Evaluating the evidence: a systematic review of reviews of the effectiveness and safety of digital interventions for ADHD

Abstract

Background

Attention-Deficit/Hyperactivity Disorder (ADHD) impacts academics, work and social relationships. Digital interventions, such as virtual reality, games, app and other, offer accessible therapeutic options, yet understanding their efficacy and potential adverse effects is crucial for safe use. The objective of this study is to identify and analyze the efficacy and adverse effects reported in systematic reviews of digital interventions for ADHD.

Methods

We conducted a systematic review of systematic reviews to assess the reported efficacy and safety of digital interventions for ADHD. We searched for relevant publications in Scopus, PubMed, PsycINFO and Cochrane Library. Both study selection and data extraction were performed in duplicate to ensure accuracy and reduce bias. This review followed PRISMA 2020 guidelines, PRISMA-harms checklist, and we used AMSTAR-2 to assess the quality and risk of bias of the included reviews.

Results

A total of 26 systematic reviews on digital interventions for ADHD were included. These reviews collectively involved 34,442 participants, with the majority focusing on children and adolescents. The digital interventions analyzed included video games, computerized cognitive training, virtual reality, apps, and others. The outcomes reported various positive effects, such as improvements in inattention and executive function, though evidence was generally low quality. Adverse effects were reported in 8 of the 26 included reviews (30,1%), and included physical discomfort, emotional reactions, and behavioral issues, such as video game addiction.

Conclusions

This systematic review of systematic reviews indicates that while digital interventions for ADHD show potential benefits, their effectiveness remains inconclusive due to low evidence quality. Adverse effects, particularly from video games, have been reported but are inconsistently documented. Future research should focus on rigorous safety assessments, standardized reporting, and long-term effectiveness.

Trial registration

This systematic review is registered in Prospero: CRD42024521084.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12888-025-06825-0.

Background

Attention-Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder characterized by persistent patterns of inattention, hyperactivity, and impulsivity that interfere with functioning or development [1]. It affects approximately 5–11% of children worldwide [2], with symptoms often continuing into adulthood [3], where the prevalence of adult ADHD is 6,76% [4]. ADHD significantly impacts individuals by impairing academic performance [5], occupational success [6], and social relationships [6], and it is associated with higher risks of comorbid health conditions [7, 8]. At society level, ADHD contributes to increased healthcare costs, educational support needs, and productivity losses, underscoring the importance of effective interventions and support systems.

Digital interventions offer solutions for addressing challenges in current healthcare systems (i.e., accountability, coverage; accessibility of health facilities; availability of human resources, commodities, and equipment; contact and continuous coverage; effective coverage; and financial coverage) [9]. Digital interventions are already being used for treating individuals with ADHD [10–13]. Digital interventions include a range of technologies, such as mobile apps, computer programs, video games, virtual reality, or robotics. These technologies have the potential to be personalized, scalable, and accessible with options for integration into daily life to treat ADHD or as a complement to other interventions, as well as to improve adherence and engage younger populations who are often more receptive to digital tools [14, 15].

In the evaluations of digital interventions, it is crucial not only to assess their efficacy but also to understand their safety and potential adverse effects or events. Adverse effects (i.e., adverse outcome that can linked to the intervention) [16] and adverse events (i.e., adverse outcome that occurs during an intervention, but is not directly caused by it, or may not be linked to it at all) [16] are a critical aspect of any intervention, as they provide a comprehensive view of the potential risks involved in using these technologies. This information is essential for informed decision-making by both clinicians and users of these technologies, as it helps to weigh the benefits against the possible risks. Moreover, understanding safety can guide the development of safer and more effective interventions, ensuring that they do not unintentionally harm users [17]. Focusing on these aspects contributes to a more holistic evaluation, ultimately aiming to enhance the quality and safety of digital mental health solutions. In the field of mental health, some research indicates that there is little risk of harm associated with certain digital interventions (e.g., chatbots) [18], while other studies suggest that there may be a gap in the reporting of safety in publications related to digital mental interventions [19–21].

With the exponential growth of research on digital interventions for health [22], particularly related to mental health [23–25], and the increasing number of systematic reviews synthesizing these studies, a systematic review focusing not only on the effectiveness but also on the safety of these interventions for ADHD is highly relevant. The research question guiding this study is: What is the effectiveness and safety of digital interventions for ADHD, as reported in systematic reviews? Understanding the potential risks associated with digital interventions helps clinicians make informed decisions, ensuring that these tools are implemented safely and effectively in clinical settings. For researchers, identifying effectiveness as well as common adverse effects and events can highlight areas that require further investigation, driving the development of safer and more effective interventions for individuals with ADHD. From a policy perspective, comprehensive knowledge of the effectiveness and safety can inform regulations and guidelines, ensuring that digital interventions for ADHD are held to high safety standards, ultimately improving patient outcomes.

Objective

The objective of this study is to identify and analyze the effectiveness and safety reported in systematic reviews of digital interventions for ADHD.

Methods

We have performed a systematic review of systematic reviews to capture the current evidence on the reported effectiveness and safety associated with digital interventions for ADHD.

This review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020 Statement) [26], the PRISMA harms checklist [27] to identify the minimal set of items to be reported when reviewing harms; and the MeaSurement Tool to Assess systematic Reviews (AMSTAR- 2) guidelines [28]. This systematic review is registered in Prospero: CRD42024521084.

Search strategy and information sources

This review is part of a larger project in which we are examining the safety of any type of intervention addressed to individuals with ADHD, as reported in systematic reviews with or without meta-analysis. With the objective of identifying the reported effectiveness and safety associated with any type of intervention for ADHD documented in systematic reviews, an electronic search was carried out on March 4 th, 2024. Additionally, a second search was conducted on March 6 th, 2025, covering the period from 2024 to 2025. These searches were carried out by the first author. The searches covered published studies comprising the terms related to ADHD in title or abstract “ADHD” or “Attention Deficit Disorder with Hyperactivity” or “Attention Deficit Disorder” or “Attention Deficit Hyperactivity Disorder” or “Attention Deficit Hyperactivity Disorders” or “Hyperkinetic Syndrome” in combination with terms related to systematic review in the title “Systematic review” or “Meta-analysis” or “Metaanalysis” and indexed in the following databases: Scopus, PubMed, PsycINFO, and the Cochrane Library. No year or language limitations were used. The full search strategy is presented in Appendix 1.

In this article, we specifically present data corresponding to systematic reviews on digital interventions for ADHD, which refer to the use of any digital technology to deliver the intervention, such as apps, computers, video games, virtual reality, robots, artificial intelligence, and others.

Eligibility criteria, selection process and data extraction

Publications were included in the review if they fulfilled following criteria:

Inclusion Criteria:

1. Systematic reviews specifically focusing on individuals with ADHD (of any age group, gender, ethnicity, and existence of comorbidity); and

2. Systematic reviews covering any digital intervention and published in any language.

Exclusion Criteria:

1. Systematic reviews that do not specifically focus on individuals with ADHD;

2. Systematic reviews that do not address digital interventions;

3. Protocols of systematic reviews;

4. Editorials, letters to the editor, errata, corrigenda, corrections, comments, retracted articles, responses, or similar materials about systematic reviews or meta-analyses.

All references captured by the search engine were uploaded to EndNote 20 and Rayyan. Duplicates were identified and removed. In order to assess eligibility, in a first step, all titles and abstracts were reviewed in duplicate: one reviewer (EG) screened all references, while two other reviewers (KD and GLC) independently screened half of the references each. In a second step, the full text of the selected studies was reviewed in duplicate; two-thirds of the references were assigned to each of the three reviewers, ensuring that each article was independently assessed in full text by two reviewers. Discrepancies were solved through discussion until reaching agreement. All selected articles were included in qualitative synthesis.

The following data were extracted into an ad hoc document: publication year, population (age, gender, ethnicity, and comorbidities), intervention (type and duration), comparator/control, and reported outcomes (effectiveness and safety). The data extraction was also performed in duplicate, with all selected articles from the full-text screening divided among the three reviewers, ensuring that each article had data extracted by two independent reviewers. No additional attempts were made to retrieve relevant data from the authors of included studies. Data extraction files were merged, and discrepancies were discussed between all authors until consensus was achieved.

Data items and synthesis of data

Following the PRISMA guidelines [26], we performed a narrative synthesis of the results and present tables and figures summarizing the outcomes (effectiveness and safety) reported in the systematic reviews. These were categorized according to the type of intervention, over time, and according to the quality of evidence of the systematic review. Additionally, we have summarized the systematic reviews that have not reported safety data. These findings were also categorized according to the type of intervention, over time, and according to the quality of evidence provided in the systematic review. All coauthors have been involved in the data synthesis.

Quality evidence assessment and risk of bias

We have used the critical appraisal tool for systematic reviews that include randomized or non-randomized studies of healthcare interventions, known as AMSTAR- 2 [28] to evaluate the quality of the evidence of the included reviews. This tool helps to rate the studies as having high, moderate, low, or critically low quality based on the assessment of 16 domains, such as the inclusion of PICO questions and criteria; existence of prior published review protocol; explained inclusion criteria; use of comprehensive literature search strategy; performed duplicate study selection; performed duplicate data extraction; or list of excluded studies with justifications of their exclusion, among others. Each included study was independently evaluated by two authors.

Results

Study selection

A total of 5,565 records were identified in the data search. After removing duplicates, 3,088 titles and abstracts were screened, and of those 540 were sought for retrieval. Of those, 40 additional articles were excluded due to lack of access to their full text, being retracted articles or duplicates (most of which resulted from overlap in the second search). The full texts of 500 systematic reviews on interventions for ADHD were obtained, and from these, a total of 26 systematic reviews addressing any type of digital intervention were included in this review [29–54]. Figure 1 presents the flowchart of the selection procedure for both search engines.

Fig. 1.

Flowchart of the selection procedure for both search engines

The list of the 474 articles systematic reviews on interventions for ADHD that do not specifically address any type of digital intervention and were therefore excluded from this review is presented in Appendix 2.

Risk of bias

The methodological quality and risk of bias of the included systematic reviews are presented in Fig. 2. Three reviews were classified as high quality [41, 46, 54], one as moderate quality [50], ten as low quality [29–31, 33, 34, 38–40, 44, 48, 49, 52, 53], and eleven were classified as critically low [29–31, 33, 34, 38–40, 44, 48, 51] using AMSTAR- 2 [28]. The risk of bias for one of the included articles could not be assessed [37], as it was published in a language not known to the authors (Slovenian) and the accuracy of the automated translation with Google Translate Documents could not be confidently relied upon.

Fig. 2.

Risk of bias of the systematic reviews evaluated with AMSTAR- 2 [28]. * AMSTAR- 2 Critical domain. H: High (No or one non-critical weakness); M: Moderate (More than one non-critical weakness); L: Low (One critical flaw with or without non-critical weaknesses); CL: Critically low (More than one critical flaw with or without non-critical weaknesses).  Yes;  No; Unclear; Ø: Not applicable

Studies classified as low or critically low mainly failed to report a list of excluded studies with reasons for exclusion (a critical domain in AMSTAR- 2), and because the authors did not address how the risk of bias might have impacted the interpretation of their results.

Description of participants, interventions and comparators

A summary of the 26 included systematic reviews is presented in Table 1. The included articles were published between 2018 and 2025 and reported the inclusion of a total of 575 articles in their reviews. The source of funding of the included reviews is reported in Appendix 3.

Table 1.

Summary of the systematic reviews included (n = 26)

Reference	Articles included in review	Participants included the reviews and their ages	Intervention type and durations (D)	Use of control groups	Outcomes-Effectiveness	Outcomes-Safety
Powell et al, 2018 [41]	14	N = 743 6,6 to 16,5 years	Several digital interventions (self-help devices, video games, internet, software, social media, apps…) for facilitating the self-management of ADHD associated difficulties D: 2 days to 3 months	Control groups: - No intervention	Technology has potential in pediatric ADHD management, but current research lacks robustness	Not reported
Gudka et al., 2025 [54]	19	N = 26,518 to 77 years	Several digital tools (web-base, apps and chatbots) for providing psychoeducation; symptom monitoring; practical interventions; healthcare and self-management information D: Not reported	Control groups: - No intervention (only used in 9/19 included articles)	Limited evidence of effectiveness was found, except for psychoeducation DHIs which may be effective in reducing ADHD symptoms	Not reported
Westwood et al, 2023 [46]	42	N = 2234 6 to 65 years	Computerized cognitive training for supporting clinical, neuropsychological and academic outcomes D: Not reported	Control groups: - Treatment as usual - Waiting list - Medication - No intervention	No effect of computerized cognitive training interventions on ADHD total or hyperactivity/impulsivity symptoms; small improvement in inattention symptoms	Not reported
Liu et al., 2024 [50]	25	N = 1780 5 to 38,9 years	Several digital tools (computers, internet, chatbot, games) for delivering CBT, cognitive or attention training, neurofeedback, and relaxation training D: 4 to 25 weeks	Control groups: - Wait list - Placebo pill	The digital interventions proved beneficial for individuals with ADHD by alleviating symptoms of ADHD, inattention, and hyperactivity/impulsivity	Not reported
Elbe et al, 2023 [35]	9	N = 285 21,1 to 39,5 years	Computerized cognitive training for supporting cognitive outcomes and symptom severity D: 4 to 12 weeks	Control groups: - Active intervention (1-back) - Wait-list	Computerized cognitive training showed a slight improvement in overall cognition, with no significant changes in symptom severity or specific cognitive skills	Not reported
Yu et al., 2024 [52]	9	N = 370 8 to 18 years	Virtual reality (games, cognitive training, and neural feedback training) for supporting attention and motor ability D: 2 to 12 weeks	Control groups: - Methylphenidate - Conventional therapy/feedback training - No intervention	Both immersive and nonimmersive virtual reality interventions significantly improved attention	Not reported
Avion, 2024 [49]	9	N = 370 6 to 17 years	Computerized-cognitive training for supporting attention, working memory, and/or combined aspects of cognitive function D: 5 to 8 weeks	Control groups: - Waiting list - Sham - Active control - Stimulant medication - Combined treatment	Computerized-cognitive training interventions did not provide enough data to determine their impact on executive and adaptive functioning	Not reported
Caselles-Pina et al, 2023 [32]	18	N = 1479 7,8 to 15,6 years	Video-games as cognitive training D: 1 to 20 weeks	Control group only mentioned for one study (Tetris)	Video games-based interventions can be used to improve ADHD symptoms and display high adherence to treatment	Headache (3 studies) Pain in the fingers (1 study) Irritability (1 study) Frustration (1 study)
Zhang et al., 2024 [53]	10	N = 683 6 to 17 years	Virtual reality-based sports games (computer-based training; motor games, and virtual reality training) for improving executive and cognitive functions D: 4 to 20 weeks	Control group: - No intervention	Virtual reality motor games significantly improved the executive and cognitive functions of children with ADHD	Not reported
He et al, 2023 [36]	31	N = 2169 4 to 17 years	Several digital interventions (RoboMemo; Cogmed working memory training; Xbox Kinect; Focus iCON mobile application; and other) as digital therapeutics D: Not reported	Control groups: - No intervention - Traditional cognitive intervention - Standard care	Digital interventions improved inattention symptoms, reduced reaction time on the continuous performance task, slightly decreased impulsive hyperactivity, and enhanced executive function	Video game therapy lead to headache, dizziness, agitation, and other adverse effects ADHD patients are at higher risk for video game and internet addiction
Rodrigo-Yaguas et al., 2022 [42]	128	N = 11,969 10 months to 36 years	Video games as cognitive rehabilitative tools D: 2 weeks to 2 years	Control groups: - No intervention - Different interventions - Usual treatment - Other comparators	Video games can be used to ameliorate ADHD symptoms while improving adherence to treatment	Playing video games for over 1 h daily may worsen ADHD symptoms Playing video games may increase attention problems The presence of electronic devices can cause sleep problems
Vitija et al, 2022 [45]	4	N = 502 2 to 18 years	Several digital interventions (mobile app; Telemental Health service; digital appointments with counserllors…) for promoting communication D: 5 to 12 weeks	Control groups: - Treatment as usual	Digital interventions bridge gaps between patients and healthcare professionals, enabling more frequent monitoring and communication	Not reported
Romero-Ayuso et al., 2021 [43]	6	N = 311 7 to 12 years	Virtual reality for supporting cognitive deficits D: 2 to 8 weeks	Control groups: - Non VR - Placebo psychotherapy training - Long-lasting methylphenamide	Virtual reality interventions led to improvements in attentional vigilance, with more correct responses and fewer omission errors, but no change in impulsivity	Children struggle to distinguish real memories from VR memories
Wilkes-Gillian et al., 2021 [47]	15	N = 106 5 to 16 years	Video-modelling for improving skills and behaviours D: 15 min to 18 months	Some studies have controls	Evidence suggests video modeling may effectively enhance social skills and behaviors in individuals with ADHD when combined with other interventions	Not reported
Black et al, 2020 [31]	32	Not reported	Several digital interventions (wearables, mobile apps and computer based) as school performance support tools D: Not reported	Not reported	Mixed conclusions on the effectiveness of utilizing these technologies as an aid of ADHD	Not reported
Păsărelu et al, 2020 [39]	1	N = 39 6 to 16 years	App for improving medication adherence D: Not reported	Control groups: - Treatment as usual (no app)	Further research is needed to determine the safety and efficacy of available ADHD apps	Not reported
Andreou et al., 2024 [29]	15	N = 310 6 to 16 years	Several digital interventions (iPad, software, Lexia SOS, augmented reality app, etc.….) for enhancing the academic achievements D: 1 lesson to 6 months	Some studies have controls	Educational software and computer games can aid children with ADHD in native language learning, supporting skills like reading, spelling, and descriptive writing	Not reported
Bernanalizadeh et al., 2024 [30]	12	N = 708 5 to 17 years	Several digital interventions (computer-based training; online behavioural parent training; virtual reality…) for supporting symptom domains D: 4 to 22 weeks	Control groups: - No intervention - Waitlist - Other comparators	Significant effect on inattention/cognitive function, hyperactivity/impulsivity, and oppositional behaviours	Not reported
Cervantes et al., 2023 [33]	36	N = 1219 3 to 35 years (social robots) and 5 to 40 (games)	Social robots and video games for supporting cognitive or behavioral rehabilitation therapies D: Not reported	Control groups (not specified)	Social robots and video games interventions have the potential to support non-pharmacological treatments of people living with ADHD	Issues related with device ergonomics
Corrigan et al., 2023 [34]	7	N = 321 7 to 18 years	Virtual reality for improving cognitive deficits D: 2 to 10 weeks	Control groups: - Computer monitor - No intervention - Traditional training - Flash games - Medication	Virtual reality interventions lead to large effect sizes in favor of VR-based interventions on outcomes of global cognitive functioning, attention, and memory	“My head become heavy” (fullness of head)
Oh et al, 2023 [38]	20	N = 1402 6 to 17 years	Serious games as digital therapeutics D: Not reported	Control groups: - Other type of games - Wait list control - Medication	Game-based interventions improved inattention and hyperactivity more than the control, while medication was more effective than game-based interventions	No serious side effects Minimal side effects
Wong et al, 2023 [48]	19	N = 1843 7 to 12 years	Several digital interventions (computer-assisted interventions, exergame training, virtual classrooms, web-based training and other) for regulating the behaviors and cognitive functions D: 4 weeks to 6 months	Control groups: - Not receiving training - Treatment as usual - Other comparators	Small and significant effect sizes in inattention, overall executive function, disruptive behavior disorder, and computer-rated visual attention	Not reported
Peñuelas-Calvo et al, 2022 [40]	22	N = 1473 8,6 to 14,7 years	Video games as digital therapeutics D: 3 to 24 weeks	Control groups: - No intervention	Effective in improving cognition and reducing ADHD symptoms	Not reported
Shou et al, 2022 [44]	6	N = 261 3 to 12 years adults mean age 38,9 years	Online interventions (including Cognitive Behavioural Therapy; Behavioural Parent Training; Online web Learning: Internet-based Cognitive Behavioural Therapy, and other) for reducing ADHD related cognitive and social disorders D: 6 to 16 weeks	Control groups: - Waiting list	Online interventions were more effective than the waiting list in improving attention deficit and social functioning in adults and children with ADHD	Not reported
Ramos-Galarza et al, 2024 [51]	40	Not reported	Several technologies (serious games; mobile apps; monitoring devices; robots; internet-based platforms; voice assistants) as digital therapeutics D: Not reported	Not reported	Increase in the patient’s motivation for the treatment process and their interest in performing the different exercises to improve their attention and self-control	Not reported
Mičič et al, 2021 [37]	25	N = 1214 5 to 17 years	Computerized cognitive training for supporting functional deficits (attention deficit, impulsivity, lack of behavioural inhibition, planning, and organisation) D: 4 weeks to 4 months	Control groups: - Placebo - Passive - Later inclussion - Partially active + placebo - Later inclusion + active	Computerized cognitive training shows a positive effect on the cognitive abilities of children with ADHD. Some studies have also shown a positive effect on reducing dysfunctional and impulsive behavior	Impulsiveness

The systematic reviews included in this review collectively involved a total of 34,442 study participants across intervention and control groups. The smallest review included 39 participants [39], while the largest included 11,969 participants [42]. Most of the studies focused on children and adolescents (ages 3 to 18), though seven reviews also included studies with adult participants [33, 35, 42, 44, 46, 50, 54].

Regarding the interventions, 10 systematic reviews addressed various types of digital interventions [29–31, 36, 41, 45, 48, 50, 51, 54]; 4 on computerized cognitive training [35, 37, 46, 49]; 4 on virtual reality-based interventions [34, 43, 52, 53]; 3 specifically focused on video games [32, 40, 42]; 1 on video modeling [47]; 1 on apps [39]; 1 on serious games [38]; 1 on online interventions [44]; and 1 on both social robots and video games [33].

The reviews reported varied intervention durations, with the shortest being from 15 min [47] and the longest lasting up to 2 years [42]. Eight reviews did not report the duration of the analyzed interventions [31, 33, 36, 38, 39, 46, 51, 54].

All systematic reviews, except two [31, 51], reported that their included studies used control groups or comparators, with waitlist and no intervention being the most commonly reported types of controls.

Description of the outcomes: effectiveness and safety

Several reviews analyzed the effectiveness of various digital interventions. The high-quality reviews focusing on various digital interventions reached different conclusions: one found that current research lacks robustness [41], while the other found limited evidence of effectiveness, except for psychoeducation DHIs, which may help reduce ADHD symptoms [54]. The moderate quality evidence review found that the digital interventions proved beneficial for individuals with ADHD by alleviating symptoms of ADHD, inattention, and hyperactivity/impulsivity [50]. Reviews considered of low and critically low evidence according to AMSTAR- 2 showed significant effects on attention, executive functions, behaviour, and learning in children with ADHD linked to the use of various digital interventions [29–31, 36, 45, 48, 51].

Regarding the computerized cognitive training, the high-quality review found no effect on ADHD total or hyperactivity/impulsivity symptoms, with a small improvement in inattention symptoms [46]. In reviews ranked of lower quality evidence computerized cognitive training showed some cognitive benefits but limited impact on ADHD symptoms and executive functioning [35, 37, 49].

Four low or critically low-quality reviews (AMSTAR- 2) found that virtual reality interventions improved cognitive functioning, attention, and memory [34]; enhanced attentional vigilance but not impulsivity [43]; improved executive and cognitive functions [53]; and significantly enhanced attention in both immersive and nonimmersive formats [52].

Additionally, low-quality reviews suggested that video games may improve ADHD symptoms, enhance cognition, and promote high adherence to treatment [32, 40, 42].

Five low or critically low quality reviews (AMSTAR- 2) found that video modeling may improve social skills [47], ADHD apps require further research [39], serious games improved inattention and hyperactivity but were less effective than medication [38], online interventions outperformed waiting list interventions in improving attention and social functioning [44], and social robots and video games may support non-drug treatments for ADHD [33].

Only 8 of the 26 included systematic reviews explicitly reported any adverse effects associated with digital interventions [32–34, 36–38, 42, 43]. The reported adverse effects of digital interventions could be considered under falling into several broad categories, such as physical discomfort (including symptoms like headache, dizziness, fullness of head, pain in the fingers, and issues related to device ergonomics) [32–34]; mental and emotional reactions, such as irritability, frustration, agitation, hangover, impulsiveness and hyperactivity [32, 33, 36, 37]; Confusion is noted in cases where children struggle to distinguish real memories from VR experiences [43]; Sleep and attention issues arise from media use, with electronic devices linked to sleep problems, and both video gaming and TV potentially worsening attention, especially in ADHD patients [42]. Additionally, behavioral and addiction risks are highlighted, with individuals with ADHD being more prone to video game and internet addiction [36], and extended video game use potentially worsening ADHD symptoms [42]. One review reported minimal or no serious side effects [38].

A comprehensive overview of the main reported outcomes regarding effectiveness and safety associated with various digital interventions for ADHD, categorized by the overall confidence in the review results, is provided in Table 2.

Table 2.

Overview of the main reported outcomes associated with the digital interventions for ADHD

Analyzed digital interventions for ADHD	Positive effects of interventions	Neutral or marginal effects of interventions	Adverse effects of interventions
Computerized cognitive training	High evidence
(to improve clinical, cognitive, academic, and functional outcomes)	• Small improvement in inattention symptoms [46]	• No effects on hyperactivity symptoms [46] • No effects on impulsivity symptoms [46]
	Low evidence
	• Improvement in overall cognition [35]	• No significant changes in symptom severity or specific cognitive skills [35] • Lack of data to determine impact on executive and adaptive functioning [49]
	Critically low evidence
	• Improved overall cognition [37] • Reduced dysfunctional behavior [37] • Reduced impulsive behavior [37]		• Impulsiveness [37]
Several digital interventions	High evidence
(to provide psychoeducation, symptom monitoring, self-management, CBT, training, and academic support)	• Psychoeducation delivered through digital interventions may help reduce ADHD symptoms [54]	• Current research lacks robustness [41] • Limited evidence of effectiveness [54]
	Moderate evidence
	• Alleviated symptoms of ADHD, inattention, and hyperactivity/impulsivity [50]
	Low evidence • Improved inattention symptoms [36] • Enhanced executive function [36] • Reduced reaction time on continuous performance tasks [36] • Decreased impulsive hyperactivity [36] • More frequent monitoring and communication with healthcare professionals [45]
	Critically low evidence
	• Aid in language learning [29] • Improved inattention symptoms [30, 48] • Improved visual attention [48] • Improved cognitive function [30] • Enhanced executive function [48] • Improved disruptive behaviour [30, 48] • Decreased impulsive hyperactivity [30] • Increased motivation for the treatment [51]	• Mixed conclusions on the effectiveness of utilizing technologies as an aid of ADHD [31]
Virtual reality	Low evidence
(to support and improve attention, motor ability, and cognitive functions)	• Improved attentional vigilance [43] • Improved executive and cognitive functions [53] • Significantly enhanced attention [52]	• No changes in impulsivity [43]	• Children struggle to distinguish real memories from VR memories [43]
	Critically low evidence
	• Improved global cognitive functioning [34] • Improved attention [34] • Improved memory [34]
Video games (for supporting cognitive training, rehabilitation, and as digital therapeutics)	Low evidence • Reduced ADHD symptoms [32, 42] • Higher adherence to treatment [32, 42]		• Headache [32] • Pain in fingers [32] • Irritability [32] • Frustration [32] • Worsening ADHD symptoms when playing > 1 h/day [42] • Increased attention problems [42] • Sleep disturbances [42]
	Critically low evidence • Reduced ADHD symptoms [40] • Improved cognition [40] • Potential to support non-pharmacological interventions [33]
Video modelling (for improving skills and behaviours)	Low evidence • Enhanced social skills [47]
Serious games	Critically low evidence
(as digital therapeutics)	• Improved inattention [38] • Improved hyperactivity [38]	• Medication more effective than serious games [38]	• Minimal side effects [38]
Social robots	Critically low evidence
(for supporting cognitive or behavioral rehabilitation therapies)	• Potential to support non-pharmacological interventions [33]		• Ergonomics [33]
Online interventions	Critically low evidence
(for reducing ADHD related cognitive and social disorders)	• Improved attention deficit [44] • Improved social functioning [44]
Apps	Critically low evidence
(for improving medication adherence)		• Inconclusive safety and efficacy of ADHD apps [39]

Discussion

Summary of findings

This systematic review includes 26 systematic reviews on digital interventions for ADHD, published between 2018 and 2025, and covering more than 34,000 patients, most of which focus on children and adolescents. The technologies tested in the studies concern digital interventions, interactive and immersive technologies, cognitive and behavioral training technologies as well as robotics and hybrid technologies including video games.

The studies showed that there is some low-quality evidence (as assessed by AMSTAR- 2) that digital interventions improve single symptoms of ADHD, such as inattention, impulsivity, and hyperactivity. Beside that they can have an effect on the treatment adherence and the monitoring and communication behavior of healthcare professionals.

The overall findings from 8 from the 26 reviews suggest that digital interventions can have a range of adverse effects, spanning physical, mental, and behavioral domains. They are related to the interaction with the technology itself (e.g. video games), the device ergonomics (e.g. in social robots) respectively such as physical discomfort. Some worsening of typical ADHD symptoms namely attention loss, mental and emotional reactions or behavioral and addiction risks were also mentioned.

Are digital interventions for ADHD effective and safe?

All the included systematic reviews examined the effectiveness of using digital health interventions in ADHD, but only twelve addressed safety, with only eight reporting any adverse effects. This could indicate a potential oversight or underreporting of safety concerns in the literature, highlighting a gap in comprehensive evaluations of both effectiveness and safety. When these findings are analyzed in further detail it is possible to identify differences in the effectiveness and the associated level of evidence of the reviews. Three of the studies considered [31, 39, 41] did not clearly identify any positive or negative outcomes associated with the use of the digital interventions (i.e., apps and several digital solutions). These reviews are the oldest considered in our analyses, published between 2018 and 2020, and showed mixed conclusions regarding the effectiveness of the interventions and highlighted the need of generating further evidence. These results could fit a scenario where the body of research in this domain was still under development.

Most of the reviews showed some evidence of positive outcomes associated with the use of digital interventions in ADHD populations. However, it is important to notice that the level of evidence for those reviews was mostly low or critically low, limiting the strength of these findings. Only two reviews showed high evidence of positive outcomes related to the use of various digital interventions to deliver psychoeducation, which was found to be beneficial to reduce ADHD symptoms [54]; and computerized cognitive training was linked to “a small improvement in inattention symptoms” [46] while it also showed moderate evidence of no impact on hyperactivity or impulsivity symptoms [46]. These findings highlight the variability in the strength of evidence across reviews and suggest that while some positive effects are observed, they are generally based on low-quality evidence.

When we consider the technologies used for delivering the interventions, it is possible to identify a similar landscape in the field of digital interventions for mental health to the one described above, where many studies on digital technologies show low or critically low evidence level [55–58]. The studies considered covered a broad range of tools and technologies used in the different interventions. However, the continuous advancement in the use of novel technologies particularly the recent use of generative artificial intelligence solutions based on the use of large language models represents an opportunity to develop new digital interventions tailored for individuals with ADHD. In most cases, these solutions might have not been fully addressed in this work due to their novelty.

Overall, our findings show that despite there have been advances in the development and use of digital health interventions for ADHD their effectiveness remains arguable as the vast majority of the systematic reviews supporting their effectiveness are considered to have low or critically low evidence. However, this does not imply that the original articles included in these reviews were of low quality. Even more, our results showed mixed and even contradictory outcomes regarding the effectiveness of the analyzed interventions, for example the use of computerized cognitive training has been associated to having no effect on impulsivity symptoms according to a review with moderate evidence [46], and also positively associated with reduced impulsive behavior or even with adverse effects related impulsiveness by a review graded as having critically low evidence [37].

The majority of the original studies included in the systematic reviews that we have analyzed focused on children/adolescent population (17/26) with a few studies (7/26) that reported a broad range of ages spanning from childhood to adult hood [33, 35, 42, 44, 46, 50, 54] and two that did not report the ages [31, 51]. It has been observed before that the severity of the ADHD-related symptomatology can decrease in some individuals during adolescence [59, 60], thus it is possible that this phenotypic variability could impact the different outcomes observed and the low level of evidence reported across the studies. However, evidence suggests that ADHD traits can become more challenging for women later in life [61, 62], particularly during transitions such as adulthood [63], where demands and responsibilities increase. This highlights the potential need for more structured, facilitated interventions for adults, rather than relying on self-directed approaches. Further evaluations focusing on adults with ADHD, especially during key life transitions, would be valuable in addressing this gap in the literature.

Safety is an important component of any heath intervention, however in the majority (18/26) of the studies included in our analyses no safety data was reported. Previous research had already highlighted the poor reporting of harms in primary studies related to diverse intervention types [27], and this issue appears to be reflected in systematic reviews as well. It is not clear whether this is due to a lack of focus in considering the safety aspects or because the interventions completely safe and free of adverse effects or adverse events. However, when the different digital interventions are analyzed in detail, it can be seen that for 5 out of the 8 types of intervention considered (see Table 2) at least one adverse effect could be associated with them. In particular, the studies that looked at the use of video game interventions [32, 42], where the ones were this information was more detailed and presented. On the other side, neither online interventions nor video-modelling [44, 47] presented information related to potential adverse effects. Focusing on the effectiveness of digital interventions is a strong attraction given the pressures and needs to support individuals with ADHD. Providing novel therapies to replace or complement pharmacological interventions, while ensuring widespread accessibility, is a mayor challenge. Digital technologies offer promising solutions by adapting and translating existing tools, such as cognitive behavioral therapy [64]. However, many of these interventions focus primarily on achieving a like-for-like replacement of existing approaches with digital platforms. While this approach is common, it risks overlooking potential adverse effects and risks associated with the transition to digital platforms. Such considerations should be addressed more systematically to ensure safety and efficacy [17]. Specifically, the increasing use of artificial intelligence and large language models in the development of digital health interventions necessitates rigorous criteria for algorithmovigilance [65, 66].

These criteria must be integrated at every stage of design, implementation, and analysis to safeguard against unforeseen consequences. In the context of ADHD and the safe use of digital health interventions, it is worth noting that the CONSORT-EHEALTH [67] checklist for reporting digital health clinical trials includes a section on “Harms.” While this section encourages reporting on potential adverse effects, its inclusion is currently only recommended rather than mandatory. Strengthening these requirements could play a critical role in ensuring the safe and effective deployment of digital interventions.

To our knowledge, no comprehensive study has been published that specifically examines the safety of any type of interventions for ADHD. However, we are currently conducting a larger project that systematically investigates the safety of all intervention types for individuals with ADHD, based on evidence from systematic reviews with or without meta-analysis. This ongoing work is registered in PROSPERO (CRD42024521084).

Managing the evidence on digital interventions safety for ADHD

Research on digital interventions for mental health has experienced exponential growth [23–25], and the volume of scientific publications on ADHD has similarly expanded significantly over the past decade [68]. This growing volume of literature can make it challenging for clinicians, educators, and other professionals working with individuals with ADHD to stay up-to-date and recommend the most effective and safest interventions. Systematic reviews and meta-analyses are considered the gold standard for synthesizing evidence that can provide professionals with crucial information to fully assess interventions, both in terms of effectiveness and safety. In this review, however, we identified that only 8 of the 26 included systematic reviews reported on the adverse effects of digital interventions for ADHD [32–34, 36–38, 42, 43], therefore leaving a critical knowledge gap for clinicians, educators, policymakers or others who rely on systematic reviews to make informed decisions. Publication bias and underreported negative outcomes in original research may contribute to the limited reporting of adverse effects in systematic reviews.

When systematic reviews cannot be relied upon to provide comprehensive reporting, especially regarding safety, identifying sources of information on possible adverse effects of digital interventions for ADHD becomes crucial. Other sources that might include safety data related to interventions are clinical trial registries (e.g., ClinicalTrials.gov), case reports, or observational studies. However, the current lack of standardization in reporting adverse effects of digital interventions [17, 69] makes it difficult to aggregate and interpret data effectively. Furthermore, clinicians working with digital interventions for ADHD may observe side effects that are not documented or shared with the scientific community. Further, individuals with ADHD and their caregivers may report adverse effects of digital interventions through informal or non-traditional channels, such as social media platforms or app store reviews. Reporting of side effects related to interventions for mental health in digital platforms has been previously reported [21, 70–73]. Although the reporting of side effects on digital platforms may lack structure, and its credibility can be questionable, it could serve as warning for potential safety concerns. This underscores the need for more research on the unintended consequences of digital interventions [74], as well as the establishment of standardized protocols for reporting adverse effects that enable comprehensive safety assessments for all stakeholders [69]. Advances in technology, including artificial intelligence, have the potential to revolutionize the identification and synthesis of safety information for ADHD digital interventions by processing vast amounts of data from diverse sources, including unstructured data posted on social media.

Strengths and limitations

In this paper, we focused on a specific condition ADHD and digital interventions addressing it. The aggregated evidence as reported in this paper was not previously available, yet crucial to recognize current limitations of digital interventions for ADHD and determine potential for future improvements, in particular regarding safety. We used the broad scope of digital technologies without focusing on specific technologies which provides an excellent overview on the current landscape in this field and which allows us to judge the differences in effectiveness and safety of the different types of technologies. However, our search was limited to only four databases, and we limited our search keywords to titles and abstracts, which may have resulted in the exclusion of relevant studies. We also did not search for grey literature, potentially limiting the breath of our findings.

Although we found adverse effects reported, their amount was limited. It is unclear whether no further adverse effects occurred or whether there is a gap in reporting. Reviews included were mainly of low or critically low quality, with many failing to report a list of excluded studies (a critical domain in AMSTAR- 2). While journal restrictions on word count or page limits may pose challenges, authors still have the option to upload these lists to repositories to enhance transparency and reproducibility. Since AMSTAR- 2 considers this a critical reporting requirement, we did not contact the authors to obtain these lists, as this information should have been included in their published reviews. The absence of this information limits the ability to fully assess the rigor of the included reviews. Further, the number of papers considered was small, at least when considering the different technologies. We relied upon the information in the included systematic reviews and did not collect the information from the original sources of the single reviews. It might be, that the data aggregated in the included reviews is incomplete. In fact, we had to acknowledge that the information in the papers was incomplete or even contradictory, and one of the included articles could not be evaluated for its risk assessment due to our lack of knowledge in which the language was written, and the accuracy of the automated translation could not be reliably confirmed.

Additionally, our review focused on previously published reviews, so newer digital interventions, including those using generative artificial intelligence or other recent technological advancements, may not have been fully addressed.

Conclusions

This review of systematic reviews on digital interventions for ADHD highlights that while there is some evidence suggesting potential benefits, the overall effectiveness of these interventions remains inconclusive, with most systematic reviews presenting low or critically low levels of evidence. While digital interventions may improve symptoms such as inattention, adherence to treatment, and communication with healthcare professionals, the findings are inconsistent, and the safety of these interventions has not been comprehensively assessed. Adverse effects, particularly those associated with video games and other technologies, have been reported, though their frequency and severity vary across studies. The lack of standardized reporting on adverse effects further complicates the ability to fully assess the safety of digital interventions for ADHD. Given the growing use of these technologies, especially those leveraging artificial intelligence and large language models, there is an urgent need for more rigorous studies, transparent reporting, and standardized protocols to ensure both the efficacy and safety of digital health interventions. Future research should prioritize evaluating the long-term effects, safety risks, and potential for widespread implementation in clinical and educational settings.

Authors’ contributions

Elia GABARRON, Kerstin DENECKE, and Guillermo LOPEZ-CAMPOS all contributed equally to the conception, design, drafting, and revision of this manuscript, and all authors approve the final version for submission.

Funding

None.

Data availability

The data used to support the findings of this study are included within the article.

Declarations

Ethics approval and consent to participate

This article is a systematic review of systematic reviews and does not involve human participants or primary data collection. Therefore, ethical approval and Consent to Participate declarations are not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.