Courses for Mental Health Professionals
Continuing Education Courses on the Internet
Home Courses Help Search

Treating Children and Adolescents with ADHD: An Overview of Empirically Based Treatments
by Russell A. Barkley, Ph.D., ABPP

6 CE Hours - $89

Last revised: 03/19/2015

Course content © copyright 2011-2015 by Russell A. Barkley, Ph.D., ABPP All rights reserved.

ContinuingEdCourses.Net dba, provider #1107, is approved to offer social work continuing education by the Association of Social Work Boards (ASWB) Approved Continuing Education (ACE) program. Organizations, not individual courses, are approved as ACE providers. State and provincial regulatory boards have the final authority to determine whether an individual course may be accepted for continuing education credit. ContinuingEdCourses.Net maintains responsibility for this course. ACE provider approval period: 3/9/2015-3/9/2021.

ContinuingEdCourses.Net dba has been approved by the National Board for Certified Counselors (NBCC) as an Approved Continuing Education Provider (ACEP), ACEP #6323. Programs that do not qualify for NBCC credit are clearly identified. ContinuingEdCourses.Net is solely responsible for all aspects of the programs.

ContinuingEdCourses.Net dba is recognized by the New York State Education Department's State Board for Social Work (NYSED-SBSW) as an approved provider of continuing education for licensed social workers #SW-0561.

Take the Course Take the Test Buy your Certificate


Learning Objectives

This is an intermediate to advanced level course. It is highly recommended that those wishing to take this course first take the course ADHD: Nature, Course, Outcomes, and Comorbidity. After completing this course, mental health professionals will be able to:

This course is adapted from the chapter by myself, Bradley Smith and Cheri Shapiro from my textbook, Mash, E. J. & Barkley, R. A. (2006) Treatment of Childhood Disorders (3rd ed.). New York: Guilford.

The materials in this course are based on the most accurate information available to the author at the time of writing. The scientific literature on ADHD grows daily, and new information may emerge that supersedes these course materials. This course will equip clinicians to have a basic understanding of the treatments for ADHD in children and adolescents.


Over the past century, numerous diagnostic labels have been given to clinically referred children having significant deficiencies in behavioral inhibition, sustained attention, resistance to distraction, and the regulation of activity level. Most recently, Attention Deficit-Hyperactivity Disorder (ADHD) (DSM-5, American Psychiatric Association, 2013) is the term used to capture this developmental disorder. Previously employed terms have been brain-injured child syndrome, hyperkinesis, hyperactive child syndrome, minimal brain dysfunction, and Attention Deficit Disorder (with or without Hyperactivity). Such relabeling every decade or so reflects a shifting emphasis in the primacy accorded certain symptom clusters within the disorder based, in part, on the substantial research conducted each year on ADHD and how investigators and theorists interpret those findings.

This course provides a critical overview of the treatments that have some efficacy for the management of ADHD as shown through scientific research. That literature is voluminous, however, and so space here permits only a brief discussion and critique of each of the major treatments. More detailed discussions of these treatments can be found in the texts by Barkley (2006) and DuPaul and Stoner (2003). In addition to this course, perspectives on state-of-the art treatment for ADHD are found in published practice guidelines from the American Academy of Child and Adolescent Psychiatry (McClellan & Werry, 2003), the American Academy of Pediatrics (2010), the American Medical Association (Goldman, 1998), and the task force on empirically validated treatments commissioned by the American Psychological Association (Pelham, Wheeler, & Chronis, 1998). My perspective on diagnosis and treatment is highly congruent with these authoritative sources, but adds the important considerations of practicality and safety of treatments to guide overall intervention planning for children and adolescents with ADHD.

This course begins with a brief overview of the nature of the disorder, its prevalence, developmental course, and etiologies. A more detailed examination of these topics can be found in the companion course on this website on the nature of ADHD in children and adolescents and in the text by Barkley (2006). Subsequently, the main purpose of this course is addressed through a critical overview of various treatments for the disorder. Minimal information on the assessment of ADHD will be provided here due to both space limitations and the availability of more detailed information on this topic to be found in the companion course, ADHD in Children: Diagnosis and Assessment.

Recent Developments

Over the past two decades there have been several important developments related to the treatment of ADHD. These developments include (a) consensus statements from major professional and scientific organizations about appropriate treatment of ADHD, (b) proliferation in the variety of medications used to treat ADHD, (c) some major additions to the research literature on treatment of adolescents with ADHD, (d) unique insights about treatment from the Multimodal Treatment Study of ADHD (The MTA Cooperative Group, 1999a, 1999b, 2004a, 2004b) and (e) increased attention to the alternative approaches to treatment development and evaluation that emphasize treatment effectiveness as opposed to just treatment efficacy. The current course places equal or greater emphasis on the practicality of treatment.

Practicality as we use the term is distinct from efficacy and effectiveness. Efficacy refers to demonstrated treatment success in controlled research studies often taking place in clinical laboratories or university settings and typically designed, supervised, and executed by experts and their students and staff. Effectiveness refers to demonstrated treatment success in naturalistic settings more typical of those clinics, hospitals, and private practices in which patients are likely to seek treatment involving controlled scientific studies supervised or instituted by the clinicians typically practicing in those settings. Practicality here refers to the ease or convenience of carrying out the treatment and hence the likelihood of its adoption in ordinary clinical practice. Classroom management or intensive all day summer treatment approaches to ADHD may be both efficacious and effective yet may not necessarily be adopted in typical classroom or community settings if the labor they require for implementation is too great or the financial cost of doing so is disproportionate within the total school or community budget. This focus on practical considerations stems from a growing concern that empirically supported treatments are being under-utilized, not only with children with ADHD but in all of children’s mental health. The consequence of this situation is that there is a two-tiered pattern of outcomes in the extant research literature. Specifically, meta-analytic reviews of the literature have found substantial evidence of beneficial effects of interventions studied in research settings but little or no evidence of beneficial effects of interventions that were delivered in typical clinical settings (Weisz, 2004).

The two-tiered system of outcomes seems to persist despite some vigorous and well-planned efforts to disseminate the empirically-validated treatments (e.g., Henggeler, Melton, Brondino, Scherer, & Hanley, 1997). The bad news is that, for a variety of reasons, clinicians in naturalistic settings did not implement empirically validated treatments with sufficient fidelity to get effective outcomes (Biglan, Mrazek, Carnine, & Flay, 2003). The good news is that several studies have found that fidelity of treatment implementation is positively correlated with beneficial outcomes (Biglan et al, 2003). Therefore, efforts to overcome barriers to effective implementation and promote treatment fidelity should result in better treatment outcomes. Consequently, this course pays close attention to barriers to dissemination of empirically supported treatments for ADHD, or the issue we call “practicality” of treatment.

To help summarize issues of effectiveness, safety, and practicality and to facilitate comparisons between treatments, we use a grading system for each of these three considerations. Our grades for empirical support of effectiveness are very similar to the ranking system proposed by Biglan et al. (2003), which is a refinement of a consensus system adopted by the Task Force on Empirically Validated Treatments by the American Psychological Association. Specifically, in this course a treatment received a grade of F if claims of effectiveness were based solely on clinical experience or weak quasi-experimental designs (e.g., single group pre- to post-test studies). The grade of D was given to treatments where there was evidence of effectiveness from one well-designed, randomized clinical trial, or an interrupted time-series that was replicated across at least three cases. The grade of C was given when there was replication of well-designed, randomized clinical trials or interrupted time-series studies, but the replication was by the same group of researchers. Treatments received a grade of B when there was replication of well-designed, randomized clinical trials or interrupted time-series studies by independent teams of researchers. Finally, a grade of A was assigned when the efficacy criteria to earn a B were met, plus there was evidence of effectiveness when the intervention is implemented in its intended setting.

Primary Symptoms of ADHD

Children having ADHD, by definition, display difficulties with attention and/or impulse control relative to normal children of the same age and sex (American Psychiatric Association, 2013; Barkley, Cook, et al., 2001). Attention is a multi-dimensional construct which can refer to problems with alertness, arousal, selective or focused attention, sustained attention, distractibility, or span of apprehension, among others (Barkley, 1988; Mirsky, 1996). Research to date suggests that, among these elements, children with ADHD most likely have their greatest difficulties with sustaining attention to tasks (persistence of responding), resisting distractions, and re-engagement of initial tasks once disrupted (Barkley, 1997c, 1997d; Douglas, 1983). These difficulties are sometimes apparent in free play settings but are much more evident in situations requiring sustained attention to dull, boring, and/or repetitive tasks (Danforth, Barkley, & Stokes, 1991; Luk, 1985). However, even when presented with apparently intrinsically interesting stimuli, such as a television program, attractive distracting stimuli, such as toys, may significantly impair attention to and comprehension of events relative to children who do not have ADHD (Lorch, Sanchez, van den Broek, Baer & Hooks, 2000). More recently, research is suggesting that the attention problems evident in ADHD are part of a larger domain of cognitive activities known as executive functioning, and especially working memory (i.e., holding information in mind that is being used to guide performance) (Barkley, 1997c, 1997d; Frazier et al., 2004; Willcutt et al., 2005). Specifically, evidence from the development of rating scales indicates that the DSM items used to define the attention deficits in ADHD load on a larger dimension containing items reflecting executive functioning and specifically working memory (Barkley, 2011a, 2011b, 2012a; Conners, 1998; Gioia, Isquith, Guy, & Kenworthy, 2000).

Often coupled with this difficulty in sustained attention is a deficiency in inhibiting behavior, or impulsiveness. Like attention, impulsiveness is also multidimensional in nature (Nigg, 1999, 2000, 2001). Rather than cognitive reflectiveness the deficit in ADHD is mainly in the capacity to inhibit or delay prepotent responses, particularly in settings in which those responses compete with rules (Barkley, 1997d). A prepotent response is that which would gain the immediate reinforcement (reward or escape) available in a given context or which has a strong history of such reinforcement in the past. Those with ADHD have difficulties with sustained inhibition of such dominant responses over time (Nigg, 1999), as well as with poor delay of gratification (Rapport, Tucker, DuPaul, Merlo, & Stoner, 1986), a steeper discounting of the value of delayed over immediate rewards (Barkley, Edwards, Lanieri, Fletcher, & Metevia, 2001b), and impaired adherence to commands to inhibit behavior in social contexts (Danforth et al., 1991). This inhibitory deficit may also include a difficulty with interrupting an already ongoing response pattern (Schachar, Tannock, & Logan, 1993), particularly when given feedback about performance and errors. In the latter case, perseverative responding may be evident despite negative feedback concerning such responding, which may be reflecting an insensitivity to errors (Sergeant & van der Meere, 1988). Overall, individuals with ADHD have poorer inhibitory control and slower inhibitory processing than normal controls (Nigg, 2001).

Numerous studies have shown that children with ADHD tend to be more active, restless, and fidgety than normal children (Porrino, Rapoport, Behar, Sceery, Ismond, & Bunney, 1983; Teicher, Ito, Glod, & Barber, 1996). As with the other symptoms, there are significant situational fluctuations in this symptom (Luk, 1985; Porrino et al., 1983). It has not always been shown convincingly that hyperactivity distinguishes ADHD from other clinic-referred groups of children (Werry, Elkind, & Reeves, 1987; Werry, Reeves, & Elkind, 1987). It may be the pervasiveness of the hyperactivity across settings that separates ADHD from other diagnostic categories of children (Taylor, 1986).

ADHD symptoms of hyperactivity have been shown to decline significantly across the elementary school years while problems with attention persist at relatively stable levels during this same period of development in children with ADHD (Barkley, 2012b; Hart, Lahey, Loeber, Applegate, & Frick, 1995). One explanation that may account for such a state of affairs is that the hyperactivity reflects an early developmental manifestation of a more central deficit in behavioral inhibition. Studies which factor analyze behavior ratings certainly show that hyperactivity and poor impulse control form a single dimension of behavior, as shown in the DSM-5 (Achenbach, 2001; DuPaul, 1991). This deficit in inhibition, of which early hyperactivity is a part, may become increasingly reflected in poor self-regulation over various developmental stages, even though the difficulties with excessive activity level may wane with maturation.

Difficulties with adherence to rules and instructions are also evident in children with ADHD (American Psychiatric Association, 2013; Barkley, 1997d). Care is taken here to exclude poor rule-governed behavior that may stem from sensory handicaps (i.e., deafness), impaired language development, or defiance or oppositional behavior. Nevertheless, children with ADHD typically show significant problems with compliance with parental and teacher commands (Danforth et al., 1991), following experimental instructions in the absence of the experimenter (Draeger, Prior, & Sanson, 1986), and with adhering to directives to defer gratification or resist temptations (Rapport et al, 1986). Like the other symptoms, rule-governed behavior is a multi-dimensional construct (Hayes, 1987). It remains to be shown which aspects of this construct are specifically impaired in ADHD.

Diagnostic Criteria

Between 1980 and the present, efforts have been made to develop more specific guidelines for the classification of children as ADHD. These efforts have been based increasingly on an empirical approach to developing a taxonomy of child psychopathology. Although guidelines appeared in the DSM-II (American Psychiatric Association, 1968), these comprised merely a single sentence along with the admonition not to grant the diagnosis if demonstrable brain injury were present. A more concerted effort at developing criteria appeared in the DSM-III (American Psychiatric Association, 1980), though still unempirical. These criteria were not examined in any field trial but were developed primarily from expert opinion. In the next revision of the DSM-III (American Psychiatric Association, 1987), an attempt was made to draw upon the results of factor analytic studies of child behavior rating scales to aid the selection of symptoms that might be included for ADHD (Spitzer, Davies, & Barkley, 1990). A small-scale field trial employing 500 children from multiple clinical sites was conducted to narrow down the potential list of symptoms, and a cutoff score on this list was chosen that best differentiated children with ADHD from other diagnostic groups.

In DSM-IV (American Psychiatric Association, 2000), the criteria were based on a better field trial and more thorough analysis of its results (Applegate et al., 1997; Lahey, Applegate, McBurnett, Biederman, Greenhill et al., 1994). Despite the increasingly empirical foundation of the DSM, there remained a few problems with these criteria having to do with developmental sensitivity to the disorder, possible gender discrimination in diagnosis, an empirically unjustified age of onset of 7-years-old, and a vexing requirement for cross-setting impairment that is confounded with the problem of poor parent-teacher agreement (Barkley, 2005). Also problematic for the generality of these criteria is the fact that the field trial used primarily male children, ages 4 to 16-years-old, who were largely of Caucasian background. Consequently, adjustments or allowances must be made when one wishes to apply the DSM criteria to females, young adults, and non-Caucasian ethnic groups. For example, the symptom of “often leaves their seat” may be less relevant to adults with ADHD. Moreover, children whose onset of symptoms was sometime during the childhood years (prior to 13) should be considered as having a valid disorder rather than adhering strictly to the DSM-IV age of onset of 7-years-old as the demarcation of a valid case of disorder (Barkley & Biederman, 1997).

Where sex differences exist, they indicate that girls with ADHD show less severe symptoms of both inattention and hyperactive-impulsive behavior, especially in school, fewer symptoms of ODD and CD, greater intellectual deficits and more symptoms of anxiety and depression than do ADHD boys (Abikoff et al., 2002; Gershon, 2002; Hartung et al., 2002). Some recent studies indicate that girls with ADHD may employ more relational aggression than non-ADHD peers (Zalecki & Hinshaw, 2004).

The publication of the DSM-5 in 2013 brings with it a few changes to the DSM-IV criteria.  While the original 18 symptoms from the DSM-IV remain in use as do the two dimensional lists for their presentation, they are followed by clarifications in parentheses to guide clinicians in applying that symptom to teens and adults.  The threshold of 6 of 9 symptoms on either of the two symptom-lists remains for application to children and teens.  But for adults, this threshold is reduced to 5 symptoms.  The requirement that symptoms occur often or more frequently, that they be developmentally inappropriate, and that they have persisted for at least the prior 6 months in DSM-IV are all carried forward to DSM-5.  The age of onset has been adjusted upward from 7 to 12-years-old, consistent with evidence that the lower onset was invalid (Barkley & Biederman, 1997), too restrictive (Barkley et al., 2008), and contributed to unreliability (Applegate et al., 1997).  Also remaining is the requirement for symptoms to be impairing across several settings and that there be impairment in major life activities.  The subtypes have been removed respecting abundant evidence that ADHD is a single disorder that varies in severity in the human population and that this is not changed by the fact that some people have more inattention than hyperactive-impulsive symptoms and vice-versa.  Yet clinicians will be provided the opportunity to specify which symptom dimension may be more predominant, as in ADHD Predominantly Inattentive Presentation.  There is a Predominantly Hyperactive-Impulsive Presentation and then the traditional Combined Presentation.  Finally, ADHD can now be diagnosed comorbidly with autistic spectrum disorders, just as it can with many other disorders; a comorbidity previously precluded in DSM-IV.

Prevalence and Sex Ratio

The The  prevalence of ADHD as defined by the DSM-IV has been found across studies to be 5-7 percent of school-age children as probably having ADHD (Breton, et al., 1999; Briggs-Gowan, Horwitz, Schwab-Stone, Leventhal, & Leaf, 2000; Polancyk et al., 2007; Willcutt, 2012). Lower rates result from rigid adherence to DSM criteria, earlier versions of the DSM in which the inattentive type did not exist, and studies that relied exclusively on parent reports [2-6%](Breton et al., 1999; Willcutt, 2012). Sex and age differences in prevalence are routinely found in research. For instance, prevalence rates may be 4% in girls and 8% in boys in the preschool age group (Gadow, Sprafkin, & Nolan, 2001) yet fall to 2% to 4% in girls and 6% to 9% in boys during the 6- to 12-year-old age period based on parent reports (Breton et al., 1999; Szatmari, Offord & Boyle, 1989). The prevalence will decrease again to 0.9 to 2% in girls and 1% to 5.6% in boys by adolescence (Breton et al., 1999; Lewinsohn, Hops, Roberts, Seeley, & Andrews, 1993; Romano, Tremblay, Vitaro, Zoccolillo, & Pagani, 2001). Importantly, if as specified in the DSM both a symptom threshold and the requirement for impairment are used, the prevalence may decrease by 20-60% from that figure based on symptom thresholds alone (Breton et al., 1999; Romano et al., 2001; Wolraich, Hannah, Baumgaertel, & Feurer, 1998).

In large part, differences in prevalence rates across studies are due to different methods of selecting samples, differences in the nature of the populations themselves (urban vs. rural, etc.), differing definitions for the disorder, and certainly to the variation in ages of the samples (Willcutt, 2012). Males are approximately three times more likely than females to have ADHD. The disorder decreases in both sexes across development. Despite all the factors affecting the prevalence of the disorder, when these are taken into account ADHD is seen to be a universal disorder occurring at similar prevalence rates across countries worldwide (Willcutt, 2012).  Thus, there is a clear consensus that ADHD is a valid, diagnosable disorder (Barkley, Cook et al., 2001).

Onset, Course, and Outcome

Studies of the developmental course and outcome of children with ADHD have been numerous and can only be summarized briefly here (see Barkley, 2005; Weiss & Hechtman, 1993 for reviews). Although some children with ADHD are reported to have been difficult in their temperament since birth or early infancy (Barkley, DuPaul, & McMurray, 1990a; Ross & Ross, 1982), the majority appear to be identifiable as deviant from normal by their caregivers between 3 and 4 years of age (Barkley, Fischer, Newby, & Breen, 1987; Loeber, Green, Lahey, Christ, & Frick, 1992). However, it may be several years later before such children are brought to the attention of professionals. Although the diagnosis of ADHD among preschoolers may be more difficult due to higher rates of disruptive behavior among the normal population at this age, a few recent studies suggest that reliable and valid diagnosis can be made for children as young as 3 years, 7 months old (Hartung et al., 2002).

During their preschool years, children with ADHD are often excessively active, mischievous, noncompliant with parental requests, and difficult to toilet train (Campbell, 1990; Hartsough & Lambert, 1985; Mash & Johnston, 1982). They may also already be manifesting some delays in academic readiness skills (Mariani & Barkley, 1997). Parental distress over child care and management is likely to reach its zenith between 3 and 6 years of age, declining thereafter as the deficits in attention and rule-following improve (Barkley, Karlsson, & Pollard, 1985; Mash & Johnston, 1983). Yet, even into the elementary school years, the stress parents report in raising children with ADHD remains considerably higher than that for parents of children in control groups (Anastopoulos, Guevremont, Shelton, & DuPaul, 1992; Bussing, Gary, Mason, Leon, Sinha, & Garvan, 2003; Fischer, 1990). Likewise, parents of teenagers with ADHD report high levels of stress and family conflict with these youth, particularly if the youth also carries a comorbid diagnosis of oppositional defiant disorder (ODD; Barkley, Anastopoulos, Guevremont, & Fletcher, 1992; Barkley, Fischer, Edelbrock, & Smallish, 1991; Edwards, Barkley, Laneri, Fletcher, & Metevia, 2001).

By entry into formal schooling (6 years of age), most children with ADHD have become recognizably deviant from normal peers in their poor sustained attention, impulsivity, and restlessness. Difficulties with aggression, defiance, or oppositional behavior may now have emerged, if they had not earlier in development (Barkley, Fischer et al., 1990, 1991). Children with ADHD developing these conduct problems or antisocial behaviors are likely to veer into a more severe path of maladjustment in later years than are those children with ADHD who do not develop aggressive/defiant behaviors or do so only to a limited degree (Barkley, Fischer, Smallish, & Fletcher, 2004). During these elementary school years, the vast majority of children with ADHD have varying degrees of poor school performance usually related to failure to finish assigned tasks in school or as homework, disruptive behavior during class activities, and poor peer relations with schoolmates. Learning disabilities in areas of reading, spelling, math, handwriting, and language, however, may also become manifest in a significant minority of children with ADHD, requiring additional special educational assistance beyond that typically needed to manage the ADHD symptoms (Barkley, 2005; Tannock, 2000).

As teenagers, a small percentage of children with ADHD will have "outgrown" their symptoms in that they now place within the broadly defined normal range in their symptom deviance. However, those who have subclinical levels of ADHD as adolescents or young adults may still be significantly impaired and might be considered to have ADHD if more developmentally appropriate diagnostic criteria were employed. Moreover, even when DSM criteria are used, the odds generally favor the continuation of the disorder as studies have estimated that 43-80 percent continue to have the disorder into adolescence (Barkley, Fischer, Edelbrock, & Smallish, 1990; Biederman, Faraone, Milberger, Guite, Mick, et al., 1996; Mannuzza, Klein, Bonagura, Malloy, Giampino, & Addalli, 1998; Weiss & Hechtman, 1993).

For adolescents with ADHD, family conflicts may continue or even increase (Barkley, Anastopoulos, et al., 1992; Fletcher, Fischer, Barkley, & Smallish, 1996) and may now center around the failure of the ADHD teen to accept responsibility for performing routine tasks, difficulties with being trusted to obey rules when away from home, and generally in the problem-solving approaches which parents and adolescents with ADHD attempt to use in resolving conflicts (e.g., authoritarian, highly emotional, excessive use of ultimatums, etc.) (Robin, 1998). Among that subset of ADHD teens who have had significant, earlier problems with aggressive and oppositional behavior, delinquency and conduct disorder are more likely to emerge, if they have not done so already, as these adolescents spend greater amounts of unsupervised time in the community (Barkley, Fischer et al., 1990; Mannuzza et al., 1998; Satterfield, Swanson, Schell, & Lee, 1994; Weiss & Hechtman, 1993). Greater-than normal substance experimentation and abuse is likely to occur within the adolescent years, mainly among youths diagnosed with ADHD and comorbid conduct disorder (Barkley, Fischer et al., 1990; Barkley, Fischer, et al., 2004; Thompson, Riggs, Mikulich, & Crowley, 1996) or bipolar disorder (Biederman et al., 1996). An increasing number of studies have replicated and extended the original report of Weiss and Hechtman (1993) suggesting that both ADHD children followed into adolescence and clinically-referred teenagers (and adults) with ADHD have a greater number of automobile accidents and speeding citations than normal teens (see Barkley, 2004a for a review; Fischer, Barkley, Smallish, & Fletcher, 2004). Research also suggests that up to 32 percent may fail to complete high school, and most fail to pursue college programs after high school (Barkley, Fischer, Smallish, & Fletcher, 2004b; Weiss & Hechtman, 1993). A greater risk of teenage pregnancy has also been noted in a recent follow-up study (Barkley et al., 2004b; Barkley et al., 2008). Certainly, the outcome of childhood ADHD in the adolescent years is far more negative than previous clinical lore had postulated.

Less research exists on the adult outcome of children with ADHD. What does exist suggests that from 8 to 66 percent may continue to have the disorder or are continuing to have symptoms of ADHD that significantly affect their life into early adulthood (Barkley, Fischer, Smallish, & Fletcher, 2002; Barkley et al., 2008) with up to 40% or more meeting DSM diagnostic criteria even 30 or more years after childhood diagnosis (Klein et al., 2012). More recent studies using more modern diagnostic criteria for ADHD appear to demonstrate higher rates of persistence of the disorder than do earlier follow-up studies. Interpersonal problems continue to plague as many as 75 percent (Weiss & Hechtman, 1993). Juvenile convictions and symptoms of adult antisocial personality may occur in 23 to 45 percent (Barkley, Fischer, Smallish & Fletcher, 2004a; Fischer, Barkley, Smallish, & Fletcher, 2004), while 20-27 percent or more may have substance use disorders (Barkley et al., 2008; Fischer et al., 2002.

Related Characteristics

Children with ADHD have a higher likelihood of having other medical, developmental, adaptive, behavioral, emotional, and academic difficulties than do peers who do not have ADHD (Barkley, 2006). Delays in intelligence, academic achievement, and motor coordination are more prevalent in children with ADHD than in matched samples of normal children or even in siblings (Barkley 2005), as are delays in adaptive functioning more generally (Greene, et al., 1996; Roizen, Blondis, Irwin, & Stein, 1994; Stein, Szumowski, Blondis, & Roizen, 1995). Problems with peer acceptance and in peer interactions are commonly documented in children with ADHD (Ernhardt & Hinshaw, 1994; Bagwell, Molina, Pelham, & Hoza, 2001; Hoza, 2007; Stroes, Alberts, & van der Meere, 2003).

As many as 87 percent of clinically diagnosed ADHD children may have at least one other disorder and 67% have at least two other disorders (Kadesjo & Gillberg, 2001). As noted earlier, children with ADHD are more likely to have co-existing oppositional defiant and conduct disorder symptoms than normal children (Angold, Costello, & Erkanli, 1999). Depression and juvenile onset bipolar disorder also appear to be more common in children with ADHD than would be expected in the general population (Biederman, Faraone, Mick, Moore, & Lelon, 1996; Jensen, Shervette, Xenakis, & Richters, 1993), especially where conduct disorder is present with ADHD (Angold et al., 1999). There is a modest increase in risk for anxiety disorder as well (Angold et al., 1999; Tannock, 2010). The severity of the ADHD symptoms may in part predict the severity of and risk for these comorbid conditions (Gabel, Schmidtz, & Fulker, 1996).

Children with ADHD appear to have more minor physical anomalies than normal children (Quinn & Rapoport, 1974) and may be smaller physically than normal children, at least during childhood (Spencer, et al.1996). They may also have more sleep difficulties than normal children (Ball & Koloian, 1995). However, prior beliefs that ADHD may have a higher than normal association with either allergies or asthma have not been corroborated by research (Biederman, Milberger, Faraone, Guite, & Warburton, 1994; McGee, Stanton, & Sears, 1993).

Research on the family interactions of children with ADHD suggests that their symptoms produce significant alterations in family functioning, particularly in those children who are also manifesting problems with oppositional and defiant behavior (Johnston & Mash, 2001). Children with ADHD have been shown to be less compliant, more negative, and less able to sustain compliance than normal children during task completion with their mothers (Danforth et al., 1991 for a review). Their mothers are more directive and negative, more lax in their discipline, less rewarding and responsive to their children's behavior, and show lower levels of maternal coping than mothers of normal children (Cunningham & Boyle, 2002; Keown & Woodward, 2002; McKee, Harvey, Danforth, Ulaszek, & Friedman, 2004), with these problems being more closely aligned with level of child conduct problems than severity of ADHD symptoms (Johnston, Murray, Hinshaw, Pelham, & Hoza, 2002; Kashdan, Jacob, Pelham, Lang, Hoza, Blumenthal, & Gnagy, 2004). There appears to be less conflict in the task-related interactions of older children with ADHD compared to younger age groups (Danforth et al., 1991). However, older children with ADHD remain deviant from same-age children in their noncompliance and parent-child conflicts, even into adolescence (Barkley et al., 1991; Edwards et al., 2001; Fletcher et al., 1996). Not surprisingly, these interaction problems are significantly greater in those ADHD teens having oppositional defiant disorder than in those without this comorbid disorder as was evident in childhood ADHD.

It is also not surprising then that parents of children with ADHD report significantly greater stress in their parental roles than do parents of normal children (DuPaul, McGoey, Eckert, & van Brakle, 2001; Harrison & Sofronoff, 2002; Johnston & Mash, 2001). These mother-ADHD child conflicts may also spill over into increased parent-child conflicts with the siblings of the ADHD child (Smith, Brown, Bunko, Blount, & Christophersen, 2002). Mothers of ADHD children routinely manifest higher levels of depression as well (Cunningham & Boyle, 2002), which is associated with even greater parenting problems as well as maternal biased reporting of children’s severity of ADHD symptoms as a function of depression related distortions (Chi & Hinshaw, 2002). Childhood ADHD is also associated with higher levels of parental ADHD, and if the child also manifests ODD or CD, parents may also manifest greater rates of mood, anxiety, and substance use disorders (Chronis, Lahey, Pelham, Kipp, Baumann, & Lee, 2003; Johnston et al., 2012).

Studies evaluating the impact of stimulant medication on these interactions suggest that the greater directive and negative behavior of the mothers of children with ADHD may be, in part, a reaction to their children's noncompliance and poor self-control rather than a cause of it (Danforth et al., 1991). Moreover, these conflicts in social interactions appear to exist in the relations of children with ADHD with their fathers (Edwards, et al., 2001; Tallmadge & Barkley, 1983) and teachers (Cunningham & Boyle, 2002; Whalen, Henker, & Dotemoto, 1980). Yet, both children and teens with ADHD do not perceive their relations with parents as being more problematic than do control children (Edwards et al., 2001; Gerdes, Hoza, & Pelham, 2003; Hoza et al., 2004). These results and others concerning the perceptions of children with ADHD concerning task performance indicate a positive illusory bias in their self-perceptions. Thus, children with ADHD do not typically suffer from low self-esteem. Rather, children with ADHD may have problems related to overly high self-esteem or at least self-perceptions of performance abilities.

The peer relations of children and teens with ADHD are also more problematic (Bagwell et al., 2001; Cunningham & Siegel, 1987) with ADHD children experiencing more rejection and fewer close friendships (Hoza, 2007). This is especially so for the subset having ODD/CD (Bagwell et al., 2001; DuPaul et al., 2001; Mikani & Hinshaw, 2003). Children with ADHD exhibit more negative behavior (DuPaul et al., 2001) and are less socially involved with non-ADHD playmates during conversations yet the children with ADHD may direct attention to their non-ADHD peers during play activities, receiving more structure in the form of praise and questioning from peers during the latter activities (Stroes, Alberts, & van der Meere, 2003). When ADHD children are presented with stories involving social problems, they have difficulty focusing on main story events, provide less relevant predictions about future outcomes, have more problems maintaining a positive outlook on future events, and do not generate as many socially acceptable solutions as non-ADHD peers (Zentall, Cassady, & Javorsky, 2001). Children with ADHD also encode fewer social cues and generate fewer responses than normal children while those with comorbid ODD/CD demonstrate a greater propensity for aggressive responses than control children (Matthys, Cuperus, & van Engeland, 1999).


The variety of proposed etiologies for ADHD are too numerous to review here. More information on etiology can be found in the other course by Barkley on this website dealing with the nature of ADHD in children and adolescents. Substantial evidence points to both neurological and genetic contributions to this disorder, and even specific brain regions and specific genes are now being implicated as contributors. So, while the exact neurochemical or neuromechanical mechanisms remain to be established for the disorder, and the suites of genes contributing to its striking heritability have yet to be completely catalogued, there is little doubt that these etiological directions hold the greatest promise for understanding the causes of the disorder.

Moreover, purely social causes of ADHD can be largely ruled out as likely contributing to most forms of ADHD and that is a major advance in itself. Social factors surely moderate types and degrees of impairments from the disorder, and even risk for comorbid ODD or CD, as well as social prejudices against those having ADHD not to mention access to services for its management. And they may even moderate severity of symptoms as perceived by caregivers. But as causes of disorder in and of themselves, social factors appear to have little research support.

Conceptualization of the Disorder

Until the 1990s, ADHD has lacked a reasonably credible scientific theory to explain its basic nature and associated symptoms and to link it with normal developmental processes. Consequently, the vast majority of research into the treatment of ADHD has remained exploratory or descriptive in nature rather than based upon any theory of the disorder. Treatments were tried principally because they had shown some efficacy for other disorders (e.g., behavior modification with the mentally retarded) or were discovered to have beneficial effects primarily by accident (e.g., stimulant medications). Thus, treatment decisions have not been guided so much by a scientific theory as by pragmatics; whatever seems to work is retained, whatever doesn’t is discarded with little guidance from any sound theoretical rationale.

The field of ADHD has reached a point, however, where the neuropsychological, neuro-imaging, and genetic studies cited above are coming to set clear limits on theorizing about the origins not only of ADHD but theories of its nature as well. Any credible theory on the nature of ADHD must now posit neuropsychological constructs that are related to the normal development of inhibition, self-regulation, and executive functioning and explain how they may go awry in ADHD. And such a theory will need to argue that these constructs arise from the functions of the prefrontal-striatal network and its interconnections with other brain regions that appear to subserve the executive functions and self-control, such as the cerebellum. Those cognitive functions will be shown to have a substantial hereditary contribution to individual differences in them given the results of twin studies on the genetic contribution to variation in ADHD symptoms.

This author has been working on just such a theoretical conceptualization of executive functioning and its extension to ADHD over the past 15 years (see Barkley, 1997d, 2006, 2012c). It is briefly discussed below, followed by its implications for the management of ADHD. Research continues on the merits of this model for ADHD but it is included here because of its far greater implications for treatment than any prior theories founded solely on ADHD arising from deficits in response inhibition (Quay, 1988), delay aversion (Sonuga-Barke et al., 1996), or arousal and energetic pools (Sergeant & van der Meere, 1988).

The model is founded on the premise that ADHD consists mainly of a developmental delay in behavioral inhibition and working memory that disrupt self-regulation; an assertion for which there is substantial research support (see Barkley, 1997d; Fraxier et al., 2004; Herevey et al., 2004; Nigg, 2001; Nigg, Goldsmith, & Sachek, 2004). This theory links behavioral inhibition to the executive functions and shows them to provide for self-regulation. Behavioral inhibition occupies a foundation in relationship to four other executive functions that are dependent upon it for their own effective execution. Self-regulation is defined as any self-directed action used to change one’s own behavior so as to alter the probability of a delayed (future) consequence. The executive functions are seen as forms of behavior-to-the-self – the actions one uses to change themselves so as to change their future.

Four executive functions were originally theorized to exist and to permit self-regulation, bringing behavior (motor control) progressively more under the control of internally represented information (forms of self-directed action), time, and the probable future and wresting it from control by the immediate external context and temporal now. Such self-control functions to maximize future consequences for the individual over merely immediate ones. The model applies only to the Combined Type of ADHD to date.

Behavioral inhibition involves: (1) the capacity to inhibit prepotent responses, creating a delay in the response to an event (response inhibition). There may be two other inhibitory processes related to it and which, at least for the moment, has been combined into a single construct concerning inhibition. These two other processes are: (2) the capacity to interrupt ongoing responses given feedback about performance, particularly those response patterns that are proving ineffective; and (3) the protection of this delay in responding, the self-directed actions occurring within it, and the goal-directed behaviors they create from interference by competing events and their prepotent responses (interference control). Through the postponement of the prepotent response and the creation of this protected period of delay, the occasion is set for the four executive functions (covert, self-directed actions) to act effectively in modifying the individual's eventual response(s) to the event. The chain of goal-directed, temporally-governed, and future-oriented behaviors set in motion by these acts of self-regulation are then protected during their performance by interference control. And even if disrupted, the individual retains the capacity or intention (via working memory) to return to the goal-directed actions until the outcome is successfully achieved or judged to be no longer necessary. The four executive functions are listed below using both their more common label in the neuropsychological literature followed by this model's redefinition of the self-directed action of which it is comprised, in parentheses:

I. Nonverbal Working Memory (Covert Self-Directed Sensing)

Nonverbal working memory is the ability to maintain mental information on-line that will be used to subsequently control a motor response. These prolonged mental representations of events, achieved by covertly sensing to the self, serve to recall past events for the sake of preparing a current response and represent hindsight or the retrospective function of working memory (Fuster, 1997). The chief forms of sensing that are being self-directed are visual imagery and private audition, or rehearing. Individuals are reactivating and using the images and sounds (and other sensory information) associated with past events to guide present and future-directed behavior.

Past events are retained in a temporal sequence and this sequence has been shown to contribute to the subjective estimation of psychological time (Michon, 1985). Analysis of temporal sequences of events for recurring patterns can be used to conjecture hypothetical future events -- the individual's best guess as to what may happen next or later in time based upon the detection of recurring patterns in past event sequences. This extension of hindsight forward into time also creates forethought or the prospective function of working memory, forming a temporally symmetrical counterpart to the retrospective function of hindsight (Fuster, 1997). And from this sense of the future likely emerges the progressively greater valuation of future consequences over immediate ones that takes place throughout child development into young adult life (Green, Fry, & Meyerson, 1994).

This self-directed sensing (seeing the past so as to see the future) gives rise to hindsight and forethought thereby permitting the individual to create a preparation or intention to act, called an anticipatory set (Fuster, 1997). In so doing, individuals are now capable of the cross-temporal organization of behavior; that is, the linking of events, responses, and their eventual consequences via their representation in working memory despite what may be considerable gaps among them in real time. Thus, self-regulation relative to time arises, at least initially, as a consequence of nonverbal working memory and the internally represented information it provides for the control and guidance of behavior over time.

II. Verbal Working Memory (Internalized, Self-Directed Speech)

During the early preschool years, speech, once developed, is initially employed for communication with others. Language is now not just a means of influencing the behavior of others but provides a means of reflection (self directed description) as well as a means for controlling one's own behavior (Berk, 1992, 1994; Diaz & Berk, 1992). Self-directed speech progresses from being public, to being sub-vocal, to finally being private, all over the course of perhaps 6-10 years. With this progressive privatization of speech comes an increasing control it permits over behavior. Self-speech now provides a tremendously increased capacity for self-control, planfulness, and goal directed behavior that further augments that being provided by the first executive function, self-directed imagery and hearing.

III. Self-Regulation of Affect/Motivation/Arousal (Self-Directed Emotion)

The occasion is now set for the self-regulation of affect, motivation, and arousal through the use of the first two executive abilities (self-sensing and self-speech). Individuals now possess the capacity to present images (and other sensory information) along with words to themselves that can be used to manipulate emotional states. That is because images and other sensory information from the past come automatically with emotional valences welded to them (how we felt about them)(Damasio, 1995). Yet it is not just one’s affect that is being managed by the use of self-speech and self-sensing (especially imagery). Emotion is, by definition, a motivational state. And so this re-presenting of words and images to the self creates a capacity for self-motivation (Fuster, 1997) because emotion and motivation are inherently linked (Ekman & Davidson, 1994; Lang, 1995). By privately manipulating and modulating emotional and motivational states, the child can induce drive or motivational states that may be required for the initiation and maintenance of goal directed behavior (Barkley, 1997d).

IV. Planning or Reconstitution (Self-Directed Play)

Bronowski (1977) reasoned that the use of images and language to represent the objects, actions, and their properties that exist in the world around us provides a means by which the world can be taken apart into pieces. These pieces can then be combined to create novel recombinations. Internal speech and imagery permit analysis (taking apart) and out of this process comes its complement – synthesis (recombination) to create entirely new ideas about the world (Bronowski, 1977) and entirely new responses to that world. It provides a means to synthesize novel behavioral sequences in the service of problem-solving and goal-directed action, particularly when obstacles are encountered in pursuit of a goal and new behaviors must be generated to solve the problem (Barkley, 1997d; Fuster, 1997). Barkley has hypothesized that, like the other executive functions, this one is also a form of self-directed behavior that, like the internalization of speech, becomes turned on the self during development and eventually is privatized.

More recently, I have expanded the theory to break out self-awareness as a separate executive function from the nonverbal working memory system and to show how these executive functions extend into everyday life over long spans of time to support social existence (self-defense, reciprocity, cooperativeness, etc.)(Barkley, 2012).  I have also shown why neuropsychological tests of EF are severely limited in their representation of the EFs especially as they are deployed in daily social life and are so poor in their ecological validity.

Over development, as the executive functions develop, they permit the construction and execution of increasingly, lengthy, complex, hierarchically organized, and novel chains of goal-directed behavior, protecting them from disruption by interference until they have been completed. This is achieved by generating internally represented information that serves to take over the control of behavior from the moment and immediate setting and direct behavior toward time and the probable or anticipated future. Such internal control over behavior not only creates a greater purposefulness or intentionality to behavior, but also a greater flexibility. The executive functions grant behavior both a more determined, persistent, reasoned, intentional, and purposive quality while permitting greater shifting of behavior as needed to achieve one’s goals – an appearance of volition, choice, and will arising from internally guided behavior (James, 1890).

The impairment in behavioral inhibition occurring in ADHD is hypothesized to disrupt the efficient execution of these executive functions thereby delimiting the capacity for self-regulation they provide. The result is impairment in the cross-temporal organization of behavior and in the guidance and control of behavior by internally represented information. This inevitably leads to a reduction in the maximization of long-term consequences for the individual. This theory, if correct, provides a much deeper insight into the nature of the disorder, and a much broader perspective on its likely impairments along with a litany of implications for its management (see below). In essence, ADHD is not so much an attention disorder from this perspective but a disorder of executive functioning – of internally guided and regulated behavior across time and toward future events – leaving the individual to be more affected by external events of the moment and more governed by concerns for immediate than delayed gratification.

Treatment Approaches

Research on the treatment of ADHD over the past two decades has focused largely on evaluating multi-modal treatment packages (MTA Cooperative Group, 1999a, 2004a). Innovations have mainly occurred in new delivery systems within psychopharmacology and even new drug development rather than in developing new psychosocial treatments. This is not to say that more information on the prevailing treatments has not been gained over this decade; that is hardly the case. It is to say that no significant breakthroughs in the psychosocial treatment of the disorder have been forthcoming. Most of the psychosocial treatment research has served to clarify the efficacy (or lack of it) of already extant treatment approaches, or their combinations.

A major problem in the ADHD treatment literature is a lack of documentation of long-term treatment effectiveness. Almost all of the research has focused on short-term effects (i.e., within 3 months), with a few studies providing intervention for up to 14 months (Shelton, Barkley et al. 2000; MTA Cooperative Group, 1999a) with follow-up evaluations going on for several years thereafter (Barkley, Shelton et al., 2002; MTA Cooperative Group, 2004a). Thus, at the time of initial writing of this course, long-term effects beyond a few years had been largely unstudied. This situation has been remedied somewhat by the Multimodal Treatment Study of ADHD, commonly called the MTA study (MTA Research Group, 1999a, 2004a),which has now followed children for 8 years after receiving 14 months of treatment (Molina et al., 2009), and the New York-Montreal multimodal treatment study (Abikoff et al., 2004a, 2004b; Hechtman et al., 2004a, 2004b). These long-term studies have shed some important insight on treatment, especially regarding the efficacy of combining psychosocial and pharmacological treatment, yet they have also shown that treatment gains do not endure once treatment is discontinued (Molina et al., 2009). This does not mean one should abandon treatment but that one should approach it like a chronic medical illness, such as diabetes, in which treatment must be continued as needed to control impairing symptoms and reduce the risk for secondary harm from an unmanaged disorder.

Another concern regarding the treatment research on ADHD has been that despite consistent findings of improvement in core symptoms of ADHD, there have been few reports of psychosocial treatment effects on key indicators of functioning such as academic achievement or social skills. For treatment of ADHD to be considered truly effective, there needs to be documentation of effectiveness on key ecological indicators of functioning in major life activities, such as school grades, sustained peer relations, etc. Again, this situation has been somewhat remedied by the MTA and, to a lesser extent, by the New York-Montreal multimodal studies.

Before venturing into a more detailed discussion of the efficacy of specific treatments for ADHD, it will be helpful to re-examine some traditional assumptions about the treatment of this disorder. They are being called into question not only by the theoretical model of Barkley discussed above, but by the results of research on etiologies (see the companion course on the nature of ADHD, ADHD: Nature, Course, Outcomes, and Comorbidity), as well as the results of follow-up studies of children who had received effective treatments for various periods of time and were then followed.

Re-examining Treatment Assumptions

Advances in research on the etiologies of ADHD and in theoretical models about the disorder seem to suggest why few treatment breakthroughs, especially in the psychosocial arena, have occurred. The information yielded from these sources increasingly points to ADHD as being a developmental disorder of probable neuro-genetic and neurological origins in which some unique environmental factors (mainly bio-hazards such as toxins and brain injuries) play a role in the expression of the disorder, though far less than do genetic ones. Therefore, unless new treatments address the underlying neurological substrates or genetic mechanisms that are contributing so strongly to it, the treatment will have fleeting or minimal impact on remedying this disorder.

We are not suggesting that prevention of ADHD is an impossible goal. For instance, some have suggested that reshaping the environments of young preschoolers, such as limiting television watching, might help to prevent some cases of ADHD (Christakis, Zimmerman, DiGiuseppe, & McCarty, 2004). However, we doubt this will prove effective given that there are serious questions concerning the direction of causality in such correlational findings (Barkley, 2004b) and that others have not replicated these initial results. Others have made a more compelling case for the reduction of environmental lead given the contribution of lead poisoning before age 3 years to the risk for later ADHD (Needleman et al., 1990; Nigg, 2006). Certainly the reduction of maternal use of alcohol and tobacco products during pregnancy would seem to be useful in view of the linkages noted earlier between these fetal neuro-toxins and risk for ADHD in the offspring of those pregnancies. This type of preventative research and related interventions should be encouraged. However, this is a course on treatment and by the time individuals meet diagnostic criteria for ADHD, we believe that they are on a chronic course and need to be treated accordingly. Therefore, the treatment of ADHD is actually symptomatic management as in diabetes. It is management of a chronic developmental condition and involves finding means to cope with, compensate for, and accommodate to the developmental deficiencies so as to reduce the numerous secondary harms that can accrue from unmanaged disorder. These means also include the provision of symptomatic relief such as that obtained by various medications.

Given the relatively greater contribution of genotype to environment in explaining individual differences in the symptoms of the disorder, it is highly likely that treatments for ADHD, while providing improvements in the symptoms, do little to change the rank ordering of such individuals relative to each other in their post-treatment levels of ADHD (see Scarr, 1992; Scarr & McCartney, 1983; Rutter, 1997; for a general discussion of this issue in developmental psychology and clinical interventions). It is also likely that such treatments, particularly in the psychosocial realm, will prove to be specific to the treatment setting and agents, showing minimal generalization to other agents or settings without actively arranging for its occurrence in those other settings or with those other caregivers.

Some of the psychosocial treatments for ADHD may have carry-over effects, mostly in the form of parents or teachers providing external structure that ameliorates ADHD-related symptoms. Ideally, these environmental adjustments will alter the developmental trajectory of the child or adolescent with ADHD. However, such interventions are not expected to produce fundamental changes in the underlying deficits of ADHD, rather they only prevent an accumulation of failures and problems secondary to ADHD. Thus, researchers and clinicians should anticipate that long-term studies are more likely to find treatment effects on problems secondary to ADHD than on deficits specific to ADHD.

The results of the MTA study lend some support to the assertions above. For instance, the study has found stronger treatment effects on core symptoms of ADHD during the intensive phases of treatment. Also, the trend in the follow-up seems to be leaning toward advantages of combined pharmacological and psychosocial treatment on constructs other than core symptoms of ADHD. However, such findings were not evident in the 24 month New York-Montreal multimodal study (Abikoff et al., 2004a; Hechtman et al., 2004a). One possible explanation for the difference is that the psychosocial treatments in the MTA were much more intensive, whereas the efficacy of the psychosocial treatments used in the New York-Montreal study was not established by the authors (this is discussed in greater detail later in the course). Thus, only strong, empirically supported treatments might be expected to have the kind of carry-over effects discussed above. Yet such treatment effects may not persist for long after the withdrawal of the formal interventions (Molina et al., 2011; Shelton et al., 2002).

The theoretical model of ADHD discussed above suggests other reasons why treatment effects may be so limited. This is largely because, according to this model, ADHD does not result from a lack of skill, knowledge, or information. It is, therefore, not going to respond well to interventions emphasizing the transfer of knowledge or of skills, as might occur in psychotherapy, social skills training, cognitive therapies, or academic tutoring. Instead, in Barkley’s (2006) model, ADHD is viewed as being a disorder of performance – of doing what one knows rather than knowing what to do. Like patients with injuries to the frontal lobes, those with ADHD find that it has partially cleaved or dissociated intellect from action, or knowledge from performance. Thus, the individual with ADHD may know how to act but may not act that way when placed in social settings where such action would be beneficial to them. The timing and timeliness of behavior is also being disrupted more in ADHD than is the basic knowledge or skill about that behavior.

From this vantage point, treatments for ADHD will be most helpful when they assist with the performance of a particular behavior at the point of performance in the natural environments where and when such behavior should be performed. A corollary of this is that the further away in space and time a treatment is from this point of performance, the less effective it is likely to be in assisting with the management of ADHD (Goldstein & Ingersoll, 1993). Not only is assistance at the “point of performance” going to prove critical to treatment efficacy, but so is assistance with the time, timing, and timeliness of behavior in those with ADHD, not just in the training of the behavior itself (Barkley, 2005). Nor will there necessarily be any lasting value or maintenance of treatment effects from such assistance if it is summarily removed within a short period of time once the individual is performing the desired behavior. The value of such treatments lies not only in providing assistance with eliciting behavior that is likely to already be in the individual’s repertoire at the point of performance where its display is critical, but in maintaining the performance of that behavior over time in that natural setting.

Disorders of performance like ADHD pose great consternation for the mental health and educational arenas of service. At the core of such problems is the vexing issue of just how to get people to behave in ways that they know are good for them yet which they seem unlikely, unable, or unwilling to perform. Conveying more knowledge does not prove as helpful as altering the motivational parameters and external cues or sources of control associated with the performance of that behavior at its appropriate point of performance. Coupled with this is the realization that such changes in behavior are maintained only so long as those environmental adjustments or accommodations are as well. To expect otherwise would seem to approach the treatment of ADHD with outdated or misguided assumptions about its essential nature.

The conceptual model of executive functioning (Barkley, 2012) as extended to ADHD introduced above brings with it many other implications for the management of ADHD (see Barkley, 2006). Some of these are briefly mentioned below:

1. If the process of regulating behavior by internally represented forms of information (working memory or the internalization of behavior) is delayed in those with ADHD, then they will be best assisted by “externalizing” those forms of information; the provision of physical representations of that information will be needed in the setting at the point of performance. Since covert or private information is weak as a source of stimulus control, making that information overt and public may assist with strengthening control of behavior by that information.

2. The organization of the individual’s behavior both within and across time is one of the ultimate disabilities rendered by the disorder. ADHD is to time what nearsightedness is to spatial vision; it has created a temporal myopia in which the individual’s behavior is governed even more than normal by events close to or within the temporal now and immediate context rather than by internal information that pertains to longer term, future events. Those with ADHD could be expected to be assisted by making time itself more externally represented, by reducing or eliminating gaps in time among the components of a behavioral contingency (event, response, outcome), and by serving to bridge such temporal gaps related to future events with the assistance of caregivers and others.

3. Given that the model hypothesizes a deficit in internally generated and represented forms of motivation that are needed to drive goal-directed behavior, those with ADHD will require the provision of externalized sources of motivation. For instance, the provision of artificial rewards, such as tokens, may be needed throughout the performance of a task or other goal-directed behavior when there is otherwise little or no such immediate consequences associated with that performance. Such artificial reward programs become for the ADHD child like prosthetic devices such as mechanical limbs to the physically disabled, allowing them to perform more effectively in some tasks and settings with which they otherwise would have considerable difficulty. The motivational disability created by ADHD makes such motivational prostheses nearly essential for most children with ADHD.

4. Given the above listed considerations, parents and teachers should reject any approach to intervention for ADHD that does not involve helping them deal with a child or adolescent with an active intervention at the point of performance. Many parents and teachers seek what might be called the “garage mechanic approach.” According to this model, a child can be dropped off someplace and be “fixed” by the “mechanic” without the parent or teacher “getting their hands dirty.” Such an approach is untenable and a hands-on approach to intervention is strongly recommended. This is true for all interventions, including pharmacotherapy.

Treatments for ADHD: Introduction

The provision of treatment services to children with ADHD has increased dramatically over the past 20 years, owing in large part to four national trends: (1) the recognition by special education laws that ADHD is an eligible condition for identification and services in public schools (circa 1991), (2) the growth of formally organized advocacy groups (such as Children and Adults with ADHD, see in the late 1980s, (3) the growth of advertising and educational efforts by pharmaceutical companies promoting new stimulant delivery systems over the past decade, and more recently new types of medication, for the management of ADHD, and (4) increased continuing education programs for educational and mental health professionals on the disorder. For instance, between 1986 and 1996, stimulant prescriptions for ADHD increased to accounting for three fourths of all physician visits for children with ADHD, with a 10-fold increase in related services such as health counseling, and a 3-fold increase in diagnostic services (Hoagwood, Kelleher, Feil, & Comer, 2000). This trend has continued into the mid-2000s. Nevertheless, this report also documented a decline in the use of follow-up care, apparently due to insurance obstacles, lengthy waiting lists, and limited access to pediatric specialists. Treatment appears to be increasingly provided by primary care professionals who are likely to utilize only medication management, with only a third of ADHD children being referred to and treated by mental health professionals (Bussing, Zima, & Belin, 1998). Those who are referred to specialists are more likely to have comorbid disorders, greater impairment, and greater family burdens. Such trends have also undoubtedly continued to the present time.

I now present the major treatment approaches employed with ADHD that have some scientifically established effectiveness. These include: (1) psychopharmacology; (2) parent training in child behavior (contingency) management methods; (3) teacher implementation of these and other child behavior management tactics; and (4) combinations of these approaches into a multi-modal therapy program. Given the weaknesses inherent in any single treatment modality, the multi-modal approach is preferred here for treating most cases of ADHD because of the inability of medication to adequately address all cases of ADHD, especially those with coexisting disorders, such as learning disabilities, anxiety, depression, or conduct disorder.


Research suggests that the three most commonly used drugs for the management of ADHD symptoms are the stimulants, the nonstimulant atomoxetine, and the antihypertensives, most recently guanfacine XR (Conner, 2006; Pliszka, 2009; Zito, Safer, DosReis, Gardner, Magder et al., 2003). Of these, the stimulants are by far the most utilized.  Use of the stimulants, however, was founded on a virtual chance discovery of their effectiveness and not on any theoretical rationale (Bradley, 1937; Winsburg, Bialer, Kupietz, & Tobias, 1972). A rationale may be emerging though in view of recent theoretical models that emphasize poor behavioral inhibition specifically and executive functioning (EF) more generally as probably being central to the nature of the disorder (Barkley, 1997, 2012c). Brain regions subserving inhibition and EF appear to be involved in the etiology of ADHD, these regions are largely dopaminergic, and stimulants (which increase extracellular dopamine) seem to produce their greatest effects within these same brain regions (see the companion course on the nature and diagnosis of ADHD in children, Attention-Deficit/Hyperactivity Disorder in Children: Diagnosis and Assessment; Nigg, 2006; Volkow, Ding, Fowler, Wang, Logan, Gatley et al., 1995; Volkow, Wang, Fowler, Logan, Angrist et al., 1997). Atomoxetine increases extracellular norepinephrine but produces an indirect increase in dopamine in the prefrontal cortex that may also explain its therapeutic benefit.  In contrast, guanfacine XR works primarily by fine tuning the alpha-2 receptors on nerve cells in the frontal cortex, or executive brain, thereby enhancing signal strength and conductivity.

Until recently, it was not clear precisely how these medications affected brain function and particularly their sites and neurochemical modes of action. It now appears as if the major therapeutic effects of the drugs are achieved through alterations in frontal-striatal activity (Volkow et al., 1997) via their impact on at least three or more neurotransmitters important to the functioning of this region and related to response inhibition, these being dopamine, norepinephrine, and epinephrine (Pliszka, McCracken, & Maas, 1996). The direct rationale, then, for employing some medications with children with ADHD may be that they directly, if only temporarily, improve the deficiencies in these neural systems related to behavioral inhibition, EF, and self-regulation.

Stimulant Medication

Since Bradley (1937) first (accidentally) discovered their successful use with behavior problem children, the stimulants have received an enormous amount of research (Faraone & Buitelaar, 2010; Greenhill, Halperin, & Abikoff, 1999) with meta-analyses indicating a clear benefit in managing the disorder in the short-term (Faraone & Buitelaar, 2010; Prasad et al., 2012; Schachter, Pham, King, Langford, & Moher, 2001) and some continuing benefit to symptomatic management (but not necessarily academic achievement) in the long term (Schachar et al., 2002). As long as clients comply with treatment, benefits can be found over as long as 5 years (Charach et al., 2004). The results of research on stimulants overwhelmingly indicate that these medications are quite effective for the management of ADHD symptoms in most children older than 5 years (Connor, 2005a; Greenhill & Osmon, 2000; Faraone & Buitelaar, 2010; Solanto, Arnsten, & Castellanos, 2001). Between 4 and 5 years of age, the drugs are equally as effective as in older children (Connor, 2002), with some 82 percent of cases responding positively (Short, Manos, Findling, & Schubel, 2004). Research in the past decade has shown these medications to be safe and effective down to age 3 years, although with somewhat fewer children responding and with slightly less degree of improvement and somewhat more side effects than is usually evident in school-age children (Greenhill et al., 2006).  Guidelines for the use of stimulant medications with ADHD children have been issued by both the American Academy of Pediatrics (2011) and the American Academy of Child and Adolescent Psychiatry (2002).

Research by Safer, Zito, and colleagues has documented dramatic increases in the overall rate of stimulant medication use among children and adolescents with ADHD (e.g., Zito et al., 2003) that have likely continued until 2009 or later. During the 1990s, stimulant prescription rates more than tripled. This was part of a general boom in the diagnosis and treatment of ADHD that probably occurred due to greater public awareness of the disorder as well as changes to special education regulations that encouraged the identification and treatment of the disorder in school settings. For example, surveys comparing physician practices in 1986 and 1999 found a three-fold increase in diagnoses of ADHD and a ten-fold increase in treatment services for ADHD. Comparing 1987 with 1997 records in the National Medical Expenditure Survey, Olfson and colleagues (Olfson et al., 2003) documented a marked expansion of access to treatment among children with ADHD, from 0.9 per 100 children to 3.4 per hundred receiving outpatient treatment. Despite this improvement in access to care, there was a decline in the intensity of treatment, as determined by number of visits and forms of treatment recommended other than medication. The authors interpret these changes as likely arising from increased access to special education services during this period, the growth of managed health care and its emphasis on brief visits and treatments, and increased public acceptance of medication use for the disorder. Such trends have likely continued to the present time.

The most rapid expansion of stimulant use has been with preschoolers (Zito et al., 2003), the poor and low-income families (Olfson et al., 2003), and adolescents (Olfson et al., 2003; Safer, Zito, & Fine, 1996), most likely due to these segments being markedly under-treated in prior years relative to elementary-age, middle and upper income children. By the end of the 1990s, several studies concluded that 2-6% of American school children (ages 5 to 15) had been treated with stimulants (Greydamus, Pratt, Sloane, & Rappley, 2003), with the largest databases placing the figure at 1.3 to 3.8% (Jensen et al., 1999; Safer & Malever, 2000; Zito et al., 2003) by the late 1990s. More recent surveys in the first decade of this century would place this percentage at 4.2% (Visser et al., 2007).  Thus, there is clearly a strong trend toward increased use of stimulants, but the exact rate of stimulant use is hard to describe due to the rapid rate of change, widely varied use among various providers and communities, and the lack of any national database that might address the issue.

The most commonly prescribed stimulants are shown in Table 1. They include methlyphenidate (e.g., Ritalin), d-amphetamine (e.g., Dexedrine), and a combination of amphetamine salts marketed under the name Adderall. Pemoline had been available for treatment of ADHD in earlier decades but its manufacture was discontinued more than a decade ago as a result of liver complications and failure in a small percentage of cases.  It receives no further attention here. Because methylphenidate and the amphetamines share similar characteristics, these drugs are discussed collectively in this section.

Methylphenidate and amphetamines are rapidly acting stimulants. In their immediate release (IR) formulations, the stimulants produce effects on behavior within 30 to 45 minutes after oral ingestion and peaking in their behavioral effects within 2 to 4 hours (Connor, 2005a). The utility of these IR formulations in managing behavior quickly dissipates within 3 to 7 hours, although minuscule amounts of the medication may remain in the blood for up to 24 hours (Greenhill & Osmon, 2000; Solanto et al., 2001). Because of their short half-life, they are often prescribed in twice or thrice daily doses. Claims that Adderall lasts longer than methylphenidate and the other stimulants have not been fully substantiated and dosing protocols for Adderall often involve twice daily administration (Greydanus et al., 2003).

An important development in treatment of ADHD is effective extended release forms of methylphenidate. These do not represent new drugs but new delivery systems for sustaining blood levels of the drug over longer periods so as to reduce dosing to once per day, where possible. Intermediate duration versions of methylphenidate that have therapeutic effects for six to eight hours include Ritalin-SR, based on a wax matrix coating, Metadate-ER, which uses a time-release pellet technology, as do Methylin-ER , Ritalin-LA, and Metadate-CD. Adderall is considered by some to be an intermediate duration stimulant. Once-daily stimulants include Dexedrine Spansules (d-amphetamine), Concerta (OROS methyphenidate, using an osmotic pump), and Adderall XR (a mixture of d- and l-amphetamines, using a time-release pellet technology) that may last up to 12 hours. More recently, lisdexamfetamine (Vyvanse) is a variation of Adderall in which the amphetamine is bound up with lysine such that the drug is only activated in the human gut and intestinal lining where an enzyme operates to cleave the lysine from the amphetamine, thereby activating the latter.  The drug is as effective as Adderall XR for children (Biederman et al., 2007).  This mechanism may extend the effects of this version of amphetamine for an additional hour or two beyond that of Adderall XR.  In September of 2012, the FDA approved a liquid form of extended release methylphenidate (Quillivant XR) that should be useful for children who may have difficulty swallowing tablets or capsules.

There is some variability in the effectiveness of these longer-acting preparations. For example, the initial version of sustained release methylphenidate (Ritalin-SR)  using a wax matrix coating had erratic effects on some children (Greydanus et al., 2003) and often reduced therapeutic efficacy relative to IR forms of the medication resulting from a truncation of the peak blood level below that required for an acceptable treatment response. This limitation has been overcome in the recently approved ER preparations, such as Concerta, Metadate CD, Adderall XR, and Vyvanse. Another unique feature of some of these new delivery packages, such as Concerta, is that they provide a steady increase in the amount of medicine delivered during the day, thus overcoming problems with diminished effect later in the day. Possibly due to the emphasis on sustained effects, some of the once daily preparations may have limited effectiveness in the first hour or so following administration. In cases of delayed effect, earlier daily administration (e.g., 6:30 instead of 7:30 AM) or a small “booster dose” of standard stimulant of the same type can be given to increase effectiveness (e.g., 5 mg of methyphenidate with Metadate CD in the morning).

Although once used predominantly for school days, there has been an increasing clinical trend toward usage throughout the week as well as school vacations, particularly for the more moderately to severely ADHD and conduct problem children. This treatment option appears to have a favorable benefit to cost ratio. Benefits have been supported by some well-designed, randomized studies. Putative costs of treatment over weekends and school holidays, mostly concern about possible growth suppression, may not be as serious as was once believed (Greydanus et al., 2003; Spencer et al., 1996).

The behavioral improvements produced by stimulants are in sustained attention, impulse control, and reduction of task- irrelevant activity, especially in settings demanding restraint of behavior (Barkley, 1977b; Connor, 2005a; Rapport & Kelly, 1993; Solanto et al., 2001). Generally noisy and disruptive behavior also diminishes with medication. Children with ADHD may become more compliant with parental and teacher commands, are better able to sustain such compliance, and often increase their cooperative behavior toward others with whom they may have to accomplish a task as a consequence of stimulant treatment (Danforth et al., 1991, for a review). Research also suggests that children with ADHD are able to perceive the medication as beneficial to the reduction of ADHD symptoms and even describe improvements in their self-esteem (DuPaul, Anastopoulos, Kwasnik, Barkley, & McMurray, 1996; Pelham, Hoza, Pillow, Gnagy, Kipp, Greiner et al., 2002), though they may report somewhat more side effects than do their parents and teachers.

Improvements in other domains of behavior in children with ADHD have also been demonstrated. Both overt and covert aggressive behavior are often reduced by stimulant treatment of children with ADHD who demonstrate abnormally high levels of pre-treatment aggressiveness (Connor et al., 2002), though the effect on overt aggression may be somewhat less if conduct disorder is present. The quality of the children's handwriting may also improve with medication (Lerer, Lerer, & Artner, 1987). Academic productivity, or the number of problems completed, and accuracy of work completion also increase, in some cases dramatically, as a function of medication (Pelham, Bender, Caddell, Booth, & Moorer, 1985; Rapport, DuPaul, Stoner, & Jones, 1986). In general, classroom behavior is significantly improved as is work productivity although there is less of an impact on academic accuracy (Prasad et al., 2012), which is usually not as problematic for children with ADHD as is productivity.  For many years, stimulants were thought not to impact academic achievement significantly (Barkley & Cunningham, 1978; Schachar et al., 2002; Schachter et al., 2001).  Since then longer term effects on academic achievement (level of difficulty of material mastered) have been documented yet the degree of improvement is small in magnitude (Langberg & Becker, 2012).

It should be strongly emphasized that the effects of stimulant medication are idiosyncratic (see Rapport et al., 1986). Although reported response rates vary across studies, many reviewers have concluded that 70-82% of children show a clinically beneficial response to any single stimulant. However, with a trial of a second stimulant the positive response rate may approach 90% (Pliszka et al., 2000). Unfortunately, there is no way to predict in advance which children will respond to which stimulant. Similarly, among the students who do respond positively to stimulants, there is no basis for predicting which dose will be best. Most children and adolescents show maximal improvement at low to moderate doses of stimulants but others are most improved at higher doses (Pelham et al., 1998; Smith et al., 1998a). In addition to this between-subject variability in doses, there is considerable variability in the domains that respond to medication. For instance, some children may improve in one domain (e.g., behavior) when treated with stimulants but show no change, or even deteriorate, in other domains (e.g., academic performance). For this reason, we strongly recommend that treatment with stimulant medication be assessed on a case-by-case basis using measures that sample a broad range of domains of functioning.

The most frequently occurring side effects of the stimulants are mild insomnia and appetite reduction, particular at the noon meal and subjective reports of stomach ache, headache, and dizziness or jitteriness (Barkley, McMurray, Edelbrock, & Robbins, 1990; Connor, 2005; Greenhill et al., 1999; Greydamus et al., 2003). These subjective side effects tend to dissipate within a few weeks of beginning medication or can be managed by reducing the dose. Temporary growth suppression (loss of 1-4 lbs in first year) may accompany stimulant treatment, but is not generally severe nor especially common (Faraone et al., 2008). It can be managed by ensuring that adequate caloric and nutritional intake is maintained by shifting the distribution of food intake to other times of the day when the child is more amenable to eating (Taylor, 1986).

Some children become irritable and prone to crying late in the afternoon as their medication may be wearing off. This may be accompanied by an increase in hyperactivity. This apparent “rebound” phenomenon appears to be rare and might be controlled by adjusting doses and dose schedules (Greydamus et al., 2003).

In approximately 1 to 2 percent of children with ADHD treated with stimulants, motor or vocal tics may occur (Connor, 2005a; Greenhill & Osmon, 2000). This is well within the base rate prevalence for tics in the normal population. In others where tics already exist, they can be mildly exacerbated by stimulant treatment in some cases, but may even be improved in others (Gadow, Sverd, Sprafkin, Nolan, & Ezor, 1995). Reviews suggest that the risk for increasing tics in such comorbid children is mostly associated with high doses of amphetamine and not with methylphenidate (Bloch et al., 2009). It now appears to be relatively safe to use stimulant medications with children with ADHD and comorbid tic disorders but to be prepared to reduce the dose or discontinue medication should the child experience a drug-related exacerbation of their tic symptoms.

To avoid potential dose-dependent side effects, we recommend a low-slow-go approach to titrating doses. That is, one begins with a low dose, slowly titrates the dose upward, yet goes further until the most appropriate dose for that child. We believe that dose should be the lowest possible level that produces satisfactory clinical improvement. This is contrary to some clinical practices that titrate doses to the highest tolerable level. Finding the lowest effective dose may be more difficult, but has the potential to save money in medication costs, to reduce risk of side effects, and perhaps improve compliance due to increased comfort on medication.

As mentioned previously, it has been difficult to establish any reliable predictors of response to stimulant medication in children with ADHD. Those characteristics having the most consistent relationship to response have been pre-treatment levels of poor sustained attention and hyperactivity (Barkley, 1976; Buitelaar, van der Gaag, Swaab-Barneveld, & Kuiper, 1995; Taylor, 1983). The more deviant is a child's level on such factors, the better their response to medication. Predictors of adverse responding have not been as well studied. Some research suggests that pre-treatment levels of anxiety are associated with poorer responding to stimulants (Buitelaar et al., 1995; DuPaul, Barkley, & McMurray, 1994; Taylor, 1983). However, Pliszka (2003) argues that response to stimulants for treating ADHD is not affected by anxiety, but children with ADHD plus anxiety may show greater benefit from psychosocial interventions than children with ADHD alone (Pliszka, 2003). The results of more recent research remains mixed on this issue, with some studies finding no relationship of anxiety to stimulant response (Abikoff et al., 2005) while others do (Moshe et al., 2012).

There is little doubt now that the stimulant medications are the most studied and most effective treatment for the symptomatic management of ADHD and its secondary consequences (Connor, 2005a; Greydanus et al., 2003). As a result, for many children with moderate to severe levels of ADHD, this may be the first treatment employed in their clinical management. And for some, where little or no significant comorbid disorders exist, it may be the only treatment required. A multi-site study, in fact, found that among stimulant responsive children with ADHD, adding various forms of psychosocial treatments, such as parent training, social skills training, psychotherapy, or academic tutoring, added no additional benefits beyond that achieved by medication alone (Abikoff et al., 2004a, 2004b; Hechtman et al., 2004a, 2004b). On the other hand, the results of the MTA study suggest that a more intensive multi-modal treatment may produce a broader range of positive results than are achieved by medication alone (e.g., Pelham, 1999).

Despite some conflicting studies and opinions, there seems to be a general consensus that stimulant treatment is not always effective (i.e., the 20% to 30% non-response rate), nor necessary (i.e., in some cases psychosocial treatment is sufficient), nor is it always sufficient (i.e., many children meet criteria for improved, but not recovered on stimulants). The issue of lack of sufficiency is particularly salient with regard to the appropriate management of the comorbid conditions often seen in ADHD, such as LD, depression, anxiety, or conduct disorder. Given that medication typically does not address all of these presenting problems shown by many children with ADHD, other treatments may be required as adjuncts.

The following issues should be considered in the decision to employ medication for the management of ADHD: (1) the age of the child; (2) duration and severity of symptoms; (3) the risk of injury to the child if untreated (either by accident or abuse) by the present severity of symptoms; (4) the success of prior treatments; (5) relatively normal levels of anxiety (perhaps); (6) the absence of stimulant abuse by the child or adolescent or their caregivers; (7) the likelihood that the parents will employ the medication responsibly, in compliance with physician recommendations; and (8) the child does not live in a group setting such as a dormitory where supervision of the medication is poor and diversion of the medication to non-ADHD students may be more likely. Some of these latter concerns related to stimulant abuse (i.e., points 6 to 8) may be somewhat ameliorated by the longer-acting preparations of stimulants, such as Concerta, that have lower abuse potential than immediate release preparations.

Several suggested paradigms for evaluating stimulant drug response in individual cases have been reported (Barkley, Fischer et al., 1988; Pelham, 1987; Rapport, Dupaul, et al., 1986). We recognize that these are not always practical or available in clinical practice but recommend them as exemplars toward which practitioners should strive to meet. The trial includes the traditional and mandatory initial medical check-up of the child to insure that there are no pre-existing conditions that might contra-indicate or complicate the medication trial, such as cardiac problems, unusually high levels of anxiety, prior history of stimulant abuse, among others. This is followed by the child’s receiving a baseline evaluation on the measures (often rating scales) to be collected across the weeks of the trial (highly recommended). Such baseline evaluations must include ratings of potential side effects of the medication given that many of these are frequently pre-existing problems with ADHD children that, if not assessed at baseline, could be misconstrued as drug side effects (Barkley et al., 1990; Connor, 2002). The child’s participation is then scheduled for a four week drug-placebo trial during which the child is tested on three different doses of medication (typically methylphenidate at 5, 10, and 15 mg given morning and noon) and a placebo (lactose powder placed in gelatin capsule) (optional). Arrangements are made to have the noon dose of medication given at school on schooldays, if this is an IR formulation of the medication. The parents, teachers, child, and clinical assistant conducting the assessments of the child are all kept blind to the order of medication doses and placebo until the end of the trial.

The major outcome variables are typically ratings completed by parents, teachers, and for children over 8, the child receiving the medication. The frequency of ratings should match the frequency of switches between dose levels in the medication trial. For example, weekly dose switches can be paired with weekly ratings, but daily dose switches need to be paired with daily ratings.

One rating scale assesses the symptoms of ADHD as well as ODD while another is used to obtain information about side effects the child may have experienced that week (see Barkley & Murphy, 2006). A third rating scale completed only by the teacher assesses work productivity and accuracy (e.g., The Academic Performance Rating Scale, for instance, see Barkley, 1990; or DuPaul & Stoner, 2003). Another potentially useful set of measures that has repeatedly been shown to be sensitive to intervention is the Child Daily Report and Parent Daily Report (Dishion & Kavanaugh, 2003 or on the internet at Furthermore, the clinical team should solicit non-standardized information relevant to impairment and other clinically or academically meaningful phenomenon. For example, teacher comments can be collected by telephone or the internet, as are parent comments during weekly clinic visits during the medication assessment.

The ratings may be supplemented by objective data (if available and practical) such as grades and direct observations of behavior in school or in a clinic room. For example, during each weekly clinic visit, the child can be given a set of math problems of appropriate grade level to perform while seated alone in a clinic playroom. Observations can be taken of the child from behind a one-way mirror or with a video camera and the observations can be coded on the Restricted Academic Situation coding sheet (see Barkley & Murphy, 1998) for behaviors related to ADHD (i.e., off-task, fidgets, plays with objects, out of seat, etc.). In addition, the amount of work attempted and the accuracy of that work can be scored. Computerized measures such as continuous performance tasks may also be used to assess response to medication, but the ecological validity of these measures is questionable (Barkley, 1991), and therefore the value of such measures may be limited. We recognize that for the busy private or clinic-based practitioner, these supplemental measures may not be available or cost effective. But the use of rating scales to evaluate ADHD and related symptoms, side effects, and even academic productivity is strongly recommended.

Different dosing schedules can be evaluated. For example, in one protocol each drug condition lasts for 7 to 10 days before the child progresses to the next drug condition. The order of the drug conditions is random except that the middle and high doses, say 10 and 15 mg of IR methylphenidate, are paired such that the 10 mg condition always precedes the 15 mg condition. This is done to reduce the possibility of unnecessary side effects being provoked by beginning the trial at an initially excessive dose. An alternative approach is to switch doses daily in a counterbalanced, random order. Compared to the former, the daily crossover design helps to better control for unusual events or spurious improvement over time. However, it incurs a much higher response burden compared to weekly ratings and may reduce compliance to the protocol. Furthermore, the daily crossover design may also miss cumulative effects.

At the end of the four-week trial, the results are tabulated and a recommendation is made concerning possible continuation of the medication and which dose seems most effective. Children not found to be responsive to this stimulant may be tried on another and if a second stimulant does not work, other medications might be considered (see Pliszka et al., 2000). Furthermore, consistent with the ‘low-slow-go” approach children should be routinely tested on a lower dose than selected in the titration trial to see if the lower dose is sufficient. If there is still room for improvement relative to the higher dose, then the higher dose is well-justified.

Effectiveness, Safety, and Practicality of Stimulant Medication

The rise in the use of stimulant medication is supported by numerous studies documenting the efficacy of these medications (Greenhill et al., 1999; Schachter et al., 2001). A meta-analytic review of 62 high-quality studies (Schachter et al. 2001) found a medium sized effect of stimulants on parent rated behavior (mean .54) and a large effect size on teacher rated behavior (.78). Later reviews have found much the same results (Faraone & Buitelaar, 2010). Most of these studies were conducted in university or medical school research programs by experts and therefore speak more to treatment efficacy than effectiveness as applied in typical community settings. Yet the substantial supportive research across labs, investigators, regions, and even countries combined with the longstanding successful and ever increasing use of this treatment in clinical practice speaks albeit indirectly to both effectiveness and practicality. According to the Biglan et al. (2003) criteria, stimulant treatment is a Grade A intervention, meaning that stimulants have a level of support most appropriate for widespread dissemination.

Stimulants used to treat ADHD are very safe (Rapport & Moffitt, 2002). At therapeutic doses, stimulants produce few negative side effects and almost all of these negative effects are symptoms youth either develop tolerance for or are reduced to a tolerable level by lowering the dose or changing to a different stimulant or non-stimulant, such as atomoxetine. Long-term negative consequences are not evident in the research literature (Greenhill et al., 1999), but the potential for mild growth suppression is an issue of ongoing investigation and debate (Faraone et al., 2008; MTA Collaborative Group, 2004b). Concerns about predisposing stimulant treated children to later substance use disorders have been refuted by more than 14 studies, despite a single study implying otherwise (see Barkley, Fischer, Smallish, & Fletcher, 2003; Wilens et al., 2003). When absorbed rapidly (e.g., inhaled nasally or injected intravenously) or taken at high doses, stimulants may result in euphoric effects and health risks similar to cocaine. Such use is uncommon and less likely with longer-acting preparations. Mortality or serious morbidity from prescribed stimulants is rare. Indeed, compared to many other commonly prescribed psychiatric medications, stimulants are among the safest drugs given to children. Nevertheless, safety may be diminished when taking stimulants in combination with other medications (e.g., clonidine) or when potent stimulants, like Adderall XR, or high doses are used with children having pre-existing cardiac abnormalities or family histories of sudden cardiac arrest.

Research specific to practical issues with stimulant medication is limited and the preliminary findings raise some questions about the effectiveness of this treatment in primary care settings. Although taking pills seems to be a simple intervention, there are some significant barriers to daily administration of stimulant medication including limited access to prescribing physicians, cost, inconvenience, uncertainty about dose or type of medication, side effects, and parent or child resistance to taking medication. Research on compliance is limited, but suggests that children and adolescents with ADHD tend to take less medication than prescribed due to missed doses and termination of treatment against medical advice (Jensen, Kettle, Roper, Sloan, Dulcan, et al., 1999). The high cost of some new formulations of stimulants ($60-100 per month) may contribute to an already tenuous compliance situation though their once-daily extended release delivery systems may counteract such a problem. The problems with compliance have high clinical liability because stimulants exert their effects only when taken as prescribed. Even the longest acting stimulants have no measurable effect 24 hours after administration, so missed doses mean the child is essentially untreated. Thus, although stimulants get high marks for safety and effectiveness, there are some practical barriers to effective use of stimulant medication.

Special Considerations with Adolescents

Although there has been much less research on stimulant treatment for adolescents than children (Smith, Waschbush, Willoughby, & Evans, 2000), there is enough research to document that stimulants have similar efficacy from childhood to adolescence (Smith et al., 1998; Smith, Pelham, Gnagy, & Yudell, 1998). However, due to increased involvement with recreational drugs that could be associated with stimulant abuse or could lead to interactions between therapeutic stimulants and recreational drugs, caution should be taken when prescribing stimulants to adolescents. Also, clinicians need to be aware of the potential for diversion of prescribed stimulant medication by an ADHD teen to other teens for recreational misuse when the ADHD teen resides in a dormitory or other group living situation with other teens.

A major threat to the effectiveness of stimulant medication is the tendency for adolescents to discontinue medication as they get older. Thus, compared to children, more vigorous monitoring and promotion of compliance is necessary when working with adolescents. To avoid premature termination of effective stimulant treatment, parents and physicians should encourage adolescents to participate in treatment decisions and self-monitoring during periodic trials of stimulants. For adolescents who do not recognize the value of taking stimulants (assuming an individualize medication trial unambiguously supports the efficacy of the stimulant), it may be necessary to negotiate behavioral contingency contracts related to the appropriate use of stimulants (see Robin, 1998, 2006). Thus, in some cases multimodal treatment with an emphasis on compliance issues may be necessary for effective stimulant treatment of ADHD in adolescents. Unfortunately, the efficacy of multimodal treatments for adolescents with ADHD has yet to be tested in a major controlled trial.

Atomoxetine (Strattera®)

Atomoxetine was approved by the U. S. Food and Drug Administration in January of 2003 for use in children with ADHD six years of age and older, and in adolescents and adults having ADHD. This makes it the first such drug so approved for use in the adult stage of the disorder, and the first new drug for ADHD in more than 25 years (pemoline having been the last such medication). The drug is a highly selective inhibitor of norepinephrine reuptake with minimal to no action at other neurotransmitter sites. Its effectiveness has established in more than 10 large-scale published studies done before or shortly following FDA approval and involving various randomized, controlled clinical trials. These samples included 3,264 children and adolescents, and 471 adults with all types of ADHD (Krotochvil et al., 2001, Michelson et al., 2001; Michelson et al., 2002; Michelson et al., 2003; Spencer et al., 1998). The clinical trials clearly established both the efficacy and safety of atomoxetine for use in the management of ADHD. Many studies have been conducted since 2003 demonstrating the safety and effectiveness of this drug for ADHD management.

Atomoxetine is not a stimulant in that it is not a dopamine agonist. It has no abuse potential as studies show that it is not preferred over placebo by stimulant abusers and does not result in symptoms of craving, dependence, or addiction. It is therefore not scheduled, whereas the stimulants are classified as schedule II agents. Consequently, atomoxetine is a more convenient medication as it can be prescribed without special prescription pads needed in the U.S. for schedule II agents, can be prescribed with refills, and can be distributed to patients by physicians as samples. The fact that it is unscheduled and has no abuse potential can make it an attractive alternative to families of ADHD children concerned about the use of Schedule II medications for their child.

Atomoxetine may assist in the management of ADHD via its inhibition of the norepinephrine transporter, thereby making more norepinephrine available in the extracellular space. This results in a secondary increase of dopamine in the prefrontal cortex. The fact that it does not increase dopamine levels in the nucleus accumbens (the primary dopamine mediated reward pathway) may explain why it does not have reinforcing or otherwise addictive properties. Because it does not appear to increase dopamine levels in the striatum that helps control motor movements, it seems to have no exacerbating effect on motor and vocal tics.

Research shows that atomoxetine reduces both inattentive and hyperactive-impulsive symptoms of ADHD in over 70% of cases. The overall effect size (degree of change in group mean scores) of atomoxetine appears to be the same as Concerta among children previously untreated with stimulants, but may have a smaller effect size than that seen with the stimulants in the treatment of individuals with ADHD who have had prior stimulant exposure (Michelson, personal communication, September 2004). In controlled studies atomoxetine has an effect size of about 0.9 to 1.0 among stimulant naïve cases but an effect size of 0.6 to 0.8 (standard deviations) in cases with prior stimulant treatment. The effect size for the stimulants ranges from 0.8 and 1.2 (Michelson, Faries, Wernicke, Kelsey, Kendrick, Sallee, et al., 2001). Peak plasma concentrations for atomoxetine occur in one to two hours after oral ingestion and persist for 6-10 hours (half-life of 4 hours). The medication may therefore be given in either once or twice daily dosing. When given in twice daily divided doses, atomoxetine shows much longer daily coverage for ADHD symptoms than do stimulant medications. In contrast to the tricyclic antidepressants  that can also affect norepinephrine re-uptake, atomoxetine demonstrates no cardiovascular toxicity or abnormalities on electrocardiogram. Atomoxetine appears to improve ODD symptoms as well in ADHD children having significant levels of these symptoms (Newcorn, Michelson, Spencer, & Milton, 2002). It also results in significant improvements in parent-child relations, peer relations, school behavior and academic performance, and co-existing internalizing symptoms, such as depression or anxiety.

Atomoxetine can be considered a first-line agent in the treatment of ADHD in children, adolescents, and adults. Whether it is the first or second choice of a starting medication will depend upon several patient and social ecological characteristics that may exist at the time of a clinical trial. For instance, in ADHD patients with comorbid anxiety, obsessive-compulsive behavior, or tic disorders or Tourette’s syndrome, atomoxetine may be a first choice agent given that the stimulants may exacerbate such pre-existing conditions. In cases where someone with a history of drug abuse resides with the child or where the child or adolescent has a substance use disorder or history of such, atomoxetine may be the preferred agent because of its absence of abuse potential. Where the child or teen with ADHD may reside in a dormitory for their school year (boarding school or college), atomoxetine might be considered ahead of stimulants because of its lack of potential for diversion to dorm-mates for their own recreational use. Obviously, in cases where prior stimulant response has been poor, atomoxetine would be the next medication in line to consider, well ahead of the tricyclic antidepressants or antihypertensive agents like clonidine that have a greater potential for more serious side effects. And because atomoxetine does not adversely impact sleep onset, it should be considered as an alternative to stimulants where stimulant-induced insomnia is significantly problematic or in cases where sleep problems are pre-existing. Also, in cases where parents are concerned about the use of a Schedule II agent in the management of their child’s behavior – often as a consequence of adverse publicity in the popular media against Ritalin and other stimulants – atomoxetine may prove useful given its unscheduled status and hence greater acceptability among such consumers. However, where there exists an urgent need to gain control over disruptive, hyperactive-impulsive, or otherwise externalizing behavior due to imminent adverse consequences (school suspension, potential abuse of the child by caregivers, etc.), or where none of these pre-existing conditions are problematic, then stimulants would be the first choice agent due to the shorter titration period and apparently greater rapidity of an onset of a therapeutic response.

Atomoxetine is prescribed by weight in young children (mg/kg) (Conner, 2005b). In children and adolescents up to 70 kg body weight atomoxetine is initiated at a total daily dose of 0.5 mg/kg. Dose titration occurs at a minimum of every 3 days to a target total daily dose of approximately 1.2 mg/kg administered either as a single daily dose in the morning or as evenly divided doses in the morning and late afternoon or early evening. In children, adolescents, and adults who weigh more than 70 kg, atomoxetine is initiated at a total daily dose of 40 mg and increased after a minimum of 3 days to a target total daily dose of approximately 80 mg, given in the morning as two evenly divided daily doses. If no treatment benefit occurs after 2-4 weeks, the dose of atomoxetine may be increased to a maximum of 100 mg/day. The total daily dose of atomoxetine in children and adolescents should not exceed 1.4 mg/kg/day or 100 mg, whichever is less. During atomoxetine initiation and dose titration, contact with the prescribing physician should occur regularly. Pulse and blood pressure should be assessed on full dose. Height and weight should be followed twice yearly.

The side effects of atomoxetine are well-documented and generally benign, like the stimulants. They include chiefly sedation, gastrointestinal disturbance (nausea), decreased appetite, and upper abdominal pain. Some weight loss may be present over the first 2-4 months of atomoxetine treatment but tends not to persist beyond the first year. Slight increases in blood pressure and heart rate may occur, as they do with the stimulants, but are typically benign unless hypertension was a pre-existing problem, in which case neither stimulants nor atomoxetine should be considered.

To summarize, atomoxetine should be considered at least a grade A treatment in terms of efficacy and safety.

Antihypertensive Medications

Two drugs originally marketed as alpha-andrenergic agonists for treating hypertension, clonidine (Catapres) and guanfacine (Tenex), have become increasingly popular for the off-label treatment of ADHD (Connor, 2005b). These drugs have primarily been used as an alternative or adjunctive medication to stimulants such as methylphenidate. Studies have found that clonidine is superior to placebo in reducing ADHD symptoms and conduct problems (see Connor, Fletcher, & Swanson, 1999 for a meta-analysis; Hazell & Stuart, 2003; Pliszka et al., 2000). Empirical support for guanfacine exists (Scahill et al., 2001) but was similarly weak, earning these drugs a grade of C (i.e., efficacy based primarily on studies with serious methodological limitations and on a few randomized, controlled studies).

The most commonly reported side effect of clonidine is drowsiness, which occurs in about 50% of cases (Connor, 2005b; Graydanus et al, 2003). This side effect is sometimes used to therapeutic advantage with children with ADHD who have difficulty falling asleep or exhibit symptom rebound after a day of taking stimulant medication. Safety concerns about the combination of stimulants and clonidine appear to be diminishing over time following some reports of deaths in the mid-1990s that were associated with, but not proven to be caused by, combing clonidine and methylphenidate. Nevertheless, there are lingering concerns about the potential electrocardiographic effects of clonidine and the potential to worsen pre-existing cardiac arrhythmias (Greydanus et al., 2003).

The need for electrocardiograms at baseline, every dose change, and every six months when taking clonidine is a serious threat to the practicality of this medication. Also, due to possible decreased glucose tolerance, monitoring of blood glucose levels at least every six months is warranted (Greydanus et al., 2003). Furthermore, there are many unpleasant side effects reported with this drug that may hamper compliance including headache, dry mouth, itchy eyes, weight gain, dizziness, and postural hypotension. In some cases treatment with clonidine may create new problems that mimic psychiatric disorders. For instance, there have been some reports of irritability, dysphoria, and attention impairment while on clonidine. Although side effects may be reduced and compliance may be improved with the availability of a patch, this delivery system runs the risk of causing a rash at the site of the patch. Finally, rapid withdrawal from clonidine may cause serious problems with symptom rebound and tachycardia. Adverse effects of guanfacine appear to be similar to those of clonidine, with possibly less sedation and more agitation and headaches (Connor, 2005b; Greydanus et al., 2003; Werry & Aman, 1999).

To summarize, compared to all other medications evaluated in this course, the risk to benefit ratio of clonidine appears to be poor, while that for guanfacine is better given its lower likelihood of affecting cardiovascular functioning.

In 2009, the FDA approved the use of an extended release form of guanfacine (Guanfacine XR or Intuniv®), an antihypertensive drug, for the management of ADHD symptoms in children. This approval was based upon recent large scale studies of the safety and effectiveness of this form of the medication in children with ADHD. That research indicates that the medication may not be as effective for the management of ADHD symptoms as are stimulants but is more effective than placebo medication. The drug may be particularly useful for the management of ADHD with comorbid mood disorders or irritability where emotional dysregulation is a prominent feature. Approval is likely to follow soon for use of the drug with adolescents and adults with ADHD.

Bupropion (Wellbutrin®)

This medication is FDA approved for treatment of depression and nicotine addiction in adults but not for ADHD. It is believed to be effective in improving attention, reducing irritability, and ameliorating depression (Connor, 2005b; Greydanus et al., 2003); however, the research on bupropion’s effects in children and adolescents should be treated with caution due to the small number of studies (two) and small sample size in those studies (a total of 36). One study with 15 adolescents found that bupropion was equivalent to methylphenidate in treating ADHD (Barrickman et al., 1995). A single-blind study with 24 adolescents with ADHD and comorbid depression found statistically significant changes in parent and child ratings, but not teacher ratings (see Connor, 2005b). In this study clinically significant change (by unblinded clinicians) was reported for about 60% of the participants. An open label study found that bupropion might be effective with adolescents with ADHD and substance use disorders (Riggs et al., 1998). Although these results are promising, and suggest that bupropion might be effective in treating ADHD in adolescents with or without comorbid depression or substance abuse, the current state of the literature supports only a grade of C for the quality of research on bupropion as a treatment for adolescents with ADHD. At best, a grade of D is warranted for treatment of children with ADHD.

Adverse effects of bupropion are rare and include nausea, anorexia, restlessness, agitation, drowsiness, headaches, exacerbation of tics, and seizures. Due to these side effects, bupropion is not appropriate for treating individuals with epilepsy or eating disorders. The risk of seizure can presumably be controlled through slow titration, taking doses more than 8 hours apart, and using sustained release formulations (Greydanus et al., 2003). Compared to tricyclic antidepressants, bupripion overdose is tolerated far better. Furthermore, there appear to be few drug to drug interactions with bupropion. The availability of once a day dosing (Wellbutrin XL) simplifies the administration of this drug.Thus, pending further research, bupropion could be the drug of choice for treating adolescents with comorbid depression or substance use disorders. However, due to the limited research, bupropion is the third or fourth medication to consider when the primary presenting problem is ADHD.

Table 1

Stimulant Preparations for ADHD

Stimulant Preparations for ADHD

Active Agent
Dose Availability

Dosing Schedule

Immediate-Release for 4- to 6- hour coverage



Adderall® Tablets

Neutral sulfate salts of dextroamphetamine saccharate and d,l-amphetamine aspartate

5, 7.5, 10, 12.5 15, 20, 30 mg

Bid to tid

Desoxyn® Tablets**
Methamphetamine HCL
5 mg
Bid to tid
Dexedrine® Tablets
Dextroamphetamine sulfate
5 mg
Bid to tid
Dextrostat® Tablets
Dextroamphetamine sulfate
5, 10 mg
Bid to tid
Focalin™ Tablets
Dexmethylphenidate HCL
2.5, 5, 10 mg
Bid to tid
Ritalin® HCL Tablets
Methylphenidate HCL
5, 10, 20 mg
Bid to tid
Intermediate-Acting for 8 hour coverage


Dexedrine® Spansule
Dextroamphetamine sustained release
5, 10,15 mg


Metadate® CD
Methylphenidate HCL extended release
20 mg


Metadate® ER
Methylphenidate HCL extended release
10, 20 mg


Ritalin-Slow Release

Methylphenidate HCL sustained release

20 mg


Quillivant XR

Methylphenidate in liquid suspension

5 mg/ml


Long-Acting for 10 to 12 hour coverage
Adderall XR® Capsules
Neutral salts of dextroamphetamine and amphetamine with dextroamphetamine saccharate and d,l-amphetamine aspartate monohydrate extended release
5, 10, 15, 20, 25, 30 mg
Concerta™ Tablets
Methylphenidate HCL extended release
18, 27, 36, 54 mg
20, 30, 40, 50, 60 70 mg

** High abuse potential

Table adapted and updated from Connor, D. F. (2006). Stimulants. In R. A. Barkley (Ed.), Attention deficit hyperactivity disorder: A handbook for diagnosis and treatment (3rd ed.). New York: Guilford. Copyright Guilford Publications, 2006. Reprinted with permission.

Behavioral Interventions – A Rationale

These methods were initially employed with ADHD children largely on an atheoretical basis – their success with mentally retarded and other developmentally disabled populations (e.g., autism) encouraged their use with behaviorally disordered children. The more recent theory of ADHD as a problem in response inhibition, EF, and self-regulation, however, and the secondary consequences this may create for their poor self-motivation to persist at assigned tasks (Barkley, 1997d, 2012c) provides a persuasive theoretically-based rationale for employing behavioral interventions with ADHD. ADHD appears to be a developmental delay in the self-regulation of behavior by internal means of representing information (working memory) and motivating goal-directed behavior. If so, then interventions that provide for more externally represented information that can serve to prompt and guide behavior and that enhance motivation via an increased density of external consequences, greater immediacy in their timing, and greater salience for the child would be useful, at least for symptomatic reduction in some settings and tasks. Such procedures for the artificial manipulation of antecedent and consequent events are precisely those provided by the behavior therapies.

A logical extension of this argument holds, however, that such socially arranged means of addressing this neurologically-based disorder of self-regulation would not likely alter the underlying neurophysiological basis for it. These techniques must be employed across situations over extended time intervals (months to years) much as prosthetic devices (e.g., hearing aides, mechanical limbs, etc.) are employed to compensate for physically handicapping conditions. Premature removal of the socially arranged stimulus prompts and motivational programs would predictably result in an eventual return to pre-treatment levels of the behavioral symptoms, just as removal of a ramp that permits physically disabled persons in wheelchairs to enter public buildings would cease allowing them such successful entry. Also, use of the behavioral techniques in only one environment would be unlikely to affect rates of ADHD symptoms in other, untreated settings unless generalization had been intentionally programmed to occur across such settings. The research reviewed below for the various behavioral techniques seems to support this interpretation.

Important to note from Barkley’s theoretical stance is that behavioral interventions are not being done chiefly to increase skills or information, as if children with ADHD were ignorant of them, but are being done to prompt internally (mentally) mediated information that can guide performance as well as enhance the motivation of these children to show what they already know. From this perspective, ADHD is a disorder of performance, not of knowledge of skills, and thus behavioral interventions are used to cue the use of those skills at key points of performance in natural settings and to motivate their display through the use of artificial consequences that ordinarily do not exist at those points of performance in natural settings.

Direct Application of Behavior Therapy Methods in the Laboratory

A number of early studies evaluated the effects of reinforcement and punishment, usually response cost, on the behavior and cognitive performance of ADHD children. These studies usually indicated that the performance of ADHD children on lab tasks measuring vigilance or impulse control can be immediately and significantly improved by the use of contingent consequences (Firestone & Douglas, 1975, 1977; Patterson, 1965; Worland, 1976). In some cases, the behavior of ADHD children approximated that of normal control children. For example, Paniagua (1987) evaluated the contribution of stimulus control to the management of ADHD children. Using a method known as correspondence training, he has attempted to establish greater control over ADHD symptoms by commands and rules previously stated publicly by the children. Correspondence refers to the degree of concordance between public statements by children as to what they will do and the actual behavior they subsequently display in that setting – in essence, the degree of agreement between "saying" and "doing".

In this paradigm, ADHD children are requested to publicly state how they will behave in an immediately subsequent situation. Their behavior in that situation is then observed after which they are reinforced or punished for the degree of correspondence. Results have suggested that under such conditions, ADHD children significantly reduce their levels of inattention and overactivity during task performance, and levels of aggressive behavior during peer interactions. However, work by Hayes et al. (1985) suggested that such self-statements must be publicly made in order to be effective because they serve as a form of public goal setting for which social consequences can be made contingent. Future research needs to show that it is the children's own statements which are serving as the controlling stimuli in such paradigms rather than the presence of the examiner during the task.

However, none of these studies examined the degree to which such changes generalized to the natural environments of the children, calling into question the clinical effectiveness of such an approach. It is highly unlikely that behavioral techniques implemented only in the clinic or laboratory would carry over into the home or school settings of these children without formal programming for such generalization. As a result, there has been no further research interest shown in the direct training of children with ADHD using behavioral means in clinical or laboratory settings. This early work remains of historical significance however. It presaged and instigated later efforts to train parents and teachers in the application of behavioral methods in home and school settings, thereby partly addressing the problem of generalization of treatment effects that limited these early laboratory demonstrations.

Direct Training of Attention in the Clinic

A few neuropsychologists have explored the direct training of attentional skills for children with ADHD in clinical settings founded largely on the success of cognitive rehabilitation and training programs for head-injured or other neurologically impaired populations (Kerns, Eso, & Thomson, 1999). One specific cognitive rehabilitation protocol for attention training is the Attention Process Training (APT) system developed for brain-injured adults by Sohlberg and Mateer (1989). The APT involves a series of cards containing drawings of family situations involving different ages, sexes, dress, and social circumstances and using different colors such that the cards can be sorted on a variety of stimulus characteristics as specified by the trainer. The APT also includes auditory tasks of a similar nature that require trainees to attend to specified stimuli on the sound track. The tasks become faster over the training sessions and include distracting stimuli. Children were reinforced for meeting specific success criterion during each session but no strategies were taught to the children as to how to attend to and succeed at these tasks. Studies to date have involved very small samples of children with ADHD (6-14) with training occurring 2-4 times per week for ½-2 hour sessions over periods of 5 to 18 weeks (Kerns et al., 1999; Semrud-Clikeman et al., 1998; Williams, 1989). These studies typically find significant improvement on both the attention training tasks (as one might expect) and on untrained comparable tasks used to assess attention given by the same examiners in the clinic lab setting. But none have been able to document significant improvements on measures of academic efficiency given in the same clinic or in parent or teacher rating scales of ADHD symptoms or related behavior at home or school. Hence, it appears that training effects are limited to the training environment and tasks with no evidence to date of generalization to untreated settings or more ecologically important activities (academic performance).

Training Parents in Child Behavior Management Methods

A plethora of research exists on parent training in child behavior modification (Kazdin, 1997; see Comer et al., 2012 and earlier Serketich & Dumas, 1996 for meta-analyses) primarily conducted with children having conduct or disruptive behavior problems. More recent studies have shown the behavioral parent training (BPT) programs to be effective for such children whether or not they have co-occurring attentional/hyperactive difficulties (Bor, Sanders, & Markie-Dadds, 2002; Comer et al., 2012; Hartman, Stage, & Webster-Stratton, 2003). A growing number of studies have examined the efficacy of this approach with children specifically diagnosed as ADHD or having high levels of hyperactive or ADHD symptoms (Chronis, Chacko, Fabiano, Wymbs, & Pelham, 2004; Lee et al., 2012). What research exists (Anastopoulos, Guevremont, Shelton, & DuPaul, 1992; Chronis et al., 2004; Sonuga-Barke, Daley, Thompson, Laver-Bradbury, & Weeks, 2001; Sonuga-Barke et al., 2002; Strayhorn & Weidman, 1989, 1991) supports the clinical efficacy of BPT with ADHD children but with some apparent caveats.

One such caveat is that most studies were of a short duration and did not examine for either generalization of treatment effects to no-treatment settings or for maintenance of treatment effects once training was discontinued (Kazdin, 1997; Eyberg, Edwards, Boggs, & Foote, 1998). One of the few studies to conduct a follow-up re-evaluation one year after treatment, however, found that the families receiving BPT were no longer different from the control group, although the child’s school behavior was rated by teachers as significantly better than the control group (Strayhorn & Weidman, 1991). Another caveat is that the studies to date have been largely university or medical school based programs implemented by expert trainers, thus attesting to the efficacy of BPT but with little research on its effectiveness in more natural clinical service delivery settings.

Another caveat is that these training programs produce larger effects on disruptive, defiant, or oppositional behavior and smaller yet still significant effects on ADHD symptoms (Comer et al., 2012; Lee et al., 2012).  Thus, BPT could hardly be considered a stand-alone treatment for children with ADHD. 

A further caveat is the high genetic contribution to the disorder insures that a number of parents undergoing training may have ADHD themselves. One study has found that maternal ADHD can significantly limit the improvement shown by children having ADHD during a BPT (Sonuga-Barke, Daley, & Thompson, 2002). Hence the treatment of parental ADHD may be a prerequisite to successful training outcomes for children with ADHD. Other forms of parental psychological maladjustment, such as maternal depression, antisocial personality and drug use, parenting stress, or marital distress, exist with greater frequency among parents of children with ADHD, could likewise limit the success of behavioral parent training, but have been largely unstudied in families with ADHD children (Chronis et al., 2004).

Studies that supplement basic BPT with additional psychological treatments designed to address such parental problems in children with ODD have shown mixed success at doing so, with apparently most families benefiting from the supplemented intervention whether or not they had these pre-existing parental problems (Chronis et al., 2004). Single-parent status and father participation in treatment have also been shown to limit the effectiveness of BPT with ODD and disruptive children but have not been directly investigated in children specifically having ADHD.

A third caveat is that parent training appears in some studies to result in more dramatic improvements in child oppositional behavior rather than in ADHD symptoms specifically, suggesting that the treatment is most useful where parent-child conflict exists in families having children with ADHD (Anastopoulos et al., 1993). Such findings may be associated with children of elementary school age or older whereas studies of preschool children with ADHD have found significant improvements in symptoms of ADHD specifically (Sonuga-Barke et al., 2001; Sonuga-Barke et al., 2002).

Those treatment techniques used to date have primarily consisted of training parents in general contingency management tactics, such as contingent application of reinforcement or punishment following appropriate/inappropriate behaviors. Reinforcement procedures have typically relied on praise, privileges, or tokens while punishment methods have usually been loss of positive attention, privileges, or tokens or formal time out from reinforcement.

Several similar, though not identical, parent training programs have been studied with children having ADHD, including Cunningham’s COPE program (Community Parent Education Program; Cunningham, Bremmer, & Secord, 1997), The Incredible Years by Webster-Stratton (1992), Eyberg’s Parent-Child Interaction Therapy (Eyberg & Boggs, 1998), and Barkley’s program, Defiant Children ( Barkley, 1997a), among others. The core methods taught in these programs are quite similar, not surprisingly, since all of the developers trained either directly with Constance Hanf at the Oregon Health Sciences University or with one of her former students. Hanf developed one of the original parent training programs for disruptive children based largely on two procedures, those being enhanced parental attention to compliant child behavior coupled with immediate time out for noncompliant behavior (see Forehand & McMahon, 1981). Though these subsequent programs vary in their format from this original program and in the procedures they have added to supplement the basic two steps, all are founded on a social learning model of disruptive child behavior (disrupted parenting and social coercion). And all have demonstrated efficacy for disruptive children, including those with ADHD (Chronis et al., 2004).

The rationale for Barkley’s version as applied to children with ADHD, however, is twofold and is not based solely on social learning theory. First, Barkley hypothesized that ADHD children may have deficits in self-regulation and executive functioning, and specifically in rule-governed behavior, or the stimulus control of behavior by commands, rules, and self-directed speech (1997d). Unlike a similar theory by Willis and Lovaas (1977), this one does not stipulate that the problem has arisen due to poor child management by parents but instead proposes a neurophysiological deficiency underlying the problem with rules. As noted earlier when discussing a theoretical rationale for behavior therapy, children with ADHD have limitations in internally represented information and motivation that instructs, guides, and supports behavior and therefore require more externally represented information and artificially arranged consequences to compensate for these executive deficits. Consequently, parents are going to need to use more explicit, systematic, externalized, and compelling forms of presenting rules and instructions to ADHD children and providing consequences for their compliance with them than are likely to be needed with normal children.

A second unique feature of Barkley’s approach to parent training is that there exists a considerable overlap of oppositional/defiant behavior with clinic-referred ADHD children and such children are recognized to have poorer adolescent and young adult outcomes (Hinshaw, 1987; Paternite & Loney, 1980; Weiss & Hechtman, 1993). ODD is recognized to at least partially originate in disrupted parenting and coercive family interactions (Barkley, 2013). Hence, appropriate training of parents must be provided for the oppositional/defiant behaviors associated with ADHD in such cases. The most useful vehicle for accomplishing both purposes seems to be training parents in behavioral techniques applied contingently for compliance or noncompliance (Barkley, 1997a, 2006).

The BPT program by Barkley, for instance, consists of 10 steps, with 1 to 2 hour weekly training sessions provided either to individual families or in groups. Each step is described in detail elsewhere (Barkley, 1997a, 2013a), but is briefly presented below:

1. Review of Information on ADHD

In the first session, the therapist provides a succinct overview of the nature, developmental course, prognosis, and etiologies of ADHD. Providing the parents with additional reading materials, such as a book for parents (Barkley, 2001, 2013b), can be a useful addition to this session. Professional DVDs are also available (Barkley, 1992a, 1992b) that present such an overview and can be loaned to parents for review at home and sharing with relatives or teachers, as needed. Such a session is essential in parent training to dispel a number of misconceptions parents often have about ADHD in children. Research suggests that just this provision of information can result not only in improved knowledge of parents about ADHD but also in improved parental perceptions of the degree of deviance of their child’s behavioral difficulties (Andrews, Swank, Foorman, & Fletcher, 1995).

2. The Causes of Oppositional/Defiant Behavior

Next, parents are provided with an in-depth discussion of those factors identified in past research as contributing to the development of defiant behavior in children. Essentially, four major contributors are discussed, these being: (a) Child characteristics, such as health, developmental disabilities, and temperament; (b) Parent characteristics similar to those described for the child; (c) Situational consequences for oppositional and coercive behavior; and (d) Stressful family events. Parents are taught that where problems exist in (a), (b), and (d), they increase the probability of children displaying bouts of coercive, defiant behavior. However, the consequences for such defiance, (c) above, seem to determine whether that behavior will be maintained or even increased in subsequent situations where commands and rules are given. Such behavior appears to primarily function as escape/avoidance learning in which oppositional behavior results in the child escaping from aversive parent interactions and task demands, negatively reinforcing the child's coercion. As in the first session, this content is covered so as to correct potential misconceptions which parents have about defiance (i.e., it is primarily attention-getting in nature). This session can be augmented by the use of two professional DVDs on the nature of oppositional defiant behavior and its management (Barkley, 1997e, 1997f).

3. Developing and Enhancing Parental Attention

Patterson (1965; 1982) has suggested that the value of verbal praise and social reinforcement to oppositional or hyperactive children is greatly reduced, making it weak as a reinforcer for compliance. In this session, parents are trained in more effective ways of attending to child behavior so as to enhance the value of their attention to their children. The technique consists of verbal narration and occasional positive statements to the child with attention being strategically deployed only when appropriate behaviors are displayed by the child. Parents are taught to reduce the amount of attention to inappropriate behaviors, including ignoring as much negative behavior as possible, while to greatly increasing their attention to ongoing prosocial and compliant child behaviors. This is a critical step because many parenting programs result in reductions of rates of negative behavior without corresponding increases in rates of positive behavior (Bor et al., 2002). One of the most effective results of parent training is to increase rates of positive behaviors that are incompatible with the negative behaviors the parents wish to terminate.

4. Attending to Child Compliance and Independent Play

This session extends the techniques developed in Session 3 to instances when parents issue direct commands to children. Parents are trained in methods of giving effective commands, such as reducing question-like commands (e.g., Why don't you pick-up your toys now?), increasing imperatives, eliminating setting activities which compete with task performance (e.g., television), reducing task complexity, etc. They are then encouraged to begin using a more effective commanding style and to pay immediate positive attention when compliance is initiated by the child. As part of this assignment, parents are asked to increase the frequency with which they give brief commands to their child this week and to reinforce each command obeyed. Research suggests that these brief commands are more likely to be obeyed thereby providing excellent training opportunities for attending to compliance. In this session, parents are also trained to provide more positive attention frequently and systematically when their children are engaged in nondisruptive activities while parents must be occupied with some other work or activity. Essentially, the method taught here amounts to a shaping procedure in which parents provide frequent praise and attention for progressively longer periods of child nondisruptive activities.

5. Establishing a Home Token Economy

As noted above in discussing the theoretical model of ADHD, children with the disorder may require more frequent, immediate, and salient consequences for appropriate behavior and compliance in order to maintain it. If this is correct, then instituting a home token economy is critical to addressing these difficulties with intrinsically generated and represented motivation by bringing more salient external consequences, more immediately, and more frequently to bear on child compliance than is typically the case. In establishing this program, the parents list most of the children's home responsibilities and privileges and then assign values of points or chips to each. The parents are encouraged to have at least 12 to 15 reinforcers on the menu so as to maintain the motivating properties of the program. Generally, plastic chips are used with children 8 or younger as they seem to value the tangible features of the token. For 9 year olds or older, points recorded in a notebook seem sufficient.

During the first week of this program, the parents are not to fine the child or remove points for misconduct. The program is for rewarding good behavior only. Parents are also asked to be liberal in awarding chips to children for even minor instances of appropriate conduct. However, chips are given only for obeying first requests. If a command must be repeated, it must still be obeyed but the opportunity to earn chips has been forfeited. Parents are also encouraged to give bonus chips for good attitude or emotional regulation in their children. For instance, if a command is obeyed quickly, without complaint, and with a positive attitude, parents may give the child additional chips beyond those typically given for that job. Where this is used, parents are to expressly note that the awarding of the additional chips is for a positive attitude. Families are encouraged to establish and maintain such programs for at least 6 to 8 weeks to allow for the newly developed interaction patterns spawned by such programs to become habitual patterns in dealing with child compliance.

6. Implementing Time Out for Noncompliance

Parents are now trained to use response cost (removal of points or chips) contingent on noncompliance. In addition, they are trained in an effective time-out-from-reinforcement technique for use with two serious forms of defiance which may continue to be problematic despite the use of the home token economy. These two misbehaviors are selected in consultation with the parents and typically involve a type of command or household rule which the child continues to defy despite parental use of previous treatment strategies. Time out is limited to these two forms of misconduct so as to keep it from being used excessively during the next week.

The time out procedure taught to parents often differs from that commonly used by them. First, the time out is to be implemented shortly after noncompliance by a child begins. Parents often wait until they are very upset with a child before instituting punishment, often repeating their commands frequently to a child in the interim. In this program, parents issue a command, wait 5 seconds, issue a warning, wait another 5 seconds, and then take the child to time out immediately should compliance not have begun to these commands or warnings. Second, children are not given control over the time out interval as they often are in many households. For instance, parents often place a child in time out then say the child can leave time out when they are quiet, ready to do as the parent asked, or when a timer signals the end of the interval. In each of these cases, determination as to when the time out interval ends is no longer under the parent’s control. This program teaches parents to simply tell the child to not leave the time out chair until the parent tells them to. Three conditions must be met by the child before time out ends and these are in a hierarchy: (1) The child must serve a minimum sentence in time out, usually 1 to 2 minutes for each year of their age; (2) the child must then become quiet for a brief period of time so as not to have disruption associated with the parents approaching the time out chair and talking to the child; and (3) the child must then agree to obey the command. Failure of the child to remain in time out until all three conditions are met is dealt with by additional punishment. The consequence is tailored to meet parental wishes but may consist of a fine within the home token system, extension of the time out interval an additional 5 or 10 minutes, or placement of the child in his or her bedroom. In the latter case, toys or other entertaining activities are previously removed from the bedroom and the door to the room may be closed and locked to preclude further escape from the punishment.

7. Extending Time Out to Additional Noncompliant Behaviors

In this session, no new material is taught to parents. Instead, any problems with previously implementing time out are reviewed and corrected. Parents may then extend their use of time out to one or two additional noncompliant behaviors with which the child may still have trouble.

8. Managing Noncompliance in Public Places

Parents are now taught to extrapolate their home management methods to troublesome public places, such as stores, church, restaurants, etc. Using a "think aloud-think ahead" paradigm, parents are taught to stop just before entering a public place, review two or three rules with the child which the child may previously have defied, explain to the child what reinforcers are available for obedience in the place, then explain what punishment may occur for disobedience, and finally assign the child an activity to perform during the outing. Parents then enter the public place and immediately begin attending to and reinforcing ongoing child compliance with the previously stated rules. Time out or response cost are used immediately for disobedience.

Time out in a public place may require slight modification from its use at home. For instance, parents may be taught to stand the child against the farthest wall from the central aisle of a store to serve as the time out location. If inconvenient, then taking the child to a restroom or having him face the side of a display cabinet may be adequate substitutes. If unavailable, then taking the child outside the building to face the front wall or returning to the car can be used for time out. When none of these locations seem appropriate, parents can be trained to use a delayed punishment contingency. In this case, the parent carries a small spiral notebook to the public place and, before entering the building, indicates that rule violations will be recorded in the book and the child will serve time out for them upon return home from this trip. This author encourages parents to keep a picture of the child sitting in time out at home with this notebook and to show it to the child before entering the public building. This serves as a reminder to the child of what may be in store should a rule be violated. Whenever time out is used in a public place, it need not be for as long an interval as at home. The author suggests that that half of the usual time out interval may be sufficient for public misbehavior given the richly reinforcing activities in public places from which the child has just been removed.

9. Improving Child School Behavior From Home

The Daily School Behavior Report Card: This session was designed to help parents assist their child’s teacher with the management of classroom behavior problems. The session focuses on training parents in the use of a home-based reward program in which children are evaluated on a daily school behavior report card by their teachers. This card serves as the means by which consequences later in the day will be dispensed at home for classroom conduct. The card can be designed to address class behavior, recess or free time behavior, or more specific behavioral or academic targets for any given child. The consequence provided at home typically consist of the rewarding or removal of tokens or points within the home token system as a function of the ratings the child has received from their teachers on this daily behavioral report card. To emphasize the importance of the school to home communication and to avoid escaping consequences if the child “loses” their school behavior report card, it is best to set up a contingency such that “no news is worse than bad news.” Thus, the most austere level in the contingency system should be when the card is incomplete or missing.

10. Managing Future Misconduct

By now, parents should have acquired an effective repertoire of child management techniques. The goal of this session is to get parents to think about how they might be implemented in the future if some other forms of noncompliance developed. The therapist challenges the parents with misbehaviors they have not seen yet and asks them to explain how they might use their recently acquired skills to manage these problems. Behavioral rehearsal (i.e., role plays) surrounding anticipated barriers to implementation of existing parenting plans or making modifications or innovations to deal with new and different behaviors are strongly recommended as a means to prepare for future misconduct.

One Month Review/Booster Session

In what is typically the final session, the concepts taught in earlier sessions are briefly reviewed, problems which have arisen in the last month are discussed, and plans made for their correction. Other sessions may be needed to deal with additional issues that persist but for most families, the previous 10 sessions appear adequate for improving rates of compliant behavior in ADHD children.

The program is intended for children ages 2 to 11 years where oppositional or defiant behavior is an issue. Studies examining the efficacy of this particular BPT with ADHD children have consistently reported significant improvements in child behavior as a function of the parents’ acquisition of these child management skills (Anastopoulos et al., 1993; Johnston, 1992; Pisterman, McGrath, Firestone, Goodman, Webster, & Mallory, 1989). Indeed, this was the program that was selected, and modified, for use in the MTA study, discussed below. Results suggest that up to 64 percent of families experience clinically significant change or recovery (normalization) of their child’s disruptive behavior as a consequence of this program (Anastopoulos et al., 1993; Sonuga-Barke et al., 2002). However, improvements in behavior may be more concentrated in the realm of aggressive and defiant child behavior than in inattentive-hyperactive symptoms (Comer et al., 2012; Johnston, 1992; Lee et al., 2012). All of these studies have relied on clinic-referred families most of whom sought the assistance of mental health professionals for their children.

In contrast to the results of research with such motivated families, This author and colleagues found that if such a clinic-based parent training program is offered to parents whose preschool children were identified at kindergarten enrollment as having significant levels of aggressive-hyperactive-impulsive behavior, most do not attend training or do not attend reliably and no treatment effect is evident (Barkley, Shelton, Crosswait, Moorehouse, Barrett et al., 2000). Moreover, no significant improvements in child behavior were found even among those who did attend at least some of the training sessions. Studies with disruptive children or those at high risk for externalizing behavior suggest that BPT may be more cost-effective, reach more severely disruptive children and more minority families, and possibly be more effective for them if they are provided as group training classes offered through neighborhood public schools in the evenings using para-professionals as trainers (Cunningham, Bremer, & Boyle, 1995; van de Wiel, Matthys, Kettanis, & Engeland, 2003). This might prove to be the case for children specifically having ADHD as well.

For teenagers with ADHD and oppositional behavior, there is little research on BPT. We have often recommended a family training program that includes the Problem Solving Communication Training Program (PSCT) developed by Robin and Foster (1989) combined with variations of Barkley’s BPT program. The efficacy of the Robin and Foster program used specifically with ADHD teenagers has been examined (Barkley, Guevremont, Anastopoulos, & Fletcher, 1992). This program was compared against the parent training program described above (Barkley, 1997a) that was modified somewhat for use with adolescents (e.g., token systems became point systems, time out was changed to grounding to the home, etc.). It was also compared against the family therapy program developed by Minuchin (Minuchin & Fishman, 1981). Families in each group received 8-10 sessions of therapy and multiple outcome measures of family conflict were collected, including videotaped parent-teen interactions. Results indicated that all three treatments produced statistically significant improvements in the various self-report ratings of family conflict but no significant improvements in the direct observations of parent-teen interactions. When statistics evaluating individual change and recovery were applied to these data, they revealed that only 5-30 percent of the families in these programs improved reliably from treatment and that only 5-20 percent had recovered (normalization) in their level of conflicts, with no significant differences among the groups in these reliable change and recovery percentages. Such results are quite disappointing and suggest that the power of treatment needs to be enhanced in various ways if it is to be of much value to most families of ADHD adolescents experiencing significant family conflict.

A subsequent study by this author and colleagues enhanced this treatment by increasing the number of sessions to 18, encouraging greater father involvement in therapy, and combining the behavioral parent training and PSCT programs together, among other changes (Barkley, Edwards et al., 2001). The study compared PSCT/BMT with PSCT alone using clinically referred teenagers having ADHD and ODD. Results were essentially the same as the initial study, with both treated groups showing improvement over time at the group level of analysis with there being no differences between them. But only 30 percent or fewer families in these groups were demonstrating clinically reliable change at the end of treatment. The combined PSCT/BMT approach was superior to the PSCT alone approach in just one respect, that being fewer dropouts from treatment. The program is available as a published clinical manual for practitioners wishing to implement this approach to therapy (Barkley, Edwards, & Robin, 1999).

To summarize the efficacy of training parents in behavior management methods, these methods receive a grade of B+ to A for use with elementary school aged children depending on which parenting approach is used. Compared to elementary school aged children, these approaches have not been as effective with preschoolers or adolescents with ADHD. Practical barriers, such as parental engagement, have seemed to limit studies with preschoolers. Innovations that improve the efficacy of parent-based interventions with teenagers with ADHD are sorely needed. Parent management training suffers from several practical limitations, such as the need for training providers, numerous demands placed on families, a commitment to use the methods consistently and persistently over a long period of time, and the need to tailor techniques in response to changing developmental and social factors. When used properly, these techniques can be very safe, but overly punitive or haphazard programs can actually make things worse. Paradoxically, some worsening of behavior (e.g., transient increases in tantrums when a parent starts ignoring the behavior) may suggest that parents are on the right track. This example illustrated the fact that the behavior methods that seem so simple at their basic conceptual level can be fraught with subtleties that confuse and frustrate parents. A limitation of these approaches is that many therapists do not have the skills and training to provide the sophisticated guidance that most parents need. Consequently, even though behavior methods can be used to get whales to jump through burning hoops, due to past failures with other providers many parents complain that their children’s behavior is completely impervious to change through behavioral methods. This is a significant limitation to the effectiveness of parent management training that must be handled delicately to recruit and retain parents in behavior management training. Intensive professional education and widespread public education to overcome ignorance and stigma associated with parent management training may be necessary for this to become truly grade A intervention for ADHD and related problems.

Training Teachers in Classroom Behavior Management

Somewhat more research has occurred on the application of behavior management methods in the classroom with ADHD children than with parent training. Moreover, there is a voluminous literature on the application of classroom management methods to disruptive child behaviors, many of which include the typical symptoms of ADHD. This research clearly indicates the effectiveness of behavioral techniques in the short-term treatment of academic performance problems in ADHD children.

A meta-analysis of the research literature on school interventions for ADHD was conducted that comprised 70 separate experiments of various within- and between-subjects designs as well as single-case designs (DuPaul & Eckert, 1997; Raggi & Chronis, 2006). The  Dupaul and Eckert review found an overall mean effect size for contingency management procedures of 0.60 for between-subject designs, nearly 1.00 for within-subject designs, and approximately 1.40 for single-case experimental designs. Interventions aimed at improving academic performance through the manipulation of the curriculum, antecedent conditions, or peer-tutoring produced approximately equal or greater effects sizes. In contrast, cognitive-behavioral treatments when used in the school setting were significantly less effective than these other two forms of interventions. Thus, despite some initial findings of rather limited impact of classroom behavior management on children with ADHD (Abikoff & Gittelman, 1984), later studies (Pelham & Hoza, 1996; Pelham, Wheeler, & Chronis, 1998; Carlson, Pelham, Milich, & Dixon, 1992), and the totality of the extant literature (DuPaul & Eckert, 1997, 1998; Raggi & Chronis, 2006) suggest that behavioral and academic interventions in the classroom can be effective in improving behavioral problems and academic performance in children with ADHD. The greatest and most reliable improvements across studies are evident with contingency management and peer-tutoring approaches while studies of curriculum modifications, strategy training, and other cognitive-behavioral approaches are less reliable (DuPaul & Eckert, 1997, 1998). Moreover, even with the most effective classroom interventions, the behavior of children with ADHD may not be fully normalized by these interventions.

As noted above in discussing laboratory applications of behavior therapy techniques, research suggests some promise in the use of stimulus control procedures with ADHD children, many of which can be readily adapted to the classroom. By reducing task length, "chunking" tasks into smaller units to fit more within the child's attention span, and setting quotas for the child to achieve within shorter time intervals some stimulus control methods may increase the success of the ADHD child with academic work (see Ayllon & Rosenbaum, 1977; DuPaul & Stoner, 2003; Pffifner, DuPaul, & Barkley, 2005). As Zentall (1985) has already documented the use of increased stimulation within the task (e.g., color, shape, texture, rate of stimulus presentation) may enhance attention to academic tasks in ADHD children. Teaching styles may play an important role in how well ADHD children attend to lectures by a teacher. More vibrant, enthusiastic teachers who move about more, engage children frequently while teaching, and allow greater participation of the children in the teaching activity may increase sustained attention to the task at hand. Zentall has also shown that permitting ADHD children to move or participate motorically while learning a task may improve attention and performance. The use of written, displayed rules and timers for setting task time limits, as already described, may further benefit ADHD children in the classroom.

A number of studies have also shown that the contingent application of reinforcers for reduced activity level or increased sustained attention can rapidly alter the levels of these ADHD symptoms (DuPaul & Eckert, 1997, 1998; DuPaul & Stoner, 2003). Usually, these programs incorporate token rewards as some research suggests that praise may not be sufficient to increase or maintain normal levels of on-task behavior in hyperactive children (Pfiffner, DuPaul, & Barkley, 2005). Early studies in this area showed that group administered rewards, where all children in class receive a reward contingent on the performance of one child, are as effective as individually administered rewards ( O'Leary, Pelham, Rosenbaum, & Price, 1976). One of the problems arising in such research, however, was the demonstration that simply reinforcing greater on-task behavior and decreased activity level did not necessarily translate into increased work productivity or accuracy (Marholin & Steinman, 1977). Since the latter are the ultimate goals of behavioral intervention in the classroom, these results were somewhat dismaying. Research now suggests that reinforcing the products of classroom behavior (i.e., number and accuracy of problems completed) not only results in increased productivity and accuracy but also indirectly in declines in off-task and hyperactive behavior (see DuPaul & Stoner, 2003).

A serious limitation to these promising results has been the lack of follow-up on the maintenance of these treatment gains over time. In addition, none of these studies examined whether generalization of behavioral control occurred in other school settings where no treatment procedures were in effect. Other studies employing a mixture of cognitive-behavioral and contingency management techniques have failed to find such generalization with ADHD children (Barkley, Copeland, & Sivage, 1980), suggesting that improvements derived from classroom management methods are quite situation specific and may not generalize or be maintained once treatment has been terminated.

The role of punishment in the management of classroom behavior in ADHD children has been less well studied. Pfiffner, O'Leary, Rosen, and Sanderson (1985) evaluated the effects of continuous and intermittent verbal reprimands and response cost on off-task classroom behaviors. They found that while each of these treatments significantly reduced disruptive and off-task behavior, the continuous use of response cost (loss of recess time) was most effective. Ayllon and Rosenbaum (1977) also report on the initial success of adding response cost contingencies to an ongoing classroom token economy. However, after less than one week, disruptive behavior returned to baseline levels despite the punishment contingency.

In a later paper, Pfiffner and O'Leary (1987) determined that the sole use of positive reinforcement for controlling ADHD behaviors in the classroom was not sufficient to maintain improved behavior in these children unless punishment in the form of response cost was added to the program. The addition of response cost further increased rates of on-task behavior and academic accuracy. These gains in behavior could then be maintained by an all-positive program once the response cost procedure was gradually withdrawn. However, abrupt withdrawal of the punishment contingency resulted in declines in on-task behavior and accuracy suggesting that the manner in which response cost techniques are implemented and then faded out of classroom management programs is important in the maintenance of initial treatment gains. In general, the efficacy of response cost procedures with ADHD children has been well-documented (DuPaul, Guevremont, & Barkley, 1992; Gordon, Thomason, & Cooper, 1990; Rapport, Murphy, & Bailey, 1982).

What conclusions can be drawn from this literature indicate that contingency management methods can produce immediate, significant, short-term improvement in the behavior, productivity, and accuracy of ADHD children in the classroom. Secondary or tangible reinforcers are more effective in reducing disruptive behavior and increasing performance than are attention or other social reinforcers. The use of positive reinforcement programs alone does not seem to result in as much improvement nor does it maintain that improvement over time as well as does the combination of token reinforcement systems with punishment, such as response cost (i.e. removal of tokens or privileges). Such findings would be expected from the theories of ADHD discussed earlier which suggest a decreased power to self-regulate motivation and a delay in the development of internalized speech and the rule-governed behavior it affords in children with this disorder. What little evidence there is, however, suggests that treatment gains are unlikely to be maintained in these children once treatment has been withdrawn, and that improvements in behavior probably do not generalize to other settings where no treatment is in effect.

Another promising method deserving of further evaluation is the use of home-based contingencies for in-class behavior and performance. Atkinson and Forehand (1979) reviewed this literature initially and found that the method offers some usefulness for managing disruptive classroom behavior. Later studies have borne out this promise (Fabiano et al., 2010). As discussed above under the author’s parent training program (Barkley,2013a), the method involves having a teacher rate a child's daily school performance, either one or more times throughout a school day. These ratings are then sent home with the child for review by the parents. The parents then dispense rewards and punishments (usually response cost) at home contingent upon the content of these daily ratings. O'Leary, Pelham, Rosenbaum, and Price (1976) employed this procedure for ten weeks with nine hyperkinetic children and documented significant improvements on teacher ratings of classroom conduct and hyperkinesis as compared to a no treatment control group. Others have similarly found such home-school behavioral report cards to be useful, either alone or in combination with parent and teacher training in behavior management, in the treatment of ADHD children (Ayllon, Garber, & Pisor, 1975; O'Leary & Pelham, 1978; Pelham, Schnedler, Bender, Nilsson, Miller, et al., 1988; The MTA Group, 1999a).

Some innovative treatment approaches, exploratory in nature, are of note. Hook and DuPaul (1999) evaluated the beneficial effects of parent tutoring on the reading performance of 4 students with ADHD using a multiple baseline across participants design. Parental tutoring of the students in reading resulted in improved reading performances both at home and in school, and stable or improved attitudes toward reading, with two of the three students reaching normal levels of reading performance at least once during the training. However, overall, these ADHD cases did not reach average performance levels consistently.

Ota and DuPaul (2002) examined the effects of using game-like math software to supplement mathematics instruction in 3 students with ADHD again using a multiple baseline across participants design. Behavioral observations and curriculum-based math probes revealed significant improvements in math performance from the software supplementation. Both of these studies employing small-scale single-case designs offer some promise of additional supplemental interventions to those discussed above. And both are now in need of replication with larger samples, a more diverse ADHD population, and examination of clinical (educational) effectiveness, not just efficacy.

Other recent innovative interventions for ADHD children in school settings which have some positive effects, include consultation-based support provided to school teachers by school psychologists and other qualified staff (Jitendra et al., 2007); a cost effective approach that deserves further study.  The Challenging Horizons Program (Evans et al., 2011) uses well-trained paraprofessionals to provide multi-method behavioral treatment, tutoring, teacher consultation, and parent training using a format of group meetings with ADHD teens after school several times per week.  The program is instituted at the school thus enhancing the likelihood of teen participation.  Results to date have been impressive.  Power et al. (2012) combined (1) parent and teacher behavioral consultations, (2) daily behavior report cards, and (3) behavioral homework interventions into a Family-School Success intervention.  Results showed significant benefits for students with ADHD in grades 2-6.  The use of peers as tutors has likewise shown some promise in improving the academic performance of students with behavioral disorders (Spencer, 2006).

To summarize the efficacy of classroom focused contingency management, this should be considered a grade A or B+ treatment depending on which techniques are used. Cognitive techniques and skills training appear to be failed interventions unless the techniques or skills are specifically reinforced at the point of performance. Behavior modification methods that use a combination of reward and punishment seem to have the best results, but there is still some uncertainty about which methods are best and for whom. It does seem to be clear that effects of these interventions are transient and need to be implemented over a very long term. Approaches that train parents to set up reinforcement contingencies based on teacher to parent communication (e.g., daily behavior reports) may be the best hope for sustained intervention across the life span. All of the significant barriers to the effectiveness of behavioral management discussed previously for parents apply to teachers. There is widespread ignorance, neglect, or poor implementation of these methods. Intensive efforts to educate and support parents and teachers in the use of these methods could make a huge difference in the functioning of many children with learning or behavior problems at school. Unfortunately, even school psychologists may not have the time or expertise to assist in setting up proper behavior management programs, so there is a quite a bit of work to be done to make this a grade A intervention that is widely available.

What follows is a list of recommendations, most from research, that can be used to help manage ADHD in school settings for children and teens with the disorder apart from or in addition to using ADHD medications.  Behind these recommendations lie 9 principles that need to be considered in planning management programs for children and teens with ADHD in school settings (Pfiffner, Barkley, & DuPaul, 2006).  Here are those 9 principles followed by the 80 suggested accommodations.

     1.  Rules and instructions provided to children with ADHD must be clear, brief, and often delivered through more visible and external modes of presentation than is required for the management of normal children.  Stating directions clearly, having the child repeat them out loud, having the child utter them softly to themselves while following through on the instruction, and displaying sets of rules or rule‑prompts (e.g. stop signs, big eyes, big ears for "stop, look, and listen" reminders) prominently throughout the classroom are essential to proper management of ADHD children.  Relying on the child's recollection of the rules as well as upon purely verbal reminders is often ineffective. 
     2.  Consequences used to manage the behavior of ADHD children must be delivered swiftly and more immediately than is needed for normal children.  Delays in consequences greatly degrade their efficacy for children with ADHD.  As will be noted throughout this chapter, the timing and strategic application of consequences with children with ADHD must be more systematic and is far more crucial to their management than in normal children.  This is not just true for rewards, but is especially so for punishment which can be kept mild and still effective by delivering it as quickly upon the misbehavior as possible – Swift, not harsh, justice is the essence of effective punishment.
     3.  Consequences must be delivered more frequently, not just more immediately, to children with ADHD in view of their motivational deficits. Behavioral tracking, or the ongoing adherence to rules after the rule has been stated and compliance initiated, appears to be problematic for children with ADHD. Frequent feedback or consequences for rule adherence seem helpful in maintaining appropriate degrees of tracking to rules over time.
     4.  The type of consequences used with children with ADHD must often be of a higher magnitude, or more powerful, than that needed to manage the behavior of normal children.  The relative insensitivity of children with ADHD to response consequences dictates that those chosen for inclusion in a behavior management program must have sufficient reinforcement value or magnitude to motivate children with ADHD to perform the desired behaviors.  Suffice it to say, then, that mere occasional praise or reprimands are simply not enough to effectively manage children with ADHD.
     5.  An appropriate and often richer degree of incentives must be provided within a setting or task to reinforce appropriate behavior before punishment can be implemented.  This means that punishment must remain within a relative balance with rewards or it is unlikely to succeed.  It is therefore imperative that powerful reinforcement programs be established first and instituted over 1 to 2 weeks before implementing punishment in order for the punishment, sparingly used, to be maximally effective.  Often children with ADHD will not improve with the use of response cost or time out if the availability of reinforcement is low in the classroom and hence removal from it is unlikely to be punitive.  "Positives before negatives" is the order of the day with children with ADHD.  When punishment fails, this is the first area which clinicians, consultations, or educators should explore for problems before instituting higher magnitude or more frequent punishment programs.
     6.  Those reinforcers or particular rewards which are employed must be changed or rotated more frequently with ADHD than normal children given the penchant of the former for more rapid habituation or satiation to response consequences, apparently rewards in particular.  This means that even though a particular reinforcer seems to be effective for the moment in motivating child compliance, it is likely that it will lose its reinforcement value more rapidly than normal over time.  Reward menus in classes, such as those used to back up token systems, must therefore be changed periodically, say every 2 to 3 weeks, to maintain the power of efficacy of the program in motivating appropriate child behavior.  Failure to do so is likely to result in the loss of power of the reward program and the premature abandonment of token technologies based on the false assumption that they simply will not work any longer.  Token systems can be maintained over an entire school year with minimal loss of power in the program provided that the reinforcers are changed frequently to accommodate to this problem of habituation.  Such rewards can be returned later to the program once they have been set aside for a while, often with the result that their reinforcement value appears to have been improved by their absence or unavailability.
     7.  Anticipation is the key with children with ADHD.  This means that teachers must be more mindful of planning ahead in managing children with this disorder, particularly during phases of transition across activities or classes, to insure that the children are cognizant of the shift in rules (and consequences) that is about to occur.  It is useful for teachers to take a moment to prompt a child to recall the rules of conduct in the upcoming situation, repeat them orally, and recall what the rewards and punishments will be in the impending situation before entering that activity or situation.  Think aloud, think ahead is the important message to educators here.  As noted later, by themselves such cognitive self‑instructions are unlikely to be of lasting benefit but when combined with contingency management procedures can be of considerable aide to the classroom management of ADHD children.
     8.  Children with ADHD must be held more publicly accountable for their behavior and goal-attainment than normal children.  The weaknesses in executive functioning associated with ADHD result in a child whose behavior is less regulated by internal information (mental representations) and less monitored via self-awareness than is the case in normal children.  Addressing such weaknesses requires that the child with ADHD be provided with more external cues about performance demands at key “points of performance” in the school, be monitored more closely by teachers, and be provided with consequences more often across the school day for behavioral control and goal attainment than would be the case in normal children.
     9.  Behavioral interventions, while successful, only work while they are being implemented and, even then, require continued monitoring and modification over time for maximal effectiveness.  One common scenario is that a student responds initially to a well-tailored program, but then over time, the response deteriorates; in other cases, a behavioral program may fail to modify the behavior at all.  This does not mean behavioral programs do not work.  Instead, such difficulties signal that the program needs to be modified.  It is likely that one of a number of common problems occurred, such as the rewards lost their value, the program was not implemented consistently, or the program was not based on a functional analysis of the factors related to the problem behavior.

Now for the 80+ suggested school accommodations for children and teens with ADHD:

Classroom Management:  Basic Considerations

Peer Tutoring (See DuPaul & Stoner, 2003; Spencer 2006).

Classroom Management: Increasing Incentives

Make Rules and Time Obvious and in Physical Forms

Train Self-Awareness

Possible Punishment Methods (check with school principal on district policies!)

Classroom Management: Tips for Teens

Figure 1.  Sample Daily School Behavior Report Card

Accommodation list adapted from Barkley, R. A. & Murphy, K. R. (2006).  Attention deficit hyperactivity disorder: A clinical workbook (3rd edition).  New York: Guilford.  Copyright by Guilford Publications.  Reprinted with permission.

Cognitive Behavioral Therapy

The provision of cognitive-behavioral treatment (CBT), or cognitive therapy, was felt previously to hold some promise for children with ADHD (Douglas, 1980; Kendall & Braswell, 1985; Meichenbaum & Goodman, 1971). Such treatment involves training children to give themselves instructions overtly in how to approach a task, strategies to employ during the task, and self-statements of evaluation and self-reinforcement at the end of the task. A few small-scale studies suggested some benefits to this form of treatment when used with children with ADHD (Fehlings, Roberts, Humphries, & Dawe, 1991). But CBT has been challenged as being seriously flawed from a conceptual (Vygotskian) point of view on which the treatment was initially founded (Diaz & Berk, 1995). Whether or not the self-statements of children with ADHD during task performance are actually deficient and in need of such correction is also open to question. And its efficacy for impulsive children or those with ADHD has been repeatedly questioned by the rather poor or limited results of empirical research (Abikoff, 1985, 1987; Abikoff & Gittelman, 1985).

Reviews of the CBT literature using meta-analyses have typically found the effect sizes to be only about a third of a standard deviation and, in many studies, even less than this (Baer & Nietzel, 1991; DuPaul & Eckert, 1997; Dush, Hirt, & Schroeder, 1989). While such treatment effects may at times rise to the level of statistical significance, they are nonetheless of only modest clinical importance and usually are to be found mainly on relatively circumscribed lab measures (Brown, Wynne, Borden, Clingerman, Geniesse, & Spunt, 1986) rather than more clinically important measures of functioning in natural settings.

A large-scale, well-controlled study of CBT conducted in the Minneapolis public school system found no effect on children with ADHD. The study involved substantial training of parents, teachers, and children, and two years of this multi-component intervention. But the researchers found no significant treatment effects on any of a variety of dependent measures at one year assessment with the exception of class observations of off-task/disruptive behavior, and no effects after two years of treatment (Bloomquist, August, & Ostrander, 1991; Braswell et al., 1997). Even the treatment effect on class observations was not maintained follow-up. Therefore, given the extant research findings of limited effect sizes in most clinical studies and the absence of treatment effects in the largest study, this treatment is given a grade of D and no further discussion of cognitive-behavioral treatments for ADHD will be presented here.

Social Skills Training

Early reviews of this social skills training (SST) as applied specifically to children with ADHD have been quite discouraging (Hinshaw, 1992; Hinshaw & Erhardt, 1991; Whalen & Henker, 1991). Children with ADHD certainly have serious difficulties in their social interactions with peers (Bagwell et al., 2001; Cunningham & Siegel, 1987; Erhardt & Hinshaw, 1994; Hubbard & Newcomb, 1991; Whalen & Henker, 1992). This seems to be especially so for that subgroup having significant levels of comorbid aggression (Hinshaw, 1992; Erhardt & Hinshaw, 1994), in which more than 50 percent of the variance in peer ratings of children whom they disliked was predicted by this behavior alone. As Hinshaw (1992) and later Hoza (2007) have summarized, the social interaction problems of children with ADHD are quite heterogeneous, and are not likely to respond to a treatment package that focuses only on social approach strategies and that treats all children with ADHD as if they shared common problems in their peer relationship difficulties. Nor is it especially clear at this time what the actual source of these peer difficulties happens to be or the mechanism by which it operates, with the exception of aggressive behavior as noted above. For instance, do children with ADHD actually lack the knowledge of proper social skills or is it that they know how to act with others but do not do so at the points of performance in social interactions where such skills would be useful to have performed? The theoretical model presented earlier would suggest that the latter is likely to be more of a problem than the former, at least for children having ADHD without significant aggression. Teaching additional skills is not so much the issue as is assisting them to perform the skills they have when it would be useful, that is at the point of performance where such skills are most likely to prove useful to the long-term social acceptance of the individual.

Those children with ADHD with comorbid aggression may well have additional problems with peer perceptions, particularly around the motives they attribute to others for their behavior, as well as in information processing about social interactions (Dodge, 1989; Milich & Dodge, 1984). This combination of both perceptual/information processing deficits along with problems performing appropriate social skills in social interactions with others may make children with ADHD with aggression particularly resistant to social skills training (Hinshaw, 1992).

Actual research on SST for ADHD contains rather mixed results. Early studies suggested that at-risk groups of children responded better (larger effect sizes) to such programs than did children with externalizing or disruptive behavioral problems (Beelman, Pfingsten, & Losel, 1994). Frankel and colleagues (Frankel, Myatt, & Cantwell, 1995) subsequently examined a social skills program for outpatient children with disruptive disorder, half of whom were diagnosed with ADHD. Nearly half of the children in the treated and wait-list control groups were receiving medication. Treatment assignment was not random, but associated with various clinical factors (date of intake, class starting date, class space available). Mothers’ ratings showed improved social skills in the treated compared to wait-list control children while teacher ratings revealed decreased aggression and withdrawal in the treated group but only among those who did not have oppositional defiant disorder. A later study by this same research team provided SST to children already receiving medication. Children with ADHD (N=35) were compared to a wait-list control group (N=12) of ADHD children who were also receiving medication. Parents were trained in strategies to help with generalization to the home setting. It is not clear how participants were assigned to treatment groups nor to what extent this study included data from participants in the earlier study. Significant benefits on both parent and teacher ratings were found in the treated compared to wait-list children with the presence of ODD having no moderating effects as it had in the earlier report (Frankel, Myatt, Cantwell, & Feinberg, 1997).

A study of young (ages 4-8 years) children having either ODD or CD, some of whom had ADHD (Webster-Stratton, Reid, & Hammond, 2001), showed positive benefits of a social skills and problem solving training program on conduct problem behavior, both on teacher ratings and on home behavioral observations, but not on parent ratings of conduct problems. Treatment effects remained at 1 year follow-up. Likewise, Pfiffner and McBurnett (1997) found evidence for the efficacy of a social skills training (SST) program for children with ADHD, but only on parent ratings and not teacher ratings. Children were randomly assigned to receive SST alone, SST supplemented with parent training in generalization strategies, or a wait-list control group. Both SST groups improved in parent rated social behavior relative to control children with improvements being maintained at a 4-month follow-up. But there was little evidence that these benefits generalized to the school setting. The addition of parent training in generalization strategies did not result in any additional benefits over social skills training alone.

All of these studies indicate some benefits of social skills training for conduct problem children, including those with ADHD, particularly when parent ratings serve as the outcome measure. Yet many suffer from significant limitations in their methods. Bear in mind that several of these studies did not employ randomized assignment to treated or untreated groups (the Frankel studies) and that parents were not blind to the intervention being received (all studies). All of the studies used wait-list control groups for their comparisons. It is not clear if such positive results would be found if efforts were made to control for therapist attention, as in attention placebo groups, or if alternative treatment approaches were also employed. Effects on teacher ratings of school social behavior are also not as encouraging as results from parent ratings but imply that some children in some studies may have demonstrated reduced social withdrawal and possibly aggression in school.

More sobering results are revealed in two other studies of ADHD samples. Sheridan and colleagues (Sheridan, Dee, Morgan, McCormack, & Walker, 1996) found no evidence that social skills training generalized to peer interactions in the school setting. A more recent study using a clinic-referred sample of 120 children having ADHD found no significant benefits for most ADHD children on a variety of measures of social functioning (Antshel & Remer, 2003). That ADHD subtype having primarily inattention may have improved in their assertion skills following treatment but not on any other measures of social interaction. Contrary to the Webster-Stratton study, children having comorbid ODD did not benefit as much as those ADHD children having no ODD. Moreover, some evidence of peer deviancy training was evident in the study. This refers to the peer reinforcement of aggressive and antisocial behavior among the children in the group such that children increase their levels of aggressive behavior as a result of group participation. Earlier studies did not examine this possibility of a detrimental impact of group social skills training. Such an adverse impact is certainly worthy of further examination in future research with children having ADHD.

At this time, social skills training for children with ADHD might be graded as a C, reflecting the inconsistent nature of the results, the limited number of studies using randomized assignment to treatment groups, the lack of blindness of parents and teachers to treatment conditions, the absence of attention-placebo or alternative treatment groups, the limited evidence of generalization to the school setting, and the fact that studies to date have mainly involved efficacy rather than effectiveness in actual clinical contexts. In addition to the questionable efficacy of social skills training, it is worth noting that there may be some risk of accelerating antisocial behavior, or deviancy training, involved in social skills training when delinquent youth are placed together in groups (Dishion et al. 1999).

Combined Interventions

Psychopharmacological and behavioral treatments are not, by themselves, typically nor completely adequate to address all of the difficulties likely to be presented by clinic-referred children or adolescents with ADHD. Optimal treatment is likely to be comprised of a combination of many of these approaches for maximal effectiveness (Carlson et al., 1992; Pelham, Wheeler, & Chronis, 1998; Phelps, Brown, & Power, 2002). However, the extent to which combined treatments are superior to medication alone is a controversial issue, especially given the relatively high cost of many psychosocial interventions. Nevertheless, findings emerging from the MTA study imply some potential advantages of combined treatment, tempered by other multi-site studies that may challenge that conclusion.

Some early research studies examined the utility of combining psychosocial and pharmacological treatment packages with interesting results. It appears that in many studies, the combination of contingency management training of parents or teachers with stimulant drug therapies is generally little better than either treatment alone for the management of ADHD symptoms (Firestone et al., 1981; Gadow, 1985; Pollard, Ward, & Barkley, 1983; Wolraich, Drummond, Salomon, O'Brien, & Sivage, 1978). Several studies also found impressive results for classroom behavior management methods (Carlson et al., 1992; DuPaul & Eckert, 1997; Pelham et al., 1988) but found that the addition of medication provided some added improvements beyond that achieved by behavior management alone. Moreover, the combination may result in the need for less intense behavioral interventions or lower doses of medication than might be the case if either intervention were used alone. Where there is an advantage to behavioral interventions, it appears to be related to functioning rather than symptom relief, such as reliably increasing rates of academic productivity and accuracy (DuPaul & Stoner, 2003). Despite some failures to obtain additive effects for these two treatments, their combination may still be advantageous given that the stimulants are not usually used in the late afternoons or evenings when parents may need effective behavior management tactics to deal with the ADHD symptoms. Moreover, a minority of children (10-25%) do not respond positively to the medications (Connor, 2005a), making behavioral interventions one of the few scientifically proven alternatives for these cases.

Several early studies have examined the combined effects of stimulant medication with cognitive-behavioral interventions. Horn, Chatoor, and Conners (1983) examined the separate and combined effects of d-amphetamine and self-instructional training with a 9 year old inpatient ADHD child. The combined program was more effective in increasing on-task behavior during class work, and decreasing teacher ratings of ADHD symptoms. However, academic productivity was improved only by the use of direct reinforcement for correct responses. In contrast, using group comparison designs, Brown, Borden, Wynne, Schleser, and Clingerman (1986) and Brown, Wynne, and Medenis (1985) found no benefits of combined drug/cognitive behavioral interventions over either treatment alone on similar domains of functioning of ADHD children. Similarly, a later study by Horn et al. (1991) did not find the combination of treatments to be superior to medication alone.

Some success for combined medication and self-evaluation procedures have been reported (Hinshaw, Henker, & Whalen, 1984a) when social skills, such as cooperation, have been targets of intervention. Yet, when these same investigators attempted to teach anger-control strategies to ADHD children to enhance self-control during peer interactions, no benefits of combined intervention were found beyond that achieved by self-control training alone (Hinshaw, Henker, & Whalen, 1984b). The self-control techniques were the most successful in teaching these children specific coping strategies to employ in provocative interactions with peers which usually lead to angry reactions from the ADHD children. Medication, in contrast, served only to lower the overall level of anger responses but did not enhance the application of specific anger control strategies. These studies suggest that each form of treatment may have highly specific and unique effects on some aspects of social behavior while not on others.

Limited research has evaluated the effects of behavioral parent training (BPT) alone and combined with child training in self-control strategies (Horn et al., 1987) on home and school behavioral problems. The results failed to find any significant advantage for the combined treatments. Both self-control training and BPT alone improved home behavior problems but neither resulted in any generalization of treatment effects to the school, where no treatment had occurred. Since a no-treatment group was not employed in this study, however, it is not possible to conclude that these effects were due to treatment rather than to nonspecific effects (e.g., maturation, therapist attention, regression effects, etc.). A later study by Horn and colleagues (Horn, Ialongo, Greenberg, Packard, & Smith-Winberry, 1990) did find such a treatment combination to be superior to either treatment used alone in producing a significantly larger number of treatment responders. Once again, however, no generalization of the results to the school setting occurred.

Satterfield, Satterfield, and Cantwell (1980) attempted to evaluate the effects of individualized multi-modality intervention provided over extensive time periods (up to several years) on the outcome of ADHD boys. Interventions included medication, behavioral parent training, individual counseling, special education, family therapy, and other programs as needed by the individual. Results suggested that such an individualized program of combined treatments continued over longer time intervals can produce improvements in social adjustment at home and school, rates of antisocial behavior, substance abuse, and academic achievement. These results seem to be sustained across at least a three year follow-up period (Satterfield, Cantwell, & Satterfield, 1979; Satterfield, Satterfield, & Cantwell, 1981; Satterfield, Satterfield, & Schell, 1987). While such treatment suggests great promise for the possible efficacy of multimodality treatment extended over years for children with ADHD, the lack of random assignment and more adequate control procedures in this series of studies limits the ability to attribute those improvements obtained in this study directly to the treatments employed. And these limitations certainly preclude establishing which of the treatment components was most effective. Still, studies such as these and others (Carlson et al., 1992; Pelham et al., 1988) raised hopes that intensive multimodality treatment can be effective for ADHD if extended over long intervals of time.

Intensive, Multi-Modal Treatment Programs

Two of the most well-known and well-regarded multi-modality intervention programs are the summer treatment programs developed by William Pelham and colleagues and conducted at Western Psychiatric Institute in Pittsburgh (Pelham, & Hoza, 1996), and the University of California-Irvine /Orange County Department of Education intervention developed by James Swanson, Linda Pfiffner, Keith McBurnett, and Dennis Cantwell (See Pffifner, DuPaul, & Barkley, 2005). The latter program incorporates a number of features of the program developed by Pelham as well as some components of the multi-modal program conducted by Stephen Hinshaw, Barbara Henker, and Carol Whalen at the University of California at Los Angeles. All of these programs rely on four major components of treatment, these being: (1) parent training in child behavior management; (2) classroom implementation of behavior modification techniques; (3) social skills training (typically around sports); and (4) stimulant medication, in some cases. While the Pelham program is conducted during the summer months in a day “camp” style program, the UCI-OCDE program is a year round day-school style program.

The Summer Day Treatment Program (STP)

This program was largely developed by Pelham and colleagues and is conducted in a day-treatment environment with a summer school/camp-like format. Daily activities include a few hours of classroom instruction which also incorporates behavior modification methods, such as token economies, response cost, and time out from reinforcement. In addition, three to four hours of sports and recreational activities are arranged each day during which behavioral management programs are operative. The program also includes parent training, peer relationship training, and a follow-up protocol to enhance the likelihood that treatment gains will be maintained after leaving the program. During their stay at the camp, some children may be tested on stimulant medication using a double-blind, placebo controlled procedure in which the child is tested on several different doses of medication while teacher ratings and behavioral observations are collected across the different camp activities. Pelham and colleagues have used this setting and larger programmatic context to conduct more focused research investigations into the effectiveness of classroom behavior management procedures alone, stimulant medication alone, and their combination in managing ADHD symptoms and improving academic performance and social behavior. Some of the components of this day-treatment program have been evaluated previously, such as classroom contingency management, and have been found to produce significant short-term improvements in children with ADHD (see DuPaul & Eckert, 1997, 1998; Pfiffner, DuPaul, & Barkley, 2005). And so they clearly seem to do so here (Carlson et al., 1992, Pelham et al., 1988). The STP program in fact was a part of the intensive multimodal treatment program for children with ADHD studied in the MTA project (see below). But other components of the program have not been so well evaluated previously for their efficacy with children having ADHD, such as social skills training. And while results from parent ratings before and after their children’s participation indicate that 86 percent believe their ADHD children improved from their participation in the program, no data have been published as yet on whether the gains made during the treatment program are maintained in the subsequent normal school and home settings after the children terminate their participation in this program.

The UCI/OCDE Program (see Pfiffner, DuPaul, & Barkley, 2005)

This program provides week-day treatment for ADHD children in kindergarten through fifth grades in a school-like atmosphere using classes of 12-15 children. The clinical interventions rely chiefly on a token economy program for the management of behavior in the classrooms and a parent training program conducted through both group and individual treatment sessions. Some training of self-monitoring, evaluation, and reinforcement also occurs as part of the class program. Children also receive daily group instruction in social skills as part of the classroom curriculum and some of these behaviors may be targeted for modification outside of the group instruction time by using consequences within the classroom token economy. Before returning to their regular public school, some children may participate in a transition school program that focuses on more advanced social skills as well as behavior modification programs to facilitate the transfer of learning to their regular school setting. Some children within this program also may be receiving stimulant medication as needed for management of their ADHD symptoms. While this program has served as an exemplar for many others, published research on its efficacy is not available. Granted, the parent training program and classroom behavior modification methods are highly similar to those used in published studies that have found them to be effective, at least in the short-term so long as they are in use (Barkley, 1997a; DuPaul & Eckert, 1997; Pelham & Sams, 1992). But the actual extent to which this particular program achieves its stated goals and, specifically, the generalization of treatment gains to non-treatment settings as well as the maintenance of those gains after children return to their public schools have not been systematically evaluated or published.

The UMASS/WPS Early Intervention Project:

Barkley, Shelton, and colleagues completed a multi-method early intervention program for kindergarten children (ages 4-6 yrs.) having significant problems with hyperactivity and aggression, at least 70 percent of whom qualified for a clinical diagnosis of ADHD (see Barkley, Shelton, Crosswait et al., 2000; Shelton, Barkley et al., 2002). This program did not utilize clinic-referred children, whose parents and even teachers may be highly motivated to cooperate with treatment. Instead, children were identified at kindergarten registration as displaying significantly high levels of hyperactive/aggressive behavior (93rd percentile) and being at high risk for both ADHD and ODD. Indeed, more than 70 percent of them met criteria for these disorders upon subsequent clinical evaluation using structured psychiatric interviews. They were randomly assigned to one of four intervention groups for their entire kindergarten year. One group received a 10 week group parent training program followed by monthly booster session group meetings. Otherwise, their children participated in the standard public school kindergarten program offered by the Worcester, MA public schools (WPS). The second group was assigned to a special enrichment kindergarten classroom in which they received accelerated instruction in academic skills, social skills training, classroom contingency management procedures (token systems and other reinforcements, response cost, time out, etc.), and cognitive therapy (self-instruction training) as part of their full-day kindergarten program. These special classes contained 12-16 hyperactive-aggressive children in each and were held in two neighborhood elementary schools in the WPS system to which they children were provided busing. Children in this special classroom also received several months of follow-up consultation to their teachers when they returned to their regular public schools for their first grade year. A third group received both the parent training and enrichment classroom treatments while a fourth group received no special services except for the initial evaluation and periodic re-evaluations. All children have been followed for two years after their participation in these treatment programs. Results indicated no beneficial effect of the parent training program, in large part because more than 60 percent of the parents did not attend the training classes regularly, if at all. The enrichment classroom produced a significant improvement in the children’s classroom behavior and social skills during the kindergarten year but did not result in any change in behavior in the home as rated by parents. Nor did it produce greater gains in academic achievement skills than had been experienced by the control groups not receiving this classroom program. Moreover, the results of the classroom appear to have attenuated during the follow-up period. Such results once again show that intensive classroom behavioral interventions can be effective in the short term for addressing the disruptive behavior of children. Yet these same results are rather sobering in view of the large investment of money, time, and staff training. Parent training programs for children at high-risk for school and home behavior problems may not be especially effective in families identified through such community screening programs, largely due to poor parental motivation and investment in the training program. And even where classroom interventions are successful in the short-term “active” treatment phase, their effects may diminish or disappear with time after children leave the treatment environment. This study suggests that the rather positive treatment outcome results for families who seek treatment and, by inference, are motivated to change themselves and their ADHD children may not be readily extrapolated to families of similarly deviant children who have not sought treatment but were identified through community screening programs.

The NIMH Collaborative Multisite Multimodal Treatment Study of Children with Attention-Deficit/Hyperactivity Disorder (MTA)

The National Institute of Mental Health (NIMH) collaborative multi-site Multimodal Treatment Study of Children with ADHD (MTA) is the first major clinical trial by NIMH with a focus on a childhood disorder (MTA Cooperative Group, 1999a, 1999b). While much research has documented the short term effectiveness of medication and behavioral interventions to treat ADHD, significant questions remain unanswered about long-term effectiveness of these interventions, alone or in combination, on the multiple functional outcome areas impacted by ADHD. Questions also remain about which types of youth with ADHD may benefit most from which types of treatment. The ambitious and groundbreaking MTA study was designed to help answer some of these major questions by randomly assigning children to four treatment groups: medication alone (MedMgt), behavior modification alone (Beh), the combination of medication and behavior modification (Comb), and community comparison (CC). In order to obtain a sufficiently large and diverse sample of youth with ADHD to begin to address these questions, a multisite study was initiated by NIMH along with funding from the U.S. Department of Education in 1992.

MTA Study Design/Methodology

Participants were randomly assigned to one of four conditions. Treatments were delivered over a 14 month period; comprehensive assessments of functioning in multiple domains were conducted at baseline prior to randomization as well as at 3, 9 and 14 months (with the 14 month assessment constituting the treatment end point assessment) (MTA Cooperative Group, 1999a). The Collaborative group published results of a 24 month follow-up, and later up to 8 years after treatment termination.

Behavioral treatments (in both Beh and Comb conditions) encompassed parent, child, and school domains. Behavioral parent training was provided by experienced training consultants and based on models by Barkley (1997a), Forehand and McMahon (1981; McMahon & Forehand, 2003). This intervention consisted of 27 group and 8 individual sessions. Child behavioral treatment consisted of an intensive summer treatment program (based on the Pelham STP model) as well as school consultation services (similar to the UCI model). The Summer Treatment Program was an intensive 8 week, 9 hour per day program; study training consultants supervised staff working with the children and continued to provide parent interventions during the summer. The same training consultants provided school consultation services (10-16 sessions of teacher consultation and establishment of a daily report card) and the staff working with the children in the summer treatment program worked in the schools in the fall as paraprofessional aids (12 weeks at half time under supervision of the training consultants and the child’s teacher). Families attended an average of 77.8% of parent training sessions, 36.2 of 40 possible STP days, 10.7 teacher consultation visits, and 47.6 (of 60) possible days with a classroom aid. Delivery of behavioral treatments was faded over the course of treatment, so that by the endpoint assessment at 14 months therapist contact with parents had ended or was reduced to once per month.

Like the intensive behavioral interventions, the medication treatments (in both MedMgt and Comb conditions) provided in the MTA occurred in a much more rigorous and intensive way than is typical in clinical practice. All medication treatment provided by the MTA included an initial 28-day double-blind, placebo-controlled titration consisting of placebo plus four different doses of methylphenidate (5, 10, 15, and 20mg) randomly given over the titration period. Three times per day dosing was used in the titration (and typically during treatment) in which the full dose was given in the morning and at lunch, as well as a half dose in the mid-afternoon. Parent and teacher daily ratings were collected during the titration; graphs portraying the results were rated by a cross-site panel of experienced clinicians. A “best dose” was chosen and the blind was broken; that dose became the initial dose for treatment. If the dose chosen was placebo, alternative medications were openly titrated until a satisfactory medication was chosen (or in the case of a robust placebo response the child was not medicated). Approximately, 89% of youth assigned to MedMgt or Comb successfully completed titration; of these, 68.5% were assigned to initial doses of methylphenidate averaging 30.5mg/day given 3 times per day, of the remaining group of youth who completed titration but not started on methylphenidate, 26 received an unblinded titration of dextroamphetamine because of unsatisfactory methylphenidate response and 32 were given no medication because of a robust placebo response. Of note is that of the 289 subjects assigned to MedMgt or Comb, 17 families refused titration and another 15 subjects did not complete titration (11 due to side effects or problems with titration) and 4 of whom inadequate amounts of titration data were gathered (MTA Cooperative Group, 1999a).

Youth assigned to the CC condition received no intervention by the MTA staff, but sought treatment as usually provided in the community. Referrals to non-MTA providers were made as necessary for these families; all CC youth and families returned for assessments at the same time as youth in the other three conditions of the study. Initially, it was thought that the CC group would provide a minimal or no treatment comparison group. However, as described later in this section, about two-thirds of the children in the CC group received medication for ADHD.

Outcomes in this study were assessed using a large number of measures in multiple domains, including verbal-report information (via interview and paper/pencil measures) by parents, teachers, and children, direct observation in the clinic and school, and computerized assessments of attention. Given the large number of measures, settings and informants used in the study, data reduction methods were conducted to condense measures into outcome domains. The major outcome domains that have received attention in the literature are: ADHD symptoms, oppositional/aggressive symptoms, social skills, internalizing symptoms, parent-child relations, parental discipline, and academic achievement.

Major findings From the MTA on ADHD Symptoms

All four MTA groups showed symptom reduction over time. In our opinion, the trends in the data favored the Comb treatment over the other three conditions, but this conclusion may depend on how those data are analyzed. When using an idiographic approach that looks at individual outcomes, there is a clear advantage of combined treatment. Swanson, Kraemer, Hinshaw, Arnold, Conners, et al, (2001), created a categorical measure of treatment outcome based on composite ADHD and ODD symptoms scores from teachers and parents using the SNAP-IV. Successful treatment was identified as scoring on average 1 or below on a composite SNAP score at the end of treatment (representing symptoms falling in the “not at all” or “just a little” range of categories at treatment endpoint). Success rates were as follows: 68% for combined treatment (Comb), 56% for medication management alone (MedMgt), 34% for behavioral treatment alone (Beh), and 25% for treatment as usual in the community (CC). A similar, but less robust, pattern of results was observed at the 24-month follow-up. Specifically, the normalization rates were 48%, 37%, 32%, and 28%, for Comb, MedMgt, Beh, and CC (MTA Cooperative Group, 2004a).

Another way to look at the MTA data is in terms of statistical significance of the group means, which is the analysis that has received the most attention in the published literature. When using this approach on the 14- and 24-month follow-up data, the MTA Collaborative Group reached the conclusion that treatments involving MedMgt (i.e., MedMgt and Comb) were superior those that did not have the intensive medication management (i.e., Beh and CC). Based on significance tests of means, the Beh and CC conditions were statistically equivalent. Likewise, the MedMgt and Comb groups were comparable, thus indicating no advantage of Comb relative to intensive MedMgt (MTA Cooperative Group, 1999a; 2004a). A few comments on these findings are warranted.

Some effects on ADHD symptoms were apparently mediated by medication effects (MTA Cooperative Group, 2004b). Therefore, it is important to note that 67% of the children in the CC group were taking medication. Thus, the CC group was an active treatment group rather than a no-treatment control. Thus, the group that received only behavior modification (Beh) was being compared to the CC group that received medications in the community. It is also important to consider the implications of the fact that there were some substantial differences in the doses of medication across the treatment groups. For instance, at the 14-month follow-up, the average daily dose for Comb was 31.2 mg while the average daily dose for MedMgt was 37.7 mg (MTA Cooperative Group, 1999a). Given that Comb and MedMgt had identical medication titration procedures, the difference in dose at 14-months suggests that the intensive behavioral intervention allowed individuals to take lower doses of medication. Lower doses are a considerable therapeutic advantage because most stimulant side effects, including the mild growth suppression observed in the MTA, are dose dependent (i.e., lower doses lessen the risk and severity of side effects; MTA Cooperative Group, 2004b).

When examining the group data, it is tempting to conclude that the MedMgt condition was superior to CC, even though most of the CC participants were medicated. Such a conclusion implies that in the package of procedures in the MedMgt protocol, which includes monthly supportive contact and decisions supported by high-quality data, is superior to routine community care. Indeed, this has been one of the major messages from the MTA Cooperative group (e.g., 2004a). However, it is noteworthy that average dose of the CC group that sought treatment in the community was 22.6 mg/day (MTA Cooperative Group, 1999a). The fact that children receiving intensive medication management in the MTA (i.e., MedMgt and Comb) were taking the equivalent of 10 mg to 15 mg more methylphenidate each day than the community control group is perplexing. In this situation, it is unclear if the higher dose or some aspect of the MedMgt intervention, such as dosing three times per day in some cases, resulted in the better outcomes.

Another consideration in comparing the Beh and Comb conditions with MedMgt and CC, is that intensive behavioral treatments were faded by study endpoint (Pelham, 1999). Due to this unequal treatment activity, it is plausible that the comparison of Beh and Comb to MedMgt at the 14-month follow-up may have been biased in favor of the MedMgt. This issue has been argued on theoretical grounds (see Pelham, 1999) and is consistent with observation that the therapeutic effect size of intensive MedMgt diminished by 50% from the intensive phase to the follow-up phase (i.e., from the 14- to 24-month follow-up; MTA Cooperative Group, 2004a).

In our reading of the MTA data, as the fading becomes an increasingly distant past event, the trend in the data seems to be for the Comb group to outperform the other groups. However, according to the MTA Collaborative Groups statistical conclusion criteria, the differences between Comb and MedMgt are not statistically significant. Moreover it appears that all treatments have declined in effectiveness at all subsequent follow-ups. Therefore, our conclusions regarding the superior efficacy of combined treatment in the MTA are open to alternative interpretation, particularly in light of a another multi-site study discussed below.

The New York-Montreal Multi-Modal Treatment Study

Although completed prior to the MTA study, the results of another multimodal treatment study involving large samples and several treatment sites have recently been reported that conflict with the findings of the MTA study concerning the benefits of combined treatment over medication management alone. The New York-Montreal (NYM) study selected 103 children with ADHD (ages 7-9 years) who were free of conduct and learning disorders and who had shown an initial positive response to methylphenidate (MPH) during a short term trial. Hence, unlike the MTA study, the NYM study focused exclusively on stimulant responsive children having far less comorbidity. These children were randomly assigned to receive 2 years of treatment in one of three treatment arms: (1) MPH alone; (2) MPH plus intensive multi-modal psychosocial treatment; or (3) MPH plus an attention-placebo psychosocial treatment. The latter approach to controlling for professional attention was not used in the MTA study. The intensive 2-year psychosocial treatment consisted of BPT, parent counseling, social skills training, psychotherapy, and extra academic assistance. Treatment contact during the first year of treatment was twice weekly, with fading of treatment to a considerable degree during the second year.

Assessments involved parent, teacher, and psychiatrist ratings, children’s self-ratings, children’s ratings of their parents, observations collected in school settings, and academic tests. The domains assessed included symptoms of ADHD and other behavioral problems (ODD), home and school functioning, social functioning, and academic performance. The results were consistent across all domains. No support was found for combining intensive psychosocial treatments of any sort with MPH in children with ADHD initially shown to be responsive to MPH (Abikoff et al., 2004a, 2004b; Hechtman et al., 2004a, 2004b). Nor was it found that MPH could be discontinued successfully in those who were receiving the combination treatment. Thus, it appears that the set of psychosocial treatments used in this study produced no incremental benefit in children shown to be strong and unambiguous responders to stimulant medication. Although the authors made some statements that there may have been improvement from MPH, the study was not designed to test for benefit from medication and uncontrolled confounds, such as maturation or regression to the mean, are plausible alternative explanations for what may seem like sustained improvement associated with MPH across the two years of treatment.

In contrast to the MTA study, this study did not include treatment within the child’s normal school setting nor did the children attend an intensive summer treatment program. Also unlike the MTA study, this study intervened over a 24 month rather than a 14 month period. Lacking in both the MTA and the New York Montreal (NYM) study was documentation that the psychosocial treatments were effective. This contrasts with the assessment of medication effects because each child received very well-controlled individualized trials that determined if medication worked. Based on the review in previous sections of this paper, several of the interventions in the NYM study are not thought to be effective for children with ADHD (e.g., social skills training and individual therapy). Furthermore, although the behavioral parent training was shown to achieve significant improvements in knowledge of behavioral methods, there was no reported change in parenting behavior (Hechtman et al., 2004). Thus, there was no evidence that the psychosocial treatments met the requirement of showing activity at the point of performance.

Overall, the results of the NYM study may not be a fair comparison because grade A treatment with methylphenidate was compared with psychosocial treatments of unknown quality. A reasonable comparison of medication and psychosocial treatment should pit equivalent quality treatments against each other (i.e., grade A medication and grade A psychosocial treatments). Such studied need to document that both treatments were delivered as intended with appropriate implementation at the point of performance. This is key with medication, because according to the NYM study poor compliance as seen with the discontinuation probe very rapidly results in deterioration. Psychosocial treatments should be evaluated with equal rigor, such as experimental analysis of the effectiveness of behavior contingencies by using reversal designs in the context in individual case studies. To our knowledge, no such study has yet been conducted, but some insights might be gained for further analysis of compliance data in the MTA and NYM studies. Until studies of the highest quality interventions and the most rigorous quality control are implanted and proper analyzed, there will be lingering questions about the relative merits of intensive multimodal treatment relative to excellent medication management for the treatment of ADHD and related problems.

Efficacy, Safety, and Practicality of Combined Treatment

Using the grading system for level of empirical validation recently described at the beginning of this course, we give the combination of stimulant medication and behavioral intervention the grade of B or lower. Let us clarify. Although the literature indicates that each of these treatments separately deserves a grade of A, what is being graded here is the superiority of their combination relative to either alone based on the evidence. Is the evidence for the combined treatment sufficient to warrant this grade? Combined treatments have been shown to be superior to unimodal treatment on some measures in some subsets of ADHD children in at least two well-designed studies by independent investigators. However, the recent NYM study that found no advantages of intensive multimodal treatment may raise some doubts. Due to the relative methodological strengths of the MTA compared to the NYM study, we believe that greater weight should be given to the MTA study. Unfortunately, the studies that support the efficacy of combined treatment were conducted in research settings that do not necessary replicate the “real world” settings in which most ADHD treatments are delivered. Thus, we are inclined to give intensive multimodal treatment a grade of B for efficacy.

The grade of B to C for intensive multimodal treatment is also intended to convey the message that the practicality of combined treatments in unknown. Indeed, there are several reasons to believe that these treatments would be very difficult to replicate in most applied settings. For instance, the acceptance/attendance data from the MTA found 81% compliance for med component but only 64% compliance for behavioral component (MTA Collaborative Group, 1999b). This suggests that there are some important issues to work out related to therapist expectations and family participation in the treatment. Moreover, the studies of combined treatment used some very unique treatments that are difficult to find in many regions of the country or to replicate in applied clinical settings, such as Pelham and colleagues Summer Treatment Program. Until barriers to access to and participation in these treatments are overcome, the effectiveness of these treatments is open to doubt.

Generally speaking, combined treatment that uses family-based behavioral interventions and stimulant medication or atomoxetine should be very safe. There are some possible safety concerns related to the multimodal treatments of ADHD that have been studied. For example, some prominent theories related to conduct problems posit that placing children with behavior problems in groups with other disruptive children could lead to some harmful effects mediated by peer facilitation of antisocial behavior (Dishion, McCord, & Poulin, 1999). This was recently found to occur in a social skills training program for ADHD children, particularly among those who were not manifesting significant conduct problems prior to treatment (Antshel & Remer, 2003). Also, this author and colleagues have twice documented an adverse effect (escalation of conflicts) during behavioral family therapy for ADHD/ODD teens on a subset of participating families (Barkley, Edwards, et al., 2001; Barkley, Guevremont et al., 1992). Researchers studying behavioral interventions typically do not examine their data for such subsets of adverse responders but should be encouraged by these results to do so.

There are side effects of the medications as well that warrant attention. Approximately 2.9% of children in the MTA reported having severe side effects, which apparently remitted with discontinuation of medication. Also, the MTA Collaborative Group (2004b) estimated that there was a growth suppression effect related to medication of approximately –1.23 cm/year and –2.48 kg/year. Thus, although the treatments studied generally seem to be effective, potential risks warrant individual monitoring for potential iatrogenic effects, both for medications and some psychosocial treatments.

Ineffective or Unproved Therapies

Numerous questionable treatments have been attempted with children with ADHD over the past century -- (see Ingersoll & Goldstein, 1993; Pelham et al, 1998; Ross & Ross, 1976, 1982 for reviews). Vestibular stimulation (Arnold, Clark, Sachs, Jakim, & Smithies, 1985), biofeedback and relaxation training (Richter, 1984), sensory-integration exercises (Vargas & Camilli, 1999) among others, have been described as potentially effective in either uncontrolled case reports, small series of case studies, or in some treatment vs. no-treatment comparisons, yet are lacking in well-controlled experimental replications of their efficacy. A meta-analysis of studies examining the benefits of physical exercise suggests that it may be preferentially beneficial for participants who are hyperactive warrants further study of this effect in better controlled research (Allison, Faith, &Franklin, 1995).

Two interventions that may be more promising, pending additional research, are EEG biofeedback (neurofeedback) and working memory training.  EEG biofeedback or neurofeedback research has shown mixed results (Lofthouse et al., 2012; Loo & McKieg, 2012).  Typically, impressive results are found by studies of low scientific rigor, such as evaluating participants merely pre- and post-treatment with no comparison to other treatments or blindness of raters to the treatment condition. But the more rigorous the methods of the study, the smaller and less likely are these results, particularly when actual EEG feedback is compared to sham EEG feedback.  Thus, while some proponents of this treatment believe it has met the test of proven utility as a treatment for ADHD, we believe that more rigorous research finds far less benefit to this intervention.  Clearly, more research is in order before the widespread adoption of this intervention for ADHD. 

The status of research on working memory training is even less clear cut.  Initial studies (Klingberg et al., 2005) were quite promising in showing that practicing working memory tasks for 45 minutes per day for weeks at a time resulted not only in improved working memory on related tasks but also on ADHD symptoms as reported by parents and teachers.  Subsequent studies, however, did not find such promising results (Rutledge et al., 2012) and often found little evidence of generalization outside the treatment setting, especially to the school environment.  Two reviews of this literature by Shipstead and colleagues have pointed out numerous shortcomings in the methodology of research into this treatment and concluded that far more research was needed to substantiate this intervention (Shipstead et al., 2010, 2012).

Many dietary treatments, such as removal of additives, colorings, or sugar from the diet or addition of high doses of vitamins, minerals, or other “health food” supplements to the diet have proven very popular (Chan, Rappaport, & Kemper, 2003) and some reviews of research claim that there is evidence for their effectiveness (Schnoll, Burshteyn, & Cea-Aravena, 2003). But a careful reading of such reviews and the existence of better controlled research finds little or no scientific support (Conners, 1980; Ingersoll & Goldstein, 1993; Milich, Wolraich, & Lindgren, 1986; Wolraich, Wilson, & White, 1995). A recent meta-analysis of restrictive elimination diets finds that while some small improvements may be gained in eliminating artificial colors from children’s diets, the degree of improvement is quite modest proving of value for only a minority of children with ADHD (Nigg, 2012). Certainly traditional psychotherapy and play therapy have not proven especially effective for ADHD (see Pelham, Wheeler, et al, 1998; and Ross & Ross, 1976).


The treatment of children and teens with ADHD is an often complex and certainly longer-term enterprise than was previously thought to be necessary. Viewed now as a chronic disorder for most children, ADHD requires treatments that must be combined and sustained in order to have a long-term impact on the quality of life and developmental outcomes of these children. Treatments appear to succeed by temporarily reducing or normalizing symptoms for as long as treatments are in effect so as to reduce the numerous secondary harms associated with unmanaged ADHD. Though numerous therapies have been proposed for this disorder, those having the greatest empirical support are contingency management methods applied in classrooms and elsewhere (summer camps), training of parents (BPT) in these same methods to be used in the home and elsewhere (community settings), psychopharmacology, particularly stimulants and atomoxetine, and to a lesser extent, the combination of behavioral treatments with medication. Evidence for CBT is lacking at this time, while that for social skills training programs paints a mixed picture that is based mainly on studies having significant methodological limitations. Better controlled and larger studies appear to show little or no treatment effects when the skills or behaviors are not cued and reinforced for occurring at the specific point of performance. Popular treatments among laypeople, such as dietary manipulations, do not have compelling evidence for their efficacy, nor do several other professionally popular treatments, such as sensory integration training.

Most cases require a combination of these more effective treatments in order to provide effective management of the disorder and its comorbid conditions. Among children who are already stimulant responsive, it is not clear to what extent intensive psychosocial treatments provide added benefit. Interventions will need to be high quality and sustained over several years (or more), and re-intervention is highly likely as new developmental transitions occur and new domains of potential impairment now become available to the individual with ADHD across the life span.


Click here for References


Take the test


© Copyright 2004-2019 by, Inc. All rights reserved.