Non-pharmacological interventions for cognitive difficulties in ADHD: A systematic review and meta-analysis

Highlights

• This meta-analysis examined non-pharmacological interventions for ADHD.

• Data were extracted from studies that used objective cognitive measures.

• Physical exercises demonstrated the highest average effect size.

• Findings highlight the positive effect of psychological interventions on ADHD.

Abstract

Attention deficit hyperactivity disorder (ADHD) is the most common neurodevelopmental disorder in children and is associated with significant risk of educational failure, interpersonal problems, mental illness, and delinquency. Despite a number of comparative and comprehensive reviews on the effects of ADHD treatments on ADHD core symptoms, evidence synthesizing the effects of ADHD interventions on cognitive difficulties is limited. In this meta-analysis, the neuropsychological effects of non-pharmacological interventions for ADHD were examined across studies published between 1980 and 2017. Data were extracted from studies that used objective cognitive measures (either computerized or pencil-and-paper), and multiple meta-analyses were conducted to compare the effectiveness across these interventions. Publication bias was assessed, as well as quality of the evidence, using Cochrane risk of bias tool for randomized control trials studies. Our final meta-analysis included 18 studies with interventions that were categorized into four categories: neurofeedback, cognitive-behavioral therapy, cognitive training, and physical exercises. Physical exercises demonstrated the highest average effect size (Morris d = 0.93). A further evaluation of cognitive functions yielded 49 effect sizes for the five categories, including attention, inhibition, flexibility, and working memory. Analyses demonstrated a homogenous, medium to large, effect size of improvement across interventions, with inhibition demonstrating the largest average effect size (Morris d = 0.685). This study highlights the positive effect of psychological interventions on ADHD cognitive symptomology and supports the inclusion of non-pharmacological interventions in conjunction with the commonly used pharmacological treatments.

Introduction

Attention deficit hyperactivity disorder (ADHD) is considered the most common neurodevelopmental disorder in children (American Psychiatric Association, 2013a,b; Goldman et al., 1998; Polanczyk et al., 2007). The diagnosis of ADHD is three times more common in boys than in girls (9.2% vs. 3.0%; (Guevara and Stein, 2001). Although most frequently diagnosed during the school years, ADHD affects individuals across the lifespan (Polanczyk et al., 2007). In the long term, ADHD is associated with a significant risk of educational failure, interpersonal problems, mental illness, and delinquency (Biederman et al., 2006). Given that the global prevalence rate of ADHD is around 5% (Polanczyk et al., 2007; Wittchen et al., 2011), it creates a substantial burden on families, as well as on health, social care, and criminal justice systems (W. E. Pelham, Foster and Robb, 2007). Etiology of ADHD is complex and multidimensional, combining genetic, psychosocial, and environmental factors (Faraone, 2000; Faraone and Biederman, 2005).

Traditionally, the primary neurocognitive deficits of ADHD were considered to be impulsivity, hyperactivity, and inattention (Halperin et al., 1990). This has been supported by several theoretical models (Willcutt, 2015), hypothesizing that ADHD may arise from dysfunctional responses to reward/punishment contingencies (Luman et al., 2005), pronounced aversion to the experience of delay (E. J. S. Sonuga-Barke et al., 1992), increased intraindividual variability in response time due to attentional fluctuation (Sergeant et al., 2003), and overall slow cognitive processing speed (McGrath et al., 2011). However, several authors have focused on a primary deficit in the executive function (EF) domain, as patients tend to display deficits in attentional and strategic flexibility, often fail to monitor and inhibit behavior effectively, and often display lower planning abilities and working memory (Alloway and Passolunghi, 2011; Barkley, 1997; Biederman et al., 2008; Brocki et al., 2008; Castellanos and Tannock, 2002; Pfiffner et al., 2018). Executive functions are a collection of higher level capacities that enable flexibility in goal-directed behavior (Welsh and Pennington, 1988), including working memory, response inhibition, and set shifting (Miyake et al., 2000). Anatomically, executive functions have been primarily localized to the prefrontal cortex and secondarily to the temporal, parietal, and limbic lobes and the striatum (Roth and Saykin, 2004).

Epidemiologic studies report that as much as 50% of referrals to child mental health clinics are for assessment and treatment of ADHD (Salomone et al., 2015). In response to this demand, a plethora of measures have been designed to assess aspects of the triad of symptoms comprising ADHD (i.e., inattention, hyperactivity, and impulsivity; see Table 1). Most commonly, these symptoms are assessed using subjective measures aimed at the behavioral manifestations of the disorder, involving different procedures across stages of development, such as parent and teacher rating scales in children (Barkley, 1991) and self-report scales in adults (Kessler et al., 2005). It is generally accepted that parent and teacher rating scales are reliable and valid components of ADHD assessments (McGough and Barkley, 2004). Less commonly used (Dulcan, 1998) but rather more objective are laboratory measures, which measure the cognitive manifestations of the disorder and purport to distinguish between clinical and non-clinical cases (Barkley, 1991; Forbes, 1998; Greenberg and Waldmant, 1993; Gualtieri and Johnson, 2005). Despite the availability of a range of computerized tests, most children are diagnosed based on clinician-determined behavioral observations from parents or teachers, which could potentially lead to inaccurate diagnosis. This may be due to the lack of clarity in behavioral descriptions, the accessibility of a recognized medicinal treatment upon diagnosis, and the overarching label of ADHD, which can be used to encompass aggression, irritability, and learning difficulties, not readily fitting into other diagnostic categories (Batstra et al., 2014). On the other hand, laboratory measures, such as the various continuous performance tasks, have been argued to have low ecological validity and suffer from poor sensitivity and specificity in diagnosing ADHD (Berger et al., 2017).

Upon diagnosis, psychostimulant medications are generally the first line of treatment for ADHD (Pliszka, 2007; Wolraich et al., 2011) acting as a moderator, most often resulting in substantial symptom reduction (Cortese et al., 2018), although parents can be reticent to give their children medication due to fears over stigma, side-effects, or the long-term effects of treatment (Muris et al., 2018). Side effects may include insomnia, lack of appetite, and headache (Newcorn et al., 2010), and there are few studies on the effects of prolonged use of ADHD medications (Kociancic et al., 2004). It has recently been reported that there were no long-term cognitive benefits of medication(e.g., reaction time or verbal working memory) after six years when participants were unmedicated at the time of testing (Schweren et al., 2018). However, greater improvements were observed when behavioral interventions, such as parent management training, were used prior to medication (William E. Pelham et al., 2016).

Although psychiatric guidelines note that treatment by medication alone is often not sufficient and recommend various social skills training and behavioral interventions, there are few guidelines as to what these non-pharmaceutical interventions should comprise (Sadock et al., 2009; Young and Myanthi Amarasinghe, 2010). Parenting books contain suggestions and guidelines for cognitive training and parenting skills (Grohol, 2018), psychiatric manuals focus on the range of neuropsychiatric medications available (American Psychiatric Association, 2013a,b; Sadock et al., 2009), and psychological guidelines emphasize a multi-systematic approach combining psychostimulant medication with parenting, school, and child psychoeducation to promote rule following and self-regulation skills (Carr, 2015). Currently there are no clear guidelines to differentiate among available interventions and inform clinicians who choose to recommend non-pharmacological treatments.

There are various non-pharmacological treatments that have been used alongside or instead of pharmacological treatments (Sharma et al., 2015). Dietary interventions have been shown to have some small improvement in ADHD symptomatology (Edmund J.S. Sonuga-Barke et al., 2013), with Omega 3 demonstrating small improvements in the emotional lability and oppositional behavior associated with ADHD (Cooper et al., 2016). Meditation-based interventions, such as mindfulness and yoga, are widely practiced, in order to improve both mental and physical health. However, there is substantial variability in the methodology of the interventions and which ADHD deficits they are design to target, and as a result no definitive conclusions have been made regarding the efficiency of such treatments (Evans et al., 2018). School-based and summer program interventions (including cognitive behavioral therapy [CBT], contingency management, and academic interventions) are considered “first-line” treatment approaches and have been demonstrated to improve academic performance (DuPaul and Eckert, 1997), although interventions, such as parental training (Lee et al., 2012) and music therapy (Maloy and Peterson, 2015), have demonstrated little long-term effects. There appears to be some support for neuro-feedback interventions in improving cognitive and self-control symptoms associated with ADHD (Arns et al., 2009; Hodgson et al., 2014; Van Doren et al., 2019), as well as for non-invasive brain stimulations (such as tDCS, and TMS) in improving cognition and clinical symptoms of inattention and impulsivity (Westwood et al., 2019). Although fewer in number, there have also been a number of studies demonstrating similar improvements following physical exercise (Cerrillo-Urbina et al., 2015; Neudecker et al., 2019), CBT, and cognitive training programs (Knouse et al., 2017).

Several systematic reviews and meta-analyses have discussed the effects of non-pharmacological interventions in ADHD (e.g., Arns et al., 2009; Fabiano et al., 2009). However, interpreting these reports, specifically in relation to the impact on core ADHD symptoms, is complicated by the inclusion of trials using nonrandomized designs, non-ADHD samples, and predominately the lack of objective cognitive measures. Furthermore, estimates of efficacy are often based on assessments made by individuals who are likely to be aware of study allocation, which may inflate effect sizes (Jadad and Enkin, 2007). The current study addresses these limitations and endeavors to lower the variability across the field by limiting the methodology to those studies using objective cognitive measures. While recognizing the importance of other outcomes (e.g., oppositional symptoms) as treatment targets for patients with ADHD, analyses of such measures were not viable in this study because of the variance and subjective nature of methodologies. The present analyses aimed to address two major research questions: (1) which leading non-pharmaceutical intervention for ADHD's cognitive symptomology is most effective? and (2) which cognitive symptoms are most amenable to change? To answer these questions, multiple meta-analyses were performed, using data retrieved from the relevant studies. Regarding the first question, we hypothesized there might not be any major preference for one intervention over the other as all interventions share a common therapeutic effect resulting in cognitive improvement. This effect might be stronger and larger than any effect of a specific intervention. As for the second question, we hypothesized that, as the cognitive symptoms are more complex, involving increasing number of cognitive process, the amenability to change should increase.

Given the range of research on single treatments and cognitive factors, studies were divided into four categories of behavioral and psychological treatments on the five key cognitive functions known to be most negatively affected in individuals with ADHD. As the amount of intervention papers measuring cognitive functions was limited, we decided to include a wider participant age range (children, adolescents, and adults), taking into count the possible variability of results. The four intervention domains included neurofeedback, using the visualization of brain activity to teach patients to increase attention and impulse control (Arns and Strehl, 2013); physical activity, including a range of aerobic exercise forms (Cerrillo-Urbina et al., 2015); cognitive training, employing adaptive schedules that are hypothesized to strengthen ADHD-deficient neuropsychological processes (e.g., working memory) (Markomichali et al., 2009); and CBT, employing skills incorporated into CBT approaches for ADHD, ranging from organization, planning, and time management skills to cognitive reappraisal strategies and mindfulness meditation skills (Lee et al., 2012). Although behavioral interventions (i.e., behavioral parent training and classroom behavioral interventions) have demonstrated some efficacy in treating ADHD, they were not included in this meta-analysis, as none of these studies included cognitive tasks as outcome measures. Furthermore, analyses covered five measures of cognitive domains: mental flexibility, the ability to switch between and hold multiple concepts (Scott, 1962); inhibition, the ability to tune out and avoid stimuli or actions that are irrelevant to the task/process at hand and override a strong internal predisposition (Macleod, 2007); attention, the behavioral and cognitive process of selectively concentrating on a discrete aspect of information (Anderson, 2005); and working memory, the function responsible for temporarily holding information available for processing (Miyake and Shah, 1999). We also created a fifth category of higher executive functions (HE) to incorporate the most complex cognitive functions of planning and reasoning (Diamond, 2013).