We classified interventions broadly into three groups depending on the underpinning theories and disciplines in which they were developed. The understanding that ADHD is a neurological imbalance in the brain is dominant in medical disciplines that argue the imbalance can be targeted by specific drugs and forms the first classification of pharmacological interventions (Durston, 2003). We based the second classification, psychosocial interventions, on theorisations that ADHD can be treated through psychological support (Young and Amarasinghe, 2010), and the final classification was entitled combination interventions which included studies where a multidisciplinary approach combining both pharmacological and psychosocial interventions is deemed most effective. Table 4 outlines our classifications and their representative underpinning theory and offers a breakdown of the sub-classifications of the interventions.

Most studies evaluated pharmacological interventions, assessing the efficacy of the three common drug treatments used to treat adult ADHD; Methylphenidate (MPH), Atomoxetine (ATX) and Lisdexamfetamine dimesylate (LDX) (k = 87). Pharmacological interventions can be further categorized according to their drug type as stimulants or anti-depressants. Caye et al. (2018) explain that psychostimulants such as Methylphenidate and Amphetamines are the first line of treatment for ADHD because they are the most researched. Second line treatments involve Atomoxetine and anti-depressants that are often prescribed if psychostimulants are contraindicated or not tolerated or when co-occurrence is present, especially in cases of ADHD and Bipolar Disorder, any substance abuse or Tourette's Syndrome (Caye et al., 2018).

A wide range of psychosocial interventions provided the basis for over a quarter (29%) of the studies which we further classified according to the theories forming the basis of the therapies including cognitive behavioral therapy, mindfulness, attention/cognition training, skills training and/or coaching, and alternative therapies.

The remaining 15 studies combined pharmacological and psychosocial interventions. Most of these studies were stimulant treatment combined with cognitive behavioral therapy (k = 9). Other psychosocial therapies combined with medication were mindfulness based cognitive therapy, group psychotherapy, and problem-focused therapy. One study advanced on the traditional two group comparison by comparing individual counseling to group psychotherapy in pharmacologically treated participants aiming to explore the most effective psychosocial treatment (Philipsen et al., 2007).

We classified the interventions classified by mode of delivery. Of the 143 studies synthesized, 21 were delivered to a group or involved a combination of group and individual delivery. All 21 studies were classified as psychosocial or a combination of pharmacological and psychosocial. Of the 21 studies, 20 were classified as effective, 12 were conducted in European countries with 12 being conducted in universities, research centers or university hospitals. Group interventions are argued to be beneficial for sharing lived experiences, coping strategies, increasing feelings of belongingness and knowledge about ADHD (Bramham et al., 2009; Jackson et al., 2014; Fullen et al., 2020). Learning in a group also increases self-efficacy through higher levels of hope and motivation (Bramham et al., 2009; Tian et al., 2018). However, only six studies directly assessed outcomes relating to self-esteem or efficacy with one study additionally measuring social functioning. Therefore, it is difficult to compare the effectiveness of group vs. individual interventions because the theorized benefits are broad and not consistently measured.

Group interventions tended to include individuals from the ADHDers support network as part of the intervention. Including significant others during the intervention is arguably effective and linked to psychoeducation whereby the ADHDer and their partners or family is empowered with knowledge about ADHD and how to better support it (Lukens and McFarlane, 2004). So far, research examining the efficacy of including family members in interventions for ADHD has focused on children and adolescents (Kaslow et al., 2012). Only five interventions in the present review involved someone in the delivery of the intervention that was not the clinician or individual with ADHD.

Of the five studies, four included the ADHDer's significant other or family member (Virta et al., 2008; Hirvikoski et al., 2011, 2017; LaLonde et al., 2013). The final study included a “support person” or “coach” for help with organizational tasks, and if the ADHDer did not have a support person from their own social network, an undergraduate student was allocated to them to adopt the support person role (Stevenson et al., 2002). All studies reported positive findings with improvements in outcomes beyond reducing symptoms like employment, maintaining relationships, organization skills, self-esteem, and knowledge about ADHD. In short, involving the support network around the ADHDer has marked effects on all outcomes emphasizing that an encouraging and supportive environment can increase the impact of an intervention.

Contexts are defined as environmental factors that affect behavior change (Denyer et al., 2008). These contextual factors can be separated into four layers: the infrastructural system, the institutional setting, interpersonal relationships and the individual themselves (Pawson and Tilley, 1997). We mapped the 143 included studies onto the four areas with referrals being infrastructural system, location as an institutional setting, clinician-patient relationship as part of the interpersonal relationships and lastly, co-occurring at the individual layer.

From the 143 studies 90 were outpatient referrals which means that the adults had received a diagnosis and were immediately referred by the psychiatrist for their first set of treatment at a specialist center or clinic (Kooij et al., 2010). Another method of recruitment was to advertise the intervention and remunerate participants with a formal diagnosis (k = 6). These two methods of recruitment attract and include participants who are already aware that they may have ADHD or have recently been diagnosed. Pawson (2013) suggests that with medication it is difficult to pinpoint the exact moment in which the intervention began, which is particularly true with outpatient referrals as basic knowledge or understanding of ADHD may exist prior to the referral for treatment. As a result of going through the diagnostic process, some level of psychoeducation, researching and learning about ADHD, might have already happened prior to medical treatment. Consequently, it is unclear whether the intervention began at the point in which the ADHDer began to learn about ADHD or at the point medication is initiated. This level of self-awareness and knowledge is difficult to measure and is likely to differ greatly between individuals. With an increase in easily available information like self-help and guidance online, it is important to consider how potential misinformation or accurate information may influence how different interventions are perceived before interventions begin.

Dropout rates varied greatly across the studies. Adherence is a central part of assessing the efficacy of an intervention (Horwitz and Horwitz, 1993). In pharmacological studies, adherence is measured through self-report during follow up sessions recording whether the participant has taken the medication or not. In psychosocial interventions, it is typically assessed by attendance. A review of medical adherence in children and adult ADHDers found non-adherence rates that ranged from 13.2 to 64% but concluded that there is minimal research addressing reasons for non-adherence in adults (Adler and Nierenberg, 2015). Qualitative research has identified forgetfulness, a challenge for ADHDers, and lack of guidance from clinicians as potential barriers to medication treatment adherence (Matheson et al., 2013). Therefore, it important for future research and practice to consider adherence as an influential factor in the efficacy of pharmacological interventions and identify and remove any potential barriers.

The most common settings for pharmacological studies were outpatient clinics or multi-center clinics in North America (k = 47). The majority of participants had been referred to the clinic, received a formal diagnosis and then received a treatment. Assessing the effectiveness of pharmaceutical interventions across multiple clinics not only provides researchers and clinicians with information about the impact of the drug in multiple countries, but also provides an indication of the prevalence of ADHD across cultures (Polanczyk et al., 2007; Wang et al., 2017). Research centers and university settings (k = 24) were more likely to study the effectiveness of psychosocial interventions, but unlike common critiques of a student sample and the lack of ecological validity, only two of these studies recruited student samples indicating good generalisability (Ward, 1993).

Prior research has drawn attention to the significance of the patient and clinician relationship as well as healthcare outcomes (Kelley et al., 2014). As a potential mechanism, it is argued that the better the quality of the relationship, the quicker the recovery and the higher the rate of adherence (Thompson and Mccabe, 2012). The relationship between the clinician-patient is often layered and dynamic making it difficult to directly assess or compare between studies (Street et al., 2009). Key components of an effective relationship are similar in medical and educational settings with research suggesting the following: good management of emotion, high patient or coachee knowledge of their own condition, client/coachee centered approaches and good communication with shared understanding (Street et al., 2009; Jackson et al., 2014; Kelley et al., 2014; Lai and McDowall, 2014). Many of the 143 included studies relied on clinician ratings of symptoms and rarely included the patient. In addition, nearly all the studies had no means of measuring the impact of the clinician/patient relationship with two measuring patient experience directly. In context, as many of the studies were in outpatient clinics, the initial appointment for the treatment was most likely the first point of contact after receiving a diagnosis for a large proportion of the ADHDers, enhancing the importance of a positive and meaningful interaction. Although no measures or understandings were assessed in these studies. We therefore identify the clinician-patient relationship as potentially influential and recommend further investigation.

As a diagnosis, ADHD is rarely present without co-occurrences, clinicians suggest that co-occurrence with depression and anxiety is particularly prevalent due to the experiences of failure, lack of support, and challenges with regulating emotion (Jensen et al., 1997). In experimental study designs, co-occurrence is considered a confounding variable because it is difficult to isolate any beneficial effects of an intervention to a specific condition (Fortin et al., 2006). Consequently, most studies excluded co-occurrences as part of their criteria (k = 135), whereas others deliberately addressed ADHD with co-occurrences like social anxiety disorder or substance use disorder (k = 8). In contrast, there is an argument as to whether removing or excluding participants with co-occurring conditions lessens the external validity because they provide an unrealistic view of ADHD (Fortin et al., 2006).

Mechanisms can be defined as the processes or underpinning methods in which the intervention operates in a specific context to produce a specific outcome that can be triggered in some contexts and not in others (Denyer et al., 2008; Dalkin et al., 2015). We understand mechanisms as the fundamental processes in which interventions are expected to be effective, relating them to the disciplines in which the interventions were developed and theorized, for example, brain chemistry, cognition, and psychoeducation (Denyer et al., 2008).

A large proportion of the studies were pharmacological in nature. From our realist synthesis and evidence-based medicine perspective, randomized controlled drug interventions are the gold standard because the mechanisms, contexts and outcomes have been explored years before the randomized control trial is carried out (Pawson, 2018). For interventions involving stimulants, the process prior to testing the stimulant in humans is extensive (Lipsky and Sharp, 2001). Therefore, the theories and mechanisms have been well-established prior to the intervention.

Cognitive models of ADHD explain a deficit in the prefrontal cortex which is responsible for executive function (Willcutt et al., 2005). Both Barkley (1997) and Brown (2006), leading researchers in ADHD, have developed theoretical models that explain ADHD as a difficulty in managing executive function resulting in impulsive and inattentive behavior. These models have been the foundations behind interventions such as cognitive behavioral therapy (CBT), cognitive remediation, and working memory or attention training. CBT was developed to address anxiety and depression by altering thought processes and behavior to avoid the repetitive negative thinking and corresponding behavior (Beck and Beck, 2011). In a Cochrane review, CBT is argued to treat ADHD by tapping into the negative thinking which has been a result of the negative experiences associated with the core symptoms (Lopez et al., 2018). ADHD is also highly co-occurring with anxiety and depression supporting the use of CBT to target co-occurring symptoms. The techniques used in CBT often include psychoeducation followed by the acquisition of techniques to address the individual challenges the person experiences (Huppert, 2009). Goal setting is an integral part of CBT and is useful in assessing effectiveness (Beck and Beck, 2011). Of the 23 studies that assessed the efficacy of CBT alone or with medication, eight were long term and half included measures of depression and/or anxiety finding positive effects in all studies (k = 11).

Mindfulness formed the intervention in seven of the studies and in many was combined with CBT (Bueno et al., 2015; Edel et al., 2017; Bachmann et al., 2018; Gu et al., 2018; Hoxhaj et al., 2018; Hepark et al., 2019; Nicastro et al., 2021). Similar to CBT, mindfulness aims to tap into cognition. However, CBT is focused on attention rather than other symptoms and aims to focus attention on the present, inner emotions and acceptance (Pirson, 2014). Most of these studies had positive effects on reducing symptoms, one found no difference in measures of memory (Bueno et al., 2015) and another found that mindfulness was effective for all participants, not just those with ADHD (Bachmann et al., 2018).

Psychoeducation aims to enhance a person's understanding of mental health by increasing the individual's knowledge and awareness of their condition and supporting them in sharing their experiences (Getachew et al., 2009). It also offers the opportunity for those supporting the individual such as family members, to help support ADHDers (Anderson et al., 1980). The idea is that self-awareness and knowledge is key in learning strategies to manage any condition or increase functioning rather than simply reduce the symptoms (Lotfi et al., 2010). In some areas, psychoeducation is argued to adopt a strengths-based approach (Lukens and McFarlane, 2004). Despite the strong evidence base as an intervention for affective disorders and preliminary evidence in interventions for children and adolescents with ADHD, there is limited research applying psychoeducation with ADHDers (Lukens and McFarlane, 2004; Dahl et al., 2020). Seven studies mentioned their use of psychoeducation typically combining it with another therapy or training (Wiggins et al., 1999; Hirvikoski et al., 2017; In de Braek et al., 2017; Bachmann et al., 2018; Hoxhaj et al., 2018; Gaur and Pallanti, 2020; Hartung et al., 2020). These studies indicated improvement in a range of outcomes including executive functioning, time management/organization skills, general functioning, and knowledge, as well as knowledge and coping in significant others. Therefore, we consider psychoeducation as a key mechanism that is worth incorporating in future design of interventions because of the variety of benefits for the ADHDer as well as significant others.

The outcomes assessed in each intervention varied greatly across the 143 studies. Initially, we classified the primary outcomes according to whether they involved a measurement of the core symptoms. Those outcomes beyond the core symptoms were further classified into what they assessed: behavior, cognition, physical/functioning, social, and person/emotion. We additionally discuss the outcomes in relation to short-term and long-term and qualitative methods.

Core outcome measures were defined as those assessing the three core symptoms of adult ADHD; impulsivity, inattention, and hyperactivity. Of the 143 studies, 108 assessed a reduction in core symptoms as their primary outcome. There are several validated measures for adult ADHD and in most of the studies a mixture of these measures was used pre and post intervention. The most popular being Conner's Adult ADHD Rating Scales (CAARS) (k = 44) and the ADHD Rating Scale (ADHD-RS) (k = 38). The authors of the ADHD-RS are prominent authors in the intervention studies and the CAARS was used to validate the measure. It is important to consider the authors as potential sources of bias because of their involvement in developing the measures.

We categorized a range of outcomes beyond core symptoms further which are displayed with relevant examples in Table 5 alongside the number of studies involving these types of measures. The vast range in outcomes suggests the impact of ADHD to all aspects of life beyond the core symptoms, from social relationships to general self-esteem. Regarding effectiveness, primary outcomes related to cognition were associated with mixed or unclear results. Outcomes relating to social and emotion/person were assessed more often in psychosocial interventions with overall positive effects of the intervention. Aside from cognitive assessments of outcomes which typically involve using technology, and the Clinical Global Impression scale (k = 41), which is purely based on the clinicians rating, many outcomes were assessed using self-report rating scales. In one study, a strength of ADHD, creativity, was measured and improved after the administration of a stimulant which is in line with the expert panel, that strengths are important to consider (Kahn, 2008). Evaluating outcomes associated with strengths supports the neurodiversity conceptualization of ADHD and challenges the pathological approach.

We classified studies as short-term if the treatment to follow-up was under 6 months (k = 111) and long-term if they were 6 months or more (k = 32). Combination treatments were more likely to be long-term than short-term (k = 6) indicating a longer follow-up. A total of 78% of pharmacological interventions were short-term with more than half of these (56%) lasting <12 weeks suggesting an immediacy to the expected effectiveness. In contrast, in psychosocial studies, therapeutic effect is assumed to be less immediate as they tend to address a wider range of symptoms and co-occurrences (Biederman et al., 2012). Long-term research into the impact of ADHD across the life span indicates that symptoms beyond the core symptoms such as functionality and anxiety become more prominent over time, yet current research does not reflect this because the majority of studies are short term and evaluating pharmacological interventions (Ingram et al., 1999).

There were two studies that evaluated the effectiveness of interventions using qualitative methods (Nordby et al., 2021) with one using both qualitative and quantitative data to support their findings (Björk et al., 2020). Both studies interviewed participants about their experience of completing the interventions and extracted themes using thematic and content analysis. Their analyses support the key mechanisms identified for effective psychosocial interventions with one theme highlighting the importance of trusting relationships with the clinician delivering the intervention (Björk et al., 2020), and the other emphasizing the sense of belonging and shared experience in group interventions (Nordby et al., 2021).