Attention-deficit/hyperactivity disorder (ADHD) is one of the most common neurodevelopmental disorders among school-age children [1], with a prevalence of 7.6% [2]. Meanwhile, according to Diagnostic and Statistical Manual of Mental Disorders Text Revision Fourth Edition [3], the prevalence of ADHD in school-age children is 3% to 7% [3]. The 3 main categories of ADHD symptoms are inattention, hyperactivity, and impulsivity, which usually manifest in the school-age period [4]. These symptoms have detrimental impacts on the quality of life and functioning, including self-esteem, academic performance, social functioning, and relationship building [5]. ADHD is usually associated with long-term disability [6]. The types of treatments vary in different stages of life [7]. Behavioral parenting training and medication are the common approaches used to improve the behaviors and self-control of school-age children with ADHD [7]. ADHD medications are associated with an increased risk of headache, anxiety, and sleep disturbances [8,9]. Behavioral therapies are generally limited by time and space [10]. Therefore, feasible nonpharmacological approaches are recommended as alternatives to regulate the behaviors, executive functions, and well-being of children with ADHD.

In the last 10 years, human-computer interaction has widely been recognized in psychiatric and mental health research [11,12]. Digital technologies, neurofeedback systems, and virtual reality for health support, care, and treatment have increasingly been adopted, thus successfully gaining psychological health advantage [13,14]. ADHD is one of the common psychiatric disorders for which technology-based treatments are often used as therapeutic tools [15,16]. The application and effectiveness of technology-based interventions have been evaluated in ADHD treatment [4,17,18]. The advantages of technology-based treatment include improved executive functions and increased physiological and mental well-being [19,20]. However, inconsistent results have been reported regarding the efficiency of technology-based intervention in school-age children with ADHD. Regarding ADHD behaviors, Dovis et al [19], Egeland et al [21], Steiner et al [22], and van der Oord et al [23] found that computer-based training improved the inattention and hyperactivity in children with ADHD. However, some studies did not find significant results between technological treatment and ADHD behaviors [24-26]. Regarding executive functions, technology-based treatments improved inhibition [27], working memory [19,28], flexibility [19,27], emotional control [19], initiation [19], planning and organization [19], organizing materials [19], monitoring [19], and metacognition [22,23]. However, several studies determined that technology-based interventions have no effect on executive functions [21,24,25]. Regarding disruptive behavior disorder, Dovis et al [19], Lim et al [29], Steiner et al [22], and van der Oord et al [23] reported that computer-assisted training and neurofeedback training regulated oppositional defiant disorder and conduct disorder, whereas Bikic et al [24] and Breider et al [30] found no significant effect. Some studies have discovered that technology-based interventions can significantly improve visual attention [20,25,31], yet numerous studies indicated no significant effects [21,24,28,32,33]. These contradictory results make it difficult to examine the effectiveness of technology-based interventions in school-age children with ADHD.

Previous reviews illustrated the effectiveness of different types of technologies on children with ADHD [34-36]. Cibrian et al [34] and Powell et al [35] summarized this topic through a narrative description. Although Powell et al [35] adopted a meta-analysis to synthesize previous studies, their research focused on the use of virtual reality among children and adolescents with ADHD.

These existing studies based on nonrandomized, cross-sectional, and observational designs have added to the knowledge base and identified the potential implications of technology in enhancing the ability and functions of children. In addition, more randomized controlled trials (RCTs) have been conducted in the recent decade, as a growing number of researchers have shown interest in the use of technology-based interventions to improve the capability and well-being of children with ADHD. A meta-analysis of RCTs can provide strong and robust evidence regarding the effectiveness of technology-based interventions in improving children with ADHD. Therefore, a systematic review and meta-analysis of the existing evidence of RCTs are needed to explicate the advantages of technology to school-age children with ADHD.

This review focused on RCTs using technologies to regulate the ability and alleviate the well-being of children with ADHD. This review was registered in the PROSPERO International Prospective Register of Systematic Reviews (registration number CRD42023446924). The studies included in this review were searched from electronic databases, including PubMed, MEDLINE, ScienceDirect, Web of Science, CINAHL (via EBSCO), PsycINFO (via OVID), and Scopus, in April 2022. The keywords used in the search engines were as follows:

Multimedia Appendix 1 shows the full search strategy.

The titles and abstracts of the selected papers were screened to identify relevant papers for this review. The analysis was performed using Population, Intervention, Control, and Outcomes framework (1) population: patients who were diagnosed with ADHD and aged between 6 and 12 years; (2) intervention: using technology (ie, the method of applying scientific knowledge for practical purposes) without restriction of technology type or frequency of intervention; (3) comparison: technology-based interventions for managing ADHD versus no interventions, interventions with placebo effect, and treatment as usual or waitlist control; and (4) outcome: behaviors, cognitive functions, and whole well-being evaluation results for patients with ADHD. Studies were selected using the following inclusion criteria: (1) articles published in English and (2) RCTs. Exclusion criteria included (1) teenagers, adults, and older adults; (2) comorbidity with autism spectrum disorder, psychosis, and affective or anxiety disorder; (3) consumption of toxic substances; (4) diagnosed with learning disorder; (5) non–peer reviewed studies; and (6) qualitative studies, reviews, cross-sectional studies, case studies, observational studies, study protocols, pre-post studies without a control group, or conference abstracts without full text. Furthermore, the references of the included papers were manually checked for eligibility. After removing duplicate articles, the studies were independently screened by 2 reviewers (KPW and BZ). The selected full-text articles were retrieved and reviewed by 4 reviewers (KPW, JQ, YJX, and BZ).

The risk of bias in each study were independently evaluated by 3 reviewers (KPW, JQ, and BZ) using the Cochrane Collaboration tool for assessing the risk of bias [37]. The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) criteria were adopted for conducting this review [38]. Multimedia Appendix 2 demonstrates the PRISMA checklist. The criteria of the tool included (1) random sequence generation, (2) allocation concealment, (3) blinding of participants and personnel, (4) blinding of outcome assessment, (5) incomplete outcome data, (6) selective reporting, and (7) other bias. In addition, judgment has 3 levels, including “low risk of bias,” “high risk of bias,” and “unclear risk of bias.” The conflicting results were settled by 4 reviewers (KPW, JQ, YJX, and BZ) through discussion.

The information of the selected studies was extracted and coded into different categories, including study characteristics (first author’s name, publication year, country, and setting), characteristics of participants (sample size, sex, and age), intervention and control condition (type of technology used, frequency, length, and duration), outcome measurement (rating scale, test, and questionnaire), and result (mean and SD).

Data processing and analysis were conducted using the Review Manager Software RevMan (version 5.4), Cochrane Collaboration. The standardized mean difference (SMD) with a 95% CI was used to compute the effect size of the continuous outcomes of the interventions. The mean value of the baseline and posttest with SDs and the number of participants in the intervention groups and control groups were selected for the effect size calculation (ie, effect size of group differences). If reported, we selected the results estimated by the analysis of covariance, which treats individual baseline scores as covariates to correct for regression to baseline imbalanced means [39]. If the analysis of covariance had not been reported, change from baseline with SDs and posttest with SDs were selected. Multiple effect sizes were included in the same study, which contradicted the assumption that the effect sizes are independent of each other in the conventional meta-analytic procedures because the effect sizes in one study may be more correlated than those in other studies [40]. The results would become biased if this dependency was not considered. Hence, effect sizes assessed by the same measures were clustered to estimate the association between technology-based interventions and the conditions of school-age children with ADHD. The random effects model was applied in the meta-analysis, given the methodological diversity across the studies. Heterogeneity was measured using the I² value (P<.10; I²>50%), and the higher the value of I², the higher the level of heterogeneity. Subgroup analysis of parent-rated, teacher-rated, computer-rated, and self-rated results was performed, where applicable. To test moderating effects, 5 study-level characteristics, including the number of sessions, sample size, setting, game elements, and types of control group, were selected to calculate the meta-regression for each moderator.

A total of 2404 articles were retrieved from electronic databases. After removing duplicates, the titles and abstracts of 1568 articles were screened. In total, 1347 articles were excluded, and the remaining 221 articles were selected for full-text screening. At this stage, we excluded reviews (n=65), experimental studies (n=31), cross-sectional studies (n=21), longitudinal studies (n=25), qualitative studies (n=24), case studies (n=19), mixed methods studies (n=8), study protocols (n=4), RCT registrations (n=3), and non-English publications (n=2). Finally, 19 RCTs that met the inclusion and exclusion criteria were identified. A flowchart of the study selection process is shown in Figure 1.

A low risk of random sequence generation was recorded from 8 trials (ie, the use of a computer random number generator and web-based system). Seven studies reported a low risk of allocation concealment. Meanwhile, 18 studies had a low risk of blinding participants and personnel. Although some studies had no complete blinding of participants and personnel [20-22,24,30,33,41,42], the reviewers determined that the outcomes were not likely to be affected by a lack of blinding. Six studies were judged to have a low risk of blinding the outcome assessment. Seven studies indicated no blinding of assessors, and 6 studies did not clearly indicate the blinding of assessors. All studies were judged to have a low risk of incomplete outcome data because the attrition rate of all studies was <20%. All studies were judged to have a low risk of selective reporting, as nearly all studies had the protocol and all studies reported the primary and secondary outcomes. Ten studies described the way to manage missing data (ie, intention-to-treat analysis) [19,22-25,27,29,30,41,43]. The risk of bias assessment is shown in Figures 2 and 3 [19,22-25, 27,29,30,41,43].

Fourteen studies used a 2-arm trials design [19-27,29-32,41-44], and the remaining 5 studies had a 3-arm trial design [19,20,22,28,33]. These studies were conducted in 11 countries: Canada [41], Denmark [24], France [33], Germany [31,42], Iran [28], the Netherlands [19,23,30], Norway [21], Singapore [29], Spain [20,32], Switzerland [27], and the United States [22,25,26,43,44].

A total of 1843 school-age children were included in this review, all of whom were diagnosed with ADHD. The number of participants included in each trial ranged from 10 to 246. Most studies were conducted in participants’ homes, and a small number of studies were conducted in schools, classrooms, and clinics. Detailed information is presented in Table 1.

Different types of technologies were adopted in the intervention group among the included studies, including computer-assisted training programs, neurofeedback training, and virtual reality. For the control groups, 6 studies used treatment or medication as the usual approach. The participants in the control groups of 3 included studies did not receive any training. Two studies used stimulants, 7 studies used placebo cognitive training, and 1 study used a digital game. Detailed information is presented in Table 2.

The intervention duration of 3 studies was 4 weeks. Meanwhile, the intervention duration of 9 studies was 5 to 8 weeks. One study had an intervention duration of 12 weeks, 2 studies had an intervention duration of 5 months, and 1 study had an intervention duration of 6 months. For the length of intervention, 1 trial was conducted for 15 minutes per session, 10 studies varied from 25 to 45 minutes per session, 2 studies varied from 45 to 60 minutes per session, and 1 study was conducted for 90 minutes per session.

The adoption of technologies in school-age children with ADHD has become prevalent in this decade and has been applied in many countries. Nevertheless, the efficacy of applying technology-based treatments for children has not been well established. To the best of our knowledge, this work is the first systematic review and meta-analysis of RCTs to assess the efficacy of technology-based interventions for school-age children, which provides strong evidence for health care practitioners. The pooled results indicated that technology-based treatment has the potential to regulate the inattention, overall executive function, and visual attention of school-age children with ADHD.

Most of the included studies were of moderate quality, and only 2 studies [25,29] fulfilled the 7 criteria for randomized trials. More than half of the included studies had an unblinded or unclear method of randomization and allocation concealment. Although the blinding of participants was unpreventably broken owing to the nature of the intervention in most studies, the level of influence on the outcomes was considered low in most studies. Only 6 studies were assessed by blinded assessors. The attrition rate in all interventions did not exceed 20%, which met the criteria of the dropout rate. Only 8 studies reported the intention-to-treat analysis. Thus, more rigorous methods should be designed to scientifically justify the effects of technology-based interventions for children with ADHD.

The pooled results indicated that technology-based treatment may have the potential to improve computer-rated inattention. No significant results were found for parent-rated and teacher-rated inattention and parent-rated and teacher-rated hyperactivity or impulsivity. This finding is consistent with that of previous studies, which showed that computer-assisted training had a positive effect on inattention [45,46]. Moreover, neurofeedback training and brain-computer interfaces training involve feeding brain activity into a computer, which maps the areas of the brain [47]. The patient controls a computer or video game by producing a brief, sustained brainwave activity in the target area, thereby training the brain regions that are aroused. These treatments include cognitive training to remediate the inattention behaviors of patients with ADHD. The insignificant finding on hyperactivity or impulsivity may be explained by the harmful effect of screen time and the frequent digital media use, which seem to increase the risk of the symptoms of hyperactivity or impulsivity. A meta-analysis conducted by Nikkelen et al [17] identified a modest association between the use of traditional digital media and ADHD behaviors. Notifications and invitation messages may pop up in these training programs. Exposure to these notifications and messages may take away the attention of a child from the task, possibly disrupting the normative development of patience and impulse control [48]. The precise descriptions of computer training, neurofeedback training, and virtual reality have not been completely demonstrated. Thus, relevant information is needed to provide further studies to assess the impact of the messages that suddenly pop up regarding the symptoms of hyperactivity and impulsivity among children with ADHD. Furthermore, it was found that a smaller sample size, nonhome setting, game elements excluded, and no treatment and nonequivalent treatment as the control group tended to have more improving effects on ADHD behavior while conducting the technology-based intervention. This suggests that fewer participants, conducting the intervention at clinic and school, intervention without game elements, and using no treatment or nontechnological intervention as control groups may increase the potential of the technology-based intervention to alleviate ADHD behaviors.

Nine categories, namely, inhibition, working memory, flexibility, emotional control, initiation, planning and organization, organizing materials, monitoring, and metacognition, were used to evaluate the executive functions. The pooled result showed a statistically significant effect of the technology-based intervention on parent-rated executive functions. A large significant effect of the control group on computer-rated working memory compared with the technological intervention was found, indicating computed-rated working memory would favor the control group instead of technological intervention. No significant effects were found from other categories related to executive functions. These findings were unexpected, as most previous studies reported positive effects of technology-based training on the improvement of executive functions [18,49,50].

Evidence indicated that technology-based interventions and computer-assisted training may regulate both parent-rated disruptive behavior disorders. This finding is consistent with the previous case study by Kotwal et al [51], who found that the frequency of parent-reported disruptive behaviors in a boy with ADHD was reduced at home and school after conducting a 3-month computer-assisted training. More number of sessions, small and large sample sizes, nonhome settings, game elements excluded, and no treatment as the control group seemed to improve the disruptive behavior disorder when adopting the technology-based intervention. It is unexpected that small and large sample sizes might moderate the association between technology-based interventions and disruptive behavior disorder. Thus, further investigations need to be conducted to evaluate the moderating effect of different sample sizes.

Participants’ ability to take in important visual information seems to be significantly improved after conducting computer-assisted training. Ordikhani-Seyedlar et al [52] used brain-computer interfaces with electroencephalograms to generate neurofeedback that significantly improved the visual attention of patients with ADHD. In addition, numerous studies have reported the positive outcomes of adopting neurofeedback training in patients with ADHD [53,54]. Alpha-beta activities in the electroencephalogram have been used to evaluate the attention level of the participants [52]. Furthermore, moderate sessions, moderate sample size, mixed setting, and game elements excluded and using equivalent treatment as a control might improve visual attention under the technology-based interventions. It is somewhat surprising that a mixed setting and equivalent treatment as the control group may have a moderating effect; thus, more RCTs should be conducted to identify the underlying reasons.

The pooled result showed a nonsignificant result on auditory attention. However, some experimental studies have reported the significant effect of biofeedback technology on strengthening auditory attention in children with ADHD [53]. Auditory attention among children with ADHD may be influenced by the intensity and frequency of sound [54,55]. White noise therapy has the potential to improve speech recognition and auditory attention in children with ADHD [56,57]. White noise, which is a random signal with equal intensity at different frequencies, is a steady, unchanging, and unobtrusive sound from certain machines such as a whirring fan and a static radio. The intensity and frequency of sound can remarkably influence the auditory attention of children with ADHD.

No statistically significant effect was found for reading fluency. Nevertheless, previous experiments reported an improvement in oral reading fluency through a computerized program, Headsprout [58]. The computerized program could enhance the engagement of students with ADHD. Hence, learning among these students becomes more effective. This observation also implies that technology, such as a computerized program, has a high potential to enhance oral reading fluency. More RCTs with a larger pediatric population are needed to evaluate the effects of technology on reading fluency among children with ADHD.

Our result regarding the effects of the technology-based intervention on adherence to stimulants was not statistically significant. The literature review conducted by Chacko et al [59] indicated that technology-based interventions could reduce environmental barriers, such as caregivers’ forgetfulness, to improve medication adherence for the pediatric population with ADHD. This nonsignificant finding may be explained by the fact that only a small number of RCTs were conducted to determine the effects of using technologies on stimulant adherence in children with ADHD.

A narrative review and pooled results found that technology had no significant effects on sensitivity to punishment and sensitivity to reward and quality of life. Few qualitative and quantitative studies have been conducted to investigate the effectiveness of technology-based interventions on sensitivity to punishment and sensitivity to reward and quality of life among children with ADHD. This circumstance indicates that more trials should be conducted.

This study is the first systematic review and meta-analysis of RCTs to examine the effectiveness of technology-based treatment for school-age children with ADHD. The strengths of this review include the use of well-defined inclusive and exclusive criteria, the application of a rigorous and well-constructed search strategy from the 7 electronic databases, and the stringent quality assessment of selected RCTs. However, this study has some limitations. First, the small sample size of the included trials implies the sparsity of RCTs in the children-computer interaction research field. Second, the pooled results may be influenced by variations in intervention designs (ie, intervention intervals and duration) because of the considerable clinical heterogeneity. Third, more than half of the included RCTs were designed using a nonblind approach owing to the unavoidable physical component of the interventions. To minimize these biases, blinded outcome assessors are strongly recommended for open-label RCTs. Fourth, the age selection was between 6 and 12 years, which may exclude preschool age children and adolescents with ADHD benefiting from technology-based treatment. Finally, RCTs published in languages other than English were excluded. Thus, language bias may be premeditated. Although RCT provide excellent internal validity and the most reliable evidence, the results are needed to be interpreted with caution owing to the inclusion of the designated population in this review. Furthermore, the use of virtual reality–based treatments for children with ADHD has been gradually implemented in the field of developmental psychology. RCTs assessing the effects of virtual reality–based treatments on children are still limited; thus, more relevant research is recommended. Moreover, the safety issues of children’s interactions with technological interventions have rarely been discussed in RCTs. Measurements of side effects can be adopted in studies to further optimize the designs of the technological treatment, which can facilitate the universality of technological therapies in patients of different ages.

This review synthesized evidence from 19 RCTs on the application of technology-based interventions in school-age children with ADHD. The results indicated that the existing RCTs were mainly of low to moderate quality, particularly in random sequence generation, allocation of concealment and blinding, and intention-to-treat analysis. Further well-designed and rigorous trials should be conducted to determine the effectiveness of technology-based interventions in children with ADHD. The pooled results indicated that children’s intervention with technological treatments has effects on computer-rated inattention, parent-rated overall executive functions, parent-rated disruptive behavior disorder, and visual attention. The number of sample sizes, setting of the intervention, exclusion of game elements, and type of control group had moderating effects on ADHD behaviors, disruptive behavior disorder, and visual attention. The potential of technology-based therapies for hyperactivity or impulsivity, inhibition, working memory, flexibility, emotional control, initiation, planning and organization, organizing materials, monitoring, metacognition, disruptive behavior disorder, visual attention, auditory attention, sensitivity to punishment and reward, quality of life, adherence to stimulants, and reading fluency requires further evaluation.