Efficacy of an Unguided Digital Intervention for Adults With Attention-Deficit/Hyperactivity Disorder

Summary

Background:

Psychotherapy for adults with ADHD is still underutilized in Germany and worldwide. Only a small percentage of these patients receive psychotherapy as recommended in the guidelines. Unguided digital interventions may help to lessen this treatment gap.

Methods:

An open-label, exploratory randomized controlled trial was conducted (trial registration: DRKS00033320). Adults with confirmed ADHD were recruited at six study centers in Germany, as well as via social media, and were randomly allotted to receive either the intervention (treatment as usual + app) or the control condition (treatment as usual). Data were collected from 2 February to 5 June 2024. The primary endpoint was patients’ health-related quality of life (QOL) at 12 weeks. The secondary endpoints included changes in the severity of ADHD symptoms, functional impairment, and symptoms of anxiety or depression. The threshold for statistical significance was set at 10% in view of the exploratory nature of the trial.

Results:

307 subjects were randomized (intervention group: n = 155, control group: n = 152). 70% were women, 29% men, and 1% was gender diverse; their mean age was 36 years (SD, 9.5). The data were analyzed according to the intention-to-treat principle with a mixed-effects model. The improvement in QOL was greater in the intervention group than in the control group (p <0.001; d = 0.54; 90% confidence interval [0.33–0.74]), despite generally low adherence to the intervention in the intervention group. 40% of subjects in the intervention group attained clinically relevant improvement in QOL, compared to 27% of control subjects. Among the secondary endpoints, ADHD symptoms improved to a greater extent in the intervention group than in the control group.

Conclusion:

Limitations include that no standardized diagnostics were conducted within the study and that no conclusions can be drawn regarding longterm effects.

Attention-deficit/hyperactivity disorder (ADHD) is a common neurobiological developmental disorder that persists into adulthood in around 60% of cases, with a global prevalence of around 2.6% (1, 2). It impacts everyday functioning, emotional well-being, and social interactions, and is associated with health risks and impaired quality of life (2, 3). Combined pharmacotherapy, psychoeducation, and psychotherapy is more effective than medication alone (2–4). Only 3–5% of patients receive cognitive behavioral therapy (CBT) in accordance with the relevant guidelines; in Germany as a whole, the rate is 5% at 1 year post diagnosis (5). This low figure is due, among other factors, to:

• Long waiting times (6–24 months)

• Misdiagnosis

• Limited availability

• Shame or fear of stigmatization (3, 5, 6–8)

About 38% of those affected receive psychotherapeutic care of some kind within the first year; however, this treatment is usually not specific to ADHD (e.g., probatory sessions, training in skills such as progressive muscle relaxation, or psychodynamic therapy) (5). Therefore, low-threshold—easily accessible, flexible, and self-directed—- evidence-based interventions are urgently needed. To be effective ADHD treatment must address both the core symptoms and the associated issues. CBT and dialectical behavior therapy (DBT) have proven effective both in reducing symptom severity and in enhancing quality of life (9–11).

Unguided digital interventions can help reduce the deficiencies in psychotherapeutic care. However, the available evidence is limited. Nasri and colleagues (12), for example, found a reduction in symptom severity after 12 weeks of a guided intervention compared with treatment as usual (TAU). Kenter et al. (13) demonstrated the efficacy of an unguided intervention regarding symptom severity and quality of life, but their study included participants with self-reported ADHD, limiting external validity.

The aim of the study reported here was to evaluate the efficacy of an unguided mobile psychological intervention (MiNDNET ADHD therapy) for adults with a confirmed diagnosis of ADHD with regard to quality of life as the primary endpoint together with secondary endpoints such as symptom severity and functional or emotional difficulties.

Methods

Study design and participants

The study was an open-label, two-arm randomized controlled trial in which an intervention group (IG) was compared with a wait list control group (WCG). Both groups also had access to TAU, i.e., the participants remained free to take up all standard health care services, including psychotherapy and pharmacotherapy. Newly initiated psychotherapeutic treatment and medication were assessed at each measurement, including changes in dosing or the drugs used.

Assessments were conducted at baseline (T0), after 6 weeks (T1; during the intervention), and after 12 weeks (T2; post-intervention). Recruitment took place at six academic study centers in Germany and via social media. A detailed description of the recruitment and screening procedures can be found in the eMethods. The inclusion and exclusion criteria are presented in the Box. All participants gave electronic informed consent.

Box. Inclusion and exclusion criteria for participation.

Inclusion criteria

• Age at least 18 years

• ADHD diagnosis according to ICD-10 or DSM-5, confirmed either by study center staff* or by written proof (medical report, certificate, or equivalent document)

• Access to a smartphone or tablet

• Availability for the 12-week study period

• Ability to comply with study protocol and assessments (sufficient German language skills: speaking, writing, comprehension)

• Current impairment by ADHD symptoms (Adult ADHD Self-Report Scale score ≥ 4)

• If prescribed: stable ADHD medication (methylphenidate, [lisdex-]amphetamine, guanfacine, atomoxetine) for ≥ 4 weeks before randomization

Exclusion criteria

• Severe mental illness: major unipolar depression, borderline personality disorder (ICD-10: F60.3), bipolar affective disorder (F31), schizophrenia, schizotypal or delusional disorder (F20-F29)

• Alcohol and/or drug abuse (ICD-10: F10-F19)

• Current suicidal tendencies

• Ongoing ADHD-specific psychotherapy

• Recent initiation or adjustment of ADHD-specific medication

The study was approved by the local psychological ethics committee (LPEK–0709). The trial was preregistered in the German Clinical Trials Register (DRKS00033320). The results are presented in accordance with the Consolidated Standards of Reporting Trials (CONSORT) recommendations. A fully completed CONSORT checklist can be found in the eChecklist.

Randomization and masking

Randomization was fully automated via the survey software Qualtrics, using a simple algorithm at the end of the baseline survey. No manual allocation occurred, ensuring allocation concealment until the time of randomization. Given the nature of the intervention, a self-managed app, no blinding of participants and investigators was possible.

Intervention

The digital ADHD therapy is an app for adults with ADHD. The app can be used in addition to the regular treatment (e.g., by primary care physicians, specialists, or psychotherapists). Users can employ the intervention independently.

The intervention aims at improving health-related quality of life. The app is based on evidence-based psychotherapeutic principles, specifically those of CBT, mindfulness-based cognitive therapy, DBT, and social skills training, together with psychoeducational content.

The intervention comprises 12 successive weekly modules for daily use. Each module contains psychoeducational videos, interactive exercises, and reflection tasks addressing topics such as attentiveness, impulsivity, organization, emotional regulation, and relapse prevention. Each module consists of seven units (i.e., daily sessions) that take 15–30 minutes to complete. Participants in the WCG were given access to the app after 12 weeks. The detailed content of the module is shown in Table 1.

Table 1. Content of the 12 weekly modules of the digital ADHD therapy app.

Module	Content
1	•Explanation of the program •Goal setting and motivation (SMART) •General psychoeducation •Resource activation
2	•Attentiveness •Self-regulation •Forgetfulness
3	•Hyperfocus detection •Regulation and utilization
4	•Motor hyperactivity •Restlessness
5	•Impulsivity •Self-control
6	•Disorganized behavior •Planning, organization, and structure
7	•Time perception •Time management (Pomodoro technique)
8	•Emotional over-reactivity •Affect lability and affect control
9	•Procrastination •Avoidance behavior
10	•Self-care •Healthy eating behavior •Social skills •Self-acceptance
11	•Sleep •Stress reduction
12	•Relapse prevention •Long-term health maintenance

The intervention with 12 successive weekly modules for daily use is currently available on prescription in Germany. The costs amount to € 441 per prescription (prior to price negotiations with the National Association of Statutory Health Insurance Funds) and are fully covered, with no additional costs for insured patients.

Measures

The primary endpoint was change in ADHD-associated quality of life (AAQoL) from T0 to T2 (14).

The secondary endpoints included changes in:

• ADHD symptoms (ASRS-1.1) (15)

• Subjective impairment (CGI-S) (16)

• Depression (PHQ-9) (17, 18)

• Anxiety (GAD-7) (19)

• Perceived stress (PSS-4) (20)

• Health competence (PHCS) (21)

• Medication adherence (MARS-D) (22)

All instruments were administered online at all three measurement times and are fully described in the eMethods.

Statistical analyses

All analyses followed a predetermined statistical analysis plan, available in the German Clinical Trials Register (DRKS00033320). Analyses followed the intention-to-treat principle (ITT) including all randomized participants (N = 307). The primary analysis used a mixed-effects model for repeated measures (MMRM) with fixed effects for group, time, and their interaction. Baseline scores were entered into the model as covariates. Within-participant correlations across repeated measurements were modeled using an unstructured covariance matrix. Results are reported as least squares mean differences (LSMD) with 90% confidence intervals (CI), reflecting the exploratory nature of this pilot trial (α = 0.10). A hierarchical testing procedure was employed to control the family-wise error rate across the primary and secondary endpoints. The sample size (n = 278; d = 0.30; 80% power) was calculated for two-sided α level of 0.10, allowing for attrition rates. To evaluate clinical relevance, thresholds for the minimal clinically important difference (MCID) were predefined and analyzed in the complete cases (CC) sample. An increase of ≥ 8 points on the AAQoL (23) and a reduction of ≥ 30% on the ASRS (12) were considered meaningful improvements. Group differences in responder status (clinically meaningfully improved vs. non-improved) were examined using χ² tests. Sensitivity analyses (CC and multiple imputation) and further information on assessment of clinical relevance are presented in the eMethods.

Results

Sample characteristics

The characteristics of the sample are shown in Table 2 and described in more detail in the eResults. Table 3 shows the psychopathological scores of the endpoints in the groups at the measurement points T0, T1, and T2.

Table 2. Participants’ demographic characteristics and psychopathological variables at the beginning of the study.

Variable	Total (N = 307)	WCG (n = 152)	IG (n = 155)
Demographic characteristics
Gender (F/M/X)*1	216/88/3 (70.4/28.7/1.0%)	98/51/3 (64.5/33.5/2.0%)	118/37/0 (76.1/23.9/0%)
Age in years)*2	36.5 (9.5)	36.6 (9.9)	36.3 (9.2)
Educational level
University degree*1	151 (49.2%)	71 (46.7%)	80 (51.6%)
Upper secondary school qualification*1	101 (32.9%)	52 (34.2%)	49 (31.6%)
Intermediate secondary school qualification*1	48 (15.6%)	24 (15.8%)	24 (15.5%)
Lower secondary school qualification*1	7 (2.3%)	5 (3.3%)	2 (1.3%)
Psychopathology *2
AAQoL (0–100)*2,*3	45.0 (8.7)	44.6 (9.1)	45.4 (8.3)
ASRS (dichot.: 0–18)*2,*3	14.7 (2.4)	14.8 (2.3)	14.5 (2.5)
CGI-S (1–7)*2,*3	5.4 (0.9)	5.3 (0.9)	5.4 (0.9)
PSS-4 (0–16)*2,*3,*4	6.9 (2.6)	7.0 (2.9)	6.7 (2.3)
PHQ-9 (0–27)*2,*3	13.1 (4.6)	13.4 (4.5)	12.9 (4.7)
GAD-7 (0–21)*2,*3	11.2 (4.3)	11.2 (4.3)	11.2 (4.2)
PHCS (8–40)*2,*3	24.4 (5.6)	24.2 (5.7)	23.6 (5.5)
Years since diagnosis)*2	2.8 (5.4)	2.8 (4.5)	2.9 (6.2)
ADHD medication
No use*1	94 (30.6%)	54 (35.5%)	40 (25.8%)
Regular use*1	213 (69.4%)	98 (64.5%)	115 (74.2%)
MARS-D (5–25)*2,*3,*5	20.7 (3.1) n = 213	20.1 (3.2) n = 98	21.2 (3.0) n = 115

^*1Expressed as n (%);

^*2expressed as mean (standard deviation);

^*3range in parentheses;

^*4lower scores = higher stress;

^*5only in patients with ADHD medication.

AAQoL, Adult ADHD Quality of Life; ASRS, Adult ADHD Self-Report Scale; CGI-S, Clinical Global Impression - Self-Report; dichot., dichotomous; GAD-7, Generalized Anxiety Disorder Scale 7; IG, intervention group; MARS-D, Medication Adherence Report Scale - German version; N, total sample size; n, sample size; PHCS, Perceived Health Competence Scale; PHQ-9, Patient Health Questionnaire 9; PSS-4, Perceived Stress Scale 4; WCG, waitlist control group

Table 3. Survey findings in the study group at all measurement times, expressed as means with standard deviations.

Endpoint	WCG			IG
		T0 n = 152	T1 n = 132 LTFU: n = 20	T2 n = 120 LTFU: n=32	T0 n = 155	T1 n = 117 LTFU: n=38	T2 n = 100 LTFU: n = 55
Quality of life (AAQoL)	44.6 (9.1)	46.1 (9.5)	47.1 (10.0)	45.4 (8.3)	49.5 (9.9)	53.3 (11.1)
Symptoms (ASRS)	14.8 (2.3)	13.3 (3.0)	13.1 (3.3)	14.5 (2.5)	12.5 (3.7)	11.0 (4.2)
Impairment (CGI-S)	5.3 (0.9)	5.2 (1.0)	5.3 (0.9)	5.4 (0.9)	5.2 (0.9)	4.7 (1.1)
Stress (PSS-4)	7.0 (2.9)	6.9 (2.9)	7.7 (1.6)	6.7 (2.3)	7.3 (3.1)	7.7 (1.6)
Depression (PHQ-9)	13.4 (4.5)	13.4 (4.7)	13.1 (5.0)	12.9 (4.7)	11.9 (4.8)	10.7 (5.0)
Anxiety (GAD-7)	11.2 (4.4)	10.9 (4.4)	10.9 (4.4)	11.2 (4.2)	10.0 (4.3)	8.9 (4.3)
Health competence (PHCS)	24.2 (6.0)	23.6 (5.3)	23.8 (5.8)	23.6 (5.5)	24.7 (5.6)	25.4 (6.4)
Medication adherence (MARS-D)* (n = 213)	20.1 (3.2) n = 98	19.2 (4.1) n = 101	19.7 (3.9) n = 93	21.2 (3.0) n = 115	21.4 (3.0) n = 98	21.4 (3.0) n = 86

^*Only in participants with ADHD medication

AAQoL, Adult ADHD Quality of Life; ASRS, Adult ADHD Self-Report Scale; CGI-S, Clinical Global Impression – Self-Report; GAD-7, Generalized Anxiety Disorder Scale 7;

LTFU, lost to follow-up; MARS-D, Medication Adherence Report Scale – German version; PHCS, Perceived Health Competence Scale; PHQ-9, Patient Health Questionnaire 9; PSS-4, Perceived Stress Scale 4; T0, baseline measurement; T1, measurement after 6 weeks (during the intervention); T2, measurement after 12 weeks (postintervention)

Primary ITT analyses

The MMRM, with an α level of 10%, showed a statistically significantly greater improvement in the primary endpoint, ADHD-related quality of life (AAQoL), in the IG than in the WCG (LSMD = 5.3; 90% CI [3.3; 7.3]; p < 0.001; d = 0.54 [0.33; 0.74]). ADHD symptom severity (ASRS: p < 0.001; d = −0.58 [−0.79; −0.37]) and functional impairment (CGI-S: p < 0.001; d = −0.69 [−0.90; −0.48]) likewise improved in the IG compared with the WCG.

No difference between the groups was observed for stress symptoms (PSS-4: p = 0.95, d = −0.01 [−0.14; 0.13]) or medication adherence (MARS-D: p = 0.14, d = 0.21 [−0.03; 0.45]). For depressive symptoms (PHQ-9: p < 0.001, d = −0.42 [−0.61; −0.22]) and anxiety (GAD-7: p < 0.001, d = −0.41 [−0.60; −0.23]), as well as for improvement in health competence (PHCS: p = 0.055, d = 0.24 [0.04; 0.44]), greater improvement was observed than in the WCG. However, these differences cannot be considered statistically significant due to the test hierarchy.

Sensitivity analyses (complete-case and multiple imputation) are reported in detail in the eResults.

Clinical relevance

Clinically meaningful improvement, defined as an AAQoL increase of ≥ 8 points, was achieved by 40% of participants (40/100) in the IG and 22% (26/120) in the WCG (χ²(1) = 8.73; p = 0.003). The ADHD symptoms (ASRS) decreased by at least 30% in 39% of participants in the IG versus 15% in the WCG (χ²(1) = 16.37; p < 0.001) (Table 4).

Table 4. Results of primary intention-to-treat analyses with a mixed model for repeated measurements at the post-intervention time point^*1.

Endpoint	LSMD	90% CI		p	Cohen’s d	90% CI
		Lower	Upper			Lower	Upper
Quality of life (AAQoL)	5.31 (5.12)	3.29 (3.04)	7.32 (7.21)	< 0.001 (< 0.001)	0.54 (0.51)	0.33 (0.30)	0.74 (0.72)
Symptoms (ASRS)	−2.02 (−2.00)	−2.76 (−2.80)	−1.27 (−1.78)	< 0.001 (< 0.001)	−0.58 (−0.55)	−0.79 (−0.78)	−0.37 (−0.33)
Impairment (CGI-S)	−0.60 (−0.64)	−0.78 (−0.85)	−0.42 (−0.43)	< 0.001 (< 0.001)	−0.69 (−0.70)	−0.90 (−0.93)	−0.48 (−0.47)
Stress (PSS-4)	−0.01 (0.23)	−0.39 (−0.38)	0.36 (0.83)	0.95 (0.70)	−0.01 (0.08)	−0.14 (−0.13)	0.13 (0.28)
Depression (PHQ-9)	−2.05 (−1.89)	−3.00 (−2.98)	−1.09 (−0.79)	< 0.001 (0.001)	−0.42 (−0.37)	−0.61 (−0.59)	−0.22 (−0.16)
Anxiety (GAD-7)	−1.86 (−1.87)	−2.69 (−2.82)	−1.02 (−0.93)	< 0.001 (< 0.001)	−0.41 (−0.41)	−0.60 (−0.62)	−0.23 (−0.20)
Health competence (PHCS)	1.20 (0.92)	0.17 (−0.26)	2.22 (2.10)	0.055 (0.13)	0.24 (0.18)	0.04 (−0.05)	0.44 (0.40)
Medication adherence (MARS-D)*2 (n = 213)	0.64 (0.27)	−0.08 (−0.54)	1.36 (1.08)	0.14 (0.09)	0.21 (0.09)	−0.03 (−0.17)	0.45 (0.34)

^*1The numbers in parentheses represent the results of sensitivity analyses on the basis of covariance analyses with multiple imputation by chained equations (ANCOVA with MICE). Least squares mean differences with 90% confidence intervals (lower and upper limits), p-value, and Cohen’s d from primary intention-to-treat analyses with a mixed model for repeated measurements post-intervention at 12 weeks (N = 307; α level set at 10%). The LSMD represent the change in between-group difference between the baseline and post-intervention measurements (positive values show a greater increase, negative values a greater decrease in the intervention group than in the control group).

^*2Only in participants with ADHD medication

AAQoL, Adult ADHD Quality of Life; ANCOVA, analysis of covariance; ASRS, Adult ADHD Self-Report Scale; CGI-S, Clinical Global Impression - Self-Report; CI, confidence interval; GAD-7, Generalized Anxiety Disorder Scale 7; LSMD, least squares mean differences; MARS-D, Medication Adherence Report Scale - German version; MICE, multiple imputation by chained equations; PHCS, Perceived Health Competence Scale; PHQ-9, Patient Health Questionnaire 9; PSS-4, Perceived Stress Scale 4

Discussion

In this preliminary trial of the efficacy of the ADHD app, the ADHD-related quality of life improved in the IG versus TAU. There were additional benefits regarding symptom severity and functional impairment, indicating a consistent benefit across multiple clinical endpoints. The evaluation of clinical relevance showed that more persons in the IG than in the WCG experienced a clinically meaningful improvement in their quality of life. However, the proportion of participants experiencing a clinically important improvement was low overall, with 40 such persons in the IG. When interpreting these results, it must be noted that the effectiveness of a self-help intervention like the evaluated app may not be comparable with that of ADHD medication, which usually shows sizeable effects on ADHD symptoms (11).

The majority of participants (69%, 213/307) were taking ADHD medication and used the app as supplemental treatment; this may have limited the clinically meaningful improvement. Furthermore, adherence to the intervention was low. Given an assumed dose-response relationship, we interpret the relatively small proportion of participants with clinically meaningful improvement as being due to the incomplete adherence. As no further analyses were conducted, however, this interpretation remains speculative.

This study adds to the limited number of high-quality trials evaluating digital interventions for ADHD (13). The moderate effects observed on quality of life and symptoms post intervention are consistent with the findings of other recent trials of digital CBT-based interventions for ADHD, including the studies by Nasri et al. (12) and Kenter et al. (13), both of which reported effects comparable with inactive control conditions. These results strengthen the evidence for the potential of low-threshold, scalable digital interventions in this population. The fact that the present trial showed efficacy without guidance is particularly advantageous in that this offers further benefits regarding availability and accessibility compared to guided web-based CBT.

On perceived stress, the IG showed no greater improvement. This may be explained by the only partial adherence to the app. On average, four modules were started in the intervention period. The recommended approach was to work through one module per week. Stress was not addressed until module 8, so most participants in the IG did not work on this topic. Low adherence is a known problem: in a meta-analysis by Karyotaki et al. (24) with over 2700 participants, almost 60% dropped out before completing half of the intervention content. While the overall attrition rate in our study was acceptable (28%, 87/307), the recommended “dosage” was not achieved. Only 20% of the participants (20/100) completed more than half of the 12 modules.

The effects found in this study could therefore be interpreted as a placebo effect. However, it must be noted that each module has extensive content. Completing the average of four modules still provided substantial therapeutic exposure, especially since the initial modules addressed core ADHD symptoms. The exploratorily observed dose-response relationship in this sample indicates that patients show greater improvement in quality of life when they engage with the app more intensively. Although this simple correlation provides only a descriptive indication of a possible dose-response relationship, it underscores the potential importance of adherence to the intervention.

Efforts should be made to further increase adherence, e.g., through automated reminders about app usage. The recommended treatment plan of one module per week might not be optimal—especially for a population with ADHD, who typically have difficulties with task management, avoidance behavior, and organizational skills. Despite the low adherence, the digital therapy was effective overall, indicating that the full “dose” of the treatment is not essential to achieve an effect. This is supported by the results of the interim analyses after 6 weeks.

No effect was found on medication adherence, which may be explained by the high medication adherence at baseline (mean MARS-D score > 20, with a maximum of 25).

Limitations

When interpreting the results presented here, it should be noted that the sample was not fully balanced in terms of sociodemographic characteristics. The high proportion of women (70%) diverges from epidemiological data in Germany, which suggest a balanced or slightly male-dominated distribution (25, 26). However, rising ADHD diagnoses among adult women (27), who are more likely to seek psychotherapy (28), may explain this overrepresentation.

Additionally, persons with ADHD often encounter difficulties at school and in further education, lowering their likelihood of attaining a higher-education qualification compared with those without ADHD (29). However, the sociodemographic pattern in this trial is not unusual in the context of studies evaluating psychotherapeutic interventions, particularly in samples of adults with ADHD. The prevalence of ADHD decreases with age (1), and gender differences in help-seeking behavior further affect clinical samples. Moreover, other studies focusing on interventions for adults with ADHD report similar levels of education (30, 31), suggesting that the observed distribution is in line with the existing research.

Nevertheless, the skewed gender and education distribution may limit generalizability of the results, particularly for older, male, or less educated subgroups, who may show different symptom courses, treatment effects, or adherence. Therefore, future studies should aim to achieve more balanced samples.

Furthermore, the participants were observed only over a 12-week period, without long-term follow-up. This limits the evaluation of lasting effects. Therefore, no claims regarding long-term effects can be made. Moreover, the assessments were based solely on the patients’ own reports; no active control group (e.g., with a sham intervention) was used. ADHD diagnoses were not reassessed using a standardized diagnostic procedure within the study. Instead, inclusion relied on previously established clinical diagnoses as documented in medical records and verified by study staff. The majority of diagnoses were relatively recent, however, and the current symptoms were ascertained using the ASRS cutoff.

Although the control group with TAU permits realistic comparison, active controls could better address nonspecific and expectancy effects (e.g., nocebo). Adverse events were not assessed systematically with standardized measures but only recorded if directly reported.

The use of a high significance threshold (α = 0.10) reflects the exploratory objectives of this study, but increases the risk of a type I error. The results should therefore be interpreted with caution and confirmed in future studies with more conservative methods (e.g., α = 0.05; active control groups).

Randomized controlled trials with longer follow-up periods and more conservative methodological approaches are necessary to strengthen the evidence for the effectiveness of self-guided digital psychotherapy for adults with ADHD. Additionally, efforts must be made to improve adherence and therefore increase the proportion of patients who experience a clinically meaningful change. Moreover, future work should apply more sophisticated dose-response analyses to explore the relationship between use and treatment effect. Furthermore, evaluation of health economic data would be helpful in drawing conclusions regarding cost effectiveness compared with standard treatment.

Conclusion

Despite some limitations, the pilot trial presented here provides preliminary but promising evidence for the efficacy of a new app for ADHD, thus addressing the urgent need for a novel intervention for adults with ADHD. Overall, the results are consistent with previous studies on web-based interventions for ADHD (12, 13). Selfguided interventions have advantages over conventional psychotherapy and guided web-based CBT in terms of availability and accessibility and therefore could greatly improve the care of adults with ADHD.

Supplementary material

eMETHODS

Recruitment and procedure

Participants were recruited between February and June 2024 at six academic study centers in Germany (University Medical Center Hamburg-Eppendorf, Charité – Universitätsmedizin Berlin, University Medical Center Frankfurt am Main, University Hospital Bonn, Asklepios Hospital Hamburg-Harburg, and Hanover Medical School) and online via advertisements on social media platforms (Instagram, Facebook, and study-related websites). Only individuals with a previously confirmed diagnosis of ADHD (ICD-10: F90.0) were eligible for participation. At the study centers, only patients with such a diagnosis were recruited. For the participants recruited online, the diagnosis was confirmed on the basis of the medical documentation submitted, the plausibility of which was verified by study staff (AB, LS) prior to inclusion. After reading the study information, patients gave their electronic informed consent to participate and were directed to the online screening and baseline data collection (T0), which was conducted on the online survey platform Qualtrics. If exclusion criteria were met, the survey terminated automatically. In this event the participants were informed about the reason for ineligibility and given the contact details of the study team for further questions. Participants who fulfilled the eligibility criteria were automatically randomized at the end of the baseline survey. Participants allocated to the IG then received an e-mail with their personal access code for the app and instructions on how to use it. Participants allocated to the WCG were informed that they would receive access to the program after the study period (12 weeks). All persons included in the study were free to continue their usual treatment during the trial. They all continued to have unrestricted access to all standard health care services, including psychotherapy and pharmacological treatment. Newly initiated psychotherapeutic treatment as well as medication was assessed at each measurement time, including information on changes in dosage and medication. Invitations to the follow-up assessments at T1 and T2 were sent automatically by e-mail. Upon completion of the post-intervention assessment, all participants received a € 10 online shopping voucher for each completed survey.

Inclusion and exclusion criteria (with the exception of a confirmed F90.0 diagnosis) were assessed exclusively via self-report. There was no direct contact between participants and the study team, except in cases of non-response to automated reminders or in the event of (technical) questions or issues in connection with the surveys or the app. Adverse events were not assessed systematically within the surveys but were handled and documented by the study team (AB and LS).

Randomization and masking

Randomization was performed automatically with the survey software Qualtrics, using a simple 1:1 allocation algorithm without stratification. Qualtrics is a secure, webbased survey platform that allows the creation, distribution, and management of online questionnaires. Participants were not masked to group assignment, as blinding is not feasible in the context of a self-guided psychological online intervention. Group allocation was performed automatically by the survey software without any manual involvement, and could not be influenced by study staff. This ensured allocation concealment up to the moment of assignment. The study investigators were not blinded to group allocation; however, the statistical analyses were prespecified in the statistical analysis plan. As all endpoints were solely self-reported and the surveys took place online, this had no influence on the results reported in this study.

Measures

All measures were used in their German versions and were assessed at all three time points (T0, T1, T2) except for the CSQ-8, which was only assessed at T2 (after the intervention period). All described measures assessed patient-reported outcomes and were implemented via online surveys on the survey platform Qualtrics.

Primary endpoint

Adult ADHD Quality of Life Scale (AAQoL)

The primary efficacy endpoint was the change in selfreported quality of life from T0 to T2. This was measured using the AAQoL (14), which was developed to assess the quality of life in adults with ADHD over the past 2 weeks. The AAQoL consists of 29 items, with responses rated on a five-point Likert scale from 1 (“not at all/almost never”) to 5 (“extremely/very often”). Negatively worded items are recoded, and all item scores are transformed to a scale ranging from 0 to 100. The item scores were then summed and divided by the number of items to calculate the total score. The AAQoL yields an overall score as well as scores for four subscales: life productivity (11 items), mental health (6 items), life perspective (7 items), and relationships (5 items). Higher scores correspond to higher quality of life. The AAQoL demonstrated satisfactory psychometric properties, with a Cronbach’s α of 0.76.

Secondary endpoints

Adult ADHD Self-Report Scale (ASRS 1.1)

The secondary efficacy endpoint was the change in self-reported health status from T0 to T2, assessed using the ASRS (15). The ASRS serves to evaluate the core symptoms of ADHD and consists of 18 items rated on a five-point Likert scale from 0 (“never”) to 4 (“very often”). An overall score can be calculated across the 18 items (0–72, dichotomized 0-18 points), as well as for two subscales, one addressing the nine symptoms of inattentiveness (inattention subscale) and the other covering the nine symptoms of hyperactivity or impulsivity (hyperactivity/impulsivity subscale). The scores of each subscale range from 0 to 36, dichotomized 0 to 9 points. Higher scores indicate a greater symptom burden. The ASRS demonstrated acceptable reliability in this study, with a Cronbach’s α of 0.61.

Clinical Global Impression Scale - Severity of Illness (CGI-S)

The subjective severity of everyday impairment due to ADHD was assessed using an adapted item from the CGI-S (16): “To what extent do you feel impaired in your daily life due to ADHD?” This was rated on a scale from 1 (“not at all”) to 7 (“extremely impaired”).

Patient Health Questionnaire (PHQ-9)

Comorbid depressive symptoms were measured using the German version of the PHQ-9 (17,18), which assesses depressive symptoms over the past 2 weeks with nine items. The total score ranges from 0 to 27. Based on recommended cut-off values, scores of 0–5 indicate “no depression”; 6–10, “mild depression”; 11–15, “moderate depression”; 16–20, “severe depression”; and 21–27, “very severe depression.” The PHQ-9 demonstrated satisfactory reliability (Cronbach’s α = 0.78).

Generalized Anxiety Disorder Scale (GAD-7)

The GAD-7 (19) consists of seven items designed to measure anxiety symptoms experienced over the past 2 weeks. Responses are rated on a four-point Likert scale from 0 (“not at all”) to 3 (“nearly every day”). A higher total score thus indicates a higher level of anxiety symptoms. The GAD-7 demonstrated good psychometric properties, with a Cronbach’s α of 0.82.

Perceived Stress Scale (PSS-4)

The PSS-4 (20) serves to assess the extent of perceived stress over the past month. The scale has four items, with responses rated on a five-point Likert scale from “never” [0] to “very often” [4]. Negative items are recoded so that lower scores indicate higher stress levels. The PSS-4 had satisfactory psychometric properties, with a Cronbach’s α of 0.71.

Perceived Health Competence Scale (PHCS)

The PHCS measures self-efficacy and competence in managing one’s own health as well as health behavior (21). The scale consists of eight items with responses rated on a five-point Likert scale from 1 (“strongly disagree”) to 5 (“strongly agree”). Higher scores reflect better health competence. The PHCS showed good psychometric properties, with a Cronbach’s α of 0.88.

Medication Adherence Report Scale (MARS-D)

Medication adherence was measured using the MARS-D (23). The questionnaire comprises five items rated on a five-point Likert scale from “always” [1] to “never” [5], with higher scores reflecting better medication adherence (22). The internal consistency of the MARS-D was satisfactory, with a Cronbach’s α of 0.69. The MARS-D was answered only by participants who indicated that they were currently taking ADHD-specific medication.

Client Satisfaction Questionnaire (CSQ-8)

Treatment satisfaction was evaluated using the CSQ-8 (e1, e2). This questionnaire consists of 8 questions, each with four possible responses, without a “neutral” option. Responses are scored from 1 to 4, with four of the eight items (items 1, 3, 6, and 7) negatively phrased. After recoding, all eight item scores are summed to create a scale score (possible range from 8 to 32). Higher scale scores indicate greater satisfaction. The CSQ-8 was answered only by participants in the IG. The CSQ-8 had excellent internal consistency (Cronbach’s α = 0.94).

Data analysis

The sample size calculation (G*Power) (e3) assumed d = 0.30, α = 0.10, and 80% power, resulting in 278 participants (139 per group); slight over-recruitment was planned to account for attrition, an exact number was not set, based on practical considerations the recruitment was terminated and deemed sufficient after an over-recruitment of around 10%. The statistical analyses were performed using IBM SPSS Statistics 27 and R Version 4.3.2. Due to the set alpha level of 10% in this trial, p-values ≤ 0.1 are considered statistically significant. A hierarchical testing strategy controlled for multiple comparisons across endpoints at T2. The test hierarchy was as follows:

• 1) AAQoL

• 2) ASRS

• 3) CGI-S

• 4) PSS-4

• 5) PHQ-9

• 6) GAD-7

• 7) PHCS

• 8) MARS-D

To assess the clinical relevance of the endpoints related to improvement of quality of life and symptom reduction (AAQoL, ASRS), these were analyzed with respect to known thresholds of minimal clinically important differences (MCID) for patients and compared across groups using χ² tests. For the AAQoL, an increase of 8 points from T0 to T2 is established as MCID (23). The MCID for the ASRS is defined as a reduction of the baseline score by 30% (12). Achievement of the MCID was evaluated using logistic regression models, which estimate the odds of achieving the MCID in the IG compared with the WCG—controlling baseline symptom severity. To evaluate the robustness of the ITT results, sensitivity analyses were conducted using multiple imputation (MI) and complete cases (CC; data from participants with complete datasets). Missing value estimation for MI was performed using multiple imputation by chained equations (MICE), generating m = 50 imputation data sets. Age, gender, education level, medication, and available total scores of the symptom questionnaires were included as predictors in the imputation model. The models were first estimated within the imputed datasets and then combined according to Rubin’s rule. Calculations for the primary and secondary endpoints in the sensitivity analyses were conducted using analyses of covariance (ANCOVAs). The respective difference scores between baseline and post-assessment (e4) (12 weeks after baseline) served as dependent variables. The respective baseline values were included as covariates in the analysis to control for regression to the mean. Standardized effect sizes for all analyses were calculated as Cohen’s d based on model-derived LSMD and pooled standard deviations. The LSMD are additionally reported as an effect size on the original scale. Values of d = 0.2 to < 0.5 indicate a small, 0.5 to < 0.8 a moderate, and > 0.8 a large effect size (e5). For exploratory purposes, all analyses were conducted in the same manner for the change at T1 (6 weeks after baseline). For further exploratory analyses, dose-response correlations between the number of modules completed and change scores were examined using Pearson correlations (two-tailed, α = 0.10).

The statistical analysis plan is available in the German Clinical Trials Register (DRKS00033320).

Negative effects

No standardized questionnaires were used to assess adverse effects. However, participants were informed in the study information materials and the consent form that they could contact the study team by e-mail or telephone at any time to report any problems, adverse experiences, or worsening symptoms. The study staff monitored all these communications throughout the trial. To further examine potential harms, reliable symptom deterioration was determined using the Reliable Change Index (RCI). The proportion of participants showing RCI-defined worsening was compared between groups using χ² tests.

eRESULTS

The data were collected between 2 February 2024 (first patient in) and 5 June 2024 (last patient out). A total of 420 persons participated in the baseline survey. At T0, 112 participants had to be excluded because they either did not meet the inclusion criteria (30 persons with an ASRS score below 4) or fulfilled an exclusion criterion (14 persons undergoing current ADHD-specific psychotherapy, 27 with a current severe depressive episode, 11 with a substance use disorder, 28 with another exclusion diagnosis, and two due to suicidal ideation). Consequently, 308 participants were randomized (IG: n = 156; WCG: n = 152). One person from the IG withdrew “for personal reasons” before data evaluation, resulting in a final inclusion of 307 participants (IG: n = 155; WCG: n = 152) in the analyses. The flow of the participants through the study is shown in the eFigure.

eFigure. Flow chart of the participants through the study.

* One person from the IG withdrew “for personal reasons” before data evaluation

ASRS, Adult ADHD Self-Report Scale; CC, complete cases; ITT, intention to treat; N, total sample size; n, sample size; T0, baseline survey; T1, interim survey after 6 weeks; T2, postintervention survey after 12 weeks

Sample

The sample was predominantly female (70.4%, 216/307; male: 28.7%, 88/307; diverse: 1.0%, 3/307). The average age of those surveyed was 36 years (mean = 36.47, SD = 9.50), and 82.1% (252/307) reported at least a university entrance qualification as their highest level of education. Almost 70% (213/307) of the participants reported that they were regularly taking medication for ADHD at the time of the baseline survey. More than one third (35.8%, 110/307) were recruited through the University Medical Center Hamburg-Eppendorf, 9.4% (29/307) through Charité Universitätsmedizin Berlin, and one person (0.3%) through Asklepios Hospital Harburg (all in Germany). Nearly half (46.9%, 144/307) were recruited online or via social media.

The remaining 7.5% indicated other recruitment sources. On average, participants were diagnosed with ADHD 2.8 years before the study (SD = 5.45, range: 0–52 years). Of the 307 participants, 38 (12.4%) were diagnosed in the year of study enrollment. A high proportion of the participants (141/307; 45.9%) had been diagnosed with ADHD 1 year prior to enrollment.

Retention and care as usual

In total, 220 participants (72%) completed the postintervention survey (T2; 12 weeks after baseline). More participants in the IG than in the WCG discontinued their study participation (55 vs. 32 dropouts, χ²(1) = 7.87, p = 0.005). No differences were found regarding psychopathology at the time of the baseline survey (primary and secondary endpoints) or age (all p > 0.14). Additionally, no differences were observed regarding gender, educational status, recruitment source, psychotherapy experience, or current medication (all p > 0.11). Reasons for dropping out could not be captured, and none of the dropouts responded to follow-up contact attempts. The means (with standard deviations) for both groups at all three time points are presented in Table 2. Concerning the use of standard care during the study period, the changes in medication status did not differ between the IG and WCG at any point (χ² tests, all p > 0.1). Moreover, there was no difference between the two groups regarding newly initiated psychotherapeutic treatment over the course of the study (χ² tests, all p > 0.1) (eFigure).

Intervention usage and satisfaction

The app was started by 92% (143/155) of participants in the IG. On average, the intervention modules were completed up to module 4 out of 12 (mean = 4.1, SD = 3.1; range: 1 – 12). Thus, participants can be categorized as partially adherent. In the post-intervention survey (T2), seven persons mistakenly indicated that they had been assigned to the control group. However, all seven had registered in the app. One of them completed the program up to module 7, two up to module 6, and one each up to module 4, module 3, module 2, and module 1. Based on the automated survey logic, the satisfaction questionnaire (CSQ-8) was not administered to these participants. Therefore, the number of CSQ-8 responses is n = 93 participants. The overall satisfaction score averaged 24 (SD = 5.1). The percentages of positive responses to all items (n = 93) can be found in eTable 1.

eTable 1. Frequencies of positive responses to the items on the patient satisfaction questionnaire by the intervention group participants.

Item	Positive responses in %
How would you rate the quality of the app? (good or excellent)	90.4
Did you get the kind of service you wanted? (yes, generally or yes, definitely)	77.4
To what extent has the app met your needs? (almost all or most of my needs)	74.2
If a friend were in need of similar help, would you recommend the app? (yes, I think so or yes, definitely)	86.1
How satisfied are you with the amount of help you received? (mostly satisfied or very satisfied)	77.5
Has the app you received here helped you to deal more effectively with your problems? (yes, somewhat or yes, a great deal)	71.0
Overall, how satisfied are you with the app? (mostly satisfied or very satisfied)	77.4
If you were to need help, would you use the app again? (yes, I think so or yes, definitely)	79.6

*Percentages are valid percentages: n = 93

Sensitivity analyses

In the sample of complete cases (CC; WCG n = 120, IG n = 100), the results of the primary ITT analyses were confirmed (α level set at 10%). Again, the IG showed greater improvements in the primary endpoint (AAQoL T0–T2, p < 0.001; see eTable 2) as well as in the secondary endpoints ASRS (p < 0.001) and CGI-S (p < 0.001). Consistent with the primary ITT analyses, no difference was observed in the CC analyses for stress symptoms (PSS-4, p = 0.97). The PHQ-9, GAD-7, and PHCS showed p-values between 0.09 and < 0.001, but this cannot be considered confirmatory due to the hierarchical testing strategy. Medication adherence (MARS-D; n = 152) showed no significant differences between the two groups (p = 0.14). The standardized effect sizes (Cohen’s d) and effect sizes on the original scale (LSMD) with confidence intervals are presented in eTable 2.

eTable 2. Results of the sensitivity analyses by means of covariance analysis with complete data sets for the change at the post-intervention survey.

Outcome	F statistic F(1;217)	LSMD	90% CI		p	Cohen’s d	90% CI
			Lower	Upper			Lower	Upper
Quality of life (AAQoL)	17.70	5.30	3.22	7.38	< 0.001	0.47	0.24	0.69
Symptoms (ASRS)	17.61	−1.94	−2.70	−1.18	< 0.001	−0.52	−0.74	−0.29
Impairment (CGI-S)	30.16	−0.62	−0.81	−0.43	< 0.001	−0.75	−0.98	−0.52
Stress (PSS-4)	0.00	−0.01	−0.37	0.36	0.97	−0.01	−0.23	0.22
Depression (PHQ-9)	12.07	−2.08	−3.07	−1.09	< 0.001	−0.35	−0.58	−0.13
Anxiety (GAD-7)	11.43	−1.78	−2.65	−0.91	< 0.001	−0.34	−0.56	−0.11
Health competence (PHCS)	2.95	1.11	0.04	2.17	0.087	0.18	−0.04	0.41
Medication adherence (MARS-D)* (n = 152)	F(1;149) = 2.18	0.68	−0.08	1.45	0.14	0.02	−0.25	0.29

Least squares mean differences with 90% confidence intervals (lower and upper limits), p-value, and Cohen’s d from the sensitivity analyses for complete data sets at the post-intervention survey after 12 weeks (n = 220; α level set at 10%). The least squares mean differences represent the between-group differences in the change between the baseline survey (T0) and post-intervention measurements after 12 weeks (T2) (positive values show a greater increase, negative values a greater decrease in the intervention group than in the control group).

^*Only in participants with ADHD medication

AAQoL, Adult ADHD Quality of Life; ASRS, Adult ADHD Self-Report Scale; CGI-S, Clinical Global Impression - Self-Report; CI, confidence interval; GAD-7, Generalized Anxiety Disorder Scale 7; LSMD, least squares mean differences; MARS-D, Medication Adherence Report Scale - German version; PHCS, Perceived Health Competence Scale; PHQ-9, Patient Health Questionnaire 9; PSS-4, Perceived Stress Scale 4

In the ITT sensitivity analyses with multiple imputation (N = 307; α level set at 10%), the results for the endpoints of the AAQoL (p < 0.001), ASRS (p < 0.001), and CGI-S (p < 0.001) were confirmed. Again, the PSS-4 showed no intergroup differences (p = 0.70). Differences between groups were again found for the PHQ-9 (p = 0.001) and GAD-7 (p < 0.001). For health competence (PHCS) no differences were found between the groups (p = 0.13). Medication adherence (MARS-D; n = 213) improved more in the IG than in the WCG at T2 (p = 0.09). The results of the ANCOVAs with multiple imputation are presented in eTable 3.

eTable 3. Results of the sensitivity analyses with the intention-to-treat sample by means of covariance analysis with multiple imputation at the postintervention survey.

Outcome	F	LSMD	90% CI		p	Cohen’s d	90% CI
			Lower	Upper			Lower	Upper
Quality of life (AAQoL)	F(1;1411.35) = 20.85	5.12	3.04	7.21	< 0.001	0.51	0.30	0.72
Symptoms (ASRS)	F(1;488.20) = 18.75	−2.00	−2.80	−1.78	< 0.001	−0.55	−0.78	−0.33
Impairment (CGI-S)	F(1;462.07) = 24.64	−0.64	−0.85	−0.43	< 0.001	−0.70	−0.93	−0.47
Stress (PSS-4)	F(1;3459.60) = 0.24	0.23	−0.38	0.83	0.70	0.08	−0.13	0.28
Depression (PHQ-9)	F(1;521.82) = 11.47	−1.89	−2.98	−0.79	0.001	−0.37	−0.59	−0.16
Anxiety (GAD-7)	F(1;822.47) = 13.36	−1.87	−2.82	−0.93	< 0.001	−0.41	−0.62	−0.20
Health competence (PHCS)	F(1;479.29)= 2.29	0.92	−0.26	2.10	0.13	0.18	−0.05	0.40
Medication adherence (MARS-D)* (n = 213)	F(1;942.23)= 2.97	0.27	−0.54	1.08	0.09	0.09	−0.17	0.34

Least squares mean differences with 90% confidence intervals (lower and upper limits), p-value, and Cohen’s d from sensitivity analyses for complete data sets at the postintervention survey after 12 weeks (N = 307; α level set at 10%). The least squares mean differences represent the between-group differences in the change between the baseline survey (T0) and post-intervention measurements after 12 weeks (positive values show a greater increase, negative values a greater decrease in the intervention group than in the control group).

^*Only in participants with ADHD medication

AAQoL, Adult ADHD Quality of Life; ASRS, Adult ADHD Self-Report Scale; CGI-S, Clinical Global Impression - Self-Report; CI, confidence interval; GAD-7, Generalized Anxiety Disorder Scale 7; LSMD, least squares mean differences; MARS-D, Medication Adherence Report Scale - German version; PHCS, Perceived Health Competence Scale; PHQ-9, Patient Health Questionnaire 9; PSS-4, Perceived Stress Scale 4 T0, baseline assessment; T2, post assessment after 12 weeks; LSMD, least squares mean difference represents the between-group difference in change from T0 to T2 (positive values = greater increase; negative values = greater decrease in intervention group vs. control group); CI, confidence interval.

Minimal clinically important difference

The MCID criterion defined by Tanaka et al. (23) was used to assess clinical relevance for the primary endpoint (AAQoL) in the CC sample. An improvement in the AAQoL sum score of at least +8 points at T2 was considered a meaningful change for the patient. In the IG, this was true for 40 participants (40.0%, 40/100) significantly more often than in the WCG (21.7%, 26/120); χ²(1) = 8.73, p = 0.003; α level set at 10%. In a logistic regression model adjusted for baseline AAQoL, participants in the IG had higher odds of achieving the MCID than those in the WCG (OR = 2.99; p < 0.001). Sensitivity analyses using a conservative imputation of missing values as non-responders yielded attenuated but consistent effects (OR = 1.81; p = 0.041).

For the ASRS, a decrease of 30% in the sum score from T0 to T2 was defined as MCID. This criterion was achieved by 39 participants from the IG (39.0%, 39/100) compared with 18 participants from the WCG (15.0%, 18/120); χ²(1) = 16.37, p < 0.001; α level set at 10%. Furthermore, a logistic regression adjusted for baseline of the CC sample of ASRS revealed that the odds of achieving MCID were higher in the IG than in the WCG (OR = 6.20; p < 0.001). Under the previously described conservative imputation, the effect remained (OR = 4.41; p < 0.001).

Dose-response relationship

To examine a possible dose-response relationship in the IG of the CC sample, a simple bivariate correlation was conducted as an exploratory, post-hoc analysis of the relationship between the difference scores of each endpoint and the number of started modules performed. This showed a small relationship between dose and increase in quality of life (n = 97; r = −0.17; p = 0.09). However, as r is small and p is borderline at the 10% level, this correlation must be interpreted with caution. Corresponding correlations were also found for the following:

• Symptom severity (n = 97; r = 0.32; p = 0.001)

• Medication adherence (n = 75; r = −0.33; p = 0.004)

• Impairment of everyday life (n = 97; r = 0.22; p = 0.03)

• Anxiety symptoms (n = 97; r = 0.19; p = 0.06)

Change at the interim survey

In all analyses (MMRM, CC, and MI), the significantly greater improvement in the IG than in the WCG regarding the primary endpoint was already evident 6 weeks after baseline (change at time T1; α level set at 10%). The complete results of the exploratory analysis of the interim assessment can be found in eTables 4–6.

eTable 4. Results from the intention-to-treat analyses with a mixed model for repeated measures at the interim survey (α level set on 10%).

Endpoint	LSMD	90% CI		p	Cohen’s d	90% CI
		Lower	Upper			Lower	Upper
Quality of life (AAQoL)	2.51	0.80	4.21	0.016	0.25	0.08	0.43
Symptoms (ASRS)	−0.60	−1.21	0.01	0.11	−0.17	−0.35	0.00
Impairment (CGI-S)	−0.10	−0.25	0.06	0.31	−0.11	−0.29	0.07
Stress (PSS-4)	0.43	−0.13	0.99	0.21	0.15	−0.45	0.34
Depression (PHQ-9)	−0.98	−1.82	−0.15	0.053	−0.20	−0.37	−0.03
Anxiety (GAD-7)	−0.66	−1.41	0.08	0.15	−0.15	−0.31	0.02
Health competence (PHCS)	0.53	−0.32	1.39	0.31	0.11	−0.06	0.28
Medication adherence (MARS-D)* (n = 213)	1.08	0.40	1.77	0.009	0.31	0.13	0.58

Least squares mean differences with 90% confidence intervals (lower and upper limits), p-value, and Cohen’s d from the intention-to-treat analyses for complete data sets at the interim survey after 6 weeks (α level set at 10%).

^*Only in participants with ADHD medication

AAQoL, Adult ADHD Quality of Life; ASRS, Adult ADHD Self-Report Scale; CGI-S, Clinical Global Impression – Self-Report; CI, confidence interval; GAD-7, Generalized Anxiety Disorder Scale 7; LSMD, least squares mean differences; MARS-D, Medication Adherence Report Scale – German version; PHCS, Perceived Health Competence Scale; PHQ-9, Patient Health Questionnaire 9; PSS-4, Perceived Stress Scale 4

eTable 6. Results of the sensitivity analyses^*1 by means of covariance analysis with multiple imputation for the change at the interim survey.

Endpoint	F	LSMD	90% CI		p	Cohen’s d	90% CI
			Lower	Upper			Lower	Upper
Quality of life (AAQoL)	F(1;1128.75) = 6.08	2.30	0.45	4.16	0.014	0.26	0.05	0.46
Symptoms (ASRS)	F(1;1409.04) = 2.56	−0.50	−1.15	0.15	0.11	−0.16	−0.37	0.05
Impairment (CGI-S)	F(1;1824.26) = 0.67	−0.12	−0.28	0.05	0.41	−0.14	−0.34	0.07
Stress (PSS-4)	F(1;1193.28) = 2.46	0.59	0.03	1.14	0.12	0.22	0.01	0.43
Depression (PHQ-9)	F(1;1653.07) = 4.24	−0.84	−1.77	0.09	0.040	−0.18	−0.39	0.02
Anxiety (GAD-7)	F(1;2131.11) = 3.04	−0.78	−1.61	0.04	0.081	−0.19	−0.39	0.01
Health competence (PHCS)	F(1;2866.54) = 1.62	0.54	−0.38	1.46	0.20	0.12	−0.08	0.32
Medication adherence (MARS-D)*2 (n = 213)	F(1;1393.92) = 6.47	0.69	−0.04	1.41	0.011	0.23	−0.01	0.48

Least squares mean differences with 90% confidence intervals (lower and upper limits), p-value, and Cohen’s d from the sensitivity analyses with multiple imputation at the interim survey after 6 weeks. The least squares mean differences represent the between-group differences in the change between the baseline survey (T0) and interim measurements after 6 weeks (T1) (positive values show a greater increase, negative values a greater decrease in the intervention group than in the control group).

^*1α level set at 10%

^*Only in participants with ADHD medication

AAQoL, Adult ADHD Quality of Life; ASRS, Adult ADHD Self-Report Scale; CGI-S, Clinical Global Impression - Self-Report; CI, confidence interval; GAD-7, Generalized Anxiety Disorder Scale 7; LSMD, least squares mean differences; MARS-D, Medication Adherence Report Scale - German version; PHCS, Perceived Health Competence Scale; PHQ-9, Patient Health Questionnaire 9; PSS-4, Perceived Stress Scale 4

Negative effects

No adverse events were reported to the study staff. Symptom worsening—defined by the reliable change index (RCI)—was rare. For quality of life (AAQoL), five participants (1.6%; WCG: 3 [2.0%], IG: 2 [1.3%]) showed reliable deterioration with no between-group difference (χ² = 0.00; p = 0.98), while no RCI-defined worsening occurred for ADHD symptoms (ASRS) in either group.

eTable 5. Results of the sensitivity analyses by means of covariance analysis with complete data sets ^*1 for the change at the interim survey.

Endpoint	F(1;246)	LSMD	90% CI		p	Cohen’s d	90% CI
			Lower	Upper			Lower	Upper
Quality of life (AAQoL)	7.17	2.79	1.07	4.50	0.008	0.27	0.06	0.48
Symptoms (ASRS)	2.72	−0.61	−1.22	0.001	0.100	−0.17	−0.38	0.04
Impairment (CGI-S)	0.70	−0.11	−0.26	0.05	0.26	−0.18	−0.38	0.03
Stress (PSS-4)	2.08	0.45	−0.07	0.96	0.15	−0.18	−0.38	0.03
Depression (PHQ-9)	4.40	−1.08	−1.92	−0.23	0.037	−0.18	−0.39	0.03
Anxiety (GAD-7)	2.62	−0.74	−1.50	0.02	0.11	−0.16	−0.37	0.05
Health competence (PHCS)	1.34	0.61	−0.26	1.47	0.25	0.08	−0.13	0.29
Medication adherence (MARS-D)* (n = 176)	F(1;173) = 7.25	1.12	0.43	1.81	0.008	0.23	−0.02	0.48

Least squares mean differences with 90% confidence intervals (lower and upper limits), p-value, and Cohen’s d from the sensitivity analyses for the interim survey after 6 weeks. The least squares mean differences represent the between-group differences in the change between the baseline survey (T0) and interim measurements after 6 weeks (T1) (positive values show a greater increase, negative values a greater decrease in the intervention group than in the control group).

^*1n = 249; α level set at 10%

^*Only in participants with ADHD medication

AAQoL, Adult ADHD Quality of Life; ASRS, Adult ADHD Self-Report Scale; CGI-S, Clinical Global Impression - Self-Report; CI, confidence interval; GAD-7, Generalized Anxiety Disorder Scale 7; LSMD, least squares mean differences; MARS-D, Medication Adherence Report Scale - German version; PHCS, Perceived Health Competence Scale; PHQ-9, Patient Health Questionnaire 9; PSS-4, Perceived Stress Scale 4

CONSORT 2025 checklist of information to include when reporting a randomised trial^*.

Section / Topic	No	CONSORT 2025 checklist item description	Reported on page no.
Title and abstract
Title and structured abstract	1a	Identification as a randomised trial	p. 1
	1b	Structured summary of the trial design, methods, results, and conclusions	p. 2-3
Open science
Trial registration	2	Name of trial registry, identifying number (with URL) and date of registration	p. 6
Protocol and statistical analysis plan	3	Where the trial protocol and statistical analysis plan can be accessed	p. 8
Data sharing	4	Where and how the individual de-identified participant data (including data dictionary), statistical code and any other materials can be accessed	p. 21
Funding and conflicts of interest	5a	Sources of funding and other support (e.g., supply of drugs), and role of funders in the design, conduct, analysis and reporting of the trial	p. 21
	5b	Financial and other conflicts of interest of the manuscript authors	p. 20
Introduction
Background and rationale	6	Scientific background and rationale	p. 4
Objectives	7	Specific objectives related to benefits and harms	p. 4
Methods
Patient and public involvement	8	Details of patient or public involvement in the design, conduct and reporting of the trial	n.a.
Trial design	9	Description of trial design including type of trial (e.g., parallel group, crossover), allocation ratio, and framework (e.g., superiority, equivalence, non-inferiority, exploratory)	p. 5
Changes to trial protocol	10	Important changes to the trial after it commenced including any outcomes or analyses that were not prespecified, with reason	n.a.
Trial setting	11	Settings (e.g., community, hospital) and locations (e.g., countries, sites) where the trial was conducted	p. 5
Eligibility criteria	12a	Eligibility criteria for participants	p.5 Box. Participants eligibility criteria
	12b	If applicable, eligibility criteria for sites and for individuals delivering the interventions (e.g., surgeons, physiotherapists)	n.a.
Intervention and comparator	13	Intervention and comparator with sufficient details to allow replication. If relevant, where additional materials describing the intervention and comparator (e.g., intervention manual) can be accessed	p.6 & Table 1
Outcomes	14	Pre-specified primary and secondary outcomes, including the specific measurement variable (e.g., systolic blood pressure), analysis metric (e.g., change from baseline, final value, time to event), method of aggregation (e.g., median, proportion), and time point for each outcome	p. 7-8
Harms	15	How harms were defined and assessed (e.g., systematically, non-systematically)	eSupplement, p.7
Sample size	16a	How sample size was determined, including all assumptions supporting the sample size calculation	p. 8
	16b	Explanation of any interim analyses and stopping guidelines	p. 5
Randomisation:
Sequence generation	17a	Who generated the random allocation sequence and the method used	p. 6
	17b	Type of randomisation and details of any restriction (e.g., stratification, blocking and block size)	p. 6
Allocation concealment mechanism	18	Mechanism used to implement the random allocation sequence (e.g., central computer/telephone; sequentially numbered, opaque, sealed containers), describing any steps to conceal the sequence until interventions were assigned	p. 6
Implementation	19	Whether the personnel who enrolled and those who assigned participants to the interventions had access to the random allocation sequence	n.a.
Blinding	20a	Who was blinded after assignment to interventions (e.g., participants, care providers, outcome assessors, data analysts)	p. 6
	20b	If blinded, how blinding was achieved and description of the similarity of interventions	—
Statistical methods	21a	Statistical methods used to compare groups for primary and secondary outcomes, including harms	p. 8
	21b	Definition of who is included in each analysis (e.g., all randomised participants), and in which group	p. 8
	21c	How missing data were handled in the analysis	p. 8
	21d	Methods for any additional analyses (e.g., subgroup and sensitivity analyses), distinguishing prespecified from post-hoc	p. 8
Results
Participant flow, including flow diagram	22a	For each group, the numbers of participants who were randomly assigned, received intended intervention, and were analysed for the primary outcome	p. 9 / Table 2
	22b	For each group, losses and exclusions after randomisation, together with reasons	eSupplement, P.8
Recruitment	23a	Dates defining the periods of recruitment and follow-up for outcomes of benefits and harms	p. 5
	23b	If relevant, why the trial ended or was stopped	n.a.
Intervention and comparator delivery	24a	Intervention and comparator as they were actually administered (e.g., where appropriate, who delivered the intervention/comparator, how participants adhered, whether they were delivered as intended [fidelity])	eSupplement, p.9-10
	24b	Concomitant care received during the trial for each group	P. 6
Baseline data	25	A table showing baseline demographic and clinical characteristics for each group	p. 9 / Table 2
Numbers analysed, outcomes and estimation	26	For each primary and secondary outcome, by group: •the number of participants included in the analysis •the number of participants with available data at the outcome time point •result for each group, and the estimated effect size and its precision (such as 95% confidence interval) •for binary outcomes, presentation of both absolute and relative effect size	p. 11-16 / Table 3 & 4
Harms	27	All harms or unintended events in each group	eSupplement, p.13
Ancillary analyses	28	Any other analyses performed, including subgroup and sensitivity analyses, distinguishing prespecified from post-hoc	eSupplement, p.11ff
Discussion
Interpretation	29	Interpretation consistent with results, balancing benefits and harms, and considering other relevant evidence	p. 16-18
Limitations	30	Trial limitations, addressing sources of potential bias, imprecision, generalisability, and, if relevant, multiplicity of analyses	p. 18-19

^*We strongly recommend reading this statement in conjunction with the CONSORT 2025 Explanation and Elaboration and/or the CONSORT 2025 Expanded Checklist for important clarifications on all the items. We also recommend reading relevant CONSORT extensions. See www.consort-spirit.org.

Citation: Hopewell S, Chan AW, Collins GS, Hróbjartsson A, Moher D, Schulz KF, et al. CONSORT 2025 Statement: updated guideline for reporting randomised trials. BMJ. 2025; 388:e081123. https://dx.doi.org/10.1136/bmj-2024-081123.

© 2025 Hopewell et al. This is an Open Access article distributed under the terms of the Creative Commons Attribution License

(https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.