We recruited 81 participants with a verified diagnosis of ADHD, currently receiving follow-up for their ADHD at District Psychiatric Center Nedre Romerike at Akershus University Hospital in the inclusion period lasting from June 2019 until April 2021, based on the following eligibility criteria, that were similar for both groups:

As part of the public, specialized healthcare system, the District Psychiatric Center Nedre Romerike offers assessment and treatment related to a range of psychiatric conditions, including ADHD. Being among the largest in Norway, the clinic served an adult population of approximately 125,000 during the inclusion epoch for this trial. Thus, based on the prevalence of 2.8% (Kooij et al., 2019), an estimated 3,500 adults diagnosed with ADHD lived in the area served by the clinic during the inclusion period.

First, we offered brief information about the study to patients potentially fulfilling the eligibility criteria. Next, we offered patients choosing to learn more about the study written information and opportunities to ask questions. Volunteers fulfilling the inclusion criteria provided written informed consent and were randomized in a 1:1 ratio to the intervention group or the control group by a person independent from the study team. See Figure 1 for the study flow chart.

The intervention consisted of: (1) Group-based GMT and (2) Individual guidance in formulating GAS goals for coping with executive problems in everyday life.

TAU consisted of individually adapted follow-up for ADHD from a team comprising psychiatrists, psychologists, social workers, psychiatric nurses, and a psycho-motoric physiotherapist. Participation in the study did not affect pharmacological treatment or access to sessions with the health care professionals in the multi-disciplinary team. The mean numbers of individual sessions provided to the participants in the intervention group in addition to the intervention, and to participants in the control group during the eight-month-long inclusion period are presented in Table 2 in the Results section. Blinding related to group allocation was not possible due to the nature of the intervention.

At baseline (T1), the participants in both groups underwent comprehensive neuropsychological assessment and responded to self-report measures regarding executive functions, psychological well-being, and intensity of ADHD symptoms. The same self-report measures were given to both groups at T2 (2 months past baseline, the time-point coinciding with completion of the psychoeducational phase), at T3 (5 months past baseline, the time-point coinciding with completion of the telephone follow-up), and finally at T4 (8 months past baseline). We scored GAS goal attainment in the intervention group between T2 and T3.

Baseline neuropsychological assessment included the widely used and well-validated instruments summarized in Table 3.

Information about current psychiatric status/comorbidity was collected from each patient’s journal. Individual goal attainment was measured by GAS (Kiresuk and Sherman, 1968), following the standard GAS procedure, thoroughly described by Turner-Stokes (2009).

The primary outcome was the BRIEF-A GEC score, and the primary time point was at 8 months (T4). Secondary outcomes included the HSCL-25 (total score and sub-scores for anxiety and depression) and the ASRS-v1.1.

The lack of previous studies on GMT in ADHD represents a challenge in estimating the required sample size. However, findings from a recent study (Cameron et al., 2020) examining the use of GMT in individuals with obsessive-compulsive disorder suggest that based on an average effect size of η² p = 0.054 for neuropsychological variables observed across all results in Cameron et al.’s study, with a critical alpha = 0.05 and 80% power, a sample size of n = 23 per group would be required to reliably detect the smallest desired effects. Based on the above, we would need a total of 46 individuals in our study, but to allow for a dropout rate of about 40%, we aimed to include at least 64 participants.

Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) Version 26. Continuous variables were described with mean and standard deviation (SD), categorical data as counts and percentages. Crude differences between groups were assessed using t-tests for continuous variables and Chi-square for pairs of categorical variables.

We used a linear mixed model for repeated measures to assess possible between-and within-group differences over time. We applied a 2×4 mixed-design, with Group (GMT, TAU) as the between-subjects factor, and Session (T1: baseline, T2: 2 months, post-GMT, T3: 5 months, post telephone follow-up, and T4: 8 months) as the within-subjects factor. In line with previous recommendations (Løvstad et al., 2016), clinically meaningful change in the BRIEF-A was operationalized as an improvement of one SD. All tests were two-sided and p-values <0.05 were considered statistically significant. All analyses were considered exploratory so no correction for multiple testing was done.

Demographic and clinical characteristics of the study population at baseline are presented in Table 4. Median age was 30.0 years (range 18–55), and 47 (58%) of the participants were women. The proportion of women in the study population was higher than expected, based on a previously-reported gender ratio of 3:1 males to females in adults with ADHD (Young et al., 2020), although the gender disparity has been reported to disappear in the adult ADHD population (Faraone et al., 2015). Furthermore, 19/81 (23.5%) were diagnosed with anxiety, 23/81 (28.4%) with depression, 4/81 (4.9%) with a personality disorder, and 2/81 (2.5% with Tourette Syndrome). Based on observations/information revealed post inclusion, an additional assessment was considered necessary for some participants, resulting in 3/81 (3.7%) being diagnosed with pervasive developmental disorder (PDD). Comorbidities were comparable between the groups. There were no significant baseline differences between the intervention and control groups regarding age, education, gender, intelligence quotient (WAIS-IV), BRIEF-A composite scores (GEC, MI, BRI), HSCL-25 scores (total, anxiety, depression) or ASRS-v1.1 total score. The proportion of employed individuals/students was significantly higher in the intervention group [25/40 (63%)] compared to the TAU group [14/41 (34%)]. The mean scores on the BRIEF-A were approximately 2–2.5 SD above the normative mean for both the intervention group and the control group.

Neuropsychological characteristics of the study sample are presented in Table 5. On a group level, the study sample achieved T-scores within the normal range on all neuropsychological tests.

Comparison with independent samples t-tests revealed no significant difference in the total number of TAU sessions provided to the intervention and control groups. However, the number of sessions with a psychiatrist was larger in the control group (mean of 5.30 sessions, as compared to 2.31 sessions in the intervention group, p < 0.001).

In the intervention group, 32 of the 41 participants attended at least five of the eight GMT sessions, which was set as the cut-off for being a GMT completer. This indicates a GMT dropout rate of 22%. One participant received a summary of the two last sessions individually, administered by video link due to changes in their job situation which made it impossible to attend in person. Those specifying reasons for not attending mentioned illness or critical events in their family or close network, work obligations, and anxiety.

Thirty-one of the 41 participants in the intervention group completed GAS goal formulation (representing a dropout rate of 24%). The participant who received the last group sessions by video-link also completed GAS goal formulation during the video-link sessions. Twenty-seven of the 31 participants (87%) completing GAS goal formulation responded to the bi-weekly telephone follow-up and were assessed for goal attainment, administered between T2 and T3.

Regarding self-report measures, the response rate varied considerably between the two groups. Whereas 30 participants (75%) in the control group responded at the five-month follow-up (T3) and 29 (73%) responded at the eight-month follow-up (T4), only 17 (42%) and 20 (49%) participants in the intervention group responded at 5 and 8 months, respectively.

A comparison of participants completing and dropping out of the GMT intervention is presented in Table 7. Chi-square tests revealed no significant differences between the dropouts and completers regarding age, gender, years of education, employment/student status, comorbid psychiatric disorders, or receiving pharmacological ADHD treatment during the study period. However, independent sample t-tests revealed a significantly higher score on the WAIS-IV Perceptual Reasoning Index (PRI) in the GMT completing group, also resulting in a higher WAIS-IV total score (FSIQ) in this group. There were no significant differences between completers and dropouts on the other WAIS-IV composite scores (Verbal Comprehension Scale, Working Memory Scale, or Processing Speed Scale).

Thirty-one participants in the intervention group (76%) succeeded in formulating individual GAS goals for coping with executive problems in everyday life. They set between one and three goals (M = 1.65), and altogether 51 goals were set. Twenty-seven of the 31 participants initially formulating goals (87%) responded to at least five of the six phone calls and were defined as completers of the GAS intervention. Goal attainment was scored following the procedure described by Turner-Stokes (2009), where a T-score of 50 represents goal attainment as expected, 1 SD is 10, and higher scores represent better performance. On a group level, an improvement of approximately two SDs occurred from baseline until phone-call number six (at T3). The mean GAS score at baseline was 31.4 (SD = 3.5), improving to 50.48 (SD = 14.8), representing a mean GAS change score of 20.7 (SD = 12. 8). The participants decided the number and contents of the goals themselves. The goals could be classified into four different categories:

The intervention did not demonstrate any additional effect in reducing perceived executive dysfunction. However, evaluation of the effects of multi-faceted interventions is challenging, and several factors may have contributed to the lack of significant differences in improvements between the groups.

Both interventions included components that could have facilitated the identification of situations where exerting executive control was required, resulting in the application of remaining reserves in executive functions. Of note, subjective executive complaints were among the inclusion criteria, and all participants had the opportunity to address these in the individual sessions. Accordingly, the differences between the intervention and TAU turned out to be less than originally planned. Future studies investigating the separate effects of interventions in multidisciplinary programs are warranted. Such studies will be particularly relevant in the ADHD population, given the typically complex and multifaceted nature of challenges pertaining to the condition.

In addition to TAU, both groups received a comprehensive neuropsychological assessment that included feedback. On a group level, all neuropsychological test scores at baseline fell within the normal range. Awareness of individual strengths and resources may have enhanced therapeutic processes and contributed to the significant improvements in mood observed in both groups. This is in line with previous studies reporting benefits of neuropsychological assessment with feedback upon self-esteem (Bennett-Levy et al., 1994), insight (Watt and Crowe, 2018), acceptance (Gruters et al., 2021), adherence to treatment and symptomatic distress (Fallows and Hilsabeck, 2013).

Of note, mean neuropsychological test scores within the normal range illustrate clearly that there is no evidence in our study of an “ADHD-specific neuropsychological profile.” Importantly, the high scores on the self-report measures of executive problems nevertheless indicate that executive problems are indeed present in this population, underscoring the importance of a broad assessment approach, not solely based on neuropsychological test scores. It is well-recognized that neuropsychological test scores are suboptimal measures of executive functions due to the high level of structure characterizing the test situations (Garcia-Molina et al., 2012).

This trial was not designed to compare pharmacological and non-pharmacological treatment effects. As we made no restrictions regarding pharmacological ADHD treatment, participants in both groups had equal options for sessions with a psychiatrist, where they could discuss whether such treatment should be initiated, terminated, or optimized. Of note, registration of the TAU content revealed a significantly higher number of sessions with a psychiatrist provided to the control group. A possible explanation for this difference can be that the less time-consuming control condition left those participants with more time and energy to focus on pharmacological issues. Simultaneously, the participants in the intervention group may have prioritized the intervention over psychiatrist sessions, either because they felt it was sufficient to meet their needs, or because it was too demanding to simultaneously focus on pharmacological issues. Further studies investigating the separate effects of pharmacological and non-pharmacological treatments in ADHD are called for.

Our findings are in line with findings from a previous RCT (Hagen et al., 2020), where patients with major depression and self-reported executive deficits took part in either GMT or computerized cognitive training. As in the current trial, improvements in executive functioning from baseline occurred in both groups, but analyses did not reveal any significant group differences in improvement (Hagen et al., 2020). Notably, as the improvements were not sustained until the two-year follow-up, the authors concluded that improvements in everyday executive functions may require maintenance or additional treatment (Hagen and Stubberud, 2021).

Analyses did reveal significant improvements in anxiety as an effect of the intervention. We consider this an important finding, providing support for a broad treatment approach in psychiatric health care, where interventions from the cognitive remediation field can complement traditional psychotherapeutic interventions, even when the aim is to improve mood. Given the considerable overlap and complex relationships between symptoms of mood disorders, ADHD, and executive problems, this makes considerable sense. Furthermore, this finding provides support for considering executive dysfunction as an important treatment target when the goal is to alleviate anxiety (Haugan et al., 2022).

The anxiety finding is particularly relevant for the approximately 30% of ADHD patients who do not respond to or tolerate psychostimulants (Bouffard et al., 2003). The finding is in line with previous findings that anxiety can improve as a result of increased attentional control/improved self-regulation (Rudea et al., 2004) and improved organization skills (Jarrett, 2016; Mohamed et al., 2020). Treatment of anxiety in the ADHD population is particularly important, as anxiety has been reported to increase the ADHD-symptom load (Reimherr et al., 2017).

GMT comprises several components that have previously been reported to reduce anxiety, including implementing strategies for executive coping in everyday life (Jarrett, 2016; Mohamed et al., 2020), mindfulness (Kabat-Zinn, 2012; Hofmann and Gómez, 2017), and the general therapeutic effects associated with being part of a group, meeting others and receiving psychoeducation (Reimherr et al., 2017).

Statistically significant reductions in self-report measures do not necessarily imply clinical significance. Moreover, it can be discussed how meaningful the significant reductions measured by the HSCL-25 were for functioning in everyday life. On a group level, the intervention group’s anxiety subscale mean score reduced from 2.14 to 1.82. On the total scale, it reduced from 2.29 to 1.97. For the HSCL-25, total scores of 1.75 and above have been suggested as an indication of clinical mental distress (Skogen et al., 2017). Cut-off scores of 1.67 for men and 1.75 for women have also been suggested (Sandanger et al., 1998). Of the 22 participants in the intervention group responding to the HSCL-25 at the eight-month follow-up, a total of 8 (37%) scored below 1.75. Thus, despite the statistically significant reductions, the scores on HSCL-25 remained within the clinical range for approximately two-thirds of the intervention participants.

The factor structure of the HSCL-25 has also been questioned, and a unidimensional model has been described as more appropriate (Skogen et al., 2017). Furthermore, it has been argued that patients do not necessarily discriminate among symptoms; improvement in one area can be experienced as improvements in other areas (Reimherr et al., 2017). On the other hand, significantly better improvements in the intervention group relative to the control group did not occur on the measures of executive functioning (BRIEF-A) or intensity of ADHD symptoms (ASRS-v1.1), indicating that the effects of the intervention were more specifically related to anxiety.

As we expected, at least a proportion of the participants in the intervention group [31/41 (76%)] successfully formulated GAS goals for coping with executive problems. Furthermore, the majority [27/31 (87%)] responded to the telephone follow-up. On a group level, an improvement of about two SDs occurred from baseline and up to the level perceived as expected and valued by the participants. The participants decided the content of their individual goals themselves. The goals could qualitatively be classified into the following four categories: Planning and organizing (22 goals), Practicing/utilizing GMT strategies in everyday life (12 goals), Physical activity (11 goals), and Pleasure activities (6 goals). The first two categories clearly relate to executive functioning, underlining the importance of executive coping for the participants.

The improvements in both executive coping and anxiety in the current study are in line with previous research reporting a reduction in anxiety as a consequence of improvement in executive coping (Jarrett, 2016; Mohamed et al., 2020). Our findings are also in line with previous suggestions, that anxiety in ADHD may arise when cognitive processing abilities are overwhelmed by the demands of the environment (Schachar et al., 1995; Haugan et al., 2022), and that emotional symptoms in ADHD may be caused by problems related to coping with daily life (Zhang et al., 2021). Improvement of anxiety as a result of guidance in individual goal setting related to activity has also been reported in older veterans (Gould et al., 2021).

The completion rates were considerably higher for the GAS goal intervention compared to the GMT intervention. One participant even completed the GAS goal intervention despite dropping out of the GMT intervention. These findings may indicate that the GAS intervention is more feasible for adults with ADHD, compared to the more time-consuming GMT intervention.

The satisfactory completion rates and successful GAS goal attainment in the current trial are in line with a previous, similar trial (Hanssen et al., 2016), concluding that GAS was a feasible and robust method in cognitive rehabilitation for patients with multiple sclerosis (MS) and cognitive/executive complaints. Beneficial results of GAS have also been reported in various other settings, including special education, rehabilitation, general medical health programs, pain management, and substance abuse treatment (Malec, 1999). GAS has, for a long time been a well-established method within the rehabilitation field (Turner-Stokes, 2009). Our results indicate that the method can apply within a psychiatric setting as well.

To our best knowledge, this was the first RCT investigating the effects of an intervention combining group-based GMT (Levine et al., 2011) and individual GAS goal-setting (Kiresuk and Sherman, 1968), aiming to help adults with ADHD cope with executive problems in everyday life. Strengths of the study included the use of a sound methodological design, data collection in a naturalistic setting in one of the largest outpatient clinics in Norway, and inclusion only of patients with a verified ADHD diagnosis. The baseline assessment revealed average general mental ability scores in the study sample. The findings can thus be generalized to adults seeking help for coping with problems related to ADHD in a psychiatric outpatient setting. It must be noted that the study only recruited participants with adequate language skills to participate in group discussions and that patients with central nervous system injury, ongoing substance abuse, or psychopathology too severe to be handled in a group-based outpatient setting were excluded. Registration of comorbidity in the study population revealed high rates of comorbid disorders (e.g., depression, anxiety), which is common in the ADHD population (McGough et al., 2005; Torgersen et al., 2013; Prevatt et al., 2015; Nelson and Liebel, 2018; Anker et al., 2019). Thus, our findings may also reflect effects not directly related to ADHD but to concurrent mental disorders. Furthermore, as inclusion was voluntary, we were only able to recruit participants willing to try out a novel treatment approach. Blinding was not possible due to the nature of the intervention. Consequently, we cannot rule out the possibility that awareness of group allocation and the purpose of the study may have affected the responses.

As participation in the study did not affect access to sessions with the health care professionals in the multi-disciplinary team, and no alternative or scam condition was provided to the control group, the total of sessions differed between the groups. This represents a methodological flaw, as the possibility that improvements could be related to other factors than the intervention (e.g., receiving attention from health care providers) cannot be ruled out.

A considerable limitation was the high attrition rates. Only 78% of the intervention group completed the main part of the intervention, GMT. Notably, similar dropout rates were reported in a recent exploratory study of GMT for patients with ADHD (Jensen et al., 2021). Comparisons between GMT completers and dropouts in the current trial did not reveal any statistically significant differences regarding age, gender, employment/student status, receiving pharmacological ADHD treatment during their study participation, or comorbid psychiatric disorders. Attrition in the current study was particularly high in the intervention group at the time points for answering self-report questionnaires at 5 and 8 months. Because of the high attrition rates in our study, there is limited power to detect small differences between the groups, increasing the chance of type-II errors.

The data does not provide the basis for determining why the response rates for the self-report measures were considerably higher in the control group. One hypothesis could be that participation in the intervention was energy-consuming, thus leaving those participants with less energy to complete and return the self-report questionnaires. This is in line with findings from a previous qualitative study, where adults with ADHD were interviewed about their experiences of participating in group-based GMT (Nordby et al., 2021). Even though most participants experienced beneficial effects, such as expanded perspectives, increased personal growth, and coping with ADHD in everyday life, some participants described taking part in the intervention as burdensome. Another possibility is that higher attrition could relate to the significantly lower proportion of occupational or academic participation in the intervention group. We cannot rule out the possibility that unknown variables might have contributed to both attrition and employment status.

One possible explanation for attrition in both groups could be that paying attention to the study was perceived as more interesting at the beginning and that completing paperwork later on was less motivating. Furthermore, we believe that the use of paper-and-pencil versions of the questionnaires was a major source of attrition. Participants in both groups frequently reported that they had lost the questionnaires or forgotten to return them. This observation lends support for digital data collection in future research.

The fact that the study was partially undertaken during the COVID-19 pandemic had implications for both intervention delivery and the everyday life of participants. In accordance with social distancing rules, we had to reduce the size of some of the groups from six to four to keep appropriate distance within the available rooms. The consequences of participants dropping out of group sessions were more noticeable in smaller groups, providing the remaining participants with fewer discussion partners and thus fewer options for shared experiences. Even more importantly, social distancing rules and other pandemic restrictions may have influenced executive coping and psychological well-being for all participants. The psychological impacts of quarantine have indeed been described as substantial and wide-ranging (Brooks et al., 2020). Yet, as inclusion and randomization were undertaken continuously, the effects of the pandemic did not differ systematically between the intervention and control groups.