Scale of the BASC-2 (2-item).

Baseline, follow-up 1 (1 mo.), follow‐up 2 (3 mo.), follow‐up 3 (6 mo.).

Substance use intervention ‘Peer Network Counselling’ (PNC) based on motivational interviewing principles to reduce social stress (1 × 20 min)

Individual, non-tech

Control condition: Participants reviewed an informational handout with a therapist covering topics related to health behaviours (e.g. exercise, nutrition/ weight management, and life skills). (1 × 20 min)

At 6 mo. follow-up PNC significantly decreased social stress (loneliness/ perceived social isolation) compared to controls, whose social stress increased (small effect size η_p2 = 0.05). PNC temporarily moderated the effect of alcohol use, but not marijuana or heavy alcohol use.

Matthews et al. (2018), USA [53]

Randomised control trial

Adolescents with Autism Spectrum Disorder (ASD)

34 (28, 35%)

10: Intervention 1 (8, 80%)

12: Intervention 2 (10, 83.33%)

12: Control group (10, 83.33%)

Authors do not report power calculations but state that “the sample size was small resulting in inadequate statistical power to detect medium and small between group differences.

However, the sample size reflects the preliminary nature of a pilot study.”

13–17 (n.s.)

(Intervention 1: 15.1, 1.29;

Intervention 2: 15.17, 1.27;

Control: 15.42, 1.08)

Parent: Social Responsiveness Scale second edition; Social Skills Improvement System; Quality of Socialisation Questionnaire – parent version. Adolescents: Test of Adolescent Social Skills; Social Interaction Anxiety Scale; Revised UCLA Loneliness Scale (20-item: [79]); Quality of Socialization Questionnaire adolescent version; Social Distance Scale; Autism Knowledge Questionnaire Pre, post, follow-up (4 mo.).

Manualised parent-assisted psychoeducational social skills intervention ‘PEERS’ teaches social and friendship skills to intellectually able adolescents with ASD using didactic lessons, role-plays and behavioural rehearsal. Parents are trained to coach their adolescents in weekly assignments.

Intervention 1: traditional PEERS curriculum;

Intervention 2: peer (typically developing adolescents) mediated PEERS curriculum ‘PwP’. (14 weeks).

Group, non-tech

Randomisation to one of two interventions or control condition: delayed traditional PEERS treatment condition (DTC) i.e. waiting list control.

Both interventions showed significantly greater improvements in social skills knowledge to DTC (d = 3.97–4.38), which was maintained at follow-up. PwP demonstrated significant increases in parent-reported social skills relative to DTC (d = 1.48) maintained at follow-up. Both interventions showed marginally significant improvement in loneliness (non-parametric) compared to DTC. With one extreme outlier in each treatment group excluded, significant reduction in loneliness in both intervention groups relative to DTC at post (d = 1.06–1.43) and maintained at follow-up.

Rohde et al. (2004), USA^a [44]

Randomised control trial

Incarcerated young males

109 (109, 100%)

39: Intervention

25.: control group 1

45: control group 2 (Sample at baseline 138)

Authors report that “with α set at .05, two-tailed, we had adequate power (>0.80) to detect medium effect sizes or larger (Cohen f > 0.33) for intervention main effects.”

Baseline: 12–22 (n.s.) (intervention and control group 1: (16.3, 1.9); control group 2: (16.8, 1.7))

Youth Self-Report; Current Suicidal ideation and Lifetime Suicidal Attempts; Life Attitudes Schedule - Short Form (24-item); Coping Skills (15-item); Rosenberg Self-Esteem Scale; UCLA Loneliness Scale (8-item: [86]); Subjective Probability Questionnaire (5-item); Social Adjustment Scale-Self Report for Youth (4-item) Knowledge of CBT Concepts (10 short answer items).

Pre and post.

CBT Coping Course (modified version of the Adolescent Coping with Depression Course, Clarke et al. (1990))

Biweekly meetings, 16 × 90 min sessions over 8 weeks

Group, non-tech

Control condition: usual care

Significant improvement in treatment group compared to controls at post-treatment in seven of the 17 examined measures: Youth Self-Report externalising scores (η² = 0.07), three measures from the Life Attitudes Scale (total score η² = 0.11, death-related η² = 0.09, self-related η² = 0.11), self-esteem (η² = 0.10), one measure of social adjustment (η² = 0.07), and CBT knowledge (η² = 0.13). All reported effect sizes are small. No significant impact was found on any of the other measures including impact on loneliness.

Smith et al. (2017), Canada [43]

Single-group intervention study

Gay and bisexual young men

19 (19,100%) (Sample at baseline = 33 (33, 100%)).

No power calculations are reported but the stated aim of the study was to assess feasibility and provide preliminary evidence of its efficacy so this should be considered a pilot study.

Baseline: 18–25 (21.91, 2.15)

Center for Epidemiological Studies – Depression; State-Trait Inventory for Cognitive and Somatic Anxiety—state version; UCLA Loneliness Scale (20-item: [80]); Rosenberg Self-Esteem Scale; Nungesser Homosexuality Attitudes Inventory Revised; Sexual Orientation Concealment Scale; Lesbian, Gay, and Bisexual Identity Scale; Condomless sex and number of sex partners (4 categories); Alcohol and drug use frequency (8 substances). Pre, post, follow-up (3mo).

Small group intervention aimed at reducing negative mental and behavioural health outcomes resulting from minority stress. ‘Project PRIDE’ (Promoting Resilience In Discriminatory Environments). Bi-weekly sessions 8 × 2.5 h (4 weeks)

Group, non-tech

N/A

Comparing pre- to post-treatment: small, significant effect sizes for reduced loneliness (d = −0.36 post, -0.35 follow-up), increased self-esteem (d = 0.27 post, 0.88 at follow-up, reduced internalised homonegativity d = −0.21 post, −0.28 follow-up, and reduced difficult process related to sexual orientation identity (d = −0.30 post, −0.29 follow-up. Most changes remained stable or increased to follow-up; self-esteem continued to significantly increase.

Stice et al. (2010), USA^b [42]

Randomised control trial

At‐risk adolescents with elevated depressive symptoms

341 (150, 44%)

89: Intervention 1

88: Intervention 2

80: Intervention 3

84: Control group.

No power calculations reported.

14–19 (15.6, 1.2)

Beck Depression Inventory; Automatic Thoughts Questionnaire; Pleasant Events Schedule; Emotional Expression (9-item);

UCLA Loneliness Scale (8-item: [86]).

Pre and post.

Cognitive Behavioural (CB) Group (Intervention 1) and Supportive Expressive (SE) Intervention Group (Intervention 2)—6 ×1 h weekly sessions. Bibliotherapy (Intervention 3) ‘Feeling Good’ by Burns, D. (1980). Controls given NIMH brochure describing major depression and recommends treatment for depressed youth – ‘Let’s Talk About Depression’, NIH and information about treatment options.

Group (interventions 1 and 2), non-tech; Individual (intervention 3), non-tech

Randomisation to one of two intervention group conditions, individual bibliotherapy, or control condition: assessment and brochure

Compared to controls both the CB (r_p = −0.32) and SE (r_p = −0.17) groups showed significant reductions in depressive symptoms, whereas there was no significant difference for the Bibliographic group. Only the CB group showed a significantly greater decrease in loneliness relative to controls (no effect size reported). CB also showed a significant decrease in negative cognitions (r_p = −0.3) and increase in pleasant activities (r_p = 21) compared to controls. Compared to controls the SE group showed no significant difference in change in loneliness but did show a significant difference in change in emotional expression (r_p = −0.3). Change in loneliness (r_p = −0.32) but not emotional expression, predicted change in depressive symptoms for the SE condition.