Digital Therapeutics for ADHD: What the Evidence Actually Establishes
Digital therapeutics for attention-deficit/hyperactivity disorder (ADHD) are software-based interventions designed to change clinically relevant attention, behavior, self-management, or functioning through structured therapeutic activities. Some are regulated software medical devices with defined indications; others are studied digital interventions delivered through games, cognitive training platforms, mobile applications, or internet-based psychotherapy. Their scientific value depends on the product, its target population, the outcome measured, the quality of controlled evidence, and the extent to which gains carry into ordinary life. The term digital therapeutic is therefore a claim requiring precise examination rather than a guarantee that a screen-based activity treats ADHD.
ADHD is a neurodevelopmental disorder diagnosed through developmental and clinical assessment of persistent inattention and/or hyperactivity-impulsivity, functional impairment, and alternative explanations. No software score independently establishes ADHD, and improvement during a computerized task does not by itself show remission of the disorder. The major clinical question is whether a specific digital intervention produces meaningful benefits beyond expectation, repeated practice, and the assistance already provided by established treatment. The current evidence supports selective adjunctive use of some tested programs and continued investigation of others, with considerable variability by technology and study design. NICE NG87 and the 2025 review of reviews establish the larger clinical and evidentiary context.
What Counts as a Digital Therapeutic for ADHD?
Regulated therapeutic software
Software as a medical device can deliver a treatment activity through an application while its manufacturer specifies the clinical indication, intended users, instructions, relevant risks, and outcome evidence. Authorization concerns a particular product and label. The United States Food and Drug Administration granted a De Novo request for EndeavorRx in June 2020, establishing a Class II category of digital therapy software for ADHD; the original indication concerned attention function measured by computerized testing in children aged 8–12 with predominantly inattentive or combined presentation and a demonstrated attention problem. Subsequent authorization expanded the labeled pediatric age range to 8–17 years. FDA De Novo record; current EndeavorRx indications.
EndeavorOTC is a distinct over-the-counter software medical device cleared through the FDA's 510(k) process in June 2024 for adults aged 18 years and older with predominantly inattentive or combined-type ADHD and a demonstrated attention issue. The actual FDA indications specify improvement in attention function measured by the Test of Variables of Attention (TOVA), while cautioning that typical behavioral symptoms may not improve and that the program is intended as part of a broader therapeutic approach. The distinction between an authorized attention-test endpoint and comprehensive ADHD treatment is central to accurate counseling. FDA 510(k) K233496.
Regulatory language matters. A De Novo authorization, 510(k) clearance, marketing eligibility, evidence from a trial, and endorsement by a professional guideline are different statements. An authorization from the FDA also applies to a defined device version, target group, and use; it cannot be transferred to the wider category of ADHD games. Software offered through an ordinary app store may be a wellness product, a research intervention, or regulated therapeutic software, depending on its actual function and applicable regulation.
Cognitive training, serious games, and adaptive attention tasks
Computerized cognitive training typically repeats exercises involving working memory, sustained attention, response inhibition, or task switching. Serious games embed similar demands in navigation, goals, feedback, and reward cycles. Adaptive programs alter their challenge level according to performance; some use statistical optimization or machine learning. These approaches may improve the practiced skill, although transfer to nontrained clinical outcomes requires independent measurement. The distinction is important because repeated success on a familiar task provides evidence of learning and engagement, while generalized improvement in classroom attention, relationships, organization, or occupational performance requires additional outcomes. Elbe et al. (2023).
Internet-delivered psychotherapy and structured behavioral programs
Digitally delivered cognitive behavioral therapy (CBT) uses psychoeducation, cognitive restructuring, planning, activity scheduling, skills practice, and behavior experiments through websites or apps, sometimes with clinician guidance. These programs target self-management, dysfunctional expectations, organization, and everyday impairments, rather than solely training a cognitive task. A program can be a genuine clinical intervention even when it has not received the specific type of FDA device authorization associated with a commercial game. Human-supported, self-guided, and blended versions have different evidence, adherence profiles, and resource requirements. Internet-delivered CBT trial (2023); self-guided intervention trial (2023).
Everyday apps, AI assistants, and digital companions
Reminders, digital calendars, task managers, timers, focus applications, wearable prompts, and general-purpose conversational systems may be useful accessibility and organizational tools. Their effectiveness for an individual can be meaningful, but convenience and consumer satisfaction do not demonstrate disorder-specific therapeutic efficacy. AI assistants that generate schedules or messages belong to a different evidence category from regulated clinical software or manualized psychotherapy. Unstructured conversations with a general chatbot do not constitute an ADHD diagnostic assessment, and a companion's conversational fluency does not establish its clinical reliability. Evidence statements must attach to the exact intervention evaluated, not to the presence of artificial intelligence in its marketing.
The Clinical Target: Attention, Symptoms, and Everyday Function
ADHD symptoms and diagnosis
Inattention, hyperactivity, and impulsivity are symptom dimensions with substantial variation across age and context. A clinical ADHD diagnosis additionally requires a developmental history, impairment, information across relevant settings, and consideration of sleep, anxiety, mood disorders, trauma, learning disorders, substance use, and other explanations. The appropriate clinician synthesizes these observations; questionnaires and computerized tests provide pieces of information. ADHD may coexist with autism and other conditions, and digital-treatment research frequently includes selected participants whose findings may not generalize to complicated clinical presentations. NICE NG87.
Sustained attention and computerized performance
The TOVA and related continuous performance tests measure particular patterns of responding during a standardized task: omissions, commissions, reaction-time variability, and other attention-related parameters. Such measures can quantify performance under controlled conditions and may be sensitive to task-specific change. They do not reproduce the competing demands of a school day, employment, family responsibilities, or social relationships. A test score can improve through specific skill acquisition, greater familiarity with test conditions, arousal, or genuine broader change. Proper studies use comparators and independently measured outcomes to identify which interpretation is justified.
Executive functions and transfer
Executive functions describe related cognitive control processes, including working memory, inhibition, planning, monitoring, and cognitive flexibility. Tests, behavior-rating inventories, and real-world tasks capture overlapping but distinct dimensions. An improvement in working-memory training is a near-transfer finding when the test resembles the trained task; a sustained decrease in missed assignments, unsafe driving, or occupational errors is farther transfer. Claims about clinically useful executive-function improvement should specify the measure, rater, duration of follow-up, and magnitude of benefit. Lee et al. (2025).
Impairment, quality of life, and patient priorities
An intervention becomes valuable when its effects address the difficulties that matter to the individual: school participation, emotional burden, meeting deadlines, reducing conflict, managing medication, maintaining routines, or improving quality of life. Symptom-rating improvement and actual functional recovery are related but not interchangeable outcomes. A software trial that establishes one endpoint should not be described as having established all of them. This principle also protects patients from paying for apparent precision in cognitive metrics that have uncertain relevance to their daily difficulties.
How ADHD Digital Interventions Are Supposed to Work
Adaptive challenge and attentional control
Adaptive cognitive-motor games present two simultaneous demands, such as visual target selection while navigating a moving environment. Algorithms regulate difficulty so a participant repeatedly practices responding under interference and shifting demands. This design seeks to engage attention allocation and cognitive control while gamification encourages adherence. The underlying therapeutic hypothesis is that repeated practice changes performance mechanisms enough to affect clinically relevant attention. A plausible hypothesis becomes established clinical evidence only when benefits surpass appropriate controls and appear on independently measured outcomes. Kollins et al. (2020)30017-0).
Repeated cognitive practice and learning
Training can produce reliable learning inside the trained environment through attention to cues, recognition of patterns, rule acquisition, increased speed, and stronger task strategies. Those gains do not necessarily change the distributed developmental and environmental processes contributing to ADHD. A valid evaluation separates trained-task accuracy, near-transfer cognition, clinical symptoms, and adaptive functioning. It also asks whether any effects persist after the program ends and whether a different task shows the same result.
Behavioral scaffolding and implementation
Digital CBT and structured self-management programs operate through another set of mechanisms. They may help people construct external reminders, sequence complex tasks, anticipate obstacles, practice planning, manage avoidance, and repeatedly review actual behavior. Their benefit may arise through improved application of strategies to real circumstances rather than a direct enhancement of working-memory capacity. This makes ecological validity especially important: meaningful improvements should be demonstrable in daily routines and functioning, ideally using information beyond a single unblinded respondent. Internet-delivered CBT trial (2023).
Engagement, rewards, and sustained use
Game mechanics can sustain attention during a session, especially when compared with repetitive exercises that participants find tedious. Yet time spent in a program, enjoyment, and completion rates are implementation outcomes, not clinical efficacy measures. A digital treatment must deliver a tolerable dose without crowding out sleep, exercise, schooling, relationships, or other care. Attrition, interruptions, and limited technological access can reduce actual effectiveness even if a research protocol produces favorable efficacy estimates.
Evidence From Randomized Trials
STARS-ADHD and the pediatric EndeavorRx evidence
The pivotal STARS-ADHD trial evaluated an adaptive digital intervention in children with ADHD using a randomized comparison with a digital control condition. The trial reported improvement in an objective computerized measure of attention, an important proof of principle for targeted digital treatment. This outcome should be interpreted together with the study's measures of symptoms, functional impairment, follow-up, and reported adverse events. An improvement on TOVA is evidence for the specified attentional endpoint; it is not equivalent to proof of broad improvement in all ADHD symptoms or functioning. Regulatory labeling likewise explicitly limits how the result should be presented. Kollins et al. (2020); FDA De Novo decision.
The intervention was designed for selected children, and eligibility criteria and structured trial conditions matter. Pediatric evidence cannot be transferred automatically to preschool children, university students, or older adults. Replication under ordinary care, clinically meaningful independent outcomes, adherence outside trials, and long-term follow-up determine how far its original findings should guide practice.
Adult regulated software and the EndeavorOTC indication
EndeavorOTC applies related software technology to a different regulatory population. The FDA's June 2024 clearance identifies adults 18 and older with primarily inattentive or combined-type ADHD and attention difficulties, with improvement in computerized attentional function as the indicated target. The instructions explicitly say that it should be incorporated within a therapeutic program rather than replacing established treatments. Adult patients may care more about time management, employment, interpersonal functioning, or household demands than a stand-alone test score; those domains require their own evidence and discussion. FDA K233496.
There is also an important distinction between an over-the-counter product's regulatory availability and unrestricted use. Safety exclusions, age restrictions, reasonable expectations, and the need for clinical assessment still apply. People with new or changing mental health symptoms should seek an assessment rather than treating the app's engagement data as a diagnosis.
Digital CBT in adults
Adult ADHD trials of digital CBT-informed treatments are heterogeneous. A 2023 randomized study compared internet-delivered CBT with internet-delivered applied relaxation and treatment as usual, which is valuable because the active comparator helps distinguish general digital participation from therapy-specific elements. Another 2023 study tested a self-guided intervention against online psychoeducation. These research designs address different questions from an attention-training video game and should be reviewed separately. Internet-delivered CBT (2023); self-guided intervention (2023).
A later randomized trial of a CBT-informed mobile app involved 154 assessed adults aged 18–55 and compared eight weeks of app use with a waitlist. The report found advantages for self-reported ADHD symptoms and quality of life, while its functional-impairment measure did not show a corresponding effect. That divergence illustrates why investigators must predefine a primary clinical outcome and include active controls when possible. Bridging the Gap trial.
A 2026 pragmatic randomized trial of the self-guided CBT- and mindfulness-based program attexis recruited 337 adults with confirmed ADHD and evaluated the intervention as an adjunct to treatment as usual. It reported greater symptom improvement after three months, with follow-up at six months. The findings support investigation of accessible adult adjunctive care, while interpretation should still consider assignment visibility, patient-reported measures, generalizability, and whether benefits persist over longer intervals. attexis trial (2026).
Another 2026 randomized trial of an unguided adult digital intervention enrolled 307 participants and reported better quality of life over 12 weeks in the intervention group. Its authors highlighted adherence challenges and limitations in diagnostic standardization and long-term inference. Those methodological details are important because a positive average treatment effect under imperfect use can coexist with modest real-world uptake or unequal outcomes among patients. Unguided digital intervention trial (2026).
Cognitive training in adults
Computerized cognitive training has a less consistent adult clinical signal. A 2023 systematic review and meta-analysis of randomized trials found a small overall cognitive benefit across nine comparisons, but did not find significant improvement in ADHD symptom severity or the specified executive-function, cognitive-speed, and working-memory subdomains. The result supports a differentiated conclusion: practicing a cognitive intervention may shift selected aggregate test outcomes without yet establishing broad adult ADHD symptom treatment. Elbe et al. (2023).
What Systematic Reviews and Meta-Analyses Show
Small average benefits across diverse interventions
A 2024 meta-analysis of 25 randomized trials with 1,780 participants reported a modest pooled reduction in overall ADHD symptoms for digital interventions (standardized mean difference –0.33; 95% confidence interval –0.51 to –0.16), with smaller effects for hyperactivity-impulsivity than for inattention. This is an average across different technologies and populations, not a predicted response for an individual app. Statistical significance does not establish long-term improvement or comparable clinical effectiveness across interventions. Liu et al. (2024).
A 2025 game-based intervention meta-analysis of eight randomized trials reported some benefits in visuospatial short-term and working-memory measures, whereas selected parent- or teacher-relevant behavioral executive-function measures did not show statistically significant improvements. This mismatch between test performance and behavior supports carefully bounded therapeutic claims. Lee et al. (2025).
The 2025 review of reviews and safety reporting
A systematic review of 26 existing reviews, spanning diverse technologies and more than 34,000 participants across their underlying bodies of research, concluded that possible benefits coexist with generally low-quality and heterogeneous evidence. Only eight of the 26 reviews reported adverse effects, including physical discomfort, emotional reactions, and behavioral difficulties. Summed participant counts across overlapping reviews should not be read as a unique pooled trial cohort. Inconsistent harm reporting is itself a reason for caution, especially for programs used by children. Review of reviews (2025).
The 2026 network meta-analysis
Hao and colleagues' 2026 network meta-analysis assembled 32 randomized trials involving 2,819 pediatric participants and compared single-task versus cognitive-motor dual-task digital therapeutics, with additional grouping by the presence of AI-integrated adaptation. AI-integrated cognitive-motor dual-task interventions ranked highly across the outcomes examined. But 40.6% of included trials were rated at high risk of bias, and prediction intervals showed substantial uncertainty for several symptom and executive-function measures. Rankings summarize a statistical network under assumptions about comparability; they are not proof that every device classified in the leading category is superior in routine clinical practice. Hao et al. (2026).
The review's classification of AI-integrated features is a research taxonomy. Clinically, an algorithm's complexity is secondary to demonstrated benefit and safety. Simple adaptive logic can be therapeutically useful, and advanced machine learning may introduce new problems involving interpretability, dataset shifts, personalization errors, and privacy. Comparative effectiveness requires direct, well-powered trials rather than technological labels alone.
Evaluating Scientific Certainty
Active controls, masking, and expectancy
A waitlist or usual-care comparator can estimate the benefit of being offered a digital program, but participants know whether they received an additional intervention. This can influence motivation, reporting, and attention to symptoms. A credible active digital control may reduce expectancy differences, although matching engagement and perceived therapeutic credibility is difficult. Independent blinded ratings and objective outcomes strengthen confidence, provided the measures are clinically interpretable.
Outcomes that were trained, tested, and lived
When a program trains a task closely resembling the test used to judge it, improved performance could reflect repeated practice or near transfer. A stronger evidence sequence shows benefit on an independent task, then on observer-rated ADHD symptoms or function, then on durable participation outcomes. These steps are not interchangeable, and evidence should be described at the highest level actually demonstrated. The FDA labels for EndeavorRx and EndeavorOTC appropriately name the particular computerized attention outcome. FDA pediatric record; FDA adult clearance.
Sample selection and age boundaries
Studies often recruit children with a particular ADHD presentation, adults with relatively stable medication, or people who can engage in scheduled sessions. Exclusion of severe comorbidity, learning differences, sensory limitations, inadequate home internet, or unstable care reduces representativeness. Evidence should specify age, diagnostic verification, medication status, severity, language, and socioeconomic context. A program validated in one group should be evaluated before making equivalent claims for another.
Follow-up, missing data, and publication bias
Many digital intervention studies are short and investigate immediate post-treatment effects. Long-term durability, maintenance dosing, late harms, real-world disengagement, and cost effectiveness remain less clear. Selective dropout can improve the apparent mean response if dissatisfied participants stop reporting outcomes. Pre-registered primary endpoints, intention-to-treat analyses, prespecified harm collection, independent replication, and open protocols strengthen confidence. The promise of an intervention does not resolve these methodological requirements.
Clinical Assessment and Selecting a Digital Intervention
A clinician begins with the diagnosis, symptom burden, functioning, goals, co-occurring conditions, and treatments already used. Clinical assessment is distinct from symptom screening, rating-scale monitoring, computerized attention tests, and any app-based recommendation score. A software intervention is considered after its intended population and therapeutic indication are matched to the person's needs. If a patient's primary goal is fewer missed assignments or improved employment functioning, the treatment should be judged on those outcomes rather than merely on in-app progress.
The program's evaluation should include the actual software name and version, regulator and jurisdiction, labeled age and presentations, contraindications, study comparators, independent outcomes, adverse effects, expected time commitment, access costs, and privacy practices. An otherwise promising intervention may be unsuitable when device access is limited, sensory needs are substantial, or engagement worsens stress and conflict. NICE NG87; FDA EndeavorOTC indications.
Digital tools do not perform the diagnostic evaluation
It is particularly important to distinguish therapeutic games from digital assessments such as QbTest. NICE HealthTech guidance HTG729 supports QbTest as an optional aid to diagnosis in 6–17-year-olds, used with a standard assessment by an appropriately qualified clinician; it does not convert QbTest into an independent diagnostic instrument or an established therapeutic intervention. That guidance also notes insufficient evidence to recommend those assessment technologies for monitoring treatment response. A digital diagnostic-support tool and digital treatment are different classes of technology and evidence. NICE HTG729; committee discussion.
Working alongside established ADHD care
ADHD management integrates age-specific psychoeducation, environmental adjustment, evidence-based medication when indicated, parent or caregiver support, school accommodations, and psychological interventions. For children, the developmental context and involvement of parents and teachers remain central. For adults, practical accommodations, structured CBT, treatment of co-occurring problems, and shared decision-making may be particularly relevant. A digital therapeutic can be selected as one component when it has relevant evidence, but its use should not delay necessary assessment or replace effective care without a sound clinical reason. NICE NG87.
Safety, Accessibility, and Ethics
Adverse effects and tolerability
Digital therapies may involve visual stimulation, repeated focused screen time, motor interactions, frustration, motion sensitivity, or discomfort. Device-specific contraindications and instructions matter, especially with photosensitive conditions, visual differences, and physical limitations. Children can become distressed when difficulty algorithms repeatedly expose failure, and households can experience conflict when adherence becomes coercive. Trials and postmarketing evaluations should actively solicit adverse experiences instead of presuming that software without medication carries no risk. Review of reviews (2025).
Attention, sleep, and the screen-time context
Therapeutic use is defined by a structured regimen, an appropriate goal, and monitored outcomes, not by an unrestricted amount of gaming. Sessions that displace sleep, learning, social activities, or movement can undermine overall well-being. During care planning, families and adults should consider session scheduling, breaks, accessibility settings, device ergonomics, and an explicit stop plan when distress or impairment increases.
Privacy, sensitive data, and algorithmic governance
A digital intervention may record reaction times, engagement, adherence, symptom ratings, and sensitive clinical details. Access controls, data retention, commercial data sharing, software-update governance, and informed consent should be transparent. For minors, parent consent, child assent where appropriate, and the effect of school-mediated access need additional attention. Adaptive personalization does not itself verify therapeutic precision or responsible data practices. Gregg and Lehmann (2024).
Accessibility and inequity
Affordability, compatibility with older devices, broadband quality, literacy, language, visual or motor accessibility, sensory preferences, and caregiver availability can all affect real-world reach. A program developed and tested under well-supported research conditions may produce different results in families with precarious access or in underrepresented populations. Study enrollment should include diverse users and report how benefit and attrition vary by context.
Conflicts of interest and independent reproduction
Device development and clinical trials can involve manufacturer sponsorship or author relationships with the company that owns the intervention. Such support does not automatically invalidate findings, but replication by independent teams, prospective trial registration, data transparency, and separate validation of patient-important outcomes are valuable. Readers should examine disclosures as part of evidence appraisal rather than inferring reliability from high production quality or institutional branding.
Children, Adolescents, and Adults: Different Clinical Questions
Children aged 8–17 may be eligible for particular authorized pediatric software, depending on exact indications and individual circumstances. For children and adolescents, the research must investigate developmental stage, school functioning, parent ratings, clinical severity, usability, and whether benefits survive beyond a short training period. Digital programs can be more feasible when the environment supports predictable sessions and feedback without turning them into a source of household conflict.
Adults have other priorities and evidence populations. The indicated adult over-the-counter product targets an attention-test outcome; digital CBT studies assess symptoms, planning, quality of life, and sometimes impairment. The choice between a cognitive game, psychoeducational program, internet CBT, and a simple digital reminder should reflect the person's goal and evidence for that exact intervention. No favorable result in one age group establishes equivalent effects in another.
For preschool-age children, older adults, and people with severe co-occurring psychiatric or neurological conditions, gaps in trial inclusion are especially relevant. Clinicians should exercise appropriate caution when a product's evidence is extrapolated beyond its studied and authorized population. Software accessibility is not a substitute for a developmental formulation of the person's needs.
Digital Therapeutics and the Artificial Era
The growing role of therapeutic software belongs to the broader transformation of healthcare through algorithmic and nonbiological systems. In clinical practice, intelligent adaptation may organize tasks, change stimulus timing, or personalize the pace of intervention. Its medical standing comes from the measurable reliability and benefit of a particular product under accountable oversight. A user-facing AI assistant, a regulated adaptive device, and a therapeutic relationship with a qualified clinician occupy different roles in care. This distinction permits innovation to proceed alongside a clear scientific account of efficacy, accountability, human decision-making, and the limits of evidence.
Practical Questions and Evidence-Based Answers
Can a video game treat ADHD?
A tested game-based intervention can have a defined therapeutic indication and may improve specific attention-related outcomes. Its evidence may not show improvements in every symptom, executive function, or everyday impairment. The clinical judgment concerns the particular program and its labeled use, not gaming as an activity. FDA EndeavorRx; Kollins et al. (2020).
Does an FDA authorization mean the software works better than medication?
An FDA authorization supports a particular device's regulated indication and supporting evidence. It does not establish superiority to stimulant or nonstimulant medications, and it does not imply that the application replaces behavioral treatment or school and workplace accommodations. Comparative treatment claims require head-to-head trials with clinically appropriate endpoints. FDA K233496; NICE NG87.
Is computerized attention improvement the same as symptom remission?
No. Computerized performance, symptom ratings, diagnosis, and functional impairment are distinct constructs. A result on TOVA or another specific task should be reported as improvement in that measured ability. Sustained behavioral and functional improvement needs separate evidence.
Do digital therapies work for adults?
Certain adult digital CBT programs have produced promising randomized-trial results, while computerized cognitive-training evidence is less consistent for overall symptoms. EndeavorOTC has a specified adult attention-function indication. Product type, control condition, follow-up duration, and patient-important outcomes determine the strength of any adult treatment claim. Elbe et al. (2023); attexis trial (2026).
Can a focus app or AI companion diagnose ADHD?
A productivity app, conversation system, or engagement score cannot establish a clinical ADHD diagnosis. Assessment relies on appropriately trained professionals, longitudinal clinical information, impairment, and differential diagnosis. An app may assist organization, monitoring, or a structured therapy program, depending on its validated purpose. NICE NG87.
Are digital therapeutics safe for children?
Safety varies by product, individual health conditions, regimen, and household context. The absence of medication-related side effects does not eliminate visual, emotional, behavioral, privacy, or accessibility risks. Parents and clinicians should use current labeling and monitor relevant harms, particularly where long-term evidence is sparse. FDA pediatric indications; 2025 review of reviews.
What evidence would make future software recommendations stronger?
Independent multisite randomized trials with credible active comparators, preregistered clinically important outcomes, blinded ratings when feasible, sufficient follow-up, transparent adverse-event collection, diverse recruitment, and real-world adherence data would clarify who benefits and by how much. Comparative cost-effectiveness and independent replication are particularly valuable for systems with recurring fees and frequent software revisions.
Clinical and Scientific Conclusion
Digital therapeutics are a developing component of ADHD care. The clearest regulated examples specify measurable improvements in attention performance in precisely defined populations. Broader research on digital CBT, cognitive training, and therapeutic games identifies some favorable outcomes, accompanied by differences in study quality, interventions, transfer to everyday life, and reported harms. The evidence supports a product-specific, outcome-specific approach: establish ADHD clinically, define the functional goal, evaluate the exact program and its regulatory indication, integrate it with appropriate established care, and measure results that matter beyond the screen. Liu et al. (2024); Hao et al. (2026); NICE NG87.
Related Articles
Cluster ADHD and Executive Function — The central scientific chapter covering ADHD diagnosis, mechanisms, lifelong development, care, and everyday functioning.
ADHD Treatment Without Medication: Behavioral and Environmental Support — Evidence-based behavioral and environmental interventions, with attention to the limits of digital self-help.
ADHD Interventions at School: Classroom Behavior and Learning Support — School-based, person-centered approaches and outcomes beyond computerized attention tasks.
Nonstimulant ADHD Medications: Atomoxetine, Guanfacine, and Other Options — Age-specific medication options and monitoring within multimodal treatment decisions.
Stimulant Treatment for ADHD: Methylphenidate and Amphetamines — Clinical trial evidence for medications, helping place software-based care in the overall treatment landscape.
References
Antshel, K. M., McBride, H., and Knouse, L. E. (2026). Bridging the Gap: Digital CBT for Adults Managing ADHD Challenges. Journal of Attention Disorders, 30(3), 370–385. https://doi.org/10.1177/10870547251384462.
Baumeister, A., Schuurmans, L., Schöttle, D., Reif, A., Kahl, K. G., Moritz, S., Karow, A., Lambert, M., Betzler, F., and Philipsen, A. (2026). Efficacy of an Unguided Digital Intervention for Adults With Attention-Deficit/Hyperactivity Disorder: A Randomized Controlled Trial. Deutsches Ärzteblatt International, 123(14), 390–396. https://doi.org/10.3238/arztebl.m2026.0065.
D'Amelio, R., Betz, L. T., Jow, S. M., Retz, W., Philipsen, A., Klein, J. P., Fassbinder, E., Jacob, G. A., and Retz-Junginger, P. (2026). Effectiveness of attexis, a digital intervention based on cognitive behavioral therapy for adults with ADHD: A randomized controlled trial. Psychological Medicine, 56, e54. https://doi.org/10.1017/S0033291726103390.
Elbe, P., Bäcklund, C., Vega-Mendoza, M., Sörman, D., Malmberg Gavelin, H., Nyberg, L., and Ljungberg, J. K. (2023). Computerized cognitive interventions for adults with ADHD: A systematic review and meta-analysis. Neuropsychology, 37(5), 519–530. https://doi.org/10.1037/neu0000890.
Food and Drug Administration. (2020). EndeavorRx De Novo decision, DEN200026. https://www.accessdata.fda.gov/cdrh_docs/pdf20/DEN200026.pdf.
Food and Drug Administration. (2024). EndeavorOTC 510(k) substantial equivalence and indications, K233496. https://www.accessdata.fda.gov/cdrh_docs/pdf23/K233496.pdf.
Gabarron, E., Denecke, K., and Lopez-Campos, G. (2025). Evaluating the evidence: A systematic review of reviews of the effectiveness and safety of digital interventions for ADHD. BMC Psychiatry, 25, 414. https://doi.org/10.1186/s12888-025-06825-0.
Gregg, A. T., and Lehmann, L. S. (2024). Privacy and Consent in Mobile Health: Solutions for Balancing Benefits and Risks. Mayo Clinic Proceedings: Digital Health. https://doi.org/10.1016/j.mcpdig.2024.05.005.
Hao, Z., Xu, H., Wang, C., Wang, B., and Qiu, Z. (2026). Multimodal Digital Therapeutics Enhanced by Task Design and AI for Attention-Deficit/Hyperactivity Disorder Core Symptoms and Executive Functions in Children and Adolescents: Systematic Review and Network Meta-Analysis of Randomized Controlled Trials. Journal of Medical Internet Research, 28, e95043. https://doi.org/10.2196/95043.
Kenter, R. M. F., Gjestad, R., Lundervold, A. J., and Nordgreen, T. (2023). A self-guided internet-delivered intervention for adults with ADHD: Results from a randomized controlled trial. Internet Interventions, 32, 100614. https://doi.org/10.1016/j.invent.2023.100614.
Kollins, S. H., DeLoss, D. J., Cañadas, E., Lutz, J., Findling, R. L., Keefe, R. S. E., Epstein, J. N., Cutler, A. J., and Faraone, S. V. (2020). A novel digital intervention for actively reducing severity of paediatric ADHD (STARS-ADHD): A randomised controlled trial. The Lancet Digital Health, 2(4), e168–e178. https://doi.org/10.1016/S2589-7500(20)30017-030017-0).
Lee, H., Lee, S., Hwang, M., and Woo, K. (2025). Effectiveness of game-based digital intervention for attention-deficit hyperactivity disorder in children and adolescents: A systematic review and meta-analysis using Beard and Wilson's conceptualization of perception in experiential learning. European Child and Adolescent Psychiatry, 34, 3853–3869. https://doi.org/10.1007/s00787-025-02788-5.
Liu, X., Yang, Y., Ye, Z., Wang, F., Zeng, K., Sun, Y., Huang, Y., and Dai, L. (2024). The effect of digital interventions on attention deficit hyperactivity disorder (ADHD): A meta-analysis of randomized controlled trials. Journal of Affective Disorders, 365, 563–577. https://doi.org/10.1016/j.jad.2024.08.156.
Nasri, B., Cassel, M., Enhärje, J., Larsson, M., Hirvikoski, T., Ginsberg, Y., Lindefors, N., and Kaldo, V. (2023). Internet delivered cognitive behavioral therapy for adults with ADHD: A randomized controlled trial. Internet Interventions, 33, 100636. https://doi.org/10.1016/j.invent.2023.100636.
National Institute for Health and Care Excellence. (2018, updated). Attention deficit hyperactivity disorder: Diagnosis and management, NG87. https://www.nice.org.uk/guidance/ng87.
National Institute for Health and Care Excellence. (2024). Digital technologies for assessing attention deficit hyperactivity disorder, HTG729. https://www.nice.org.uk/guidance/htg729.