Social Functioning in Children With ADHD Treated With Long-Term Methylphenidate and Multimodal Psychosocial Treatment

S P E C I A L S E C T I O N

Social Functioning in Children With ADHD TreatedWith Long-Term Methylphenidate and Multimodal

Psychosocial TreatmentHOWARD ABIKOFF, PH.D., LILY HECHTMAN, M.D., RACHEL G. KLEIN, PH.D.,

RICHARD GALLAGHER, PH.D., KAREN FLEISS, PSY.D., JOY ETCOVITCH, M.A., LORNE COUSINS, PH.D.,

BRIAN GREENFIELD, M.D., DIANE MARTIN, M.A., AND SIMCHA POLLACK, PH.D.

ABSTRACT

Objective: To test that methylphenidate combined with intensive multimodal psychosocial intervention, which includes

social skills training, significantly enhances social functioning in children with attention-deficit/hyperactivity disorder

(ADHD) compared with methylphenidate alone and methylphenidate plus nonspecific psychosocial treatment (attention

control). Method: One hundred three children with ADHD (ages 7–9), free of conduct and learning disorders, who

responded to short-term methylphenidate were randomized for 2 years to receive (1) methylphenidate alone, (2)

methylphenidate plus multimodal psychosocial treatment that included social skills training, or (3) methylphenidate plus

attention control treatment. Assessments included parent, child, and teacher ratings of social function and direct school

observations in gym. Results: No advantage was found on any measure of social functioning for the combination

treatment over methylphenidate alone or methylphenidate plus attention control. Significant improvement occurred

across all treatments and continued over 2 years. Conclusions: In young children with ADHD, there is no support for

clinic-based social skills training as part of a long-term psychosocial intervention to improve social behavior. Significant

benefits from methylphenidate were stable over 2 years. J. Am. Acad. Child Adolesc. Psychiatry, 2004;43(7):820–829.

Key Words: attention-deficit/hyperactivity disorder, social function, school observations, long-term stimulant treatment,

psychosocial treatment.

Social deficits, even early in life, have been demon-strated among children with attention-deficit/hyper-activity disorder (ADHD) (e.g., Alessandri, 1992;Whalen and Henker, 1985, 1991). Because social im-pairment and peer status are predictors of long-term

outcome (Greene et al., 1997; Ollendick et al., 1992;Parker and Asher, 1987), reversing social deficits is animportant clinical goal.Stimulant treatment reduces negative social behav-

iors (Gadow et al., 1995; Gillberg et al., 1997; Hin-shaw et al., 1989; Klein and Abikoff, 1997; Whalen etal., 1987) but does not enhance prosocial behavior(Buhrmester et al., 1992; Hinshaw et al., 1989), animportant facet of peer acceptance (Barton, 1986).Problematically, gains are not maintained when medi-cation is discontinued (Abikoff and Gittelman, 1985).Although psychosocial treatment has been proposed

for improving the social behavior of children withADHD, there is limited support for its efficacy. Cog-nitive interpersonal problem solving has been unsuc-cessful (Abikoff, 1991). Pfiffner and McBurnett (1997)found that short-term social skills training combinedwith parent training was superior to a wait-list controlbut only on parent ratings.

Accepted January 30, 2004.Drs. Abikoff, Klein, Gallagher, and Fleiss are with the NYU Child Study

Center, New York University School of Medicine, New York; Drs. Hechtmanand Greenfield are with the Department of Psychiatry, McGill University andMontreal Children’s Hospital, Montreal, Quebec, Canada, Ms. Etcovitch iswith Montreal Children’s Hospital, Dr. Cousins is with McGill University andthe Summit School, Montreal, Quebec, Canada; Ms. Martin is with NassauCommunity College, Garden City, NY; and Dr. Pollack is with the Departmentof Computer Information Systems and Decision Science, St. John’s University,Queens, NY.

Correspondence to Dr. Abikoff, NYU Child Study Center, 215 LexingtonAvenue, 13th Floor, New York, NY 10016; e-mail: [email protected]/04/4307–0820©2004 by the American Academy of Child

and Adolescent Psychiatry.DOI: 10.1097/01.chi.0000128797.91601.1a

J . AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:7, JULY 2004820

Combined pharmacological and social skills inter-vention studies are scarce. Klein and Abikoff (1997)found that children with ADHD treated with methyl-phenidate alone or methylphenidate combined withclinical behavior therapy were viewed by teachers asmore popular than children on behavior therapy alone.The popularity ratings of the two medication groupswere identical.The current investigation evaluates the adjunctive

efficacy of multimodal psychosocial treatment (MPT)in children with ADHD. It tests whether 2 years ofmethylphenidate combined with a comprehensive psy-chosocial treatment confers significantly better func-tion than methylphenidate alone. We report here onchildren’s social functioning. Other outcomes, i.e.,children’s symptomatology, academic performance,parenting, are presented in companion articles (Abikoffet al., 2004; Hechtman et al., 2004a,b).We hypothesized that, relative to methylphenidate

alone, children on the combination would have bettersocial function and experience more positive and lessnegative social behavior from peers and teachers. Ad-ditionally, we predicted that the superiority of the com-bination would result from the social skills training andnot from nonspecific aspects of treatment. Therefore, itwas hypothesized that the combination of methylphe-nidate and social skills training would be superior tomethylphenidate plus attention control psychosocialtreatment (ACT).Furthermore, it was hypothesized that relative ad-

vantages associated with 1 year of combined treatmentwould persist. Hence, we predicted superiority ofmethylphenidate and social skills training during a sec-ond year of maintenance treatment. We hypothesizedthat treatment groups would demonstrate differentfunctional patterns over time. Specifically, significantincremental improvement during year II was predictedwith combined treatment, relative to methylphenidatealone and methylphenidate plus an attention control.In the latter two groups, a flattening or attenuation oftreatment effects was predicted. Finally, we predictedthat the superiority of combined treatment would haveclinical significance as indicated by greater normalizationof social behavior relative to the other two treatments.

METHOD

Details of the design and its rationale are presented in Klein et al.(2004). Briefly, the study was conducted at two large medical cen-

ters (New York and Montreal) between 1990 and 1995. Medica-tion-free boys and girls, 7.0 to 9.9 years of age, met diagnostic andseverity criteria for ADHD (N = 103). Because treatment included2 years of methylphenidate, children had to exhibit meaningfulbenefit from a 5-week clinical trial of methylphenidate.

Treatments

Children were randomly assigned for 2 years to (1) methylphe-nidate alone (M) (n = 34), (2) methylphenidate plus MPT (M +MPT) (n = 34), or (3) methylphenidate plus ACT (M + ACT)(n = 35). Treatment was initiated between October and April andbridged 3 school years. Children were switched, single blind, tomatching placebo after 1 year of treatment.

Multimodal Psychosocial Treatment.MPT integrated several treat-ment components that aimed to optimize improvement acrossfunctional domains. Children received individualized academic as-sistance, organizational skills training, individual psychotherapy,social skills training, and, when necessary, reading remediation.Parents received parent management training. Each component wasdelivered once weekly during the first year and once monthly dur-ing the second year. A 75% attendance rate was required. Teacherscompleted daily report cards, which formed the basis of a home-based reinforcement program for school behaviors and academicperformance. Treatment modules were fully manual based beforestudy initiation (manuals are available from the authors). Proce-dures specific to each domain are detailed in papers that reportthese outcomes (Abikoff et al., 2004; Hechtman et al., 2004a,b).

Social Skills Training. At the time of study initiation, no evi-dence-based social skills training for children with ADHD wasavailable. However, two social competency programs adapted forthis study had been reported to improve social skills of “sociallyhandicapped children”: the Getting Along With Others: TeachingSocial Effectiveness to Children Program (Jackson et al., 1983) andthe Walker Social Skills Curriculum: The ACCEPTS Program(Walker et al., 1983). The programs use direct instruction, mod-eling, behavioral rehearsal, feedback, and social reinforcement.These components were modified for relevance to children withADHD and to enhance generalization.In groups of four, children received social skills training that

addressed five principal functions: basic interaction skills, gettingalong with others, contacts with adults at home and school, con-versational skills, and problem situations. Different behaviors weretargeted weekly (e.g., joining conversations, waiting for one’s turn,group cooperation). In-session components included modeling,role-playing, and viewing videotapes of previous sessions in whichappropriate and inappropriate social behaviors were identified.Several strategies were implemented to facilitate generalization:

(1) homework assignments, (2) teaching parents to foster and praisetargeted social behaviors, (3) praising children during academicremediation for targeted social skills, and (4) including social be-haviors on the daily school report card.A behavior modification system initially used concrete rewards

for appropriate behaviors. At the end of each 20-minute interval,children were rated on general behavior (following rules, goodsportsmanship) and program skills (social skills taught). Ratings (+3to −3) carried an equivalent number of points. Children were toldthe specific behaviors rated and point total after each interval. Atthe end of 60-minute sessions, children were informed of the day’spoint total. Beginning with the seventh session, consequences weredelivered at home.External reinforcement was replaced with self-evaluation at the

seventh session to increase children’s awareness of their social be-

TREATMENT OF SOCIAL FUNCTIONING IN ADHD

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 43:7, JULY 2004 821

havior. These procedures were implemented in phases to maximizeaccuracy of children’s self-evaluations (Hinshaw et al., 1984;Turkewitz et al., 1975).

Attention Control Psychosocial Treatment. The manual-based at-tention control interventions paralleled all formal aspects of MPTbut excluded social skills training.

Peer Activities. A peer activity group served as the attention con-trol for social skills training. Groups of four children worked ontasks and played with minimal structure. No reinforcement, rein-forced self-evaluation, or social skills training were used.

Maintenance Treatment During Year II

Medication and MPT and ACT visits continued monthly duringthe second year.

Measures

Children were evaluated twice before experimental treatment.First, all assessments were conducted when the child was medica-tion free, before the initiation of the short-term methylphenidatetrial; second, child ratings and school observations were obtained atthe end of the methylphenidate trial. All measures were repeatedafter 6, 12, 18, and 24 months of treatment to identify the timingof possible treatment differences.

Rating Scales. Parents completed the Social Skills Rating Scaleand children completed the child version. The total score served asthe outcome measure. The Social Skills Rating Scale has high in-ternal consistency and test–retest reliability as well as adequate con-struct and concurrent validity (Gresham and Elliott, 1990).Teachers completed the Taxonomy of Problem Situations

(Dodge et al., 1985). Its total score possesses high internal consis-tency and acceptable stability and differentiates socially rejected andaccepted children (Dodge et al., 1985).

Direct School Observations. Observations were conducted in gym,twice at each assessment, using the Social Interaction ObservationCode (Revised) (Abikoff et al., 1985). It records spontaneouslyinitiated and reactive positive, negative, and neutral social behaviorsas well as peers’ and teachers’ behavior initiated toward the childand his/her response. Trained observers, blind to treatment anddiagnosis, observed the study child and a same-sex classmate ofunremarkable comportment identified by teachers (“comparison”or “normal”). (Because comparison children were anonymous andwere unaware of being observed, parental consent was not requiredby the institutional review board or schools.) Whenever study chil-dren changed classes, another comparison was observed. Observa-tions lasted 30 minutes, 15 minutes for each child, with 1-minuteobservations alternating between study and comparison children(10-second observe, 5-second record). For each 10-second interval,only the first occurrence of observed behaviors was recorded.Scores were averaged across observations at each assessment. We

analyzed (1) positive, negative, and neutral social behaviors initiatedby study children and directed toward peers and teachers (referredto as “others”); (2) positive, negative, and neutral social behaviorsinitiated by others and directed toward study children; (3) studychildren’s positive, negative, and neutral responses to behaviorsdirected toward them; and (4) positive, negative, and neutral re-sponses of others to study children. Rates reflect the percentage ofintervals in which initiations and responses occurred.Interobserver agreement was conducted in approximately 15% of

observations. Each site’s observer trainer functioned as the “stan-dard.” Phi coefficients for initiated behaviors ranged from 0.74 to

0.96. The percentage of agreements for initiated behaviors rangedfrom 79% to 87% and from 76% to 92% for social responses.

Data Analyses

Analyses of variance tested for group differences in pretreatmentvalues. None were significant except for socioeconomic status.There were no significant group × site or group × site × timeinteractions. Consequently, data were collapsed across sites.Repeated measures over time for dependent variables were mod-

eled as a mixed-model analysis of covariance implemented in ProcMixed (SAS v8.1, Cary, NC), controlling for socioeconomic status.Empirical data exploration indicated that an unstructured covari-ance model best fit the data. Model parameter estimates and theirstandard errors were generated through maximum likelihood func-tions.Differential treatment effects in year I compared pretreatment

and medication baseline with status at 6 and 12 months. For dif-ferential maintenance effects, Proc Mixed analyses (covarying so-cioeconomic status) compared 12-, 18-, and 24-month data fordifferential change patterns. These tests yield main effects for groupand time and group × time interaction effects. The latter are themain interest of the study.To control for multiple tests, α was set at p < .01, two tailed; p

values between .05 and .01 are indicated as trends in the tables.Tables with F values are available from the authors.

RESULTS

The parent Social Skills Rating Scale mean of 76.2for study children at pretreatment (Table 1) indicatesclinically significant decrement from the norm (100 ±15) (Gresham and Elliot, 1990). In contrast, studychildren rated themselves as average. That their judg-ment was less than accurate is documented by objectiveschool observations. Compared with normal children,study children were more than twice as likely to initiate(1.89% of intervals versus 0.84%, t = 4.61, p < .000)and receive negative social behaviors (1.52% versus0.86%, t = 3.73, p < .000). However, positive and neu-tral behavior rates did not differ.

Year I Treatment Effects Relative to Pretreatment

Rating Scales. As Table 1 shows, ratings of socialcomportment by parents and children reveal no advan-tage of M + MPT over M alone or M + ACT. Teacherratings also failed to show the superiority of M + MPT.All informants’ ratings (parents, children, and teach-

ers) were significantly improved between pretreatmentand 6 months, regardless of treatment received (p <.01–.001) (Table 1).

School Observations in Gym. There were no signifi-cant group or group × time interactions in rates of

ABIKOFF ET AL.


positive, negative, or neutral social behaviors initiatedby study children (Table 2). Over time, all groups ini-tiated significantly fewer positive and negative behav-iors (p <.000).With regard to social behaviors directed toward

study children, no significant treatment differenceswere found for positive, negative, or neutral behaviors(Table 3). Thus, classmates and teachers did not behavedifferentially toward subjects as a function of treatmentdelivered. Across all treatments, there were diminishedpositive, negative, and neutral behaviors directed towardsubjects. Similarly, over time, fewer social behaviors weredirected toward normal children (Table 3, time effects).As shown in Table 4, compared with children on M

or M + ACT, children on M + MPT did not responddifferentially to social behavior directed at them byothers (positive, negative, or neutral).

A significant time effect on study children’s positiveresponses to neutral social behavior reflects an overallreduction from pretreatment to 12 months (from 2%to 0.3%, p < .01).There were no significant treatment differences in

others’ positive or neutral responses to study children’spositive behaviors (Table 5). Significant group ef-fects for others’ neutral responses to study children’sneutral behaviors were due to others exhibiting moreneutral responses to children in theM + ACT (41%) thanthe M + MPT (35%) (p < .05) and M (30%) groups(p < .001).Positive, negative, and neutral social behaviors to-

ward study children diminished significantly duringyear I (p < .01–.001). These changes occurred betweenpretreatment and 6 months, without further decreasesbetween 6 and 12 months.

TABLE 1Ratings of Children’s Social Skills by Parents, Children, and Teachers

Measure

Treatment Groups

M M + MPT M + ACT

Mean SD Mean SD Mean SD

SSRS-Pa,b

Pretreatment 78.1 16.1 75.7 20.4 75.7 22.46 mo 80.8 19.5 88.3 14.8 75.8 20.912 mo 78.5 23.6 87.5 20.8 88.0 18.118 mo 83.9 25.4 83.8 20.7 88.8 16.624 mo 89.5 22.8 88.3 20.1 91.4 18.9

SSRS-Cc

Pretreatment 102.2 30.6 96.3 20.9 103.7 28.5Medication baseline 105.5 19.1 102.2 22.1 109.9 27.56 mo 106.3 23.9 112.6 19.9 109.2 18.512 mo 111.6 24.3 108.0 20.1 111.9 19.518 mo 113.1 24.7 107.1 20.9 109.3 21.824 mo 106.9 17.7 108.4 26.1 110.8 21.0

TOPSd

Pretreatment 3.3 0.8 3.3 0.7 3.4 0.66 mo 2.4 0.9 2.3 0.8 2.4 0.812 mo 2.0 0.7 2.4 0.8 2.3 0.818 mo 2.3 0.9 2.2 0.7 2.3 1.024 mo 2.2 0.5 2.4 0.8 2.3 0.8

Note: M = methylphenidate; MPT = multimodal psychosocial treatment; ACT = attention control psychosocial treatment; SSRS = socialskills rating scale (P, parent; C, child); TOPS = taxonomy of problem situations (teachers).

a Group × time interaction: not significant. Trend at year I, p < .013; at 6 months, M + MPT versus M + ACT, p < .001; at 12 months,M + ACT versus M, p < .05.

b Time effects: pretreatment versus 6 months, p = .01, versus 12 months, p < .001.c Time effects: pretreatment versus 6 and 12 months, p < .005, no significant group × time interaction.d Time effects: pretreatment versus 6 and 12 months, p < .0001, no significant group × time interaction.



Year I Treatment Effects From Medication Baseline

These analyses examine treatment differences rela-tive to ratings obtained after the 5-week methylpheni-date trial. Relative to medication baseline, nosignificant treatment differences in children’s self-ratings were obtained during year I (Table 1). Simi-larly, direct observations in gym yielded no differentialtreatment effects (Tables 2 and 3).A significant time effect (p < .005) on positive be-

haviors initiated by study children reflects a significantdecrease from medication baseline (0.57%) to 12months (0.26%).

Year II Treatment Effects

During year I, we failed to obtain evidence of thesuperiority of M + MPT over M alone or M + ACT.Therefore, we cannot examine the hypothesis that ad-vantages of intensive M + MPT are maintained over a

second year. Nonetheless, year II findings indicatewhether M + MPT led to incrementally superior func-tion relative to other treatments.The hypothesis of differential improvement in social

functioning with maintenance treatment was not sup-ported for ratings of social behavior by parents, chil-dren, or teachers (Table 1). Observations of children’ssocial behavior in gym also failed to produce signi-ficant treatment differences regarding initiation ofpositive, negative, or neutral behaviors, nor were theresignificant treatment differences in the frequency ofclassmates’ or teachers’ positive, negative, or neutralsocial behaviors directed toward the treated children(Table 3).Analyses of the response of others to study children

yielded a significant group effect in positive responseto positive behaviors, indicating higher positive re-sponses to children in M +MPT (46%) than M + ACT

TABLE 2Social Behaviors Initiated by Subjects and Normal Comparisons During Gym

Initiated Behavior

Normal

Treatment Groups

M M + MPT M + ACT

Mean SD Mean SD Mean SD Mean SD

Positivea,b

Pretreatment 0.9 1.3 1.2 1.8 1.0 1.3 1.0 1.6Medication baseline — — 0.3 0.6 0.7 0.9 0.7 0.96 mo 0.6 1.2 0.4 1.1 0.5 0.8 0.4 0.812 mo 0.4 0.7 0.2 0.5 0.3 0.6 0.2 0.418 mo 0.2 0.4 0.2 0.3 0.2 0.3 0.3 0.424 mo 0.3 0.5 0.2 0.4 0.3 0.8 0.2 0.4

Negativea,c

Pretreatment 0.8 1.4 1.8 1.7 2.2 1.9 1.7 1.9Medication baseline — — 1.2 1.4 1.0 1.2 0.8 1.06 mo 0.5 0.6 0.9 1.2 .60 0.8 0.8 1.012 mo 0.6 1.0 0.5 1.2 1.2 1.6 1.1 2.118 mo 0.6 0.9 1.0 1.5 0.6 0.9 0.6 0.724 mo 0.4 0.7 0.9 9.9 0.5 0.7 0.5 0.7

Neutrald


Note: Group × time interactions: none significant.a Time effects in subjects: pretreatment versus 6 and 12 months, p < .000; medication baseline versus 12 months, p < .001.b Time effects in normal subjects: pretreatment versus 12 months, p < .000.c Time effects in normal subjects: pretreatment versus 6 months, p < .02.d Time effects in normal subjects: pretreatment versus 12 months, p < .03; 6 versus 12 months, p < .03.

ABIKOFF ET AL.


(27%) (p < .005) and M (25%) (p < .001). There wasalso a significant group × time interaction in the rateof others’ neutral responses to study children’s neutralbehaviors (p < .005). At 24 months, children in theM group received higher rates of neutral responses(52%) than children in M + MPT (34.1%) and M +ACT (35.1%) groups. In view of no other treat-ment differences, this finding has little clinical signifi-cance.No significant time effects occurred during year II.

Normalizing Treatment Effects

In the absence of significant treatment advantages ofthe combination treatment, there is no basis for exam-ining treatment differences in normalization of socialbehavior. However, it is informative to ascertainwhether behavioral normalization occurred acrosstreatments. To do so, children’s scores were collapsed

across the three treatments and compared with those ofpaired classmates via dependent t tests.The significant pretreatment differences in negative

behaviors between comparison and study children nolonger occurred with treatment. Nevertheless, duringyear I, subjects still tended to initiate more negativebehavior (p = .03). However, at year II, subjects andcomparisons were indistinguishable (p = .33). Impor-tantly, during years I and II, study children no longerreceived more negative behavior than comparisons (p =.82 and .85, respectively). At 18 and 24 months, studychildren received more positive behavior than compari-sons (p = .004 and .04, respectively).

DISCUSSION

The problematic social functioning of children withADHD is considered to result from an interplay of

TABLE 3Observed Social Behaviors Directed Toward Subjects and Normal Comparisons During Gym

Directed Behavior

Normal

Treatment Groups

M M + MPT M + ACT

Mean SD Mean SD Mean SD Mean SD

Positivea,e


Negativeb,d


Neutralc,f


Note: Group × time interactions: none significant.Time effects in subjects: apretreatment versus 6 months, p = .01, versus 12 months, p < .000, 6 versus 12 months, p < .05; bpretreatment

versus 6 and 12 months, p < .000; cpretreatment versus 6 months, p < .002, versus 12 months, p < .01.Time effects in normal subjects: dpretreatment versus 6 months, p < .002, versus 12 months, p < .03; e12 versus 18 and 24 months, p <

.000; f12 versus 18 months, p < .002; versus 24 months, p < .03.



factors not fully ameliorated by stimulant medication.Deficient social-cognitive skills (e.g., inattention to andmisinterpretation of social cues), inappropriate behav-iors, and poor interpersonal skills are assumed to havean adverse affect. The social skills training program thatwe implemented addressed these dysfunctions via mul-tiple behavioral and generalization procedures. How-ever, despite the duration and scope of social trainingthat was integrated in a broad MPT, we obtained noadvantages in social functioning with the interventionover and above those associated with stimulant treat-ment alone except for the lone finding in year II ofgreater positive responses to positive behaviors by peersand teachers of those in the M + MPT group (Table 4).

A second year of maintenance treatment did not resultin further improvement, but improvements seen inyear I were maintained.Our results are similar to those of the MTA (mul-

timodal treatment of ADHD) study (MTA Coopera-tive Group, 1999), in which medication and combinedtreatments did not differ on teacher- and parent-ratedsocial skills. In the MTA study, pretreatment parentratings of social skills were in the normal range,whereas they were clearly deviant in this study. Thus,the social behavior of study children allowed muchroom for improvement.Blinded observations collected in gym were intended

to avoid rater bias and treatment allegiance effects that

TABLE 4Summary of Findings Regarding ADHD Children’s Response Patterns to Others’ Behavior

ADHD Children’sResponse

Behaviors of Othersa Directed Toward ADHD Children

Positive

SignificantContrasts

Negative


Neutral


Year Ip ≤

Year IIp ≤

Year Ip ≤

Year IIp ≤

Year Ip ≤

Year IIp ≤

PositiveGroup NS NS — NR NR — NS NS —Time NS NS — NR NR — .011 NS Pre vs. 12 mo, p < .01Group × time NS NS — NR NR — NS .03 M vs. M + ACT, p < .01

M + MPT vs. M + ACT,p < .03

M vs. M + MPT, p < .05Time (normalsubjects) NS NS — NR NR — NS NS —

NegativeGroup NC NC — NS NS — NS NS —Time NC NC — NS NS — NS NS —Group × time NC NC — NS NS — NS NS —Time (normalsubjects) NC NC — NS NS — NS NS —

NeutralGroup NS NS — NS NS — NS NS —Time .04 NS 6 vs. 12 mo, p < .03 NS NS — .04 NS 6 vs. 12 mo, p < .02Group × time NS NS — NS NS — NS NS —Time (normalsubjects) NS NS — NS NS — NS NS —

NoneGroup NC NS — NS NS — NS NS —Time NC NS — NS NS — .015 NS Pre vs. 6 mo, p < .02Group × time NC NS — NS NS — NS NS 6 vs. 12 mo, p < .02Time (normalsubjects) NC NS — NS NS — NS NS —

Note: ADHD = attention-deficit/hyperactivity disorder; NS = not significant; NR = not relevant; NC = could not be computed; Pre =pretreatment.

a Others: all classmates and teachers.

ABIKOFF ET AL.


can influence parents and teachers who serve astreatment agents. Observations also enabled us to ex-amine both discrete and transactional social behaviorsin real-world settings. In accord with scale find-ings, observations failed to yield differential treatmenteffects in social behaviors initiated by or directed to-

ward the treated children. Similarly, the social skillsintervention did not affect children’s responses to oth-ers, despite efforts that specifically targeted such trans-actions.Children in all treatments improved over time on

parent and teacher ratings of social functioning. Im-

TABLE 5Summary of Findings Regarding Others’ Responses to ADHD Children’s Behavior

Responses toADHD Children

Behavior of ADHD Children Toward Othersa

Positive


Negative


Neutral


Year Ip ≤

Year IIp ≤

Year Ip ≤

Year IIp ≤

Year Ip ≤

Year IIp ≤

PositiveGroup NS .007 M + MPT > M +

ACT < .005NR NR — NS NS —

Time NS NS M + MPT >M < .001

NR NR — NS NS —

Group × time NS NS — NR NR — NS NS —Timeb (normalsubjects)

NS NS — NR NR — .011 NS Pre vs. 12 mo, p < .01

NegativeGroup — — — NC NC — NS NS —Time — — — NC NC — NS NS —Group × time — — — NC NC — NS NS —Timec (normalsubjects)

NS NS — NC NC — NS NS —

NeutralGroup NS NC — NC NC — .004 NS M vs. M + MPT, p < .05

M + MPT vs. M + ACT,p < .001

Time NS NC — NC NC — NS NS —Group × time NS NC — NC NC — NS .005 M vs. M + ACT, 24 mo,

p < .03M + MPT vs. M + ACT,12 mo, p < .002, 24 mo,< .05

Timeb (normalsubjects)


NoneGroup NS NC — NC NC — .003 NS M + MPT vs. M + ACT,

p < .001Time NS NC — NC NC — NS NS —Group × time .03 NC — NC NC — NS .04 M vs. M + ACT, 12 mo,

< .05M + MPT vs. M + ACT,12 mo, p < .001

Timeb (normalsubjects)


Note: ADHD = attention-deficit/hyperactivity disorder; NS = not significant; NR = not relevant; NC = could not be computed; Pre =pretreatment.

a Others: all classmates and teachers.b Time effect in normal children.



provement over time is difficult to interpret withoutuntreated control groups to rule out maturationalchanges. Nevertheless, time effects yield heuristic in-formation about the clinical course of stimulant-treatedchildren. Findings indicate a broad impact of methyl-phenidate in reducing negative social interactions, in-cluding those initiated by treated children and thosedirected toward them. Findings complement those re-ported with acute medication (Cunningham et al.,1985). Findings are also consonant with reports ofemanative medication effects on adults’ social behaviorstoward stimulant-treated children with ADHD (Bark-ley and Cunningham, 1979; Humphries et al., 1978;Whalen et al., 1980).It is possible that reductions in negative social inter-

actions over 2 years represent maturational change be-cause similar decreases occurred in comparisons.However, the diminution in negative interactions intreated children was clinically meaningful because theyno longer differed from the classroom norm. At theleast, findings rule out worsening of social functioningin children with ADHD undergoing extended stimu-lant treatment.There were significant reductions in positive social

interactions initiated by the treated children. Is this anindication of general suppression of social behaviorwith methylphenidate treatment? Several findings ar-gue against it. First, comparison children showed acorresponding decrease in positive behaviors. Second,children with ADHD and comparison children did notdiffer significantly in positive social behavior rates atany point. Others have reported similar findings (Kleinand Young, 1979; Pelham and Bender, 1982), suggest-ing that measures of prosocial behaviors may not cap-ture the social impairment of children with ADHD.Importantly, there were no significant changes withtreatment in rates of neutral social interactions initi-ated, a finding that argues against a general dampeningof social interaction.There were significant reductions over time in nega-

tive, positive, and neutral social behaviors directed to-ward study children. These appear to reflect a generaldiminution of social interactions with age, at least ingym classes. Rates of behaviors directed toward studychildren were in the “normal” range because they didnot differ from rates directed toward comparisons. Thelone exception was that significantly higher rates ofpositive behaviors directed toward study children by

peers and teachers than toward comparisons occurredonly at 18-months.Analyses of change in year I relative to status at

medication baseline and of change from 12 to 24months failed to indicate time effects, with the excep-tion of decreases in observed negative social behaviors.This stability suggests that social effects of stimulantmedication occur very early and that, at least for mea-sures obtained at medication baseline, no additionalimprovements accrue with continued treatment. Inlight of this finding, it could be maintained that be-havioral improvement was maximized early with medi-cation and that ceiling effects precluded detectingfurther improvement. Parent ratings argue against thisbecause the mean rating across all children at the end ofyear I was 84.8, far below the norm. Although therewere no further benefits with an additional year ofmaintenance treatment, there was no attenuation ofefficacy in social functioning. This finding parallels re-sults reported in companion articles (Abikoff et al.,2004; Hechtman et al., 2004a,b) regarding mainte-nance of gains across numerous functional domainswith long-term stimulant treatment. As noted, parentratings indicate that children’s social functioning athome remained compromised. It is unknown whichaspects of social behavior remain dysfunctional, butmedication, either alone or combined with the socialskills training provided here, is not sufficient to elimi-nate them.

Limitations

Measures of social behavior consisted of parent,child, and teacher ratings and school observations.Other measures, such as sociometric ratings and friend-ship indices, might have yielded treatment effects.Similarly, it is conceivable that social skills interven-tions for children with ADHD are not targeting ap-propriate social skills. It is also possible that social skillsare learned during training but do not generalize toreal-world settings, possibly due to a lack of reinforce-ment in natural settings or an underlying disturbancein the ability of children with ADHD to generalizelearned social behaviors.

Clinical Implications

The study does not support the expectation thatstimulant-treated children with ADHD benefit from

ABIKOFF ET AL.


clinic-based social skills training. The benefits of stimu-lant treatment on social function do not diminish overtime.

Disclosure: Dr. Abikoff is a member of the ADHD Advisory Board anda principal investigator in clinical trials, Shire Pharmaceutical Co.,and a member of the Metadate CD Advisory Board of Celltech Phar-maceuticals. He is a recipient of an investigator-initiated grant fromMcNeil Consumer and Specialty Pharmaceuticals. Dr. Hechtman re-ceived research funding from Eli Lilly, Janssen Ortho, Purdue, ShirePharmaceutical Co., and GlaxoSmithKline Beecham and is on thespeakers roster of Shire Pharmaceutical Co., Janssen Ortho, and EliLilly. Dr Klein is a member of the ADHD Advisory Board of ShirePharmaceutical Co.

REFERENCES

Abikoff H (1991), Cognitive training in ADHD children: less to it thanmeets the eye. J Learn Disabil 24:205–209

Abikoff H, Gittelman R (1985), Hyperactive children treated with stimu-lants. Is cognitive training a useful adjunct? Arch Gen Psychiatry 42:953–961

Abikoff H, Hechtman L, Klein RG et al. (2004), Symptomatic improve-ment in children with long-term methylphenidate and multimodal psy-chosocial treatment. J Am Acad Child Adolesc Psychiatry 43:802–811

Abikoff H, Martin D, Gittelman R (1985), Social interaction observationcode. Psychopharmacol Bull 21:869–873

Alessandri SM (1992), Attention, play, and social behavior in ADHD pre-schoolers. J Abnorm Child Psychol 20:289–308

Barkley RA, Cunningham CE (1979), The effects of Ritalin on the mother-child interactions of hyperactive children. Arch Gen Psychiatry 36:201–208

Barton EJ (1986) Modification of children’s pro-social behavior. In: Chil-dren’s Social Behavior: Development, Assessment, and Modification, StrainPS, Guralnick MJ, Walker HM, eds. Orlando, FL: Academic Press, pp331–372

Buhrmester D, Whalen CK, Henker B, MacDonald V, Hinshaw SP (1992),Prosocial behavior in hyperactive boys: effects of stimulant medicationand comparison with normal boys. J Abnorm Child Psychol 20:103–121

Cunningham CE, Siegel LS, Offord DR (1985), A developmental dose-response analysis of the effects of methylphenidate on the peer interac-tions of attention deficit disordered boys. J Child Psychol Psychiatry26:955–971

Dodge KA, McClaskey CL, Feldman E (1985), Situational approach to theassessment of social competence in children. J Consult Clin Psychol53:344–353

Gadow KD, Nolan E, Sprafkin J, Sverd J (1995), School observations ofchildren with attention deficit hyperactivity disorder and comorbid ticdisorder: effects of methylphenidate treatment. Dev Behav Pediatr16:167–176

Gillberg C, Melander H, von Knorring A et al. (1997), Long-term centralstimulant treatment of children with attention-deficit hyperactivity dis-order: a randomized double-blind placebo-controlled trial. Arch GenPsychiatry 54:857–864

Greene RW, Biederman J, Faraone SV, Sienna M, Garcia-Jetton J (1997),Adolescent outcome of boys with attention-deficit/hyperactivity disor-

der and social disability: results from a 4-year longitudinal follow-upstudy. J Consult Clin Psychol 65:758–767

Gresham FM, Elliott SN (1990), Social Skills Rating System: Manual. CirclePines, MN: American Guidance Service

Hechtman L, Abikoff H, Klein RG et al. (2004a), Academic achievementand emotional status of children with ADHD treated with long-termmethylphenidate and multimodal psychosocial treatment. J Am AcadChild Adolesc Psychiatry 43:812–819

Hechtman L, Abikoff H, Klein et al. (2004b), Children with ADHDtreated with long-term methylphenidate and multimodal psychosocialtreatment: impact on parental practices. J Am Acad Child Adolesc Psy-chiatry 43:830–838

Hinshaw SP, Henker B, Whalen CK (1984), Self-control in hyperactiveboys in anger-inducing situations: effects of cognitive-behavioral train-ing and of methylphenidate. J Abnorm Child Psychol 12:55–77

Hinshaw SP, Henker B, Whalen CK, Erhardt D, Dunnington RE (1989),Aggressive, prosocial, and nonsocial behavior in hyperactive boys: doseeffects of methylphenidate in naturalistic settings. J Consult Clin Psychol57:636–643

Humphries T, Kinsbourne M, Swanson J (1978), Stimulant effects oncooperation and social interaction between hyperactive children andtheir mothers. J Child Psychol Psychiatry 19:13–22

Jackson NF, Jackson DA, Monroe C (1983), Getting Along With Others:Teaching Social Effectiveness to Children. Champaign, IL: Research Press

Klein RG, Abikoff H (1997), Behavior therapy and methylphenidate in thetreatment of children with ADHD. J Atten Disord 2:89–114

Klein RG, Abikoff H, Hechtman L, Weiss G (2004), Design and rationaleof controlled study of long-term methylphenidate and multimodal psy-chosocial treatment in children with ADHD. J Am Acad Child AdolescPsychiatry 43:792–801

Klein AR, Young RD (1979), Hyperactive boys in their classroom: assess-ment of teacher and peer perception, interactions, and classroom be-haviors. J Abnorm Child Psychol 7:425–442

MTA Cooperative Group (1999), A 14-month randomized clinical trial oftreatment strategies for attention-deficit/hyperactivity disorder. ArchGen Psychiatry 56:1088–1096

Ollendick TH, Weist MD, Borden MC, Green RW (1992), Sociometricstatus and academic, behavioral, and psychological adjustment: a five-year longitudinal study. J Consult Clin Psychol 60:80–87

Parker JG, Asher SR (1987), Peer relations and later personal adjustment:are low-accepted children at risk? Psychol Bull 102:357–389

Pelham WE, Bender ME (1982), Peer relationships in hyperactive children:description and treatment. Adv Learn Behav Disabil 1:365–436

Pfiffner LJ, McBurnett K (1997), Social skills training with parent organi-zation: treatment effects for children with attention deficit disorder. JConsult Clin Psychol 65:749–757

Turkewitz H, O’Leary KD, Ironsmith M (1975), Generalization and main-tenance of appropriate behavior through self-control. J Consult ClinPsychol 43:577–583

Walker HM, McConnell S, Holmes D, Todis B, Walker J, Golden N(1983), The Walker Social Skills Curriculum: The Accepts Program. Aus-tin, TX: Pro Ed Resources

Whalen CK, Henker B (1985), The social worlds of hyperactive (ADDH)children. Clin Psychol Rev 5:447–478

Whalen CK, Henker B (1991), Therapies for hyperactive children: com-parisons, combinations, and compromises. J Consult Clin Psychol59:126–137

Whalen CK, Henker B, Dotemoto S (1980), Methylphenidate and hyper-activity: effects on teacher behaviors. Science 208:1280–1282

Whalen CK, Henker B, Swanson JM, Granger D, Kliewer W, Spencer J(1987), Natural social behaviors in hyperactive children: dose effects ofmethylphenidate. J Consult Clin Psychol 55:187–193



Documents

Social Functioning in Children With ADHD Treated With Long-Term Methylphenidate and Multimodal Psychosocial Treatment