A systematic review of mindfulness based interventions for children and young people with ADHD and their parents

Journal articles were required to meet the following criteria in order to be included in the present review. These criteria were defined in accordance with the PICO requirements: Population, phenomenon of interest, design, evaluation outcomes and research type (Murdoch University, 2019).

Participants under 18 years of age with/without their parents. The children met the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) or the International Classification of Diseases (ICD-10) criteria for ADHD. Studies were excluded if ADHD was not the primary diagnosis.

The following interventions were included: mindfulness based stress reduction (MBSR; Kabat-Zinn, 1990), mindfulness training for parents (MP; Kabat-Zinn & Kabat-Zinn, 1997), mindfulness-based cognitive therapy (MBCT; Segal et al., 2002), acceptance and commitment therapy (ACT; Hayes et al., 2004) and mindfulness-based positive behaviour support (MBPBS; Singh et al., 2015). Interventions were delivered by therapists trained in mindfulness based approaches. Studies that used mindfulness only as a component of the intervention (for example, an adjunct to behavioural parent training, BPT) were excluded.

The studies were selected according to the following criteria of eligibility: pre-post quantitative assessment design conducted in any setting, with or without a control group. The studies included in this review were published in English, Italian, French, Spanish and Portuguese in peer-reviewed journals or dissertations.

Studies that assessed the effectiveness of outcomes through post-treatment follow-up. The informants of this follow-up were either the child with ADHD and/or the parent and/or teacher. Relevant outcomes included behavioural and psychological changes in children with ADHD and/or parents after MBIs.

The review assessed existing research and followed the preferred reporting items for systematic reviews and meta-analysis (PRISMA) guidelines for systematic reviews (Moher et al., 2009). Studies were collected searching the following databases: PsychINFO (1806–present), EMBASE (1947–present), MEDLINE (1946–present), CINAHL (1937–present), ERIC (1966–present), ASSIA (1987–present) and Social Services Abstracts (1979–present). In order to expand the searches to the grey literature and to prevent publication bias, Proquest Dissertation and Thesis (1743–present) and Cochrane Central were also searched. The last search was completed on 15th September 2017.

The main search keywords were the following: 1) mindfulness OR MBCT OR MBPBS OR MBSR OR acceptance; 2) attention deficit hyperactivity disorder OR ADHD OR attention deficit

disorder; 3) child* OR youth OR adolescent* OR young people, AND/OR parent* OR mother* OR father* OR caregiver* OR carer*. Following the database searches, Google Scholar was used to find the grey literature. The reference list of articles and dissertations on MBI and ADHD were also closely scrutinized. The process of article selection is illustrated in Figure 1.

Figure 1

The quality of the studies included was assessed using a bespoke quality appraisal method tool, ‘Quality Indicators for Group Experimental and Quasi-Experimental Research in Special Education’ (Gersten et al., 2005). Moreover, elements from the Cochrane EPOC checklist were added to the quality criteria (Greenhalgh, et al., 2005). The criteria provided a comprehensive analysis of the quality of the studies and comprises the following checklist items: research question and design, participants/sampling, treatment fidelity, outcome measures, data analysis, follow-up, attrition and generalizability.

As suggested by Coull and Morris (2011), the first author (I.T.) rated each outcome (research question and design, participants/sampling, treatment fidelity, outcome measures, data analysis, follow-up, attrition and generalizability ) as: ‘well-covered’ (2 points); ‘adequately addressed’ (1 point); and ‘poorly addressed’, ‘not addressed’, ‘not reported’ and ‘not applicable’ (all 0 points). Each study quality has been then categorized as being: excellent, very good, reasonable and limited. Twenty percent of the articles reached an agreement of 100% on methodological quality with a second independent rater (N.F.).

The initial search strategy produced 1424 articles, to which two studies were added from other resources. After the duplicates were removed, a total of 1244 articles remained (the search strategy is elucidated in Figure 1). Through the title screening process, 1225 studies were considered non-eligible to be included in the systematic review. The abstract and the full text of 17 studies were carefully examined. Two dissertations were not available in full-text format, so they were excluded.

Upon additional comprehensive reviewing of these articles, 7 studies were excluded from the sample following full-text review, due to: not being an empirically validated mindfulness program but a meditation practice instead (Grosswald et al., 2008; Harrison et al., 2004; Kratter, 1983; Moretti-Altuna, 1987); being a mindful parenting psychoeducation program with no therapist guidance (Anderson & Guthery, 2015); the primary diagnosis was a learning disability and ADHD was a comorbidity (Haydicky et al., 2012); the results were mixed between children and adults with ADHD (Zylowska, et al., 2008). The final review was based on 10 studies.

Ten studies ( see Table 1) were included in this systematic review. Four studies employed a multiple baseline across participants’ design (Carboni, 2012; Carboni, et al., 2013; Shecter, 2013; Singh, et al., 2010). Five studies used a quasi-experimental design, and pre- and post-test (Haydicky, et al., 2013; Haydicky, 2014; van de Weijer-Bergsma, et al., 2012; van der Oord, et al., 2012; Worth, 2013). Only one study (Sidhu, 2013) employed a randomized pre-test and post-test with control group design.

This review included studies with a total of 134 children diagnosed with ADHD. In six studies, the children’s parents were also receiving mindfulness treatment (n = 89). Children were between the age of 8 and 18 years, with the majority of participants being male. According to the data available, the parents were predominantly mothers from a middle-high socio-economic status.

Nine studies reported that some of the participants were on medication for ADHD (range 10–100%). Four studies reported that children had a comorbid disorder (including: learning disability, depression, anxiety, tic disorder and ODD). In terms of outcome measures, 8 studies utilized well validated tools. Two studies have used measures that were developed by the primary investigator (Shecter, 2013), and single-item rating scales whose psychometric properties were not established (Singh, et al., 2010). Half of the studies were published in peer-reviewed journals whilst the remaining were a part of doctoral theses.

Core symptoms of ADHD

Nine studies evaluated whether MBI improved the symptoms of ADHD within a classroom environment and in a home setting through parental/teacher and self-reports, and through computerized tests. The reporting results are consistent in terms of the reduction of inattention. In the only study with a control group design (Sidhu, 2013), it was found that the mindfulness training group showed statistically significant improvements in the attention span as measured in both BASC 2 (ηp2 = 0.147, medium effect size) and Conners Parent Rating Scale – Revised (ηp2 = 0.32, large effect size).

Carboni (2012) and Carboni, et al. (2013) also found an increase in on-task behaviours in classroom setting according to behavioural observations. Where reported, the significant change in attention, reported by parents and teachers, showed a medium to large effect size (Haydicky, et al., 2013; Haydicky, 2014; Sidhu, 2013; van der Oord, et al., 2012; van de Weijer-Bergsma, et al., 2012; Worth, 2013). These results were maintained with a large effect size (d = 0.80) at follow-up in the study by van der Oord et al. (2012).

Some reporting differences were noted among participants. Adolescents did not perceive changes in their inattention level after the MBI in the studies by: Haydicky et al. (2013), Haydicky (2014), and van de Weijen-Bergsma, et al. (2012). Moreover, in the study by van de Weijer-Bergsma, et al. (2012) at the 8-week follow up, the mothers did not report any significant reduction in attention problems, whilst the fathers and adolescents saw a difference with a large effect size (d = 1.5 and d = 0.9, respectively). Shecter (2013) noted that 14 parents (64%) reported that their adolescents experienced less difficulty paying attention after completing the BMI, whilst only four adolescents indicated reductions (44%).

Sidhu (2013) used a Test of Variables of Attention (TOVA) and found that the mindfulness training group showed a significant improvement in inattention as compared to the control group in the infrequent target condition with a medium to large effect size (ηp2 = 0.21). This was also demonstrated in the reaction time and the commission errors tests with a large effect size.

Worth (2013) reported a statistically significant reduction in attentional shifting with a large effect size in both post-intervention scores: numbers only (r2 = 0.76) and numbers and letters (r2 = 0.72) on the Trail Making Test (TMT). Similar results were found in the Response-Distractor Inhibition Test

(SCWT), and in the Color-Word scores a large effect size (r2 = 0.81) at post-intervention.

Van de Weijer-Bergsma, et al. (2012) have found conflicting results when a computerized attention test for baseline speed (BS) was used, with no improvement in reaction time at 8 and 16-weeks follow-up. However, a significant reduction in reaction speed between pre- and post-test with a large effect size (d = 0.9) was detected.

Van der Oord, et al. (2012) have found a reduction in hyperactivity and impulsivity with a medium effect size (d = 56), which was maintained at follow up (d = 0.59). Carboni (2012) and Carboni et al. (2013) have found that their BASC-2 teacher and parent ratings have indicated an improvement in the ability to inhibit behaviour and a decrease in hyperactive behaviour.

However, these results are not consistent with other studies where parents did not report significant changes in hyperactivity and impulsivity (Haydicky et al., 2013; Worth, 2013). Also, Shecter (2013) noted that, according to parent reports at follow-up, the MBI had a reductive effect on most of the adolescents’ hyperactivity (71%) and impulsivity (64%) symptoms. However, only 44% adolescents reported reduced hyperactivity levels from participating in the MBI.

Three studies have studied the impact of MBI on executive functioning. Haydicky et al. (2013) and Haydicky (2014) showed a significant effect of MBI on learning problems (p = 0.049) and executive function (p = 0.010). Using the BRIEF test, van de Weijer-Bergsma et al. (2012) found a borderline reduction in metacognitive problems at pre- and post-test (p = 0.08, d = 1.0) and at 8-week follow up (p = 0.01, d = 1.8), as reported by the fathers. Tutors and mothers did not see any significant change.

Haydicky et al. (2013) and Haydicky (2014) reported that parents found a significant reduction in conduct problems between pre-and post-test, with a medium to large effect size (d = 0.7). Van de Weijer-Bergsma et al. (2012) found that at post-test, there was a significant reduction in externalizing problems (d = 0.2) reported by the fathers but not by the adolescents, mothers or tutors. At the 8-week follow-up, the reduction reported by the fathers was maintained (d = 0.3) and the adolescents reported a borderline significant reduction (d = 0.5).

Six studies analysed the impact of MBI on parental stress and the results were consistently positive. In Haydicky et al. (2013) and Haydicky, (2014) parents reported a significant reduction in parenting stress with a large effect size at follow up compared to post-test (d = 0.81). Moreover, the authors found improvements in family functioning (p = 0.43). Shecter (2013) reported that 85% parents showed improvements in terms of parental stress in the last half of treatment and/or across follow up.

Van der Oord, et al. (2012) noted a significant reduction in parental stress with a medium effect size (d = 0.57) and over-reactivity with a large effect size (d = 0.85). Van de Weijer-Bergsma et al. (2012) found a significant reduction in parenting stress between pre- and post-test reported by the fathers (d = 0.07), which was maintained at the 8-week follow-up (d = 1.1). There was a borderline significant reduction in the parental over-reactivity reported by the mothers between pre- and post-test (d = 1.0), which was not maintained at the follow-up.

Singh et al. (2010) found an increase in child compliance and in satisfaction among parents regarding the interaction with the child, further increased during follow-up. Moreover, Shecter (2013) reported a positive effect of MBI on 71% of parents’ and 78% of adolescents’ self-reported distress from conflict.

Five studies were rated as reasonable according to a bespoke quality appraisal method tool adapted from the article published by Gersten et al. (2005) and with elements from the Cochrane EPOC checklist (Greenhalgh et al., 2005). The significant methodological limitations were related to: outcome measures, data analysis, completeness of data set and follow-up, and attrition (Table 2).

The sample used for the studies is indeed representative of the population of interest as all the participants have received a diagnosis of ADHD. However, most of the participants have a reported comorbid diagnosis, which might lead to confounding factor. As most of the studies have utilized a small sample, this indicates an important methodological limitation that might hinder the subsequent external validity and generalizability of the results (Shadish, 2002).

Another main methodological limitation is that in 9 studies out of 10, there is a lack of randomized control group, which might again hinder the external validity of the studies. As discussed, five studies used a quasi-experimental design where the internal validity can be at stake. The groups that have received the MBI and the control groups may not be comparable at baseline. The random assignment may allow the participants to have the same opportunity to be allocated to any of the groups and the observed treatment intervention features would be related to chance rather than to methodical considerations. Quasi-experimental design studies are not always able to establish the causal relationship between the intervention and the observed results, particularly if there are significant confounding variables (Gribbons & Herman, 1997).

The lack of randomized control group might have affected the four studies that have used a multiple baseline design. These studies might have incurred in experimenter biases, which might have hindered the objectivity of their causal effects. It is not clear in all the studies if an inter-rater agreement of two independent observers has been met (Iversen, 2013). Moreover, due to the very small sample (between 2 and 9 children), the results might not be generalized as some characteristics differ more across individuals.

Most of the studies are uncontrolled with no comparison with other treatments to determine whether the improvements are caused by the MBI. In three studies, just one therapist delivered the intervention and the interpersonal skills might have led to the changes rather than the intervention. The MBI were clearly described, but in some of the studies, there was no reference to treatment fidelity and compliance.

Another concern regarding the methodological quality is that most of the studies relied on self-report and parental data to evaluate the effectiveness of the intervention. According to Cassady (2001) self-reporting measures might create potential biases including: selective memory, exaggeration and misattribution. Moreover, in the case of children with ADHD this is particularly significant. Boys with ADHD were found to overestimate with positive illusory bias, compared to the teacher related criterion (Hoza, et al., 2002).

Therefore, teachers’ reports might be more reliable than self-reports. Performance based and observational measures (Carboni, 2012; Carboni et al., 2013; Sidhu, 2013; van de Weijer-Bergsma et al., 2012; Worth, 2013) are generally deemed more reliable in assessing the treatment effects. Therefore, the conflicting evidence regarding the effectiveness of the core symptoms of ADHD and the promising results regarding parental stress and family functioning might be treated with caution due to the methodological limitations of the studies.

The strengths of the method used in this review is the use of grey literature, which included five doctoral dissertations. According to a Cochrane Database Systematic Review, published trials tend to display a larger treatment effect than grey trials. This poses significant implications for people that write systematic reviews who ought to make sure that they detect grey trials, in order to reduce the risk of introducing bias (Hopewell, et al, 2007). As the studies included in this systematic review were clinically diverse and there was a heterogeneity of measures, it was deemed difficult to conduct a meta-analysis. Genuine differences in effects may have been concealed (Green & Higgins, 2008). The search criteria included studies published in five languages (English, Italian, French, Spanish and Portuguese); however, after close scrutiny of nine databases, the selected studies were solely in English.

As noted, there were some serious methodological weaknesses in the studies. This implies that the outcomes of the studies should be treated with caution, and that consequently, this review also has several limitations. Due to the subject in question being in its infancy, 10 articles were included in this review. The main pitfall that has been found relates to all the studies having used a small sample. As discussed, this might create problems as the small sample size of the studies may affect their statistical power. This can subsequently influence the probability that a true effect is suggested by a nominally statistically significant finding (Button et al., 2013). Therefore, as the quantitative analysis might not have been adequately powered and due to the lack of randomized control groups, the external validity of the studies might not be reached. These studies used a quasi-experimental and multiple baseline designs, which were critiqued for potentially hindering the objectivity of the causal effects and generalizability of the results.

Due to all the studies being based in Europe and North America with families that were predominantly Caucasian and highly educated, this may also reduce the generalizability of the results to other socio-economic and cultural backgrounds (Sue, 1999). The psychometric properties of the outcome measures of two studies (Shecter, 2013; Singh, et al., 2010) have not been established, which poses questions regarding the accuracy of the results of this evidence.

Another important limitation was that 5 studies have partially relied on self-report measures. These might not be considered as objective measures due to the tendency of adolescents with ADHD to underrate the severity of their symptoms (Hoza, et al., 2002). In most of the studies, the main informants were mothers. Only one study (van de Weijer-Bergsma et al., 2012) relied on paternal and maternal reports. Remarkably, the mothers did not report any significant improvements in any child measure, whilst the fathers reported improvements in every measure. This echoes a meta-analysis where it was found that there was a significant difference between mothers and fathers in reporting both internalizing and externalizing problems in children (Duhig, et al., 2000).

Moreover, according to a study by de Nijs et al. (2004), the between-observer agreement (parents and teachers) on the presence of ADHD symptoms within the same situation at home or school, was low. It was found that one of the causes for informant disagreement on ADHD symptoms was parenting stress (van der Oord, et al., 2006). Therefore, reports regarding the symptomatology of ADHD should be collected by both parents and teachers in home and school environments, respectively, in order to provide a comprehensive picture of the child.

Finally, the quality of the studies was assessed according to a bespoke quality criteria based on the article published by Gersten et al. (2005) and with elements from the Cochrane EPOC checklist (Greenhalgh et al., 2005). In order to enhance the quality and robustness of research findings, the existing evidence ought to address the methodological limitations in terms of: participant sampling, research question and design, fidelity, outcome measures, data analysis, follow-up, attrition and generalizability.

One difficulty for future research on MBI would be to isolate the treatment effects on the core symptoms of ADHD. This is because of the high comorbidity rates and the first line treatment for ADHD being pharmacotherapy. It is important to stress that this line of research is young, and stronger evidence is needed before MBI could be recommended as an alternative to the usual treatment for children/adolescents with ADHD and their parents. Future research might benefit from the following recommendations.

Research design: It is recommended the replication of the studies with larger samples more representative of populations from different socio-economic and cultural backgrounds in order to address validity risks. Moreover, it would be beneficial to apply a random assignment to test intervention effects, including active control conditions.

Participant and intervention characteristics: It would be beneficial to have a more detailed understanding of the participants’ characteristics, particularly in terms of comorbidity and use of medication, which might affect the impact of the MBI. Moreover, in order to have a clear idea of the MBI treatment fidelity, the intervention characteristics should be clearly identified.

Power calculations, varied outcome measures and follow-up: power calculations and pre-post intervention measurements should be delivered. Moreover, most of the evidence assessed in this review relied on data collected by either child/parent or parent/teacher. It would be helpful to integrate multi-informant (including both parents) and multi-method assessments in order to reduce measurement errors. Moreover, the use of validated psychometric measures would be helpful in order to demonstrate a reasonable consistency in results. Moreover, the use of qualitative data might enrich the analysis due to the complexity of the ADHD diagnosis, which the use of solely quantitative data might not capture. It is also recommended for future research to collect follow-up data in order to establish if there are sustained advantages from mindfulness based practices being integrated in lifestyles.

This review highlights that, even though the evidence is conflicting, there is still scope for MBI to be considered as an adjunct to evidence based treatments for ADHD. Behavioral Parent Training has been recommended for children with ADHD by the National Institute for Health and Clinical Excellence (2008). Due to the significant evidence that concludes that MBI have been effective in decreasing parental stress, there is a solid basis for recommending mindfulness to be incorporated in parenting programs. By applying some principles of Mindful Parenting within the context of parenting programs, it is hoped that this would increase child compliance. Taking into account the problems that some children encounter in terms of side effects and being unresponsive to medications, MBI offer techniques that can be introduced included in an individual lifestyle. Children with ADHD may consequently learn some elements of the life-long skills of self-regulation and self-control in the home environment.

In conclusion, there is some evidence that MBI may have the potential to improve the core symptoms of ADHD but research using a more robust methodological criterion is vital before clinical recommendations can be made. However, the evidence on the impact of MBI show positive results in terms of decreasing parental stress and family functioning and could be seen as a potential adjunct to current evidence based treatment.

ADHD is one of the most diagnosed disorders among children and it is crucial to find ways to help this vulnerable group and their parents. This systematic review has found some evidence that MBI may have the potential to improve the core symptoms of ADHD, however there are conflicting results in terms of decreasing the hyperactivity levels. Moreover, the existing evidence demonstrated positive results in terms of decreasing parental stress and family functioning. MBI could be seen as a potential adjunct to current evidence based treatments, including becoming a component of parenting programs. In Positive Parenting Programs (including Triple P), one of the main components is for parents to use their self-care skills. MBI could become that extra adjunct that would allow parents to become less over-reactive and stressed with their children with ADHD.

The current research has several methodological considerations that should be tackled before clinical recommendations can be made. These include: research design that would include randomization with a larger sample more representative of populations from different socio-economic and cultural backgrounds; more detailed understanding of the participants’ characteristics, particularly in terms of co-morbidity and use of medication; power calculations, varied outcome measures and follow-up.