Adolescence is a period of rapid physical, psychological and social development. Approximately 75% of mental illnesses emerge during this period, Reference Kim-Cohen, Caspi, Moffitt, Harrington, Milne and Poulton1 with many persisting into adulthood, producing significant long-term consequences for an individual’s social adjustment, physical health, overall functioning (e.g. sleep) and quality of life. Reference Clarke, Pote and Sorgenfrei2 In the UK, 1 in 6 adolescents (aged 11–16 years) have been identified as having a probable mental illness – a figure that has steadily increased in the past decade. Reference Newlove-Delgado, Williams, Robertson, McManus, Sadler and Vizard3 While the exact reasons for this increase are still uncertain, its effects on adolescents and society are a major concern for practitioners, researchers and policy-makers alike. Reference Gunnell, Kidger and Elvidge4 Current attempts to support this population are largely designed to target specific conditions (e.g. depression Reference Knapp and Wong5 ). However, 60% of adolescents with one diagnosable mental illness have one or more additional conditions. 6 Mental health comorbidity – the presence of two or more mental illnesses in an individual – is the rule rather than the exception, and has been associated with greater clinical severity and a poorer overall quality of life. Reference Dalgleish, Black, Johnston and Bevan7 Transdiagnostic interventions are designed to be directly effective across several mental illnesses, altering psychopathological processes common to multiple conditions (e.g. emotion regulation). Reference Roefs, Fried, Kindt, Martijn, Elzinga and Evers8 Emerging evidence suggests that this approach is effective in targeting diverse psychopathologies, activating a range of related, beneficial developmental cascades, including improvement in social and academic outcomes. Reference Forbes, Rapee and Krueger9 A transdiagnostic approach to treatment is also considered to be time- and cost-effective compared with disorder-specific strategies, and may offer a more sustainable alternative to treatments currently available to this population. Reference Dalgleish, Black, Johnston and Bevan7
Emotion regulation as a transdiagnostic mechanism
Emotion regulation has received increased attention in recent years as a transdiagnostic mechanism and clinical target in psychological treatment. Reference Aldao, Gee, De Los Reyes and Seager10 Although the concept of emotion regulation remains unclear by definition, it can be broadly understood as a goal-directed, multidimensional process wherein an individual monitors, evaluates and shapes their emotions when they have them, and how they internally experience or outwardly express them. Reference Aldao, Gee, De Los Reyes and Seager10,Reference D’Agostino, Covanti, Rossi Monti and Starcevic11 There have been several different conceptualisations of emotion regulation, but by far the most influential is the Process Model of Emotion Regulation. Reference Gross12 According to this model, an individual recognises an emotion regulation goal (e.g. to communicate to others; to modify behaviour), selects and finally implements specific emotion regulation strategies. Reference Gross13 Gross defined a set of five distinct emotion regulation processes occurring at different points in an emotional experience (see Fig. 1): situation selection, situation modification, attentional deployment, cognitive change and response modulation. Reference Gross13 Each can be understood to influence an individual’s emotional response in a way that can be interpreted as either adaptive (e.g. problem-solving, acceptance) or maladaptive (e.g. withdrawal, suppression Reference Gross12 ), depending on the context.
Gross’s Process Model of Emotion Regulation. Reference Gross12 Situation selection can be understood as an individual’s efforts to alter the likelihood of being in an emotion-evoking situation. Situation modification involves modifying a situation at the time to change its emotional impact. Attentional deployment involves directing attention towards or away from an emotion or its causes. Cognitive change enables reappraisal of a situation to change its emotional significance. Response modulation includes any efforts to modify the behavioural, experiential and physiological elements of an emotional response. Reference Gross12
Maladaptive patterns of emotional experience or expression are typically understood as emotion dysregulation and have physiological, cognitive and social consequences. Reference Gross12,Reference Gross13 Evidence demonstrates that emotion dysregulation is present across a range of psychopathologies, including both internalising (e.g. generalised anxiety disorder, major depressive disorder, dysthymia) and externalising disorders (e.g. attention-deficit hyperactivity disorder (ADHD), conduct disorder, oppositional defiant disorder). Reference Helland, Mellblom, Kjøbli, Wentzel-Larsen, Espenes and Engell14 Recent findings also indicate a significant shift in emotion regulation between ages 13 and 15 years (e.g. access to strategies, use of adaptive versus maladaptive strategies), suggesting that adolescence is a particularly vulnerable period in the development of emotion regulation. Reference Zimmermann and Iwanski15 Therefore, interventions targeting emotion regulation as a transdiagnostic construct central to the development and maintenance of psychopathology may reduce the risk and severity of adolescent psychopathology.
Emotion regulation as a clinical target
Existing psychological interventions adopt different approaches to improving emotion regulation. Some focus on reducing the use of emotion regulation strategies that may be understood as maladaptive, such as rumination (e.g. rumination-focused cognitive behavioural therapy, RF-CBT), Reference Watkins, Scott, Wingrove, Rimes, Bathurst and Steiner16 while others focus on increasing the use of strategies that may be understood as adaptive, such as acceptance (e.g. acceptance and commitment therapy (ACT Reference Hayes, Strosahl and Wilson17 )). Others move beyond modifying the use of specific strategies and instead focus on developing wider emotion regulation skills (e.g. identifying and labelling emotions, understanding the context in which emotions occur, applying distress tolerance techniques – for example, dialectical behaviour therapy (DBT Reference Linehan18 )). Much of the research to date has focused on the effectiveness of in-person emotion regulation interventions, despite a growing number of self-directed digital solutions (e.g. mobile apps) available for adolescent emotion regulation and psychopathology. Reference Moltrecht, Deighton, Patalay and Edbrooke-Childs19 Such interventions are led by the service user, with little to no support from anyone else (e.g. therapist, parent/carer), and aim to widen access to support for adolescents. Some attempts have been made to examine the effectiveness of digital interventions targeting emotion regulation in adolescents, and emerging findings demonstrate that, in general, such interventions (e.g. digital games, virtual reality therapies) are effective in improving emotion regulation. Reference Reynard, Dias, Mitic, Schrank and Woodcock20 However, the effectiveness of interventions delivered in this self-directed format is, as yet, unclear. It is thought that this approach has a greater capacity for innovation and engagement with adolescents, Reference Moltrecht, Patalay, Bear, Deighton and Edbrooke-Childs21 and the potential to extend effective care cost-effectively and sustainably, Reference Moltrecht, Patalay, Deighton and Edbrooke-Childs22 but more research is needed to determine how such interventions can be applied at scale to support this population.
Review objectives
This systematic review investigates evidence on current self-directed digital interventions developed for adolescents (aged 11–18 years), and their effects on emotion (dys-)regulation, psychopathology and functioning (e.g. academic achievement). The review provides an important extension to existing work that has thus far demonstrated the effectiveness of in-person or therapist-supported interventions available for young people (aged 6–24 years), Reference Helland, Mellblom, Kjøbli, Wentzel-Larsen, Espenes and Engell14,Reference Moltrecht, Deighton, Patalay and Edbrooke-Childs19 as well as the utility of a broad spectrum of digital emotion regulation interventions for adolescents. Reference Reynard, Dias, Mitic, Schrank and Woodcock20 This review takes a more specific focus to develop evidence on a burgeoning number of self-directed, scalable, digital mental health interventions available to adolescents with or without diagnosed psychopathology. This is a timely contribution, employing a systematic review approach to provide a robust insight into the existing evidence base that underpins the growing number of publicly available emotion regulation mobile apps developed to provide mental health support at scale to young people. Reference Eisenstadt, Liverpool, Infanti, Ciuvat and Carlsson23
Specifically, we sought to answer the following research questions:
-
(a) Are current self-directed digital interventions that target emotion regulation acceptable and feasible for use within an adolescent population?
-
(b) Do self-directed digital interventions that target emotion regulation in adolescents have benefits for psychopathology and overall functioning (e.g. academic achievement)?
-
(c) How do the components of current self-directed digital interventions map onto existing theory and models of emotion regulation?
Method
This systematic review was registered in PROSPERO, the International Prospective Register of Systematic Reviews (no. CRD42022385547). The primary amendment to this protocol Reference Thomson, Lawrence, Oliver, Wright and Hosang24 is the decision not to undertake a meta-analysis due to the methodological and clinical heterogeneity of the included studies. We also decided to include a new database (ACM Digital Library) after the review was registered. No further amendments were made. The review was conducted according to the procedure and requirements described in the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines. Reference Page, McKenzie, Bossuyt, Boutron, Hoffmann and Mulrow25 The completed PRISMA checklist is provided in Supplementary File 1 available at https://doi.org/10.1192/bjo.2025.10888.
Search strategy
The search strategy (provided in full in Supplementary File 1) was designed to identify all studies examining one or more self-directed digital mental health interventions for adolescents and that include at least one component to target emotion (dys-)regulation (i.e. strategy or skill). It searched for synonyms for the following three concepts: adolescents, emotion (dys-)regulation and self-directed digital interventions. The strings were combined based on the population (i.e. adolescents aged 11–18 years), intervention format (i.e. self-directed digital interventions) and intervention target (i.e. emotion regulation) of interest. Limitations were placed on the publication date (i.e. from 2010) and language (i.e. English). A body of research addressing this topic was identified before the search through word of mouth, key reference lists and simple searches to check iterations of the strategy. The search was revised until it was sufficiently sensitive to capture the pre-identified studies.
Information sources
Ten electronic databases were searched for studies published from 1 January 2010 to 13 November 2024: MEDLINE, PsycInfo, Global Health, Scopus, Web of Science Core Collection, EBSCO CINAHL, EBSCO ERIC, Ovid Embase, The Cochrane Central Register of Controlled Trials (CENTRAL) and ACM Digital Library. ‘Grey’ literature, such as preprints and theses, was also included in this review (databases: HMIC, EThOS, PsyArXiv, Trip, ClinicalTrials.gov). This search was updated periodically to identify any new relevant research from the selected databases (most recent search, 13 November 2024); subsequently, no further studies were identified for inclusion.
Eligibility criteria
Studies were screened according to set eligibility criteria (Table 1). Studies were eligible for inclusion if the intervention included at least one component to target emotion regulation (i.e. an emotion regulation strategy or skill), and if the study included a valid measure of emotion regulation. Studies that did not measure emotion regulation using a validated measure were excluded from this review, to protect against the inclusion of evidence that is not robust. Also excluded were studies that measured a related but distinct construct (e.g. self-regulation, coping), and those that did not provide any evidence on the effects of the intervention on emotion regulation. Studies must have reported primary research with direct contact or observation of individuals, but there were no further restrictions on study design, setting or geographical location.
Eligibility criteria for screening
Study selection
All identified studies were exported to Endnote 20.5 for macOS (Clarivate, London, UK; https://endnote.com/downloads/), and any duplicates were removed following a specified method Reference Falconer26 before being imported into Rayyan Reference Ouzzani, Hammady, Fedorowicz and Elmagarmid27 for further de-duplication and screening. Abstracts and titles of identified studies were screened by A.T. as a primary screener based on the eligibility criteria. A total of 50% of articles were jointly assigned for second screening at the title and abstract stage (E.L. screened 10%, E.S. screened 15%, J.L. screened 25%). There was strong agreement between screener pairs (E.L. + A.T., κ = 0.99; E.S. + A.T., κ = 0.99; J.L. + A.T., κ = 0.98). Disagreements (n = 35) regarding the inclusion of a study were discussed, and research articles reviewed, until a consensus was reached. Those studies that met the eligibility criteria entered the full-text screening stage for further checks against the eligibility criteria. Full-text screening allowed for the identification of those interventions that did not measure emotion (dys-)regulation directly. All articles were independently screened by A.T. and E.S. at the full-text stage. There was strong agreement between screener pairs (κ = 0.86). Disagreements (n = 1) regarding the inclusion of a study were discussed, and research articles reviewed, until a consensus was reached.
Data extraction and management
Data were extracted and collated from eligible studies by two independent reviewers (A.T. and E.L.), and tracked in Microsoft Excel using a structured coding form and associated coding manual. Information relating to study characteristics (e.g. author(s), publication date), participant characteristics (e.g. age, gender), digital intervention characteristics (e.g. name, focus) and relevant clinical and emotion dysregulation outcomes were extracted from each study. Data were first extracted on 16 August 2023. Study investigators were contacted for missing/unreported data or additional details, as required.
Outcomes
The primary outcome of this review is the change in emotion (dys-)regulation occurring as a result of participation in a self-directed digital intervention that addresses emotion (dys-)regulation. Emotion (dys-)regulation must be assessed and, where possible, using a valid and appropriate item, scale or measure (e.g. Child Social Behaviour Questionnaire), Reference Luteijn, Jackson, Volkmar and Minderaa28 including through clinical interviews or self-reported measures. An existing review of emotion (dys-)regulation assessment Reference Adrian, Zeman, Erdley, Lisa and Sim29 and similar reviews Reference Aldao, Nolen-Hoeksema and Schweizer30,Reference Sloan, Hall, Moulding, Bryce, Mildred and Staiger31 were used as guidance to determine a measure’s eligibility. The measure must have been cited as valid by at least one of these reviews to be adjudged eligible. Effect sizes were extracted for measures of emotion regulation (or calculated when data were available) and interpreted according to Cohen’s conventions. Reference Cohen32 Other outcomes of interest in this review were the change in psychopathology and functioning (e.g. academic achievement). Symptoms of psychopathology were assessed by any available valid and appropriate measure, including through clinical interview or self-reported measures. Where data were available, information about the acceptability and feasibility of the interventions was also collated.
Quality and risk of bias assessment
Information to determine any study bias was also collated. Two researchers (A.T. and E.L.) independently assessed the methodological quality of the included studies using the Effective Public Health Practice Project quality assessment tool (EPHPP). EPHPP is applicable to a range of quantitative study designs (e.g. case–control studies), and has been judged particularly suitable for systematic reviews on the effects of interventions/treatments. Reference Deeks, Dinnes, D’Amico, Sowden, Sakarovitch and Song33 Evidence has shown that EPHPP has good content and construct validity. Reference Deeks, Dinnes, D’Amico, Sowden, Sakarovitch and Song33,Reference Thomas, Ciliska, Dobbins and Micucci34
Data synthesis
Meta-analyses could not be undertaken due to the heterogeneity of interventions, study designs and outcome measures. There were also too few studies for synthesis into comparable groups. As such, a narrative synthesis of the results was conducted, guided by recommendations from Popay and colleagues. Reference Popay, Roberts, Sowden, Petticrew, Arai and Rodgers35 The effects of identified interventions are summarised based on the specific components employed to target emotion regulation, drawing attention to notable differences in the conceptualisations of emotion regulation as a clinical target (i.e. emotion regulation strategies, emotion regulation skills and deficits). Specific attention is also given to the acceptability and feasibility of the self-directed digital intervention format. Data from these studies, including significance and direction of effects, are presented in summary tables. Given the range of outcome measures and statistical approaches applied across studies, results should not be compared across interventions.
Results
A total of 9094 records were retrieved from the database searches (see Fig. 2 for the study selection process). Following the removal of 3113 duplicate records, the abstracts and titles of the remaining records were screened according to the eligibility criteria. At the abstract and title screening stage, 5907 records were excluded. In total, 74 papers entered the full-text screening stage, of which 6 met the eligibility criteria and provided sufficient data. Reference Hilt and Swords36–Reference Kuosmanen, Fleming, Newell and Barry41
Study selection.
Characteristics of included studies
The characteristics of the included studies are summarised in Table 2. Most (k = 5) were conducted in the USA and employed a sample of between 80 and 618 adolescents aged 12–20 years. Studies were a mixture of brief cohort (one group pre- + post-test) studies (k = 3 Reference Hilt and Swords36,Reference Mrazek, Mrazek, Reese, Kirk, Gougis and Delegard39,Reference Schnitker, Shubert, Ratchford, Lumpkin and Houltberg40 ) and randomised controlled studies (RCTs) (k = 3 Reference Hilt, Swords and Webb37,Reference Houck, Modrowski, Hadley, Barker, Myers and Bala38,Reference Kuosmanen, Fleming, Newell and Barry41 ), comparing the intervention with either an active or wait-list control. Two studies tested the same intervention – the first was a cohort study to test preliminary feasibility and acceptability Reference Hilt and Swords36 and the second was a pilot RCT to confirm the intervention’s effects. Reference Hilt, Swords and Webb37 Across most studies (k = 5), investigating the effects of a self-directed digital intervention in improving emotion regulation was the focus. However, for one study, although the primary focus was on mental health symptoms (depression and anxiety), emotion regulation was captured as a secondary outcome. Reference Kuosmanen, Fleming, Newell and Barry41 A range of measures were used to capture change in emotion regulation as a result of using the intervention (Table 2). Only half of the studies Reference Hilt and Swords36,Reference Hilt, Swords and Webb37,Reference Kuosmanen, Fleming, Newell and Barry41 investigated the impact of the intervention on improving psychopathology, but most measured at least one related functional outcome (i.e. stress, worry, life satisfaction, self-control), although the types of outcomes measured differed across studies. Reference Hilt and Swords36,Reference Houck, Modrowski, Hadley, Barker, Myers and Bala38–Reference Kuosmanen, Fleming, Newell and Barry41 Results from the assessment of risk of bias are shown in Table 2. The quality of most of the included studies was weak or moderate. The uncontrolled study design was one of the primary reasons for the lower quality ratings, as well as the lack of blinding and missing information about confounders in some studies. Full ratings for each included study are included in Supplementary File 2.
Study characteristics
ADS, Affect Dysregulation Scale; ASRI-P, Adolescent Self-Regulatory Inventory–Parent report; CDI, Children’s Depression Inventory; CEMS, Children’s Emotion Management Scales; CRSQ, Children’s Response Styles Questionnaire; DERS, Difficulties in Emotion Regulation Scale; ERBS-R, Emotion Regulation Behaviours Scale – Revised; ERC, Emotion Regulation Checklist; ERQ, Emotion Regulation Questionnaire; ERQ-CA, Emotion Regulation Questionnaire for Children and Adolescents; GAD-7, Generalised Anxiety Disorder Rating Scale; MASC, Multidimensional Anxiety Scale for Children; PSC, Paediatric Symptom Checklist; RCT, randomised controlled trial; SMFQ, Short Mood and Feeling Questionnaire.
a. Parent report.
Intervention characteristics
The characteristics of the included interventions are summarised in Table 3. All interventions were brief and were in the early stages of development and testing. The active intervention period in all studies lasted between 14 and 49 days, although some studies allowed participants to engage with the intervention beyond this point. Reference Hilt and Swords36,Reference Hilt, Swords and Webb37,Reference Kuosmanen, Fleming, Newell and Barry41 Some interventions offered shorter daily activities (1–15 min Reference Hilt and Swords36,Reference Hilt, Swords and Webb37,Reference Schnitker, Shubert, Ratchford, Lumpkin and Houltberg40 ), while others comprised longer modules (20–45 min each) that participants completed once per week. Reference Houck, Modrowski, Hadley, Barker, Myers and Bala38,Reference Mrazek, Mrazek, Reese, Kirk, Gougis and Delegard39,Reference Kuosmanen, Fleming, Newell and Barry41
Intervention characteristics
CBT, cognitive–behavioural therapy; CD-ROM, compact disc read-only memory; iTRAC, talking about risk and adolescent choices; MBAT, mindfulness-based attention training; MBSR, mindfulness-based stress reduction.
a. Participants retained access to the intervention following termination of the active intervention period.
b. The duration was longer for some participants due to practical limitations (i.e. missed sessions).
All interventions were self-directed and delivered digitally, although the format varied among studies. Half of the interventions included (k = 3) were app based and could be accessed via smartphones whenever and wherever the participant chose. Reference Hilt and Swords36,Reference Hilt, Swords and Webb37,Reference Schnitker, Shubert, Ratchford, Lumpkin and Houltberg40 The remaining interventions, while self-directed, were delivered during a dedicated intervention session via tablet or computer in a classroom at school. Reference Houck, Modrowski, Hadley, Barker, Myers and Bala38,Reference Mrazek, Mrazek, Reese, Kirk, Gougis and Delegard39,Reference Kuosmanen, Fleming, Newell and Barry41
Effects of the interventions
Full details of the interventions and study findings are included in Supplementary File 2. The type of intervention, intervention components, outcomes reported, effect sizes and any significant results are summarised in Table 4.
Intervention components and outcomes reported
A, anxiety symptoms; AC, access; AW, awareness; C, coping; CL, clarity; CR, cognitive reappraisal; CT, control; D, depressive symptoms; DT, distress tolerance; ER, emotion regulation; ES, expressive suppression; IS, internalising symptoms; LS, life satisfaction; MBAT, mindfulness-based attention training; MBSR, mindfulness-based stress reduction; P, patience; PD, perceived level of demand; S, stress; SC, self-control; SM, stress management; W, worry; WB, well-being.
*, outcome reported but significance not measured; **, significant improvement in outcome found (P < 0.05); ***, significant deterioration in outcome found (P < 0.05); ****, no significant improvement in outcome found.
Interventions targeting antecedent-focused emotion regulation strategies: mindfulness and mood-monitoring
The most common component across most identified interventions was mindfulness – an approach that focuses on observing, describing, acting with awareness, non-judging and non-reactivity to emotional experiences. Reference Baer, Smith, Hopkins, Krietemeyer and Toney42 A total of four of the five interventions included a mindfulness component (i.e. breathing work, meditation) to support young people in regulating their emotions. Reference Hilt and Swords36,Reference Hilt, Swords and Webb37,Reference Mrazek, Mrazek, Reese, Kirk, Gougis and Delegard39–Reference Kuosmanen, Fleming, Newell and Barry41 Significant improvements to emotion regulation were observed in three of these interventions, Reference Hilt and Swords36,Reference Hilt, Swords and Webb37,Reference Mrazek, Mrazek, Reese, Kirk, Gougis and Delegard39,Reference Kuosmanen, Fleming, Newell and Barry41 while the fourth observed no significant improvements in emotion regulation over time. Reference Schnitker, Shubert, Ratchford, Lumpkin and Houltberg40 Despite the significant improvements observed, overall effect sizes were small to moderate (see Table 4). Notably, some adverse effects were observed. In the study by Hilt and colleagues, Reference Hilt, Swords and Webb37 a mindfulness-based intervention with a mood-monitoring component was compared with a mood-monitoring-only control. While significant reductions in rumination were observed overall (t[69.86] = 3.53, P < 0.001), a quarter of adolescents in the mindfulness condition experienced clinically significant worsening of rumination at the end of the intervention period. Reference Hilt, Swords and Webb37
Psychopathology was included as an outcome of just two of the interventions across three studies. The first intervention was tested across a cohort study Reference Hilt and Swords36 and a pilot RCT; Reference Hilt, Swords and Webb37 it included mindfulness and mood-monitoring components. Overall, although the intervention reduced anxiety symptoms in both studies, this reduction was non-significant in the pilot RCT. Reference Hilt, Swords and Webb37 In contrast, while the authors observed a significant reduction in depressive symptoms in the pilot RCT, Reference Hilt, Swords and Webb37 there was no significant impact of the intervention on depressive symptoms in the cohort study. Reference Hilt and Swords36
Interventions targeting antecedent-focused RCTs (strategies: problem-solving and behavioural avoidance)
In addition to mindfulness activities, the intervention by Mrazek and colleagues Reference Mrazek, Mrazek, Reese, Kirk, Gougis and Delegard39 included exercises to target other antecedent-focused emotion regulation strategies, such as problem-solving or behavioural avoidance. In this cohort study, the intervention was found to significantly improve emotion regulation in adolescents but the effect size was small overall (d = 0.29). After participating in the intervention, adolescents also showed significant changes in perceived stress management. Despite this, there were non-significant changes in actual stress and overall life satisfaction among adolescents. Reference Mrazek, Mrazek, Reese, Kirk, Gougis and Delegard39 Psychopathology was not included as an outcome in this study.
Interventions targeting multiple RCTs (strategies and related skills (e.g. emotional awareness))
Two interventions identified Reference Houck, Modrowski, Hadley, Barker, Myers and Bala38,Reference Kuosmanen, Fleming, Newell and Barry41 had a wider focus, targeting several different antecedent- and response-focused emotion regulation strategies (e.g. acceptance, expressive suppression). The first Reference Houck, Modrowski, Hadley, Barker, Myers and Bala38 included activities to develop the skills related to emotion regulation, such as emotional awareness and emotion controllability beliefs. This intervention was tested in a pilot RCT with a wait-list control, and was found to be beneficial in improving emotion regulation overall. Participants who completed this intervention perceived themselves as having better emotional competence and endorsed greater use of the emotion regulation strategies targeted by the intervention. Participants reported greater belief that emotions can be changed, greater awareness of their emotions, self-efficacy for managing emotions and perceived access to emotion regulation strategies. There were unexpected effects for behavioural measures of distress tolerance: participants who completed the intervention persisted for less time on tasks and showed reduced distress tolerance compared with wait-list controls. Psychopathology was not included as an outcome in this study.
The second intervention Reference Kuosmanen, Fleming, Newell and Barry41 targeted a broad spectrum of emotion regulation strategies and abilities through a cognitive behavioural therapy (CBT) and behavioural activation approach. Although one module was dedicated to emotion regulation, different strategies and abilities relevant to emotion regulation were included throughout (i.e. problem-solving, cognitive reappraisal, emotional awareness). The intervention was tested in a cluster RCT and demonstrated mixed effects for improving emotion regulation among participants. It was shown to be significant in improving expressive suppression when compared with a wait-list control. However, a non-significant decrease in cognitive reappraisal was observed. There were no significant intervention effects observed for either depressive or anxiety symptoms.
Feasibility and acceptability of the interventions
The acceptability and feasibility of these interventions are illustrated in Table 5. Because most interventions were in the earlier stages of their development, for many studies, understanding the acceptability and feasibility of the intervention was very much the focus.
Acceptability of the interventions
Most studies (k = 4) measured intervention acceptability among adolescents; to do so, various quantitative and qualitative methods were employed. Overall, the authors commented that their interventions were acceptable for use among adolescents, in keeping with previous evidence. Reference Moltrecht, Patalay, Deighton and Edbrooke-Childs22,Reference Liverpool, Mota, Sales, Čuš, Carletto and Hancheva43 Many adolescents felt that the interventions were simple to use and that the exercises were clear, helpful and easy to understand. One study also measured parents’ perceptions, and found that they similarly reported the beneficial impact of the intervention on their children. Reference Hilt and Swords36
In one of the studies, differences were observed in terms of the extent to which different intervention components were perceived as helpful. Reference Kuosmanen, Fleming, Newell and Barry41 Participants reported that the intervention was most helpful in confronting problems and recognising negative thoughts, but least helpful in changing behaviour. In this particular intervention, the least utilised activities were relaxation and mindfulness techniques.
Importantly, some participants had negative experiences with some of the interventions, either finding them not helpful at all or reporting that they had negatively impacted their mental health or ability to regulate their emotions. One study reported four adverse events among their participants; Reference Hilt, Swords and Webb37 all four events involved increased awareness of negative thoughts and feelings during the intervention period. Two participants reported this in each condition (mindfulness + mood-monitoring versus mood-monitoring alone); this may be related to monitoring negative emotions.
Discussion
The aim of this systematic review was to synthesise evidence on current self-directed, digital interventions that target emotion regulation, and to explore their benefits and acceptability among adolescents (aged 11–18 years). Overall, the evidence base was small. Just six papers were identified, which measured the effects of a total of five brief interventions for emotion regulation. There was wide clinical and methodological diversity across interventions and studies; each of the interventions included different components to target emotion regulation. These components have been mapped onto the Process Model of Emotion Regulation Reference Gross12 to better illustrate the diversity in approaches to targeting emotion regulation in adolescence (Fig. 3). Although Gross’s Process Model of Emotion Regulation Reference Gross12 suggests dynamic variability in the consequences of each of the proposed emotion regulation strategies (e.g. cognitive reappraisal), in practice, researchers and clinicians alike have tended to interpret these as either adaptive (e.g. mindfulness, acceptance) or maladaptive (e.g. withdrawal, suppression Reference Bonanno and Burton44 ). This has similarly directed intervention efforts, with the focus often being on developing interventions aimed at increasing the use of ‘adaptive’ strategies and/or decreasing the use of ‘maladaptive’ strategies.
Model of emotion regulation intervention targets. This figure shows Gross’s (1998) Process Model of Emotion Regulation. Examples of related emotion regulation skills are listed alongside. The included interventions are numbered from one to six. [1] Hilt & Swords, 2021, [2] Hilt et al, 2023, [3] Houck et al, 2022, [4] Mrazek et al, 2019, [5] Schnitker et al, 2021, [6] Kuosmanen et al, 2017. Those shown in the red box [1–2, 4–5] included intervention components that primarily targeted attention processes in emotion regulation. Those shown in blue [3, 6] targeted each emotion regulation process, from situation selection to response modulation, as well as related emotion regulation skills shown in the figure. NB: []*significant improvement in emotion regulation found in the present systematic review (P < 0.05).
This was reflected in the interventions identified in this review. The majority focused on increasing the use of adaptive strategy mindfulness to improve emotion regulation. Mindfulness-based interventions have gathered increasing attention within the field, Reference Goldin and Gross45 and are increasingly employed in schools to support adolescents with their emotional well-being and mental health. Reference Emerson, De Diaz, Sherwood, Waters and Farrell46 Although mindfulness was offered as a component across four of the five identified digital interventions, findings on its benefits were mixed. Despite some significant improvements in emotion regulation being observed, effect sizes were small overall. One study also reported that participants engaged with mindfulness techniques the least when compared with other intervention activities. Reference Kuosmanen, Fleming, Newell and Barry41 Of particular note in one study Reference Hilt, Swords and Webb37 was the finding that a quarter of adolescents who completed the mindfulness-based intervention experienced clinically significant worsening of rumination. This would align with existing evidence that mindfulness is not always a helpful strategy, particularly when applied universally to the adolescent population. Reference Kuyken, Ball, Crane, Ganguli, Jones and Montero-Marin47
This study also reported four adverse events among their participants; all four events involved increased awareness of negative feelings during the intervention period, supporting previous findings in adults that mood monitoring, in the absence of adequate strategies to process the emotions identified, can sometimes be harmful rather than helpful. Reference Palmier-Claus, Shryane, Taylor, Lewis and Drake48,Reference Faurholt-Jepsen, Frost, Ritz, Christensen, Jacoby and Mikkelsen49 Indeed, there is a growing consensus that the adaptiveness of any given emotion regulation strategy is context dependent rather than universal and depends on the appropriateness of the strategy to the specific situation in which it is used. Reference Bonanno and Burton44 There has been increased evidence that the use and functional benefits of any specific type of emotion regulation strategy will tend to vary across people and situations and, by extension, the most effective use of emotion regulation strategies is likely to be one that is most flexible. Reference Bonanno and Burton44,Reference Aldao and Nolen-Hoeksema50 Accordingly, interventions such as those proposed by Houck and colleagues Reference Houck, Modrowski, Hadley, Barker, Myers and Bala38 and Kuosmanen and colleagues, Reference Kuosmanen, Fleming, Newell and Barry41 which include activities to develop the related skills required to access certain strategies and use them flexibly, may be the most appropriate approach to supporting adolescents with emotion regulation and their mental health.
Few studies have measured the impact of targeting emotion regulation on mental health outcomes. The impact of these interventions on adolescent psychopathology was measured in just three studies, two of which investigated the same intervention. Reference Hilt and Swords36,Reference Hilt, Swords and Webb37 Overall findings were mixed: while reductions in symptoms of anxiety and depression were observed, these were non-significant across studies. Given the growing evidence on the importance of emotion regulation for psychopathology in adolescence, Reference Moltrecht, Deighton, Patalay and Edbrooke-Childs19 its omission as an outcome in most studies is surprising and limits current attempts to better understand its role in supporting adolescents. However, this may reflect the fact that most interventions were in the earlier stages of their development and, therefore, understanding the feasibility of the interventions in targeting emotion regulation was the primary focus for most studies. Reference Skivington, Matthews, Simpson, Craig, Baird and Blazeby51 Greater research is needed to understand the effects of targeting emotion regulation through self-directed digital interventions on adolescent psychopathology.
Clinical implications
This review identified a total of 6 studies that investigated the effects of 5 self-directed, digital interventions available to adolescents (aged 11–18 years). Despite large methodological and clinical heterogeneity, preliminary evidence from these studies suggests that these interventions showed benefits for emotion regulation. The self-directed and digital format was also highly feasible and acceptable to adolescents, a finding that is in keeping with previous research on digital interventions for young people. Reference Liverpool, Mota, Sales, Čuš, Carletto and Hancheva43 Although the evidence base was small, this review highlights some important considerations for researchers and practitioners working to support adolescents with their mental health.
Primarily, although most of the interventions employed mindfulness as a strategy to improve emotion regulation, this was not a universally acceptable or beneficial approach. Given the deluge of literature on mindfulness-based interventions, such a finding suggests that there is a need to look elsewhere for additional solutions to support young people with their emotion regulation and mental health. One such recommendation would be to focus on developing research into other emotion regulation strategies that receive arguably less attention (e.g. cognitive reappraisal) and the related skills required to access these strategies and use them flexibly. Individual, social, cultural and environmental factors are known to be highly influential in the success of different emotion regulation strategies. Reference Bonanno and Burton44 As such, the focus of interventions should move beyond teaching adolescents different strategies, to also equipping them with the skills they need to apply these confidently (e.g. emotional awareness, self-efficacy). A further key finding of note from this review was the strong feasibility and acceptability of the self-directed digital format for delivering this type of intervention to adolescents. Digital health is a growing field, and provides opportunities to broaden access to care in a cost-effective and sustainable manner. Particularly among adolescents, this is known to be a highly acceptable format for the delivery of mental health interventions. Reference Liverpool, Mota, Sales, Čuš, Carletto and Hancheva43
Limitations of this review
The findings of this review should be interpreted with caution due to several limitations. Primarily, the lack of studies identified limits any conclusions that could be drawn. Despite the planned focus of this review on psychopathology, as it transpired, just three studies measured this as an outcome and findings were mixed. Moreover, only symptoms of anxiety and depression were measured, making it difficult to draw conclusions about the effects of emotion regulation as a transdiagnostic mechanism. There was also significant clinical and methodological heterogeneity among identified studies, meaning that direct comparison of interventions was not possible, nor making conclusions about those most effective in improving emotion regulation and psychopathology. However, we argue that insight can be gleaned from understanding the different approaches currently taken to improve emotion regulation in adolescents and their effects within individual studies, but acknowledge that, with the lack of available evidence, no definitive conclusions should be drawn about the effectiveness of any individual approach in comparison with another.
Only 50% of studies were dually screened at title and abstract stage. Although rater pairs showed strong agreement in those articles that were screened, the lack of 100% dual screening may have contributed to the decreased number of identified studies.
There was considerable variation across the measures used to capture change in emotion regulation, as expected given the large variation in the different conceptualisations of emotion regulation and its subsequent measurement. Reference Gross52 Although some validated measures exist, these are not employed consistently across studies of emotion regulation, meaning that it is difficult to compare the effects of different interventions. Similarly, functional outcomes such as academic achievement or life satisfaction were different across all of the included studies and thus could not be compared.
There were further limitations within the studies included in this review. The quality of the studies was limited, and many of the interventions were at the early stages of testing. Most studies were pre- to post-cohort studies and did not include a comparison group, which contributed to the increased risk of bias. Similarly, many of the included studies had short follow-up time periods, meaning that any change in effects observed over time was not captured here. Furthermore, although the participants included in these studies were ethnically diverse, most were conducted in the USA, limiting conclusions that can be drawn about the benefits of these interventions globally. As such, many findings that were reported in this review should be interpreted in light of these shortcomings.
Suggestions for future research
As the prevalence of mental health problems among adolescents increases, more research must be done to determine the impact of self-directed digital interventions targeting transdiagnostic mechanisms like emotion regulation, particularly when applied at scale to support adolescents with their mental health. Recommendations from this research echo a common suggestion within the emotion regulation literature – to better refine and clarify the concept of emotion regulation as it applies to research and practice. In this review, several different measures of emotion regulation were employed, dependent on the authors’ understanding of emotion regulation as a concept to be targeted in an intervention. The lack of conceptual clarity surrounding emotion regulation, and the way in which its change can be interpreted or measured, mean that it is difficult to compare and synergise emotion regulation interventions and illuminate those components that are most beneficial in improving emotion regulation and mental health. Clarifying the definition of emotion regulation is an essential step to a more synchronous approach in the field, and ultimately to better understanding of how this mechanism can be targeted to support adolescents with their mental health.
Supplementary material
The supplementary material is available online at https://doi.org/10.1192/bjo.2025.10888
Data availability
Data availability is not applicable to this article because no new data were created or analysed in this study.
Acknowledgements
The authors thank all authors who shared their data and missing information with us upon request. We also thank Professor Jennifer Lau for her practical support with this study, including assistance with screening.
Author contributions
A.T., G.H. and B.O. conceived the initial idea for the systematic review. A.T., E.L. and E.S. screened the studies at the title and abstract stage. A.T. and E.S. screened the studies at the full-text stage. A.T. and E.L. extracted data from the included studies and carried out the quality assessment. A.T. drafted the manuscript, with G.H. and B.O. providing critical insight. A.T., G.H., B.O., E.L., E.S. and B.W. contributed to the revision of the manuscript and approved the final version.
Funding
This work is funded by the UK Research and Innovation, Economic and Social Research Council, London Interdisciplinary Social Science Doctoral Training Partnership (grant no. ES/P000703/1). The funders had no role in study design, data collection and analysis, the decision to publish or the preparation of the manuscript.
Declaration of interest
None.
Open access
eLetters
No eLetters have been published for this article.