Effects of Universal Digital Mental Health Interventions for Children and Youth on Psychological Outcomes – A Systematic Review and Meta-Analysis

Effects of Universal Digital Mental Health Interventions for Children and Youth on Psychological Outcomes – A Systematic Review and Meta-Analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Effects of Universal Digital Mental Health Interventions for Children and Youth on Psychological Outcomes – A Systematic Review and Meta-Analysis Dr. Kaitlin Di Pierdomenico, Oana Bucsea, Haleh Hashemi, Arianna Leguia, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6811900/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Children and youth are facing rising mental health challenges in an increasingly digital world. Universal Tier 1 digital mental health interventions (DMHIs) – low-intensity, high-reach supports for non-clinical populations – have emerged as scalable tools to meet early-stage needs, but their effectiveness remains unclear. This systematic review and meta-analysis synthesized evidence from 51 studies (N = 30,474) to assess the impact of DMHIs across emotional, behavioural, social, and cognitive outcomes, and examined whether delivery format (hybrid vs. virtual) and outcome timing (post-intervention vs. follow-up) influenced results. Nine of 16 calculated effect sizes found statistically significant, small effects. Hybrid interventions – those incorporating any in-person component – showed more consistent and sustained benefits, with moderate-certainty evidence at follow-up. However, most findings were low certainty. Parents of younger children and marginalized children and youth were notably understudied. Future research must prioritize marginalized youth, broaden outcome evaluation, and deliver large-scale, high-quality trials to inform the inclusive and sustainable implementation of digital mental health supports. INPLASY registration number: INPLASY202490026 Biological sciences/Psychology Health sciences/Health care Figures Figure 1 Figure 2 Figure 3 Introduction Children and youth are facing unprecedented mental health challenges while simultaneously navigating an increasingly digital world [https://www.who.int/publications/i/item/9789240106796]. Digital technologies have become deeply integrated into their daily lives, shaping how they socialize, learn, and access information. This shift presents both risks and opportunities. While concerns persist about excessive screen time, physical and mental health implications, and online safety [1], digital platforms also offer promising avenues for expanding access to mental health resources, particularly for people who experience barriers to traditional care, such as long wait times, stigma, and geographic limitations. Digital mental health interventions (DMHIs) have emerged as scalable and flexible tools capable of delivering widespread, early-stage mental health support. These interventions – ranging from mobile apps and web-based programs to chatbots and virtual sessions – aim to provide timely, accessible care to children and youth before more intensive services are needed. The move toward digital intervention delivery accelerated during the COVID-19 pandemic, which disrupted in-person care and prompted the rapid expansion of e-mental health resources in many countries [https://mentalhealthcommission.ca/wp-content/uploads/2024/09/An-E-Mental-Health-Strategy-for-Canada-FINAL.pdf]. Despite their growing availability and adoption, critical gaps remain in the evidence base grounding the effectiveness, sustainability, and impact of DMHIs across core psychological domains (e.g., emotional, behavioural, social, cognitive). Many interventions have been implemented without rigorous evaluation, raising concerns about quality, safety, and long-term outcomes. Recent policy guidance – from the World Health Organization’s Strategic Mental Health Action Plans [https://www.who.int/publications/i/item/9789240106796] to national e-mental health strategies [https://mentalhealthcommission.ca/wp-content/uploads/2024/09/An-E-Mental-Health-Strategy-for-Canada-FINAL.pdf, https://www.mqhealth.org.au/__data/assets/pdf_file/0011/1010630/emstrat.pdf] – has emphasized the urgent need for digital interventions that are evidence-informed, accessible, equitable, and integrated within broader mental health systems. While previous reviews have mapped the landscape of DMHIs for children and youth [2, 3, 4, 5, 6, 7], they have typically focused on clinical samples, single-outcome measures, or short-term effects. This systematic review and meta-analysis addresses these gaps by evaluating the effectiveness of universal digital mental health interventions across four psychological domains and two time periods (immediate and follow-up), with the aim of informing more effective care delivery and guiding future policy development. To address these gaps, the current meta-analysis focuses specifically on universal (Tier 1) DMHIs – low-intensity, high-reach interventions designed for non-clinical child and youth populations. It builds directly on our prior scoping review (Di Pierdomenico et al., under review), which found that most interventions used hybrid formats (combining digital tools with in-person components), targeted multiple psychological outcomes, were commonly online psychoeducational programs, and often included independent child/youth-led components. The scoping review also identified several important gaps, including the underrepresentation of parents of younger children (particularly those aged 0-4), limited work with marginalized populations, and poor reporting of racial and cultural demographics. Building on the comprehensive understanding of what has currently been done in the area of universal DMHIs via the aforementioned scoping review, this meta-analysis will focus on synthesizing psychological outcomes and evaluating how both intervention format and assessment timing relate to overall impact. It has two primary aims: Understand the efficacy of universal DMHIs by synthesizing outcomes across four psychological outcome domains (emotional, behavioural, social, and cognitive), two intervention formats (hybrid, virtual), and two outcome assessment time periods (immediate post-intervention, follow-up). Assess the certainty of the evidence for each effect size using the GRADE framework [https://gdt.gradepro.org/app/handbook/handbook.html] to inform future research, policy, and practice. By systematically analyzing intervention effectiveness, this meta-analysis offers critical insights to guide the design and implementation of more inclusive, evidence-informed DMHIs. These findings can help ensure that effective mental health supports reach diverse child and youth populations in meaningful and sustainable ways. Results Study selection We identified 21,729 records through database searches, along with 7 additional records through snowball searching, all of which were categorized as database sources. After removing duplicates, 6,537 records remained for title and abstract screening. A total of 100 full-text articles were assessed for eligibility, and 49 were excluded – most commonly due to the presence of a mental health diagnosis in the comparison groups or being the wrong article type (see Supplementary Table 1). Ultimately, 51 studies were included for review and data extraction. A PRISMA flow diagram is presented in Fig. 1 . Study characteristics Our prior scoping review (Di Pierdomenico et al., under review) provided a comprehensive description of study characteristics (see Supplementary Materials [Table 2 in Di Pierdomenico et al., under review]) and informed the analysis plan for the current meta-analysis. In brief, most studies were conducted in high-income countries, published within the last three years, and focused on children aged 5 to 18, with limited reporting of racial or cultural demographics. Interventions were broadly aimed at mental health promotion, prevention, or a combination of both, and often targeted multiple psychological outcomes. Most used hybrid delivery models – typically in school settings – and relied on self-report measures. Formats ranged from online programs and apps to games and virtual communication tools. Intervention structures varied, with many using self-directed formats supported by some level of facilitation, most often by research staff or mental health professionals. Appraisal of methodological quality Study quality was assessed using two tools: the Cochrane Risk of Bias 2 (RoB 2) [ 8 ] for randomized controlled trials (RCTs) and the ROBINS-I tool (Risk of Bias in Non-randomized Studies of Interventions) [ 9 ]. Summary plots of the risk of bias assessments are provided in Fig. 2 and Fig. 3 . Of the 51 studies included in this meta-analysis, 35 (69%) were classified as RCTs and 16 (31%) as non-RCTs. Among the RCTs, 8 studies (22.9%) were rated as having an overall low risk of bias, 11 studies (31.4%) as having unclear risk, and 16 studies (45.7%) as having a high overall risk of bias. The most common sources of bias among RCTs were related to selection bias due to lack of allocation concealment (n = 1), performance bias due to lack of blinding of participants and personnel (n = 13), detection bias (n = 9), and attrition bias (n = 5). For the 16 non-randomized studies, 13 employed pre-post single-group designs, while 3 were non-randomized quasi-experimental designs with comparison groups [ 10 , 11 , 12 ]. To ensure consistency, both the Cochrane RoB 2 tool (via RevMan), and the ROBINS-I framework were used. The ROBINS-I assessments revealed that 14 studies (87.5%) presented with a serious overall risk of bias, and 2 studies (12.5%) were rated as having a moderate risk of bias. Serious risk of bias was observed across the seven ROBINS-I domains, including: confounding (n = 14), classification of interventions (n = 2), participant selection (n = 3), deviations from intended interventions (n = 1), outcome measurement (n = 2), and selection of the reported results (n = 1). The Cochrane RoB 2 evaluations of non-RCTs were consistent with the ROBINS-I findings, confirming poor methodological quality in the majority of non-randomized studies. Results We conducted 16 separate meta-analyses across 51 studies (30,474 participants; accounting for within-study duplicates) grouped by four psychological outcome domains (emotional, behavioural, social, and cognitive), two delivery formats (hybrid and virtual), and two assessment time periods (immediate post-intervention, within one month; and follow-up, one to six months post-intervention). Forest plots corresponding to each meta-analysis, along with associated quality appraisals informing the GRADE tables, are available in Supplementary Figs. 1–16. GRADE tables are presented by outcome domain, intervention format, and assessment time period. Standardized mean differences (SMDs) were calculated to accommodate variation in outcome measures across studies. In all analyses, negative SMD values indicate reductions in adverse psychological outcomes post-intervention, representing a positive effect of the intervention. To assess confidence in each effect estimate, we applied the GRADE approach [ https://training.cochrane.org/grade-approach ], evaluating the certainty of evidence based on risk of bias, inconsistency, indirectness, imprecision, and other considerations. 1. Emotional Outcomes Emotional outcomes, referring to how participants feel, were examined in four separate meta-analyses (See Table 1 ). Four effect sizes were calculated, each based on data from 5 to 21 studies. Three of the four effects were significant (small magnitude), with the certainty of evidence ranging from very low (three effect sizes) to moderate (one effect size). Table 1 GRADE Summary of Findings Table (Emotional) Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Emotional [placebo] Relative (95% CI) Emotion Outcome x Hybrid Format x Immediate Post-Intervention (< 1-month) Time Period 21 randomised trials (16/21) very serious a serious b not serious not serious none 3111/3111 (100.0%) 3156/3156 (100.0%) *SMD -0.18 (-0.26 to -0.11) ⨁◯◯◯ Very low a,b CRITICAL Emotion Outcome x Virtual Format x Immediate Post-Intervention (< 1-month) Time Period 15 randomised trials (10/15) very serious c very serious d not serious not serious none 1471/1471 (100.0%) 1948/1948 (100.0%) *SMD -0.24 (-0.47 to -0.01) ⨁◯◯◯ Very low c,d CRITICAL Emotion Outcome x Hybrid Format x Follow-Up (≥ 1-month to ≤ 6-months) Time Period 13 randomised trials (13/13) not serious e serious f not serious not serious none 2636/2636 (100.0%) 3285/3285 (100.0%) *SMD -0.19 (-0.31 to -0.07) ⨁⨁⨁◯ Moderate e,f CRITICAL Emotion Outcome x Virtual Format x Follow-Up (≥ 1-month to ≤ 6-months) Time Period 5 randomised trials (3/5) very serious g not serious h not serious serious i none 599/599 (100.0%) 707/707 (100.0%) SMD -0.05 (-0.16 to 0.06) ⨁◯◯◯ Very low g,h,i CRITICAL CI : confidence interval; SMD : standardized mean difference Explanations *Indicates significant effect size a. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (11/21 had high risk of bias overall, 7/21 had unclear risk of bias overall, and 3/21 had low risk of bias overall) because of potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors, and attrition. b. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 53%. c. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (12/15 had high risk of bias overall, 2/15 had unclear risk of bias overall, and 1/15 had low risk of bias overall) because of potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors, and attrition. d. We downgraded the certainty of evidence by two levels for inconsistency due to heterogeneity among studies, with an I² statistic of 84%. e. Most included studies had unclear risk of bias (6/13 had unclear risk of bias overall, 4/13 had low risk of bias overall, and 3/13 had high risk of bias overall) because of potential issues with methods, i.e., blinding of personnel and outcome assessors, and attrition. f. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 67%. g. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (3/5 had high risk of bias overall and 2/5 had unclear risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of participants and personnel, and attrition. h. I² statistic of 0%. We downgraded the certainty of evidence by one level for imprecision, as the confidence interval (-0.16 to 0.06) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate. 2. Behavioural Outcomes Behavioural outcomes, or how participants act, were examined in four separate meta-analyses (See Table 2 ). Four effect sizes were calculated, each based on data from 4 to 9 studies. Three of the four effects were significant (small magnitude), with the certainty of evidence ranging from very low (3 effect sizes) to moderate (1 effect size). Table 2 GRADE Summary of Findings Table (Behavioural) Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Behavioural [placebo] Relative (95% CI) Behaviour Outcome x Hybrid Format x Immediate Post-Intervention (< 1-month) Time Period 8 non-randomised studies (5/8) very serious a not serious b not serious not serious none 415/415 (100.0%) 510/510 (100.0%) *SMD − 0.15 (-0.27 to -0.03) ⨁◯◯◯ Very low a,b CRITICAL Behaviour Outcome x Virtual Format x Immediate Post-Intervention (< 1-month) Time Period 9 non-randomised studies (6/9) very serious c serious d not serious not serious none 954/954 (100.0%) 1018/1018 (100.0%) *SMD − 0.29 (-0.52 to -0.07) ⨁◯◯◯ Very low c,d CRITICAL Behaviour Outcome x Hybrid Format x Follow-Up (≥ 1-month to ≤ 6-months) Time Period 4 randomised trials (4/4) serious e not serious f not serious not serious none 375/375 (100.0%) 451/451 (100.0%) *SMD − 0.16 (-0.29 to -0.03) ⨁⨁⨁◯ Moderate e,f CRITICAL Behaviour Outcome x Virtual Format x Follow-Up (≥ 1-month to ≤ 6-months) Time Period 4 non-randomised studies (3/4) very serious g serious h not serious very serious i none 166/166 (100.0%) 180/180 (100.0%) SMD -0.40 (-0.84 to 0.03) ⨁◯◯◯ Very low g,h,i CRITICAL CI : confidence interval; SMD : standardised mean difference Explanations *Indicates significant effect size a. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (7/8 had high risk of bias overall and 1/8 had low risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition. b. I² statistic of 0%. c. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (7/9 had high risk of bias overall, 1/9 had unclear risk of bias overall, and 1/9 had low risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition. d. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 72%. e. We downgraded the certainty of evidence by one level for serious study limitations as most included studies had high or unclear risk of bias (2/4 had high risk of bias overall, 1/4 had unclear risk of bias overall, and 1/4 had low risk of bias overall) because of potential issues with methods, i.e., blinding of personnel and outcome assessors. Approximately 34% of the overall weight came from studies with high risk of bias, and over 50% came from a study with unclear risk. Only a small portion (14.5%) was contributed by a study at low risk of bias. f. I² statistic of 0%. g. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (3/4 had high risk of bias overall and 1/4 had unclear risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, attrition, and reporting bias. h. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 70%. We downgraded the certainty of evidence by two levels for imprecision, as the confidence interval (–0.84 to 0.03) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate. Additionally, two of the four included studies had small experimental sample sizes (< 35 participants), which may further reduce the overall precision of the pooled result. 3. Social Outcomes Social outcomes, referring to how participants connect with others, were examined in four separate meta-analyses (See Table 3 ). Four effect sizes were calculated, each based on data from 3 to 8 studies. Two of the four effects were significant, with the certainty of evidence ranging from very low (3 effect sizes) to moderate (1 effect size). Table 3 GRADE Summary of Findings Table (Social) Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Social [placebo] Relative (95% CI) Social Outcome x Hybrid Format x Immediate Post-Intervention (< 1-month) Time Period 8 randomised trials (5/8) very serious a serious b not serious not serious none 1416/1416 (100.0%) 1284/1284 (100.0%) *SMD − 0.15 (-0.27 to -0.03) ⨁◯◯◯ Very low a,b CRITICAL Social Outcome x Virtual Format x Immediate Post-Intervention (< 1-month) Time Period 5 randomised trials (3/5) very serious c very serious d not serious serious e none 534/534 (100.0%) 506/506 (100.0%) SMD -0.07 (-0.48 to 0.34) ⨁◯◯◯ Very low c,d,e CRITICAL Social Outcome x Hybrid Format x Follow-Up (≥ 1-month to ≤ 6-months) Time Period 3 randomised trials (3/3) serious f not serious g not serious not serious none 1046/1046 (100.0%) 910/910 (100.0%) *SMD − 0.11 (-0.20 to -0.02) ⨁⨁⨁◯ Moderate f,g CRITICAL Social Outcome x Virtual Format x Follow-Up (≥ 1-month to ≤ 6-months) Time Period 4 randomised trials (3/4) very serious h not serious i not serious serious j none 532/532 (100.0%) 505/505 (100.0%) SMD -0.08 (-0.21 to 0.05) ⨁◯◯◯ Very low h,i,j CRITICAL CI : confidence interval; SMD : standardised mean difference Explanations *Indicates significant effect size a. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (5/8 had high risk of bias overall and 3/8 had unclear risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors. b. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 67%. c. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (4/5 had high risk of bias overall and 1/5 had low risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition. d. We downgraded the certainty of evidence by two levels for inconsistency due to heterogeneity among studies, with an I² statistic of 89%. e. We downgraded the certainty of evidence by one level for imprecision, as the confidence interval (–0.48 to 0.34) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate. f. We downgraded the certainty of evidence by one level for serious study limitations as most included studies had unclear risk of bias (2/3 had unclear risk of bias overall and 1/3 had low risk of bias overall) because of potential issues with methods, i.e. blinding of personnel and outcome assessors, and attrition. g. I² statistic of 0%. h. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (2/4 had high risk of bias overall and 2/4 had unclear risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition. I² statistic of 11%. j. We downgraded the certainty of evidence by one level for imprecision, as the confidence interval (–0.21 to 0.05) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate. 4. Cognitive Outcomes Cognitive outcomes, capturing how participants think or solve problems, were examined in four separate meta-analyses (See Table 4 ). Of the four effect sizes conducted, only one showed a statistically significant effect (based on data from 4 to 9 studies; small effect size). The certainty of evidence in these results ranged from very low (3 effect sizes) to low (1 effect size). Table 4 GRADE Summary of Findings Table (Cognitive) Certainty assessment № of patients Effect Certainty Importance № of studies Study design Risk of bias Inconsistency Indirectness Imprecision Other considerations Cognitive [placebo] Relative (95% CI) Cognitive Outcome x Hybrid Format x Immediate Post-Intervention (< 1-month) Time Period 9 randomised trials (6/9) very serious a serious b not serious serious c none 1116/1116 (100.0%) 1030/1030 (100.0%) SMD -0.09 (-0.25 to 0.08) ⨁◯◯◯ Very low a,b,c CRITICAL Cognitive Outcome x Virtual Format x Immediate Post-Intervention (< 1-month) Time Period 7 randomised trials (5/7) very serious d very serious e not serious serious f none 702/702 (100.0%) 713/713 (100.0%) SMD -0.25 (-0.53 to 0.02) ⨁◯◯◯ Very low d,e,f CRITICAL Cognitive Outcome x Hybrid Format x Follow-Up (≥ 1-month to ≤ 6-months) Time Period 4 randomised trials (4/4) serious g serious h not serious serious i none 564/564 (100.0%) 580/580 (100.0%) SMD -0.05 (-0.29 to 0.20) ⨁◯◯◯ Very low g,h,i CRITICAL Cognitive Outcome x Virtual Format x Follow-Up (≥ 1-month to ≤ 6-months) Time Period 4 randomised trials (2/4) very serious j not serious k not serious not serious none 226/226 (100.0%) 228/228 (100.0%) *SMD − 0.30 (-0.49 to -0.12) ⨁⨁◯◯ Low j,k CRITICAL CI : confidence interval; SMD : standardised mean difference Explanations *Indicates significant effect size a. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (4/9 had high risk of bias overall and 5/9 had unclear risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition. b. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 72%. c. We downgraded the certainty of evidence by one level for imprecision, as the confidence interval (–0.25 to 0.08) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate. d. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (5/7 had high risk of bias overall, 1/7 had unclear risk of bias overall, and 1/7 had low risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition. e. We downgraded the certainty of evidence by two levels for inconsistency due to heterogeneity among studies, with an I² statistic of 78%. f. We downgraded the certainty of evidence by one level for imprecision, as the confidence interval (–0.53 to 0.02) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate. g. We downgraded the certainty of evidence by one level for serious study limitations as all included studies (4/4) had unclear risk of bias because of potential issues with methods, i.e. blinding of personnel and outcome assessors, and attrition. h. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I² statistic of 75 We downgraded the certainty of evidence by one level for imprecision, as the confidence interval (–0.29 to 0.20) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate. j. We downgraded the certainty of evidence by two levels for serious study limitations as all included studies (4/4) had high risk of bias because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition. k. I² statistic of 0%. Final Results Summary Discussion This meta-analysis examined the effectiveness of universal DMHIs for children and youth across emotional, behavioural, social, and cognitive outcomes. Of the 16 meta-analyses conducted, nine yielded statistically significant – but small – effects. The largest effect sizes, though still considered small, were observed for virtual interventions targeting emotional and behavioural outcomes immediately post-intervention, and for cognitive outcomes at follow-up. However, hybrid interventions showed the most consistent pattern of effectiveness, with small but significant effects observed across emotional, behavioural, and social outcomes at both the immediate post-intervention and follow-up time periods. The highest level of evidence certainty – rated moderate according to the GRADE framework – was observed for hybrid interventions at follow-up. All other findings were supported by low or very low certainty, largely due to methodological limitations, inconsistency across studies, and imprecise effect estimates. Six of the nine statistically significant outcomes were found in hybrid interventions, underscoring their broader impact and greater potential for sustained effects compared to fully virtual formats. This pattern likely reflects the added structure, relational support, and accountability that hybrid interventions provide. These programs often take place within school or community contexts that encourage adult oversight and peer interaction, even when self-led. For example, in Pisani et al. (2024) [ 13 ], the self-guided Text4Strength texting program was embedded within a broader school-based initiative, balancing independent engagement with opportunities for connection and support. In contrast, fully virtual interventions were more often delivered in unsupervised settings, which may reduce engagement and sustained impact. These observations align with prior research showing that DMHIs delivered within guided or structured environments tend to produce stronger engagement and adherence [ 3 , 4 ]. For large-scale implementation, embedding digital interventions into existing institutions may help optimize both reach and effectiveness. The clearest and most consistent effects in this meta-analysis emerged for emotional and behavioural outcomes, followed by social outcomes – mirroring a broader pattern in the DMHI literature. For years, DMHIs have largely focused on reducing symptoms of anxiety, stress, and externalizing behaviours in children and youth [ 5 , 6 ]. These emotional and behavioural domains are not only more frequently targeted in intervention design, particularly in cognitive behavioural therapy (CBT)-based programs, but are also more readily measured through validated self-report tools. As a result, effects in these areas are easier to detect, both statistically and practically. In contrast, cognitive outcomes appear to have been relegated to secondary status – understudied, inconsistently measured, and overshadowed by more established clinical targets. This reflects persistent gaps in the literature, as noted in previous reviews [ 14 , 15 ], and highlights a critical limitation in the current evidence base. Importantly, the absence of consistent findings in this domain does not imply that digital interventions are ineffective for cognitive development; in fact, the largest effect size – albeit small – was observed for the cognitive outcome. Expanding the focus of DMHIs to directly and intentionally engage with social connection and cognitive skill-building is not just a research imperative; it is essential to understanding how these tools can support the full spectrum of child and youth development. Until high-quality studies fill these gaps, practitioners should be cautious about assuming that improvements in anxiety or mood will automatically extend to academic performance or interpersonal functioning. Although virtual interventions showed encouraging results – particularly for emotional and behavioural outcomes immediately after use, and for cognitive outcomes at follow-up – the strength of this evidence remains limited. These findings were rated as low or very low certainty, primarily due to small sample sizes, high risk of bias, and notable inconsistency across studies. These limitations speak to a broader and ongoing challenge in the digital mental health landscape: ensuring that virtual tools are not only accessible, but also effective when delivered outside of structured, supportive environments. Unlike hybrid interventions, which are often embedded within school or community settings that provide some form of supervision, many virtual programs are accessed independently and without the scaffolding of adult oversight or peer engagement. These contextual differences may explain why hybrid interventions tended to produce more lasting effects across multiple outcome domains. While many digital tools boast features like interactivity, gamification, reminders, or personalization – elements shown to improve user engagement – these features are inconsistently incorporated and unevenly evaluated. As the field continues to evolve, the literature increasingly reflects a shared understanding: virtual interventions hold great promise, but their real-world effectiveness depends on thoughtful design and intentional implementation strategies [ 16 ]. Without those, even the most innovative platforms may struggle to reach their full potential. Perhaps one of the most pressing stories told by this meta-analysis is not just what the evidence reveals, but who is missing from the evidence altogether. Adding to this concern is the overwhelming dominance of studies conducted in high-income, Western countries. While these regions have been instrumental in advancing digital innovation, their prominence risks sidelining the unique needs and lived experiences of communities in lower-resource settings. If digital mental health tools are to fulfill their promise, they must be adaptable – not only across devices, but across cultures, languages, and systems of care. To bridge this divide, future research must move beyond simply expanding access. It must actively prioritize equity and contextual relevance. This means improving demographic transparency, co-designing interventions with underrepresented communities, and ensuring that digital tools are tested and refined in diverse real-world settings. Only then can we begin to realize a vision of digital mental health that is for all. Conclusion This meta-analysis offers emerging evidence that DMHIs can positively support emotional, behavioural, social, and cognitive well-being in children and youth. However, the promise of these tools remains largely unrealized. Small effect sizes and low quality evidence highlight an urgent need to better understand how to design and deliver interventions that can achieve stronger, more lasting impacts – especially as rates of mental distress among children and youth continue to rise. Across outcomes, hybrid interventions stood out – particularly at follow-up – for producing more consistent effects in emotional, behavioural, and social domains. These findings, supported by moderate-certainty evidence, suggest that embedding digital tools within structured settings such as schools or community programs may enhance and sustain their impact. By contrast, although virtual interventions showed promise in certain domains, their effectiveness was supported by lower-certainty evidence, largely due to limitations in study design, inconsistency, and imprecision. Taken together with findings from our prior scoping review, these results reinforce both the promise and the limitations of the current evidence base. Key gaps – such as the underrepresentation of younger children, lack of demographic transparency, and limited inclusion of marginalized populations – raise ongoing concerns about equity, reach, and contextual fit. To increase the magnitude of effect, future research must go beyond tool development to focus on the elements that support effective use. This includes embedding digital interventions in real-world environments, designing for sustained engagement, and ensuring interventions are not only developmentally relevant and culturally responsive, but also grounded in strong methodological rigor. Evaluating outcomes that reflect the full range of children’s lived experiences – and doing so in a way that is transparent, powered, and scalable – will be critical for moving from potential to impact. Limitations While this meta-analysis offers a comprehensive synthesis of the existing evidence on universal DMHIs for children and youth, several limitations should be acknowledged. First, the overall certainty of evidence was low to very low across most outcomes, with only a few analyses rated as moderate certainty. No findings were supported by high-certainty evidence. This limits the confidence with which conclusions can be drawn. The low certainty was primarily due to small sample sizes, heterogeneity in intervention delivery, inconsistency across studies, and the methodological quality of trials, many of which had a high or unclear risk of bias. As recommended by GRADE and Cochrane guidelines, these factors significantly reduce the reliability of pooled effect estimates. Despite using the best available data, the findings remain subject to considerable uncertainty and should be interpreted with caution. Second, although intervention format and setting were clearly reported across studies, there was considerable variation in how hybrid and virtual interventions were implemented and evaluated. For example, interventions varied in the degree of facilitator involvement and delivery structure. Additionally, the inclusion of both randomized and non-randomized study designs may have introduced further heterogeneity. These factors complicate direct comparisons and may have contributed to imprecision in effect estimates. Third, this review included only peer-reviewed literature, which may have resulted in the exclusion of relevant unpublished or grey literature, potentially introducing publication bias. Similarly, eligibility criteria were narrowly defined to ensure comparability of outcomes, but this may have led to the exclusion of studies with valuable but differently framed data. Taken together, these limitations extend the gaps identified in our prior scoping review, which found persistent underrepresentation of early childhood populations and historically marginalized communities in DMHI research. Furthermore, the predominance of studies conducted in high-income countries, with limited representation from low-resource settings, raises concerns about generalizability and global applicability. Despite these constraints, this meta-analysis provides a valuable overview of how DMHIs are currently being implemented in universal child and youth populations. Addressing these limitations through larger, higher-quality trials, more inclusive research designs, and better reporting standards will be essential for advancing the field and scaling equitable, evidence-based DMHIs. Methods Registration and protocol Before its initiation, our systematic review was registered at the International Platform of Registered Systematic Review and Meta-analysis Protocols (INPLASY Protocol 6765). Our reporting adheres to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) statement reporting guidelines [ 17 ] (See Supplementary Table 2). Search strategy and eligibility criteria A comprehensive systematic search was conducted to identify peer-reviewed studies evaluating DMHIs for universal populations of children and youth. The search strategy was developed in collaboration with an academic librarian and focused on three core components: (1) universal scope, (2) digital intervention format, and (3) child and youth mental health outcomes. Searches were conducted in three electronic databases: MEDLINE, PsycInfo, and Embase. An initial scan of Google Scholar helped identify key articles and inform the final list of keywords, which was refined to ensure compatibility across platforms. The initial search was conducted between June 26 and July 5, 2024, and an updated search using the same strategy was performed on January 6, 2025, to capture studies published between July 3, 2024, and January 6, 2025. A full list of search terms is provided in Supplementary Table 3. To be eligible for inclusion, studies had to meet the following criteria: Publication Type: Only full-text, peer-reviewed original research articles were included, as these represent the highest standard for evidence-based practice. Book chapters, case studies, conference abstracts, and dissertations were excluded. Participants: Studies were eligible if the minimum or mean age of participants was between 0 and 18 years. Studies with some participants over 18 were included if the average age fell within this range. Studies exclusively involving adults or not specifying participant age were excluded. Mental Health Status: Only studies involving non-clinical populations were included. Studies focusing on participants with acute psychiatric or medical conditions (e.g., hospitalization for suicide attempt or cancer treatment) or those recently discharged from such care were excluded. Intervention Format: Eligible studies included digital mental health interventions delivered either entirely online (virtual) or through a hybrid model combining digital tools with in-person components, such as school-based facilitation. Studies that relied solely on in-person interventions without a digital component were excluded. Outcomes: Studies were included if they reported at least one outcome related to emotional, behavioural, social, or cognitive well-being. Studies measuring only physiological outcomes without a corresponding psychological component were excluded. When multiple outcomes were reported, the most psychometrically robust measure per domain was prioritized for extraction. Study selection All search results were imported into Covidence, a systematic review management platform, to facilitate de-duplication, title and abstract screening, full-text review, and data extraction. Records with complete indexing were automatically categorized by database source, while entries with incomplete metadata (e.g., manually imported records or non-indexed citations) were labeled as “unspecified sources.” Screening was carried out by a four-person review team. A random 25% subset of records was double-screened, with each record independently reviewed by two team members to assess consistency in screening decisions. The remaining records were single-screened. A fifth senior author provided oversight, helped resolve conflicts, and supported decision-making. Full-text articles deemed potentially eligible were retrieved and independently assessed by two reviewers. Disagreements were resolved through discussion or, if needed, in consultation with the senior author. Reference lists from included studies and relevant reviews were also manually searched to identify any additional eligible articles. Agreement between reviewers was high across both screening stages, with Cohen’s kappa values ranging from 0.84 to 0.98 for title and abstract screening, and 0.90 to 0.96 for full-text review. Data extraction Data extraction was conducted using Microsoft Excel. Variables were initially selected based on the PICO framework (Population, Intervention, Comparator, Outcomes) and refined iteratively as familiarity with the included studies developed. For each study, reviewers extracted information on study design, publication year, and participant characteristics, including age range and lived experience of marginalization. Intervention characteristics were also captured, including delivery format (hybrid or virtual), setting (e.g., school, home, community), facilitation type (self-led, facilitator-led, or both), total number of sessions, duration in hours, assessment time periods (post-intervention or follow-up), intervention type (psychoeducational, skills-building, or both), and any reported information on facilitator training or qualifications. Mental health outcomes were categorized into emotional, behavioural, social, or cognitive domains, and the type of measurement tools and reporting methods were recorded (e.g., whether measures were validated or unvalidated, and whether data were collected via self-report, caregiver-report, or teacher-report). The data extraction template was piloted and refined to ensure clarity and consistency across reviewers. For all included full-text articles, data were independently extracted by two reviewers, with discrepancies resolved through discussion. Descriptive statistics, including frequencies, medians, and ranges, were used to summarize study and intervention characteristics. Data were presented in both tabular and graphical formats where appropriate. Inter-rater agreement during title and abstract screening and full-text review was calculated to assess consistency. Where meta-analysis was feasible, SMDs were calculated to account for variability in outcome measures across studies. Random-effects models were used to estimate pooled effect sizes, and heterogeneity was assessed using the I² statistic and Chi² test, following Cochrane guidelines. Risk of bias assessment To assess the methodological quality of included studies, two reviewers independently applied established appraisal tools. For randomized studies, the Cochrane Risk of Bias 2 (RoB 2) tool was used, classifying risk as low, some concerns, or high. The ROBINS-I tool (Risk of Bias in Non-randomized Studies of Interventions) tool was used for non-randomized studies, categorizing risk of bias as low, moderate, or serious. Discrepancies between reviewers were resolved through discussion until consensus was reached. A detailed summary of risk of bias assessments is provided in the methodological quality section above. Meta-analysis All meta-analyses were conducted using the Generic Inverse Variance method in Review Manager (RevMan), version 5.4 [ https://training.cochrane.org/online-learning/core-software-cochrane-reviews/revman ]. This approach was applied consistently across all 16 outcome-level meta-analyses, which included studies with varying designs (e.g., RCTs, quasi-experimental, and single-group pre-post). Effect sizes were reported as SMDs with 95% confidence intervals (CIs). For non-randomized studies pre-post studies, we handled the data based on how each individual study reported their pre- and post-intervention measurements. When change scores or sufficient data were reported, we treated the data as paired by entering the same N for pre and post. When only pre and post means, SDs, and sample sizes were reported (without change scores or pre-post correlations), we entered the data as reported [ 18 , 11 , 12 , 19 , 20 , 10 ], treating pre and post as separate groups in the analysis. Assessment of the certainty of the evidence We assessed confidence in the findings using the GRADE framework, which evaluates five domains: risk of bias, inconsistency, indirectness, imprecision, and other considerations. All domains were assessed except other considerations, which includes publication bias; this was not formally evaluated due to the limited number of studies per outcome, as standard methods for detecting publication bias (e.g., funnel plots) are unreliable with fewer than 10 studies. Given the limited evidence base for many outcomes, particular attention was paid to risk of bias, inconsistency, imprecision, and indirectness. Risk of bias was evaluated using the Cochrane RoB 2 tool for randomized trials and the ROBINS-I tool for non-randomized studies. Outcomes were downgraded when most contributing studies were rated as high or unclear risk of bias. Inconsistency (heterogeneity) was evaluated using the I² statistic, with thresholds based on GRADE guidance: Not serious (I² 50%) – High variability, indicating substantial inconsistency across study results Interpretation of I² was supported by examining the direction and overlap of confidence intervals, as well as the number and quality of contributing studies. Imprecision was evaluated based on the width of confidence intervals, the total number of participants, and the number of studies contributing to each pooled estimate. Outcomes with wide CIs or limited data were downgraded accordingly. Indirectness was not a major concern, due to the specificity of our inclusion criteria. By focusing exclusively on universal Tier 1 DMHIs for children and youth, the included studies were closely aligned with the population, intervention, and outcomes of interest. We also conducted sensitivity analyses using a leave-one-out approach to assess the influence of individual studies on pooled estimates and to test the stability of the results. Each outcome was ultimately assigned a certainty rating of high, moderate, low, or very low, reflecting the overall strength of evidence and our confidence in the findings. Declarations Data availability All data generated or analyzed during this study is available from the corresponding author upon reasonable request. Code availability No custom code was used for the analyses. All meta-analyses were conducted using Review Manager (RevMan) version 5.4, available from the Cochrane Collaboration: [ https://training.cochrane.org/online-learning/core-software ]. Competing Interests The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Author Contribution KDP co-conceptualized the study, led the systematic review and meta-analysis, and drafted the initial manuscript. RPR conceptualized the study, assisted with the analysis, and supported the development of the initial manuscript. OB, HH, and AL assisted with the analysis and offered editorial feedback on the submitted manuscript. AL contributed to the development of the study’s conceptual foundation. RC, as the statistical consultant for the meta-analysis, assisted with the analysis and provided editorial feedback on the submitted manuscript. Acknowledgement We would like to thank the Jackman Foundation and York University (funding dedicated to DIVERT Mental Health collaboration, a CIHR funded Health Research Training Platform) for funding that supported this research project. References Organisation for Economic Co-operation and Development. Children and young people’s mental health in the digital age: shaping the future . OECD Publishing; https://doi.org/10.1787/9789264318706-en (2018). Liverpool, S. et al. Engaging children and young people in digital mental health interventions: systematic review of modes of delivery, facilitators, and barriers. J. Med. Internet Res . 22, e16317 (2020). https://doi.org/10.2196/16317 Clarke, A. M., Kuosmanen, T. & Barry, M. M. A systematic review of online youth mental health promotion and prevention interventions. J. Youth Adolesc . 44, 90–113 (2015). Hollis, C. et al. Annual Research Review: digital health interventions for children and young people with mental health problems - a systematic and meta-review. J. Child Psychol. Psychiatry 58, 474–503 (2017). https://doi.org/10.1111/jcpp.12663 Grist, R. et al. Technology delivered interventions for depression and anxiety in children and adolescents: a systematic review and meta-analysis. Clin. Child Fam. Psychol. Rev. 22, 147–171 (2019). Garrido, S. et al. What works and what doesn’t work? A systematic review of digital mental health interventions for depression and anxiety in young people. Front. Psychiatry 10, 759 (2019). Piers, C., Monks, H., McLoughlin, L. T. & Grové, C. Can digital mental health interventions bridge the ‘digital divide’ for socioeconomically and digitally marginalised youth? A systematic review. Child Adolesc. Ment. Health 28, 54–65 (2023). https://doi.org/10.1111/camh.12620 Sterne, J. A. C. et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 366, l4898; https://doi.org/10.1136/bmj.l4898 (2019). Sterne, J. A. C. et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ 255, i4919; https://doi.org/10.1136/bmj.i4919 (2016). Hassen, H. M. et al. Effectiveness and implementation outcome measures of mental health curriculum intervention using social media to improve the mental health literacy of adolescents. J. Multidiscip. Healthc. 15, 979–997 (2022). Peuters, C. et al. A mobile healthy lifestyle intervention to promote mental health in adolescence: A mixed-methods evaluation. BMC public health 24, 44 (2024). Sousa, P. et al. Controlled trial of an mHealth intervention to promote healthy behaviours in adolescence (TeenPower): effectiveness analysis. J. Adv. Nurs. 76, 1057–1068 (2020). Pisani, A.R. et al. Text4Strength: a brief automated text messaging program to promote help-seeking and reduce mental health stigma among high school students. J. Child Psychol. Psychiatry 65, 58–69 (2024). Fleming, T. et al. Beyond the trial: systematic review of real-world uptake and engagement with digital self-help interventions for depression, anxiety, or emotional distress. J. Med. Internet Res . 20, e199 (2018). Orlowski, S. et al. A systematic review of mental health interventions for social and emotional well-being in Australian indigenous communities. BMC Public Health 16, 1039 (2016). Beames, J. R. et al. Effect of engagement with digital interventions on mental health outcomes: a systematic review and meta-analysis. Front. Digit. Health 3, 764079 (2021). Page, M. J. et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 372, n71 (2021). Goodman, R., & Newman, D. Testing a digital storytelling intervention to reduce stress in adolescent females. Storytelling, Self, Society , 10, 177–193 (2014). Subotic-Kerry, M. et al. Examining the impact of a universal positive psychology program on mental health outcomes among Australian secondary students during the COVID-19 pandemic. Child Adolesc. Psychiatry Ment. Health 17, 70 (2023). Morawska, A. & Sanders, M. R. Self-administered behavioural family intervention for parents of toddlers: Effectiveness and dissemination. Behav. Res. Ther. 44, 1839–1848 (2006). Additional Declarations No competing interests reported. Supplementary Files SupplementaryDataTable2ReprintDiPierdomenicoetal.underreview.pdf SupplementaryDataMetaAnalysis.pdf editorialpolicychecklist.pdf Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 14 Sep, 2025 Reviews received at journal 11 Sep, 2025 Reviewers agreed at journal 19 Aug, 2025 Reviewers agreed at journal 18 Aug, 2025 Reviews received at journal 02 Jul, 2025 Reviewers agreed at journal 09 Jun, 2025 Reviewers invited by journal 09 Jun, 2025 Editor assigned by journal 06 Jun, 2025 Submission checks completed at journal 06 Jun, 2025 First submitted to journal 03 Jun, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6811900","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":468654477,"identity":"f7cc9f34-dbf7-432d-b9b0-3e94db19ca7a","order_by":0,"name":"Dr. Kaitlin Di Pierdomenico","email":"","orcid":"","institution":"York University","correspondingAuthor":false,"prefix":"Dr.","firstName":"Kaitlin","middleName":"Di","lastName":"Pierdomenico","suffix":""},{"id":468654478,"identity":"1709faec-842b-48c2-b9f5-b6367dd87c8f","order_by":1,"name":"Oana Bucsea","email":"","orcid":"","institution":"York University","correspondingAuthor":false,"prefix":"","firstName":"Oana","middleName":"","lastName":"Bucsea","suffix":""},{"id":468654479,"identity":"a7d67bd2-f2f2-4ef3-a481-b8435d1c093a","order_by":2,"name":"Haleh Hashemi","email":"","orcid":"","institution":"York University","correspondingAuthor":false,"prefix":"","firstName":"Haleh","middleName":"","lastName":"Hashemi","suffix":""},{"id":468654480,"identity":"657b5642-606a-47bd-83a8-f138c5f86021","order_by":3,"name":"Arianna Leguia","email":"","orcid":"","institution":"York University","correspondingAuthor":false,"prefix":"","firstName":"Arianna","middleName":"","lastName":"Leguia","suffix":""},{"id":468654481,"identity":"8ee0466d-8d8a-4783-aeec-48240197b5de","order_by":4,"name":"Anne Lovegrove","email":"","orcid":"","institution":"Strong Minds Strong Kids, Psychology Canada","correspondingAuthor":false,"prefix":"","firstName":"Anne","middleName":"","lastName":"Lovegrove","suffix":""},{"id":468654482,"identity":"974ca7da-e07b-44de-a71b-9699bf19c0a8","order_by":5,"name":"Dr. Robert Cribbie","email":"","orcid":"","institution":"York University","correspondingAuthor":false,"prefix":"Dr.","firstName":"Robert","middleName":"","lastName":"Cribbie","suffix":""},{"id":468654483,"identity":"1982aafa-ed56-4819-b629-2bccb5aa6ffd","order_by":6,"name":"Dr. Rebecca Pillai Riddell","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAmklEQVRIiWNgGAWjYHACxgdAwoAkLcwGJGthkyBNi/ns3mfVPBX3jA0OMD/8QJQWmTvHzW7znCk2MzjAZixBlBYJiTS227xtCTYGB3gYiNdSDNXC/INoLcxALUCH8bARaYvMMWbJOWcSjCUPs5lZEKdFuo3xw5uKBMO+482PbxClBeRlJh4Qg5k49RAtjMR5exSMglEwCkYsAAALliYMWncrDgAAAABJRU5ErkJggg==","orcid":"","institution":"York University","correspondingAuthor":true,"prefix":"Dr.","firstName":"Rebecca","middleName":"Pillai","lastName":"Riddell","suffix":""}],"badges":[],"createdAt":"2025-06-03 13:38:35","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6811900/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6811900/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":84534312,"identity":"283927f7-b2b3-48c3-9490-485c206ca4da","added_by":"auto","created_at":"2025-06-13 06:47:18","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":104972,"visible":true,"origin":"","legend":"\u003cp\u003ePresents the PRISMA flowchart outlining the study selection process. Data were extracted using Covidence. Studies were excluded for the following reasons: (1) the sample did not fall within the target age range, defined as children and youth aged 0 to 18 years, based on either minimum or mean age; (2) the source was not a peer-reviewed journal article, including book chapters, case studies, conference abstracts, dissertations, or other grey literature; (3) the study focused on participants with a diagnosed mental health condition, those admitted to or discharged from hospital (e.g., for suicide attempts or cancer treatment), or those targeting psychosis prevention or conducted in pediatric clinical settings involving comorbid chronic illness; (4) the study did not assess emotional, behavioural, cognitive, or social mental health outcomes, and instead reported only physiological outcomes; (5) the intervention lacked a digital component and therefore did not meet the inclusion criteria; and (6) the study lacked sufficient data for extraction.\u003c/p\u003e","description":"","filename":"Picture1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6811900/v1/8bb86ef7b6b0443a58162670.jpg"},{"id":84535001,"identity":"dda3ab10-627d-452e-8c35-ec13de810b94","added_by":"auto","created_at":"2025-06-13 06:55:18","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":42384,"visible":true,"origin":"","legend":"\u003cp\u003eQuality assessment of included articles using the Cochrane Risk of Bias 2 (RoB 2) tool for randomized studies (n=35).\u003c/p\u003e","description":"","filename":"Picture2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6811900/v1/c2242266b160126f96f6b755.jpg"},{"id":84534316,"identity":"54b5429a-9ffb-4aac-b70a-649d0fac5f0f","added_by":"auto","created_at":"2025-06-13 06:47:19","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":46208,"visible":true,"origin":"","legend":"\u003cp\u003eRisk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool for non-randomized studies (n=16).\u003c/p\u003e","description":"","filename":"Picture3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6811900/v1/27921df7fa6fdd1668d5dca8.jpg"},{"id":84536404,"identity":"5426e1dd-086b-48e7-aae8-d04ad0d28885","added_by":"auto","created_at":"2025-06-13 07:19:20","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1954775,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6811900/v1/d6ae9a63-de1c-4728-8014-06f74023f00e.pdf"},{"id":84534317,"identity":"b0348954-bfc3-49ed-ab1e-493bb15a7fcf","added_by":"auto","created_at":"2025-06-13 06:47:19","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":315585,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryDataTable2ReprintDiPierdomenicoetal.underreview.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6811900/v1/5edbfad36e9028bfcd3001fc.pdf"},{"id":84534331,"identity":"72dd0c69-b1d7-45cd-ac00-e6f39b6a0b86","added_by":"auto","created_at":"2025-06-13 06:47:19","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":5540780,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryDataMetaAnalysis.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6811900/v1/66120140a86a41cb2af549ef.pdf"},{"id":84534320,"identity":"dc4322cf-d1f6-433c-b426-5a07a362a1f2","added_by":"auto","created_at":"2025-06-13 06:47:19","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":1664034,"visible":true,"origin":"","legend":"","description":"","filename":"editorialpolicychecklist.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6811900/v1/b04fd03df71a111f47174894.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Effects of Universal Digital Mental Health Interventions for Children and Youth on Psychological Outcomes – A Systematic Review and Meta-Analysis","fulltext":[{"header":"Introduction","content":"\u003cp\u003eChildren and youth are facing unprecedented mental health challenges while simultaneously navigating an increasingly digital world [https://www.who.int/publications/i/item/9789240106796]. Digital technologies have become deeply integrated into their daily lives, shaping how they socialize, learn, and access information. This shift presents both risks and opportunities. While concerns persist about excessive screen time, physical and mental health implications, and online safety [1], digital platforms also offer promising avenues for expanding access to mental health resources, particularly for people who experience barriers to traditional care, such as long wait times, stigma, and geographic limitations.\u003c/p\u003e\n\u003cp\u003eDigital mental health interventions (DMHIs) have emerged as scalable and flexible tools capable of delivering widespread, early-stage mental health support. These interventions \u0026ndash; ranging from mobile apps and web-based programs to chatbots and virtual sessions \u0026ndash; aim to provide timely, accessible care to children and youth before more intensive services are needed. The move toward digital intervention delivery accelerated during the COVID-19 pandemic, which disrupted in-person care and prompted the rapid expansion of e-mental health resources in many countries [https://mentalhealthcommission.ca/wp-content/uploads/2024/09/An-E-Mental-Health-Strategy-for-Canada-FINAL.pdf].\u003c/p\u003e\n\u003cp\u003eDespite their growing availability and adoption, critical gaps remain in the evidence base grounding the effectiveness, sustainability, and impact of DMHIs across core psychological domains (e.g., emotional, behavioural, social, cognitive). Many interventions have been implemented without rigorous evaluation, raising concerns about quality, safety, and long-term outcomes. Recent policy guidance \u0026ndash; from the World Health Organization\u0026rsquo;s \u003cem\u003eStrategic Mental Health Action Plans\u003c/em\u003e [https://www.who.int/publications/i/item/9789240106796] to national e-mental health strategies [https://mentalhealthcommission.ca/wp-content/uploads/2024/09/An-E-Mental-Health-Strategy-for-Canada-FINAL.pdf, https://www.mqhealth.org.au/__data/assets/pdf_file/0011/1010630/emstrat.pdf] \u0026ndash; has emphasized the urgent need for digital interventions that are evidence-informed, accessible, equitable, and integrated within broader mental health systems.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhile previous reviews have mapped the landscape of DMHIs for children and youth [2, 3, 4, 5, 6, 7], they have typically focused on clinical samples, single-outcome measures, or short-term effects. This systematic review and meta-analysis addresses these gaps by evaluating the effectiveness of universal digital mental health interventions across four psychological domains and two time periods (immediate and follow-up), with the aim of informing more effective care delivery and guiding future policy development.\u003c/p\u003e\n\u003cp\u003eTo address these gaps, the current meta-analysis focuses specifically on universal (Tier 1) DMHIs \u0026ndash; low-intensity, high-reach interventions designed for non-clinical child and youth populations. It builds directly on our prior scoping review (Di Pierdomenico et al., under review),\u0026nbsp;which found that most interventions used hybrid formats (combining digital tools with in-person components), targeted multiple psychological outcomes, were commonly online psychoeducational programs, and often included independent child/youth-led components. The scoping review also identified several important gaps, including the underrepresentation of parents of younger children (particularly those aged 0-4), limited work with marginalized populations, and poor reporting of racial and cultural demographics. Building on the comprehensive understanding of what has currently been done in the area of universal DMHIs via the aforementioned scoping review, this meta-analysis will focus on synthesizing psychological outcomes and evaluating how both intervention format and assessment timing relate to overall impact. It has two primary aims:\u0026nbsp;\u003c/p\u003e\n\u003col\u003e\n \u003cli\u003eUnderstand the efficacy of universal DMHIs by synthesizing outcomes across four psychological outcome domains (emotional, behavioural, social, and cognitive), two intervention formats (hybrid, virtual), and two outcome assessment time periods (immediate post-intervention, follow-up).\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eAssess the certainty of the evidence for each effect size using the GRADE framework [https://gdt.gradepro.org/app/handbook/handbook.html] to inform future research, policy, and practice.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eBy systematically analyzing intervention effectiveness, this meta-analysis offers critical insights to guide the design and implementation of more inclusive, evidence-informed DMHIs. These findings can help ensure that effective mental health supports reach diverse child and youth populations in meaningful and sustainable ways.\u003c/p\u003e"},{"header":"Results ","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n \u003ch2\u003eStudy selection\u003c/h2\u003e\n \u003cp\u003eWe identified 21,729 records through database searches, along with 7 additional records through snowball searching, all of which were categorized as database sources. After removing duplicates, 6,537 records remained for title and abstract screening. A total of 100 full-text articles were assessed for eligibility, and 49 were excluded \u0026ndash; most commonly due to the presence of a mental health diagnosis in the comparison groups or being the wrong article type (see Supplementary Table 1). Ultimately, 51 studies were included for review and data extraction. A PRISMA flow diagram is presented in Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eStudy characteristics\u003c/h3\u003e\n\u003cp\u003eOur prior scoping review (Di Pierdomenico et al., under review) provided a comprehensive description of study characteristics (see Supplementary Materials [Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e in Di Pierdomenico et al., under review]) and informed the analysis plan for the current meta-analysis. In brief, most studies were conducted in high-income countries, published within the last three years, and focused on children aged 5 to 18, with limited reporting of racial or cultural demographics. Interventions were broadly aimed at mental health promotion, prevention, or a combination of both, and often targeted multiple psychological outcomes. Most used hybrid delivery models \u0026ndash; typically in school settings \u0026ndash; and relied on self-report measures. Formats ranged from online programs and apps to games and virtual communication tools. Intervention structures varied, with many using self-directed formats supported by some level of facilitation, most often by research staff or mental health professionals.\u003c/p\u003e\n\u003ch3\u003eAppraisal of methodological quality\u003c/h3\u003e\n\u003cp\u003eStudy quality was assessed using two tools: the Cochrane Risk of Bias 2 (RoB 2) [\u003cspan class=\"CitationRef\"\u003e8\u003c/span\u003e] for randomized controlled trials (RCTs) and the ROBINS-I tool (Risk of Bias in Non-randomized Studies of Interventions) [\u003cspan class=\"CitationRef\"\u003e9\u003c/span\u003e]. Summary plots of the risk of bias assessments are provided in Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e and Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e. Of the 51 studies included in this meta-analysis, 35 (69%) were classified as RCTs and 16 (31%) as non-RCTs. Among the RCTs, 8 studies (22.9%) were rated as having an overall low risk of bias, 11 studies (31.4%) as having unclear risk, and 16 studies (45.7%) as having a high overall risk of bias. The most common sources of bias among RCTs were related to selection bias due to lack of allocation concealment (n\u0026thinsp;=\u0026thinsp;1), performance bias due to lack of blinding of participants and personnel (n\u0026thinsp;=\u0026thinsp;13), detection bias (n\u0026thinsp;=\u0026thinsp;9), and attrition bias (n\u0026thinsp;=\u0026thinsp;5).\u003c/p\u003e\n\u003cp\u003eFor the 16 non-randomized studies, 13 employed pre-post single-group designs, while 3 were non-randomized quasi-experimental designs with comparison groups [\u003cspan class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan class=\"CitationRef\"\u003e12\u003c/span\u003e]. To ensure consistency, both the Cochrane RoB 2 tool (via RevMan), and the ROBINS-I framework were used. The ROBINS-I assessments revealed that 14 studies (87.5%) presented with a serious overall risk of bias, and 2 studies (12.5%) were rated as having a moderate risk of bias. Serious risk of bias was observed across the seven ROBINS-I domains, including: confounding (n\u0026thinsp;=\u0026thinsp;14), classification of interventions (n\u0026thinsp;=\u0026thinsp;2), participant selection (n\u0026thinsp;=\u0026thinsp;3), deviations from intended interventions (n\u0026thinsp;=\u0026thinsp;1), outcome measurement (n\u0026thinsp;=\u0026thinsp;2), and selection of the reported results (n\u0026thinsp;=\u0026thinsp;1). The Cochrane RoB 2 evaluations of non-RCTs were consistent with the ROBINS-I findings, confirming poor methodological quality in the majority of non-randomized studies.\u003c/p\u003e\n\u003ch3\u003eResults\u003c/h3\u003e\n\u003cp\u003eWe conducted 16 separate meta-analyses across 51 studies (30,474 participants; accounting for within-study duplicates) grouped by four psychological outcome domains (emotional, behavioural, social, and cognitive), two delivery formats (hybrid and virtual), and two assessment time periods (immediate post-intervention, within one month; and follow-up, one to six months post-intervention). Forest plots corresponding to each meta-analysis, along with associated quality appraisals informing the GRADE tables, are available in Supplementary Figs.\u0026nbsp;1\u0026ndash;16. GRADE tables are presented by outcome domain, intervention format, and assessment time period. Standardized mean differences (SMDs) were calculated to accommodate variation in outcome measures across studies.\u003c/p\u003e\n\u003cp\u003eIn all analyses, negative SMD values indicate reductions in adverse psychological outcomes post-intervention, representing a positive effect of the intervention. To assess confidence in each effect estimate, we applied the GRADE approach [\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://training.cochrane.org/grade-approach\u003c/span\u003e\u003c/span\u003e], evaluating the certainty of evidence based on risk of bias, inconsistency, indirectness, imprecision, and other considerations.\u003c/p\u003e\n\u003ch3\u003e1. Emotional Outcomes\u003c/h3\u003e\n\u003cp\u003eEmotional outcomes, referring to how participants feel, were examined in four separate meta-analyses (See Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). Four effect sizes were calculated, each based on data from 5 to 21 studies. Three of the four effects were significant (small magnitude), with the certainty of evidence ranging from \u003cem\u003every low\u003c/em\u003e (three effect sizes) to \u003cem\u003emoderate\u003c/em\u003e (one effect size).\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eGRADE Summary of Findings Table (Emotional)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"7\"\u003e\n \u003cp\u003eCertainty assessment\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e№ of patients\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEffect\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCertainty\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eImportance\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e№ of studies\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStudy design\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRisk of bias\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eInconsistency\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eIndirectness\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eImprecision\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eOther considerations\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEmotional\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e[placebo]\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRelative\u003c/p\u003e\n \u003cp\u003e(95% CI)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"12\"\u003e\n \u003cp\u003eEmotion Outcome x Hybrid Format x Immediate Post-Intervention (\u0026lt;\u0026thinsp;1-month) Time Period\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(16/21)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3111/3111 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3156/3156 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e*SMD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e-0.18\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.26 to -0.11)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003ea,b\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"12\"\u003e\n \u003cp\u003e\u003cstrong\u003eEmotion Outcome x Virtual Format x Immediate Post-Intervention (\u0026lt;\u0026thinsp;1-month) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(10/15)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ed\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1471/1471 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1948/1948 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e*SMD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e-0.24\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.47 to -0.01)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003ec,d\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"12\"\u003e\n \u003cp\u003e\u003cstrong\u003eEmotion Outcome x Hybrid Format x Follow-Up (\u0026ge;\u0026thinsp;1-month to \u0026le;\u0026thinsp;6-months) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(13/13)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003csup\u003ee\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003ef\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2636/2636 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3285/3285 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e*SMD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e-0.19\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.31 to -0.07)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e⨁⨁⨁◯\u003c/p\u003e\n \u003cp\u003eModerate\u003csup\u003ee,f\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"12\"\u003e\n \u003cp\u003e\u003cstrong\u003eEmotion Outcome x Virtual Format x Follow-Up (\u0026ge;\u0026thinsp;1-month to \u0026le;\u0026thinsp;6-months) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(3/5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003eg\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003csup\u003eh\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003ei\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e599/599 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e707/707 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSMD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e-0.05\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.16 to 0.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003eg,h,i\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003eCI\u003c/strong\u003e: confidence interval; \u003cstrong\u003eSMD\u003c/strong\u003e: standardized mean difference\u003c/p\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n \u003ch2\u003eExplanations\u003c/h2\u003e\n \u003cp\u003e*Indicates significant effect size\u003c/p\u003e\u003cspan\u003e\n \u003cp\u003ea. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (11/21 had high risk of bias overall, 7/21 had unclear risk of bias overall, and 3/21 had low risk of bias overall) because of potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003eb. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I\u0026sup2; statistic of 53%.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ec. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (12/15 had high risk of bias overall, 2/15 had unclear risk of bias overall, and 1/15 had low risk of bias overall) because of potential issues with methods, i.e. allocation concealment, blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ed. We downgraded the certainty of evidence by two levels for inconsistency due to heterogeneity among studies, with an I\u0026sup2; statistic of 84%.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ee. Most included studies had unclear risk of bias (6/13 had unclear risk of bias overall, 4/13 had low risk of bias overall, and 3/13 had high risk of bias overall) because of potential issues with methods, i.e., blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ef. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I\u0026sup2; statistic of 67%.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003eg. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (3/5 had high risk of bias overall and 2/5 had unclear risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of participants and personnel, and attrition.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003eh. I\u0026sup2; statistic of 0%.\u003c/p\u003e\n \u003c/span\u003e\u003cspan\u003e\n \u003cp\u003eWe downgraded the certainty of evidence by one level for imprecision, as the confidence interval (-0.16 to 0.06) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate.\u003c/p\u003e\n \u003c/span\u003e\n\u003c/div\u003e\n\u003ch3\u003e2. Behavioural Outcomes\u003c/h3\u003e\n\u003cp\u003eBehavioural outcomes, or how participants act, were examined in four separate meta-analyses (See Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e). Four effect sizes were calculated, each based on data from 4 to 9 studies. Three of the four effects were significant (small magnitude), with the certainty of evidence ranging from \u003cem\u003every low\u003c/em\u003e (3 effect sizes) to \u003cem\u003emoderate\u003c/em\u003e (1 effect size).\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u0026nbsp;\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eGRADE Summary of Findings Table (Behavioural)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"7\"\u003e\n \u003cp\u003eCertainty assessment\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e№ of patients\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEffect\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eCertainty\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eImportance\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"1\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e№ of studies\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStudy design\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRisk of bias\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eInconsistency\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eIndirectness\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eImprecision\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eOther considerations\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eBehavioural\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e[placebo]\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRelative\u003c/p\u003e\n \u003cp\u003e(95% CI)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"4\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"13\"\u003e\n \u003cp\u003eBehaviour Outcome x Hybrid Format x Immediate Post-Intervention (\u0026lt;\u0026thinsp;1-month) Time Period\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enon-randomised studies\u003c/p\u003e\n \u003cp\u003e(5/8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e415/415 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e510/510 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e*SMD \u0026minus;\u0026thinsp;0.15\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.27 to\u003c/p\u003e\n \u003cp\u003e-0.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003ea,b\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"13\"\u003e\n \u003cp\u003e\u003cstrong\u003eBehaviour Outcome x Virtual Format x Immediate Post-Intervention (\u0026lt;\u0026thinsp;1-month) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enon-randomised studies\u003c/p\u003e\n \u003cp\u003e(6/9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003ed\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e954/954 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1018/1018 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e*SMD \u0026minus;\u0026thinsp;0.29\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.52 to\u003c/p\u003e\n \u003cp\u003e-0.07)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003ec,d\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"13\"\u003e\n \u003cp\u003e\u003cstrong\u003eBehaviour Outcome x Hybrid Format x Follow-Up (\u0026ge;\u0026thinsp;1-month to \u0026le;\u0026thinsp;6-months) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(4/4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003ee\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003csup\u003ef\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e375/375 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e451/451 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e*SMD \u0026minus;\u0026thinsp;0.16\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.29 to\u003c/p\u003e\n \u003cp\u003e-0.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁⨁⨁◯\u003c/p\u003e\n \u003cp\u003eModerate\u003csup\u003ee,f\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"13\"\u003e\n \u003cp\u003e\u003cstrong\u003eBehaviour Outcome x Virtual Format x Follow-Up (\u0026ge;\u0026thinsp;1-month to \u0026le;\u0026thinsp;6-months) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enon-randomised studies\u003c/p\u003e\n \u003cp\u003e(3/4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003eg\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003eh\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ei\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e166/166 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e180/180 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSMD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e-0.40\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.84 to 0.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003eg,h,i\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003eCI\u003c/strong\u003e: confidence interval; \u003cstrong\u003eSMD\u003c/strong\u003e: standardised mean difference\u003c/p\u003e\n\u003ch3\u003eExplanations\u003c/h3\u003e\n\u003cp\u003e*Indicates significant effect size\u003c/p\u003e\n\u003cp\u003ea. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (7/8 had high risk of bias overall and 1/8 had low risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n\u003cp\u003e\u003cspan\u003e\u003c/span\u003e\u003c/p\u003e\n\u003cp\u003eb. I\u0026sup2; statistic of 0%.\u003c/p\u003e\u003cspan\u003e\n \u003cp\u003ec. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (7/9 had high risk of bias overall, 1/9 had unclear risk of bias overall, and 1/9 had low risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003ed. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I\u0026sup2; statistic of 72%.\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003ee. We downgraded the certainty of evidence by one level for serious study limitations as most included studies had high or unclear risk of bias (2/4 had high risk of bias overall, 1/4 had unclear risk of bias overall, and 1/4 had low risk of bias overall) because of potential issues with methods, i.e., blinding of personnel and outcome assessors. Approximately 34% of the overall weight came from studies with high risk of bias, and over 50% came from a study with unclear risk. Only a small portion (14.5%) was contributed by a study at low risk of bias.\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003ef. I\u0026sup2; statistic of 0%.\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003eg. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (3/4 had high risk of bias overall and 1/4 had unclear risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, attrition, and reporting bias.\u003c/p\u003e\n\u003c/span\u003e\u003cspan\u003e\n \u003cp\u003eh. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I\u0026sup2; statistic of 70%.\u003c/p\u003e\n\u003c/span\u003e\n\u003cp\u003eWe downgraded the certainty of evidence by two levels for imprecision, as the confidence interval (\u0026ndash;0.84 to 0.03) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate. Additionally, two of the four included studies had small experimental sample sizes (\u0026lt;\u0026thinsp;35 participants), which may further reduce the overall precision of the pooled result.\u003c/p\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003e3. Social Outcomes\u003c/h2\u003e\n \u003cp\u003eSocial outcomes, referring to how participants connect with others, were examined in four separate meta-analyses (See Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e). Four effect sizes were calculated, each based on data from 3 to 8 studies. Two of the four effects were significant, with the certainty of evidence ranging from \u003cem\u003every low\u003c/em\u003e (3 effect sizes) to \u003cem\u003emoderate\u003c/em\u003e (1 effect size).\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u0026nbsp;\u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eGRADE Summary of Findings Table (Social)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"7\"\u003e\n \u003cp\u003eCertainty assessment\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e№ of patients\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEffect\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eCertainty\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eImportance\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e№ of studies\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStudy design\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRisk of bias\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eInconsistency\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eIndirectness\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eImprecision\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eOther considerations\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSocial\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e[placebo]\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRelative\u003c/p\u003e\n \u003cp\u003e(95% CI)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"1\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"14\"\u003e\n \u003cp\u003eSocial Outcome x Hybrid Format x Immediate Post-Intervention (\u0026lt;\u0026thinsp;1-month) Time Period\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(5/8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1416/1416 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1284/1284 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e*SMD \u0026minus;\u0026thinsp;0.15\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.27 to\u003c/p\u003e\n \u003cp\u003e-0.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003ea,b\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"14\"\u003e\n \u003cp\u003e\u003cstrong\u003eSocial Outcome x Virtual Format x Immediate Post-Intervention (\u0026lt;\u0026thinsp;1-month) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(3/5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ed\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003ee\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e534/534 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e506/506 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSMD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e-0.07\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.48 to 0.34)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003ec,d,e\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"14\"\u003e\n \u003cp\u003e\u003cstrong\u003eSocial Outcome x Hybrid Format x Follow-Up (\u0026ge;\u0026thinsp;1-month to \u0026le;\u0026thinsp;6-months) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(3/3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003ef\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003csup\u003eg\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1046/1046 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e910/910 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e*SMD \u0026minus;\u0026thinsp;0.11\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.20 to\u003c/p\u003e\n \u003cp\u003e-0.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁⨁⨁◯\u003c/p\u003e\n \u003cp\u003eModerate\u003csup\u003ef,g\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"14\"\u003e\n \u003cp\u003e\u003cstrong\u003eSocial Outcome x Virtual Format x Follow-Up (\u0026ge;\u0026thinsp;1-month to \u0026le;\u0026thinsp;6-months) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(3/4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003eh\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003csup\u003ei\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003ej\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e532/532 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e505/505 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSMD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e-0.08\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.21 to 0.05)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003eh,i,j\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003e\u003cstrong\u003eCI\u003c/strong\u003e: confidence interval; \u003cstrong\u003eSMD\u003c/strong\u003e: standardised mean difference\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003ch2\u003eExplanations\u003c/h2\u003e\n \u003cp\u003e*Indicates significant effect size\u003c/p\u003e\u003cspan\u003e\n \u003cp\u003ea. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (5/8 had high risk of bias overall and 3/8 had unclear risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003eb. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I\u0026sup2; statistic of 67%.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ec. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (4/5 had high risk of bias overall and 1/5 had low risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ed. We downgraded the certainty of evidence by two levels for inconsistency due to heterogeneity among studies, with an I\u0026sup2; statistic of 89%.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ee. We downgraded the certainty of evidence by one level for imprecision, as the confidence interval (\u0026ndash;0.48 to 0.34) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ef. We downgraded the certainty of evidence by one level for serious study limitations as most included studies had unclear risk of bias (2/3 had unclear risk of bias overall and 1/3 had low risk of bias overall) because of potential issues with methods, i.e. blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003eg. I\u0026sup2; statistic of 0%.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003eh. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (2/4 had high risk of bias overall and 2/4 had unclear risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n \u003c/span\u003e\u003cspan\u003e\n \u003cp\u003eI\u0026sup2; statistic of 11%.\u003c/p\u003e\n \u003c/span\u003e\u003cspan\u003e\n \u003cp\u003ej. We downgraded the certainty of evidence by one level for imprecision, as the confidence interval (\u0026ndash;0.21 to 0.05) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate.\u003c/p\u003e\n \u003c/span\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\n \u003ch2\u003e4. Cognitive Outcomes\u003c/h2\u003e\n \u003cp\u003eCognitive outcomes, capturing how participants think or solve problems, were examined in four separate meta-analyses (See Table \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). Of the four effect sizes conducted, only one showed a statistically significant effect (based on data from 4 to 9 studies; small effect size). The certainty of evidence in these results ranged from \u003cem\u003every low\u003c/em\u003e (3 effect sizes) to \u003cem\u003elow\u003c/em\u003e (1 effect size).\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u0026nbsp;\u003ctable id=\"Tab4\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eGRADE Summary of Findings Table (Cognitive)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"7\"\u003e\n \u003cp\u003eCertainty assessment\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e№ of patients\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eEffect\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003eCertainty\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eImportance\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e№ of studies\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStudy design\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRisk of bias\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eInconsistency\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eIndirectness\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eImprecision\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eOther considerations\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCognitive\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e[placebo]\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRelative\u003c/p\u003e\n \u003cp\u003e(95% CI)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"4\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colspan=\"15\"\u003e\n \u003cp\u003eCognitive Outcome x Hybrid Format x Immediate Post-Intervention (\u0026lt;\u0026thinsp;1-month) Time Period\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"1\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(6/9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1116/1116 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1030/1030 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eSMD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e-0.09\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.25 to 0.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003ea,b,c\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"15\"\u003e\n \u003cp\u003e\u003cstrong\u003eCognitive Outcome x Virtual Format x Immediate Post-Intervention (\u0026lt;\u0026thinsp;1-month) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(5/7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ed\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ee\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003ef\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e702/702 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e713/713 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eSMD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e-0.25\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.53 to 0.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003ed,e,f\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"15\"\u003e\n \u003cp\u003e\u003cstrong\u003eCognitive Outcome x Hybrid Format x Follow-Up (\u0026ge;\u0026thinsp;1-month to \u0026le;\u0026thinsp;6-months) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(4/4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003eg\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003eh\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eserious\u003csup\u003ei\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e564/564 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e580/580 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eSMD\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e-0.05\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.29 to 0.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁◯◯◯\u003c/p\u003e\n \u003cp\u003eVery low\u003csup\u003eg,h,i\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"15\"\u003e\n \u003cp\u003e\u003cstrong\u003eCognitive Outcome x Virtual Format x Follow-Up (\u0026ge;\u0026thinsp;1-month to \u0026le;\u0026thinsp;6-months) Time Period\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003erandomised trials\u003c/p\u003e\n \u003cp\u003e(2/4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003every serious\u003csup\u003ej\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003csup\u003ek\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enot serious\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003enone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e226/226 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e228/228 (100.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003e*SMD \u0026minus;\u0026thinsp;0.30\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e(-0.49 to\u003c/p\u003e\n \u003cp\u003e-0.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003e⨁⨁◯◯\u003c/p\u003e\n \u003cp\u003eLow\u003csup\u003ej,k\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eCRITICAL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003e\u003cstrong\u003eCI\u003c/strong\u003e: confidence interval; \u003cstrong\u003eSMD\u003c/strong\u003e: standardised mean difference\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\n \u003ch2\u003eExplanations\u003c/h2\u003e\n \u003cp\u003e*Indicates significant effect size\u003c/p\u003e\u003cspan\u003e\n \u003cp\u003ea. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (4/9 had high risk of bias overall and 5/9 had unclear risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003eb. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I\u0026sup2; statistic of 72%.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ec. We downgraded the certainty of evidence by one level for imprecision, as the confidence interval (\u0026ndash;0.25 to 0.08) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ed. We downgraded the certainty of evidence by two levels for serious study limitations as most included studies had high risk of bias (5/7 had high risk of bias overall, 1/7 had unclear risk of bias overall, and 1/7 had low risk of bias overall) because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ee. We downgraded the certainty of evidence by two levels for inconsistency due to heterogeneity among studies, with an I\u0026sup2; statistic of 78%.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ef. We downgraded the certainty of evidence by one level for imprecision, as the confidence interval (\u0026ndash;0.53 to 0.02) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003eg. We downgraded the certainty of evidence by one level for serious study limitations as all included studies (4/4) had unclear risk of bias because of potential issues with methods, i.e. blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003eh. We downgraded the certainty of evidence by one level for inconsistency due to heterogeneity among studies, with an I\u0026sup2; statistic of 75\u003c/p\u003e\n \u003c/span\u003e\u003cspan\u003e\n \u003cp\u003eWe downgraded the certainty of evidence by one level for imprecision, as the confidence interval (\u0026ndash;0.29 to 0.20) includes values favouring both the intervention and control groups, limiting the conclusiveness of the effect estimate.\u003c/p\u003e\n \u003c/span\u003e\u003cspan\u003e\n \u003cp\u003ej. We downgraded the certainty of evidence by two levels for serious study limitations as all included studies (4/4) had high risk of bias because of potential issues with methods, i.e. random sequence generation, allocation concealment, blinding of personnel and outcome assessors, and attrition.\u003c/p\u003e\n \u003c/span\u003e \u003cspan\u003e\n \u003cp\u003ek. I\u0026sup2; statistic of 0%.\u003c/p\u003e\n \u003c/span\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\n \u003ch2\u003eFinal Results Summary\u003c/h2\u003e\n \u003cdiv class=\"gridtable\"\u003e\u003cimg src=\"https://myfiles.space/user_files/69519_bce2c0439cd956a6/69519_custom_files/img1749796459.png\"\u003e\u003cbr\u003e\u003c/div\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis meta-analysis examined the effectiveness of universal DMHIs for children and youth across emotional, behavioural, social, and cognitive outcomes. Of the 16 meta-analyses conducted, nine yielded statistically significant \u0026ndash; but small \u0026ndash; effects. The largest effect sizes, though still considered small, were observed for virtual interventions targeting emotional and behavioural outcomes immediately post-intervention, and for cognitive outcomes at follow-up. However, hybrid interventions showed the most consistent pattern of effectiveness, with small but significant effects observed across emotional, behavioural, and social outcomes at both the immediate post-intervention and follow-up time periods.\u003c/p\u003e \u003cp\u003eThe highest level of evidence certainty \u0026ndash; rated moderate according to the GRADE framework \u0026ndash; was observed for hybrid interventions at follow-up. All other findings were supported by low or very low certainty, largely due to methodological limitations, inconsistency across studies, and imprecise effect estimates. Six of the nine statistically significant outcomes were found in hybrid interventions, underscoring their broader impact and greater potential for sustained effects compared to fully virtual formats.\u003c/p\u003e \u003cp\u003eThis pattern likely reflects the added structure, relational support, and accountability that hybrid interventions provide. These programs often take place within school or community contexts that encourage adult oversight and peer interaction, even when self-led. For example, in Pisani et al. (2024) [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], the self-guided Text4Strength texting program was embedded within a broader school-based initiative, balancing independent engagement with opportunities for connection and support. In contrast, fully virtual interventions were more often delivered in unsupervised settings, which may reduce engagement and sustained impact. These observations align with prior research showing that DMHIs delivered within guided or structured environments tend to produce stronger engagement and adherence [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. For large-scale implementation, embedding digital interventions into existing institutions may help optimize both reach and effectiveness.\u003c/p\u003e \u003cp\u003eThe clearest and most consistent effects in this meta-analysis emerged for emotional and behavioural outcomes, followed by social outcomes \u0026ndash; mirroring a broader pattern in the DMHI literature. For years, DMHIs have largely focused on reducing symptoms of anxiety, stress, and externalizing behaviours in children and youth [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. These emotional and behavioural domains are not only more frequently targeted in intervention design, particularly in cognitive behavioural therapy (CBT)-based programs, but are also more readily measured through validated self-report tools. As a result, effects in these areas are easier to detect, both statistically and practically.\u003c/p\u003e \u003cp\u003eIn contrast, cognitive outcomes appear to have been relegated to secondary status \u0026ndash; understudied, inconsistently measured, and overshadowed by more established clinical targets. This reflects persistent gaps in the literature, as noted in previous reviews [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], and highlights a critical limitation in the current evidence base. Importantly, the absence of consistent findings in this domain does not imply that digital interventions are ineffective for cognitive development; in fact, the largest effect size \u0026ndash; albeit small \u0026ndash; was observed for the cognitive outcome.\u003c/p\u003e \u003cp\u003eExpanding the focus of DMHIs to directly and intentionally engage with social connection and cognitive skill-building is not just a research imperative; it is essential to understanding how these tools can support the full spectrum of child and youth development. Until high-quality studies fill these gaps, practitioners should be cautious about assuming that improvements in anxiety or mood will automatically extend to academic performance or interpersonal functioning.\u003c/p\u003e \u003cp\u003eAlthough virtual interventions showed encouraging results \u0026ndash; particularly for emotional and behavioural outcomes immediately after use, and for cognitive outcomes at follow-up \u0026ndash; the strength of this evidence remains limited. These findings were rated as low or very low certainty, primarily due to small sample sizes, high risk of bias, and notable inconsistency across studies. These limitations speak to a broader and ongoing challenge in the digital mental health landscape: ensuring that virtual tools are not only accessible, but also effective when delivered outside of structured, supportive environments.\u003c/p\u003e \u003cp\u003eUnlike hybrid interventions, which are often embedded within school or community settings that provide some form of supervision, many virtual programs are accessed independently and without the scaffolding of adult oversight or peer engagement. These contextual differences may explain why hybrid interventions tended to produce more lasting effects across multiple outcome domains.\u003c/p\u003e \u003cp\u003eWhile many digital tools boast features like interactivity, gamification, reminders, or personalization \u0026ndash; elements shown to improve user engagement \u0026ndash; these features are inconsistently incorporated and unevenly evaluated. As the field continues to evolve, the literature increasingly reflects a shared understanding: virtual interventions hold great promise, but their real-world effectiveness depends on thoughtful design and intentional implementation strategies [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Without those, even the most innovative platforms may struggle to reach their full potential.\u003c/p\u003e \u003cp\u003ePerhaps one of the most pressing stories told by this meta-analysis is not just what the evidence reveals, but who is missing from the evidence altogether. Adding to this concern is the overwhelming dominance of studies conducted in high-income, Western countries. While these regions have been instrumental in advancing digital innovation, their prominence risks sidelining the unique needs and lived experiences of communities in lower-resource settings. If digital mental health tools are to fulfill their promise, they must be adaptable \u0026ndash; not only across devices, but across cultures, languages, and systems of care.\u003c/p\u003e \u003cp\u003eTo bridge this divide, future research must move beyond simply expanding access. It must actively prioritize equity and contextual relevance. This means improving demographic transparency, co-designing interventions with underrepresented communities, and ensuring that digital tools are tested and refined in diverse real-world settings. Only then can we begin to realize a vision of digital mental health that is for all.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis meta-analysis offers emerging evidence that DMHIs can positively support emotional, behavioural, social, and cognitive well-being in children and youth. However, the promise of these tools remains largely unrealized. Small effect sizes and low quality evidence highlight an urgent need to better understand how to design and deliver interventions that can achieve stronger, more lasting impacts \u0026ndash; especially as rates of mental distress among children and youth continue to rise.\u003c/p\u003e \u003cp\u003eAcross outcomes, hybrid interventions stood out \u0026ndash; particularly at follow-up \u0026ndash; for producing more consistent effects in emotional, behavioural, and social domains. These findings, supported by moderate-certainty evidence, suggest that embedding digital tools within structured settings such as schools or community programs may enhance and sustain their impact. By contrast, although virtual interventions showed promise in certain domains, their effectiveness was supported by lower-certainty evidence, largely due to limitations in study design, inconsistency, and imprecision.\u003c/p\u003e \u003cp\u003e Taken together with findings from our prior scoping review, these results reinforce both the promise and the limitations of the current evidence base. Key gaps \u0026ndash; such as the underrepresentation of younger children, lack of demographic transparency, and limited inclusion of marginalized populations \u0026ndash; raise ongoing concerns about equity, reach, and contextual fit.\u003c/p\u003e \u003cp\u003eTo increase the magnitude of effect, future research must go beyond tool development to focus on the elements that support effective use. This includes embedding digital interventions in real-world environments, designing for sustained engagement, and ensuring interventions are not only developmentally relevant and culturally responsive, but also grounded in strong methodological rigor. Evaluating outcomes that reflect the full range of children\u0026rsquo;s lived experiences \u0026ndash; and doing so in a way that is transparent, powered, and scalable \u0026ndash; will be critical for moving from potential to impact.\u003c/p\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eWhile this meta-analysis offers a comprehensive synthesis of the existing evidence on universal DMHIs for children and youth, several limitations should be acknowledged.\u003c/p\u003e \u003cp\u003eFirst, the overall certainty of evidence was low to very low across most outcomes, with only a few analyses rated as moderate certainty. No findings were supported by high-certainty evidence. This limits the confidence with which conclusions can be drawn. The low certainty was primarily due to small sample sizes, heterogeneity in intervention delivery, inconsistency across studies, and the methodological quality of trials, many of which had a high or unclear risk of bias. As recommended by GRADE and Cochrane guidelines, these factors significantly reduce the reliability of pooled effect estimates. Despite using the best available data, the findings remain subject to considerable uncertainty and should be interpreted with caution.\u003c/p\u003e \u003cp\u003eSecond, although intervention format and setting were clearly reported across studies, there was considerable variation in how hybrid and virtual interventions were implemented and evaluated. For example, interventions varied in the degree of facilitator involvement and delivery structure. Additionally, the inclusion of both randomized and non-randomized study designs may have introduced further heterogeneity. These factors complicate direct comparisons and may have contributed to imprecision in effect estimates.\u003c/p\u003e \u003cp\u003eThird, this review included only peer-reviewed literature, which may have resulted in the exclusion of relevant unpublished or grey literature, potentially introducing publication bias. Similarly, eligibility criteria were narrowly defined to ensure comparability of outcomes, but this may have led to the exclusion of studies with valuable but differently framed data.\u003c/p\u003e \u003cp\u003e Taken together, these limitations extend the gaps identified in our prior scoping review, which found persistent underrepresentation of early childhood populations and historically marginalized communities in DMHI research. Furthermore, the predominance of studies conducted in high-income countries, with limited representation from low-resource settings, raises concerns about generalizability and global applicability.\u003c/p\u003e \u003cp\u003eDespite these constraints, this meta-analysis provides a valuable overview of how DMHIs are currently being implemented in universal child and youth populations. Addressing these limitations through larger, higher-quality trials, more inclusive research designs, and better reporting standards will be essential for advancing the field and scaling equitable, evidence-based DMHIs.\u003c/p\u003e \u003c/div\u003e "},{"header":"Methods","content":"\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003cdiv id=\"Sec20\" class=\"Section3\"\u003e \u003ch2\u003eRegistration and protocol\u003c/h2\u003e \u003cp\u003e Before its initiation, our systematic review was registered at the International Platform of Registered Systematic Review and Meta-analysis Protocols (INPLASY Protocol 6765). Our reporting adheres to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) statement reporting guidelines [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] (See Supplementary Table\u0026nbsp;2).\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eSearch strategy and eligibility criteria\u003c/h2\u003e \u003cp\u003eA comprehensive systematic search was conducted to identify peer-reviewed studies evaluating DMHIs for universal populations of children and youth. The search strategy was developed in collaboration with an academic librarian and focused on three core components: (1) universal scope, (2) digital intervention format, and (3) child and youth mental health outcomes.\u003c/p\u003e \u003cp\u003eSearches were conducted in three electronic databases: MEDLINE, PsycInfo, and Embase. An initial scan of Google Scholar helped identify key articles and inform the final list of keywords, which was refined to ensure compatibility across platforms. The initial search was conducted between June 26 and July 5, 2024, and an updated search using the same strategy was performed on January 6, 2025, to capture studies published between July 3, 2024, and January 6, 2025. A full list of search terms is provided in Supplementary Table\u0026nbsp;3.\u003c/p\u003e \u003cp\u003eTo be eligible for inclusion, studies had to meet the following criteria:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003ePublication Type: Only full-text, peer-reviewed original research articles were included, as these represent the highest standard for evidence-based practice. Book chapters, case studies, conference abstracts, and dissertations were excluded.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eParticipants: Studies were eligible if the minimum or mean age of participants was between 0 and 18 years. Studies with some participants over 18 were included if the average age fell within this range. Studies exclusively involving adults or not specifying participant age were excluded.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eMental Health Status: Only studies involving non-clinical populations were included. Studies focusing on participants with acute psychiatric or medical conditions (e.g., hospitalization for suicide attempt or cancer treatment) or those recently discharged from such care were excluded.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eIntervention Format: Eligible studies included digital mental health interventions delivered either entirely online (virtual) or through a hybrid model combining digital tools with in-person components, such as school-based facilitation. Studies that relied solely on in-person interventions without a digital component were excluded.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eOutcomes: Studies were included if they reported at least one outcome related to emotional, behavioural, social, or cognitive well-being. Studies measuring only physiological outcomes without a corresponding psychological component were excluded. When multiple outcomes were reported, the most psychometrically robust measure per domain was prioritized for extraction.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eStudy selection\u003c/h2\u003e \u003cp\u003eAll search results were imported into Covidence, a systematic review management platform, to facilitate de-duplication, title and abstract screening, full-text review, and data extraction. Records with complete indexing were automatically categorized by database source, while entries with incomplete metadata (e.g., manually imported records or non-indexed citations) were labeled as \u0026ldquo;unspecified sources.\u0026rdquo;\u003c/p\u003e \u003cp\u003e Screening was carried out by a four-person review team. A random 25% subset of records was double-screened, with each record independently reviewed by two team members to assess consistency in screening decisions. The remaining records were single-screened. A fifth senior author provided oversight, helped resolve conflicts, and supported decision-making.\u003c/p\u003e \u003cp\u003eFull-text articles deemed potentially eligible were retrieved and independently assessed by two reviewers. Disagreements were resolved through discussion or, if needed, in consultation with the senior author. Reference lists from included studies and relevant reviews were also manually searched to identify any additional eligible articles.\u003c/p\u003e \u003cp\u003e Agreement between reviewers was high across both screening stages, with Cohen\u0026rsquo;s kappa values ranging from 0.84 to 0.98 for title and abstract screening, and 0.90 to 0.96 for full-text review.\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003eData extraction\u003c/h2\u003e \u003cp\u003eData extraction was conducted using Microsoft Excel. Variables were initially selected based on the PICO framework (Population, Intervention, Comparator, Outcomes) and refined iteratively as familiarity with the included studies developed. For each study, reviewers extracted information on study design, publication year, and participant characteristics, including age range and lived experience of marginalization. Intervention characteristics were also captured, including delivery format (hybrid or virtual), setting (e.g., school, home, community), facilitation type (self-led, facilitator-led, or both), total number of sessions, duration in hours, assessment time periods (post-intervention or follow-up), intervention type (psychoeducational, skills-building, or both), and any reported information on facilitator training or qualifications. Mental health outcomes were categorized into emotional, behavioural, social, or cognitive domains, and the type of measurement tools and reporting methods were recorded (e.g., whether measures were validated or unvalidated, and whether data were collected via self-report, caregiver-report, or teacher-report).\u003c/p\u003e \u003cp\u003e The data extraction template was piloted and refined to ensure clarity and consistency across reviewers. For all included full-text articles, data were independently extracted by two reviewers, with discrepancies resolved through discussion. Descriptive statistics, including frequencies, medians, and ranges, were used to summarize study and intervention characteristics. Data were presented in both tabular and graphical formats where appropriate. Inter-rater agreement during title and abstract screening and full-text review was calculated to assess consistency.\u003c/p\u003e \u003cp\u003eWhere meta-analysis was feasible, SMDs were calculated to account for variability in outcome measures across studies. Random-effects models were used to estimate pooled effect sizes, and heterogeneity was assessed using the I\u0026sup2; statistic and Chi\u0026sup2; test, following Cochrane guidelines.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eRisk of bias assessment\u003c/h2\u003e \u003cp\u003eTo assess the methodological quality of included studies, two reviewers independently applied established appraisal tools. For randomized studies, the Cochrane Risk of Bias 2 (RoB 2) tool was used, classifying risk as low, some concerns, or high. The ROBINS-I tool (Risk of Bias in Non-randomized Studies of Interventions) tool was used for non-randomized studies, categorizing risk of bias as low, moderate, or serious. Discrepancies between reviewers were resolved through discussion until consensus was reached. A detailed summary of risk of bias assessments is provided in the methodological quality section above.\u003c/p\u003e \u003cdiv id=\"Sec25\" class=\"Section3\"\u003e \u003ch2\u003eMeta-analysis\u003c/h2\u003e \u003cp\u003eAll meta-analyses were conducted using the Generic Inverse Variance method in Review Manager (RevMan), version 5.4 [\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://training.cochrane.org/online-learning/core-software-cochrane-reviews/revman\u003c/span\u003e\u003cspan address=\"https://training.cochrane.org/online-learning/core-software-cochrane-reviews/revman\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e]. This approach was applied consistently across all 16 outcome-level meta-analyses, which included studies with varying designs (e.g., RCTs, quasi-experimental, and single-group pre-post). Effect sizes were reported as SMDs with 95% confidence intervals (CIs).\u003c/p\u003e \u003cp\u003eFor non-randomized studies pre-post studies, we handled the data based on how each individual study reported their pre- and post-intervention measurements. When change scores or sufficient data were reported, we treated the data as paired by entering the same N for pre and post. When only pre and post means, SDs, and sample sizes were reported (without change scores or pre-post correlations), we entered the data as reported [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], treating pre and post as separate groups in the analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section3\"\u003e \u003ch2\u003eAssessment of the certainty of the evidence\u003c/h2\u003e \u003cp\u003eWe assessed confidence in the findings using the GRADE framework, which evaluates five domains: risk of bias, inconsistency, indirectness, imprecision, and other considerations. All domains were assessed except other considerations, which includes publication bias; this was not formally evaluated due to the limited number of studies per outcome, as standard methods for detecting publication bias (e.g., funnel plots) are unreliable with fewer than 10 studies.\u003c/p\u003e \u003cp\u003eGiven the limited evidence base for many outcomes, particular attention was paid to risk of bias, inconsistency, imprecision, and indirectness.\u003c/p\u003e \u003cp\u003eRisk of bias was evaluated using the Cochrane RoB 2 tool for randomized trials and the ROBINS-I tool for non-randomized studies. Outcomes were downgraded when most contributing studies were rated as high or unclear risk of bias.\u003c/p\u003e \u003cp\u003eInconsistency (heterogeneity) was evaluated using the I\u0026sup2; statistic, with thresholds based on GRADE guidance:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eNot serious (I\u0026sup2; \u0026lt; 25%) \u0026ndash; Low variability between studies\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eSerious (I\u0026sup2; = 25\u0026ndash;50%) \u0026ndash; Moderate variability\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eVery serious (I\u0026sup2; \u0026gt; 50%) \u0026ndash; High variability, indicating substantial inconsistency across study results\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eInterpretation of I\u0026sup2; was supported by examining the direction and overlap of confidence intervals, as well as the number and quality of contributing studies.\u003c/p\u003e \u003cp\u003eImprecision was evaluated based on the width of confidence intervals, the total number of participants, and the number of studies contributing to each pooled estimate. Outcomes with wide CIs or limited data were downgraded accordingly.\u003c/p\u003e \u003cp\u003eIndirectness was not a major concern, due to the specificity of our inclusion criteria. By focusing exclusively on universal Tier 1 DMHIs for children and youth, the included studies were closely aligned with the population, intervention, and outcomes of interest.\u003c/p\u003e \u003cp\u003eWe also conducted sensitivity analyses using a leave-one-out approach to assess the influence of individual studies on pooled estimates and to test the stability of the results.\u003c/p\u003e \u003cp\u003eEach outcome was ultimately assigned a certainty rating of high, moderate, low, or very low, reflecting the overall strength of evidence and our confidence in the findings.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":" \u003cdiv id=\"Sec27\" class=\"Section3\"\u003e \u003ch2\u003eData availability\u003c/h2\u003e \u003cp\u003eAll data generated or analyzed during this study is available from the corresponding author upon reasonable request.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003eCode availability\u003c/h2\u003e \u003cp\u003eNo custom code was used for the analyses. All meta-analyses were conducted using Review Manager (RevMan) version 5.4, available from the Cochrane Collaboration: [\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://training.cochrane.org/online-learning/core-software\u003c/span\u003e\u003cspan address=\"https://training.cochrane.org/online-learning/core-software\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e\u003cp\u003e \u003ch2\u003eCompeting Interests\u003c/h2\u003e \u003cp\u003eThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eKDP co-conceptualized the study, led the systematic review and meta-analysis, and drafted the initial manuscript. RPR conceptualized the study, assisted with the analysis, and supported the development of the initial manuscript. OB, HH, and AL assisted with the analysis and offered editorial feedback on the submitted manuscript. AL contributed to the development of the study\u0026rsquo;s conceptual foundation. RC, as the statistical consultant for the meta-analysis, assisted with the analysis and provided editorial feedback on the submitted manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eWe would like to thank the Jackman Foundation and York University (funding dedicated to DIVERT Mental Health collaboration, a CIHR funded Health Research Training Platform) for funding that supported this research project.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eOrganisation for Economic Co-operation and Development. \u003cem\u003eChildren and young people\u0026rsquo;s mental health in the digital age: shaping the future\u003c/em\u003e. OECD Publishing; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1787/9789264318706-en\u003c/span\u003e\u003c/span\u003e (2018).\u003c/li\u003e\n\u003cli\u003eLiverpool, S. et al. Engaging children and young people in digital mental health interventions: systematic review of modes of delivery, facilitators, and barriers. \u003cem\u003eJ. Med. Internet Res\u003c/em\u003e. 22, e16317 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2196/16317\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n\u003cli\u003eClarke, A. M., Kuosmanen, T. \u0026amp; Barry, M. M. A systematic review of online youth mental health promotion and prevention interventions. \u003cem\u003eJ. Youth Adolesc\u003c/em\u003e. 44, 90\u0026ndash;113 (2015).\u003c/li\u003e\n\u003cli\u003eHollis, C. et al. Annual Research Review: digital health interventions for children and young people with mental health problems - a systematic and meta-review. \u003cem\u003eJ. Child Psychol. Psychiatry\u003c/em\u003e 58, 474\u0026ndash;503 (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/jcpp.12663\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n\u003cli\u003eGrist, R. et al. Technology delivered interventions for depression and anxiety in children and adolescents: a systematic review and meta-analysis. \u003cem\u003eClin. Child Fam. Psychol. Rev.\u003c/em\u003e 22, 147\u0026ndash;171 (2019).\u003c/li\u003e\n\u003cli\u003eGarrido, S. et al. What works and what doesn\u0026rsquo;t work? A systematic review of digital mental health interventions for depression and anxiety in young people. \u003cem\u003eFront. Psychiatry\u003c/em\u003e 10, 759 (2019).\u003c/li\u003e\n\u003cli\u003ePiers, C., Monks, H., McLoughlin, L. T. \u0026amp; Grov\u0026eacute;, C. Can digital mental health interventions bridge the \u0026lsquo;digital divide\u0026rsquo; for socioeconomically and digitally marginalised youth? A systematic review. \u003cem\u003eChild Adolesc. Ment. Health\u003c/em\u003e 28, 54\u0026ndash;65 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/camh.12620\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n\u003cli\u003eSterne, J. A. C. et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. \u003cem\u003eBMJ\u003c/em\u003e 366, l4898; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/bmj.l4898\u003c/span\u003e\u003c/span\u003e (2019).\u003c/li\u003e\n\u003cli\u003eSterne, J. A. C. et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. \u003cem\u003eBMJ\u003c/em\u003e 255, i4919; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/bmj.i4919\u003c/span\u003e\u003c/span\u003e (2016).\u003c/li\u003e\n\u003cli\u003eHassen, H. M. et al. Effectiveness and implementation outcome measures of mental health curriculum intervention using social media to improve the mental health literacy of adolescents. \u003cem\u003eJ. Multidiscip. Healthc.\u003c/em\u003e 15, 979\u0026ndash;997 (2022).\u003c/li\u003e\n\u003cli\u003ePeuters, C. et al. A mobile healthy lifestyle intervention to promote mental health in adolescence: A mixed-methods evaluation. \u003cem\u003eBMC public health\u003c/em\u003e 24, 44 (2024).\u003c/li\u003e\n\u003cli\u003eSousa, P. et al. Controlled trial of an mHealth intervention to promote healthy behaviours in adolescence (TeenPower): effectiveness analysis. \u003cem\u003eJ. Adv. Nurs.\u003c/em\u003e 76, 1057\u0026ndash;1068 (2020).\u003c/li\u003e\n\u003cli\u003ePisani, A.R. et al. Text4Strength: a brief automated text messaging program to promote help-seeking and reduce mental health stigma among high school students. \u003cem\u003eJ. Child Psychol. Psychiatry\u003c/em\u003e 65, 58\u0026ndash;69 (2024).\u003c/li\u003e\n\u003cli\u003eFleming, T. et al. Beyond the trial: systematic review of real-world uptake and engagement with digital self-help interventions for depression, anxiety, or emotional distress. \u003cem\u003eJ. Med. Internet Res\u003c/em\u003e. 20, e199 (2018).\u003c/li\u003e\n\u003cli\u003eOrlowski, S. et al. A systematic review of mental health interventions for social and emotional well-being in Australian indigenous communities. \u003cem\u003eBMC Public Health\u003c/em\u003e 16, 1039 (2016).\u003c/li\u003e\n\u003cli\u003eBeames, J. R. et al. Effect of engagement with digital interventions on mental health outcomes: a systematic review and meta-analysis. \u003cem\u003eFront. Digit. Health\u003c/em\u003e 3, 764079 (2021).\u003c/li\u003e\n\u003cli\u003ePage, M. J. et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. \u003cem\u003eBMJ\u003c/em\u003e 372, n71 (2021).\u003c/li\u003e\n\u003cli\u003eGoodman, R., \u0026amp; Newman, D. Testing a digital storytelling intervention to reduce stress in adolescent females. \u003cem\u003eStorytelling, Self, Society\u003c/em\u003e, 10, 177\u0026ndash;193 (2014).\u003c/li\u003e\n\u003cli\u003eSubotic-Kerry, M. et al. Examining the impact of a universal positive psychology program on mental health outcomes among Australian secondary students during the COVID-19 pandemic. \u003cem\u003eChild Adolesc. Psychiatry Ment. Health\u003c/em\u003e 17, 70 (2023).\u003c/li\u003e\n\u003cli\u003eMorawska, A. \u0026amp; Sanders, M. R. Self-administered behavioural family intervention for parents of toddlers: Effectiveness and dissemination. \u003cem\u003eBehav. Res. Ther.\u003c/em\u003e 44, 1839\u0026ndash;1848 (2006).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"npj-digital-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjdigitalmed","sideBox":"Learn more about [npj Digital Medicine](http://www.nature.com/npjdigitalmed/)","snPcode":"41746","submissionUrl":"https://submission.springernature.com/new-submission/41746/3","title":"npj Digital Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6811900/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6811900/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eChildren and youth are facing rising mental health challenges in an increasingly digital world. Universal Tier 1 digital mental health interventions (DMHIs) \u0026ndash; low-intensity, high-reach supports for non-clinical populations \u0026ndash; have emerged as scalable tools to meet early-stage needs, but their effectiveness remains unclear. This systematic review and meta-analysis synthesized evidence from 51 studies (N\u0026thinsp;=\u0026thinsp;30,474) to assess the impact of DMHIs across emotional, behavioural, social, and cognitive outcomes, and examined whether delivery format (hybrid vs. virtual) and outcome timing (post-intervention vs. follow-up) influenced results. Nine of 16 calculated effect sizes found statistically significant, small effects. Hybrid interventions \u0026ndash; those incorporating any in-person component \u0026ndash; showed more consistent and sustained benefits, with moderate-certainty evidence at follow-up. However, most findings were low certainty. Parents of younger children and marginalized children and youth were notably understudied. Future research must prioritize marginalized youth, broaden outcome evaluation, and deliver large-scale, high-quality trials to inform the inclusive and sustainable implementation of digital mental health supports.\u003c/p\u003e \u003cp\u003eINPLASY registration number: INPLASY202490026\u003c/p\u003e","manuscriptTitle":"Effects of Universal Digital Mental Health Interventions for Children and Youth on Psychological Outcomes – A Systematic Review and Meta-Analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-13 06:47:14","doi":"10.21203/rs.3.rs-6811900/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-09-14T23:08:31+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-11T10:09:17+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"117598194981315185547176442112099902853","date":"2025-08-19T08:33:41+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"233343053775606061294213260573581294697","date":"2025-08-18T21:52:56+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-02T17:46:20+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"48094362780413488610152492411324507656","date":"2025-06-09T13:29:18+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-06-09T04:16:19+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-06-06T15:35:20+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-06-06T04:46:11+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Digital Medicine","date":"2025-06-03T13:35:39+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"npj-digital-medicine","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjdigitalmed","sideBox":"Learn more about [npj Digital Medicine](http://www.nature.com/npjdigitalmed/)","snPcode":"41746","submissionUrl":"https://submission.springernature.com/new-submission/41746/3","title":"npj Digital Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"2606a267-38d9-47de-b60e-02ef97a217d4","owner":[],"postedDate":"June 13th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":49756233,"name":"Biological sciences/Psychology"},{"id":49756234,"name":"Health sciences/Health care"}],"tags":[],"updatedAt":"2026-05-02T14:53:37+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-13 06:47:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6811900","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6811900","identity":"rs-6811900","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}