Hyperactivity Disorder

SPECIAL SECTION

Executive Functioning Among Finnish Adolescents With Attention-Deficit/Hyperactivity Disorder SANDRA K. LOO, PH.D., LORIE A. HUMPHREY, PH.D., TERTTU TAPIO, M.A., IRMA K. MOILANEN, M.D., PH.D., JAMES J. McGOUGH, M.D., JAMES T. McCRACKEN, M.D., MAY H. YANG, M.A., JEFF DANG, M.P.H., ANJA TAANILA, PH.D., HANNA EBELING, M.D., PH.D., MARJO-RIITTA JA¨RVELIN, M.D., PH.D., AND SUSAN L. SMALLEY, PH.D.

ABSTRACT Objective: The present study examined cognitive functioning in a population sample of adolescents with and without attention-deficit/hyperactivity disorder (ADHD) from the Northern Finland Birth Cohort 1986. Method: The sample consisted of 457 adolescents ages 16 to 18 who were assessed using a battery of cognitive tasks. Performance according to diagnostic group (control, behavior disorder, and ADHD) and sex was compared. Then, the effect of executive function deficit (EFD) was assessed by diagnostic group status on behavioral and cognitive measures. Results: When compared to non-ADHD groups, adolescents with ADHD exhibited deficits on almost all of the cognitive measures. The behavior disorder group obtained scores that were generally intermediate between the ADHD and control groups, but exhibited deficits in intelligence and executive function similar to the ADHD group. Approximately half the ADHD sample had EFD; however, the type and presence of EFDs were not differentially related to cognitive performance as a function of diagnosis. Conclusions: These findings indicate that EFDs are more frequent in ADHD than control or behavior disorder groups. EFDs are a general risk factor for poor cognitive functioning across multiple domains, irrespective of diagnostic status. J. Am. Acad. Child Adolesc. Psychiatry, 2007;46(12):1594Y1604. Key Words: executive function, neuropsychological, impairment, working memory, inhibition.

Cognitive functioning deficits have been well documented among children with attention-deficit/hyperactivity disorder (ADHD) and are apparent across the life span from early childhood to middle adulthood (Seidman, 2006). These problems in neuropsychologiAccepted July 26, 2007. Drs. Loo, Humphrey, McGough, McCracken, and Smalley and Ms. Yang and Mr. Dang are with the University of California, Los Angeles; Drs. Moilanen, Taanila, Ebeling, and Ja¨rvelin and Ms. Tapio are with the University of Oulu; and Dr. Ja¨rvelin is with the Imperial College London. This research was supported by National Institute of Mental Health grants MH063706 (Smalley, Ja¨rvelin), and MH01966 (McGough), the Juselius Foundation, and the Academy of Finland. Article Plus (online only) materials for this article appear on the Journal’s Web site: www.jaacap.com. Correspondence to Dr. Susan L. Smalley, UCLA Semel Institute, Room 47438, 760 Westwood Plaza, Los Angeles, CA 90095-1759; e-mail: ssmalley@ mednet.ucla.edu. 0890-8567/07/4612-1594Ó2007 by the American Academy of Child and Adolescent Psychiatry. DOI: 10.1097/chi.0b013e3181575014

1594

cal functioning converge nicely with brain imaging findings on the pathophysiology of ADHD as well as on the neurobiology of specific cognitive processes such as attention, inhibition, and working memory (WM). From all of these lines of inquiry comes a picture of atypical activity in the frontal corticostriatal network and frontoparietal attention network in ADHD resulting in impaired performance on measures of both executive and nonexecutive cognitive functions. Although there is no one definition of executive function (EF), we use it here to describe higher cognitive control processes used to attain a future goal by problem solving and carrying out a solution (Pennington and Ozonoff, 1996). Although EFs may encompass many different cognitive processes, most would agree that response inhibition (RIN), interference control, WM, and set-shifting are components of EF. We adopt that definition here and acknowledge that data are presented on several other cognitive mechanisms that

J. AM. ACAD. CHILD ADOLE SC. PSYCHIATRY, 46 :12, DECE MBER 2007

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

EXECUTIVE FUNCTIONING IN ADHD

others may consider an EF but for our purposes are considered non-EF processes. A wealth of literature suggests that EFs are associated with ADHD. For example, a meta-analysis of 83 studies found that EFs have moderate effect sizes (ESs) in ADHD ranging from 0.46 to 0.69 (Willcutt et al., 2005). Although the predominant focus has been on EFs, non-EFs such as response variability have also received considerable support as being associated with ADHD (Castellanos et al., 2006). EF deficits (EFDs) appear to continue across the life span into adolescence and adulthood, although the majority of studies have focused on children between the ages of 7 and 12. Studies of adults with ADHD have found that EFDs not only continue into adulthood (Schoechlin and Engel, 2005), but may have a larger ESs than in childhood (Lijffijt et al., 2005). There are surprisingly few studies that have focused on adolescents with ADHD, although existing studies are consistent in implicating similar deficits in adolescents (Seidman et al., 1997). Although consistent group differences according to diagnosis have been shown, the diagnostic efficiency of cognitive measures is ~70% (Doyle et al., 2000) due to a significant proportion of children diagnosed with ADHD exhibiting normal performance on various measures. Recent studies have shown that approximately 50% of children with ADHD have normal performance on any given cognitive task or process (Nigg et al., 2005). In addition, EFDs are not specific to ADHD and are seen in many other psychiatric disorders such as other disruptive behavior disorders, autism, and schizophrenia. These findings suggest that group level differences may be driven by a subgroup of ADHD children with significant cognitive impairment and that cognitive deficits are neither necessary nor sufficient to cause ADHD. This has led many to question the concept of EFDs as core in ADHD and to reformulate the question to address whether there are neuropsychologically impaired subgroups within ADHD. Biederman and colleagues (2004) examined the impact of EFDs on multiple domains of functioning and found that children with ADHD who had impaired scores on at least two EF tests had significantly more grade retentions, lower academic achievement scores, and higher rates of learning disabilities when compared to ADHD children without EFDs. ADHD children with and with-

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 46:12, DECEMBER 2007

out EF impairment were similar, however, in symptom severity, psychiatric comorbidity, and social functioning. This concept of neuropsychologically impaired subtypes remains exploratory and requires further evaluation; this study is a step in that direction. In the Northern Finnish Birth Cohort (NFBC) 1986 sample, we have a unique opportunity to examine cognitive functioning in one of the largest samples of adolescents with and without ADHD. The purpose of the present study is twofold. First, we characterize cognitive performance across diagnostic groups by testing the hypothesis that adolescents with ADHD exhibit more cognitive deficits compared to those without ADHD (i.e., controls and a behavior disorder group). Furthermore, based on previous findings that sex may play a role in the neuropsychological heterogeneity of ADHD (Nigg et al., 2002), we will test whether girls with ADHD have more severe cognitive deficits than boys with ADHD. Second, we sought to examine the relationship of EFDs and diagnostic status on behavioral and non-EF cognitive measures. We hypothesized that the EF impairment within the ADHD group would be uniquely associated with differences in cognitive functioning such that the ADHD group with EFDs would exhibit significantly worse performance on these measures. METHOD Participants The NFBC 1986 is a population cohort of 9,432 children born alive and part of a program of longitudinal study of health and well-being in northern Finland (Ja¨rvelin et al., 1993). Participants in the present study represent a nested case-control subset of 457 adolescents, ages 16 to 18 years, who were directly assessed for ADHD as part of the NFBC-ADHD study (Smalley et al., 2007). All of the participants signed informed consent based on the University of Oulu and University of California, Los Angeles institutional review boards. Sampling methods and diagnostic procedures are given in the accompanying article by Smalley et al. (2007). Briefly, the cohort was screened for ADHD using the Strengths and Weaknesses of ADHD-Symptoms and NormalBehavior scales (SWAN) (Swanson et al., 2001). Among the screened sample, 457 adolescents participated in a clinical assessment to diagnose ADHD based on a direct psychiatric interview and best estimate procedure. Based on these criteria, 188 subjects were diagnosed with lifetime probable or definite ADHD, 166 were controls (SWAN-defined controls, no ADHD), and 103 were considered a behavior disorder (BD) group (BD, SWAN-defined cases, no ADHD diagnosis). A complete description of the three groups is given in Smalley et al. (2007). A comprehensive cognitive battery was administered over 3 hours by master`s degreeYlevel Finnish psychologists. None of the

1595

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

LOO ET AL.

children were taking psychostimulant therapy due to any condition. All of the subjects completed the testing successfully. Tests were chosen if they measured cognitive processes previously reported to be deficient in ADHD, and if Finnish versions were not available, then the tests had to be easily adaptable for Finnish subjects (i.e., computerized tasks). Tests estimating intelligence were given first to control for the impact of inattention on those results; all of the other tests were given in a counterbalanced fashion to minimize order effects. Cognitive Measures Intelligence was estimated using the WAIS-R (Wechsler, 1992) Vocabulary and Block Design subtests. Subtests from the Reading Association in Finland (Text reading time, Word reading time) were given to estimate language-based academic abilities. The Stop-Signal task (Logan et al., 1997) and the Attentional Network Task (ANT; Fan et al., 2002) were administered as measures of RIN. The Stop-Signal Reaction Time is calculated by subtracting the mean delay required by a subject to correctly inhibit 50% of responses from the mean reaction time on go-trials. Two other measures are derived from the Stop-Signal task: the Go reaction time (Go-RT) and the SD of the reaction time (SDRT), which are measures of response speed and variability. The Attention Network Task is a flanker test (using arrows) that also reports cued reaction times. Several conditions exist, including congruent flankers (all pointing left), incongruent flankers (target pointing right while flankers point left), or a neutral condition (no arrows, only lines). The Bconflict^ variable measures the effect of interference and is defined as the difference between the subject`s reaction time to congruent versus incongruent flankers. The Conners Continuous Performance Test II (CPT; Conners, 2000) was administered as a measure of sustained attention. The CPT generates several dependent measures including errors of omission and commission, reaction time, and variability, as well as $ (response style) and d-prime (stimulus discrimination). The rest of the cognitive assessment included tests from larger, standardized test batteries. The two subtests making up the WAIS-III Processing Speed Index were given (Digit-Symbol/ Coding and Symbol Search). In addition, the two subtests from the WAIS-III were administered (Digit Span and Arithmetic), as well as the Spatial Span and Letter-Number-Sequences subtest from the Wechsler Memory Scale (Wechsler, 1981). The Verbal Fluency Test and the Fingertip Tapping Test from the NEPSY (Korkman et al., 1998) were also given. Finally, the Trailmaking test (Reitan, 1979) was administered, though was modified to include an alphabet-only trial between Trails A and B. Data Analyses All of the analyses were conducted using SPSS, version 12.1 (SPSS, 2003). All of the data were checked for normality (skewness and kurtosis) and log-transformed when these indicators suggested non-normal distribution (word reading time, text reading time). Due to multiple comparisons, a p value of .01 was used as a threshold for statistical significance to balance type I and II error; findings with a p value of .05 to .02 were considered a trend toward significance. Analysis of variance (ANOVA) were conducted for each dependent variable and post hoc (Tukey honestly significant difference) tests to determine the pattern of diagnostic group differences were conducted when the ANOVA was significant. In addition, ESs were calculated using Cohen`s f statistic and in-

1596

terpreted according to the following guidelines: 0.10 = small effect, 0.25 = medium effect, and 0.40 = large effect (Cohen, 1988). As suggested by Barkley (1997), results are presented both with and without estimated intelligence as a covariate. A principal components analysis (PCA) with oblique rotation was used with the EF tasks from the cognitive battery for the purposes of data reduction. Scores on each component were created using linear combinations of the standardized variables weighted by their eigenvector. The cutoff for EFD was defined as the extreme 10th percentile of scores (directionality depended on whether higher or lower scores indicated poor performance) in the normal control sample following a method used by Nigg (2005). EFD was then compared across diagnostic groups, ADHD subtypes, and sex using x2 statistical tests. RESULTS Cognitive Functioning by Diagnosis

ESs and results of the analyses by diagnosis and sex are presented in Table 1. Means and SDs for diagnostic groups and gender are available on the Journal’s Web site at www.jaacap.com through the Article Plus feature. There were numerous diagnostic group differences in cognitive functioning, even when controlling for IQ, and these include reading fluency, WM, inhibition tasks, response variability, and set-shifting. ESs were generally small to medium, with the largest ESs emerging for estimated Full Scale IQ (0.35) and WAIS Digit-Symbol/ Coding (0.37). Pairwise comparisons indicated better cognitive functioning in the control group when compared to both the ADHD and BD groups on most measures. Reading fluency and processing speed (WAIS Coding) were the only tasks that were significantly different across all three diagnostic groups. The BD group obtained scores that were generally intermediate between the ADHD and control groups on all measures, but they had deficits similar to those of the ADHD group in intellectual functioning and EFs such as interference control and WM. To evaluate the role of ADHD symptomatology on cognitive test performance in the BD group, we compared controls versus BD group and BD group versus ADHD separately while adjusting for the number of ADHD symptoms. Group differences between control and BD groups did not change for any test examined. Comparisons between the ADHD and BD groups reveal similar results; however, several group differences became nonsignificant after controlling for ADHD symptom severity: text and word reading time, WAIS Digit Span and CPT Commission Errors, and d-prime. One test remained statistically significant (Trails B: F1,282 = 7.07, p = .008), whereas two others became

J. AM. ACAD. CHILD ADOLE SC. PSYCHIATRY, 46 :12, DECE MBER 2007

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

EXECUTIVE FUNCTIONING IN ADHD

trends toward significance (Trails Alphabet A: F1,284 = 5.34, p = .02 and WAIS Digit-Symbol/Coding: F1,285 = 4.70, p = .03). Sex differences were evident for several cognitive tasks, with females exhibiting better performance on verbal fluency, fine motor control, reading fluency, setshifting, and processing speed than males, whereas males were relatively stronger on measures of visual orienting, interference control, verbal WM, and computational skill. Most sex differences were robust to using intelligence as a covariate; however, all of the ESs were small according to Cohen’s f . One significant sex  diagnostic group interaction effect emerged on the WAIS Coding subtest (F2,446 = 4.73, p = .009), which remained significant after controlling for the effects of IQ ( p = .01). The interaction effect suggests that the discrepancy between girls with and without ADHD on Coding was significantly larger than the discrepancy between their male counterparts. No other sex  diagnostic group interaction reached statistical significance. Behavioral and Cognitive Functioning Within EFD Subgroups

Principal Components Analysis of EF Tasks. PCA was performed on EF measures and two components with eigenvalues >1 were retained for further analysis. CPT commission errors loaded equally poorly (loading = 0.3) on each component and was therefore removed from the PCA. Based on the estimated weights, these components were described as reflecting primarily WM and RIN. Eigenvectors on specific tasks were WAIS-III Digit Span (WM = 0.75, RIN = j0.23), Trails B (WM = j0.68, RIN = 0.29), WMS Spatial Span (WM = 0.58, RIN = j0.03), WMS LetterNumber Sequencing (WM = 0.79, RIN = j0.37), WAIS-III Arithmetic (WM = 0.73, RIN = j0.18), Stop-Signal Reaction Time (WM = j0.16, RIN = 0.84), and ANT Conflict (WM = j0.29, RIN = 0.68). Using the component score distributions in the control sample, we defined cutoff scores for RIN and WM based on poor performance at or exceeding the 10th percentile. This cutoff was then applied to the whole sample to identify individuals with significant EFDs (i.e., RIN, WM, or both). There was significant overlap (x2 = 19.63, p < .0001) between RIN and WM deficits. As a result, we compared groups that had no

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 46:12, DECEMBER 2007

impairment (No EFD), pure inhibition deficit (RIN), pure WM deficit, and both inhibition and WM deficit (Both). Frequency of EFD. Among diagnostic groups, EFD (RIN, WM, or Both) was significantly more likely in the ADHD group than in the control sample. Half (52%) of the ADHD sample were classified as having an EFD compared with 29% in the BD group and 18% in the controls (x2 = 51.8, p < .0001; note that the 18% in controls reflects the either/or rule for defining EFD as RIN and WM alone represents 10% by definition). The ADHD group had significantly higher frequencies of WM deficits with and without RIN than the BD and control groups; rates of pure WM deficits were seen in 30% of the ADHD group compared to 21% of the BD group and 8% of the control group. Similarly, 12% of the ADHD group had deficits on both the WM and RIN components, whereas 7% of the BD and 2% of the control groups had deficits on both. The frequency of pure RIN deficit was not particularly different across diagnostic groups with 9% of the ADHD group, 8% of the controls, and 1% of the BD group showing RIN deficit. The distribution of EFDs across ADHD subtypes did not reach statistical significance (e.g., Combined type 48%, Inattentive type 60%, Hyperactive-Impulsive type 35%; x2 = 6.48, p = .37). Similarly, sex differences in EFDs were not significant and indicated that more than half of males and females in the total sample have no EFD (65% versus 68%, respectively; x2 = 3.78, p = .38). Behavioral and Cognitive Correlates of EFDs. We tested for the impact of specific EFDs (No EFD, RIN, WM, or Both), association of EFD by diagnostic group status (control, BD, and ADHD), and the effect of EFD specifically within the ADHD group on behavioral and cognitive measures. To reduce the number of comparisons, only measures with significant diagnostic group differences were included for these analyses; the results are summarized in Table 2. Means and SDs for the EFD groups are available on the Journal’s Web site at www.jaacap.com through the ArticlePlus feature. Impact of Specific EFDs. When examining specific EFDs (No EFD, RIN, WM, or Both) across the whole sample, the No EFD group had significantly lower SWAN scores ( p < .0001) than the WM and Both groups, which did not differ from each other. The RIN group did not differ from the No EFD group on any of the SWAN ratings. Cognitively, individuals classified

1597

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

General intelligence: WAIS Vocabulary Block design Estimated FSIQ Reading fluencya Text reading time Word reading time Stop-Signal taska Stop Signal Reaction Time Go Reaction Time SDRT WAIS subtests Arithmetic Digit-Symbol/Coding Symbol search Digit Span Wechsler Memory Scale Spatial Span LNS Trailmaking Testa Trails A time Trails Alphabet A time Trails B time

1598 0.29 0.31 0.35 0.27 0.27 0.14 0.00 0.23 0.29 0.37 0.25 0.25 0.18 0.29 0.14 0.20 0.27

17.11y 17.39y 5.05** <1 10.75y 18.30y 30.15y 13.35y 15.23y 6.93*** 18.71y 5.15** 9.90y 16.46y

Effect Size Cohen`s f

19.05y 21.89y 27.97y

ANOVA F2,450

C
C9BD,AD C9BD,AD

C9BD,AD C9BD9AD C9BD,AD C,BD9AD

C
C
C9BD,AD C9BD,AD C9BD,AD

Pairwise Comparisons

1.68 3.33 5.86**

1.70 5.45**

6.22** 14.88y 3.64* 6.63**

2.28 <1 3.87*

7.10*** 10.02y

V V V

IQ Adjusted F2,449

<1 19.48y 14.39***

<1 <1

5.42* 42.24y 12.05*** 1.27

<1 <1 <1

20.31y 8.80**

2.11 2.99 5.06*

ANOVA F1,447

TABLE 1 Cognitive Performance by Diagnostic Group and Sex Diagnostic Status

0.00 0.20 0.18

0.00 0.00

0.10 0.31 0.18 0.10

0.00 0.00 0.00

0.20 0.14

0.10 0.10 0.10

Effect Size Cohen`s f

Sex

V M9F M9F

V V

M9F M
Y Y Y

M9F M9F

M
Direction of Difference

<1 14.27*** 9.09**

<1 <1

13.81y 38.13y 7.58** <1

<1 <1 <1

12.87*** 5.13*

V V V

IQ Adjusted F1,446

LOO ET AL.

J. AM. ACAD. CHILD ADOLE SC. PSYCHIATRY, 46 :12, DECE MBER 2007

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

0.10 0.10 0.10 0.00 0.00 0.10 0.18 0.18 0.20 0.00 0.00 0.20 0.23 0.10 0.25 0.18 0.14

2.47 1.76 1.54 <1 <1 2.25 7.98*** 5.67** 9.91y <1 <1 7.17*** 9.40y <1 11.02y 6.47*** 2.73

C9BD,AD C,AD9BD V C9BD,AD C,AD9BD V

C
V V V V V V C>AD C9AD

3.42* 5.72** 1.04 5.39** 4.21* 1.94

3.90* 1.10 <1

1.57 1.48 <1 <1 <1 <1 <1 <1

3.37 1.22 1.94 <1 4.40* <1

3.48 6.55** 7.06**

6.88** 4.06* 5.87** 6.70** 3.22 15.40y 23.49y 24.98y

0.10 0.00 0.10 0.00 0.10 0.00

0.10 0.14 0.14

0.14 0.10 0.10 0.10 0.10 0.18 0.23 0.23

V V V V M9F V

V M
M
1.51 <1 1.84 <1 2.64 1.27

7.74** 5.30* 5.92*

8.57** 4.83* 4.76* 5.69* 2.32 10.32*** 16.26y 18.30y

Note: Cohen`s f represents effect size and is interpreted according to the following guidelines: 0.10 = small effect, 0.25 = medium effect, and 0.40 = large effect. IQ = estimated intelligence; FSIQ = Full Scale IQ; SDRT = SD of the reaction time; LNS = letter number sequencing; FT = finger tapping; VF = verbal fluency; ANT = Attention Network Task; CPT = Continuous Performance Task; C = control group; BD = behavioral disorder group; AD = ADHD group; M = males; F = females. Tests that are in italics were used in the principal components analysis of executive function tasks. a Denotes that higher score indicates poor performance. Means and SDs are available online. * p e .05; **p e .01,*** p e .001, y p e .0001.

NEPSY FT repetitions preferreda FT repetitions nonpreferreda FT sequences preferreda FT sequences nonpreferreda FT total timea VF category VF phonological VF total score ANTa Conflict Alerting Orientinga Conners CPT Omission errors Commission errors Reaction time Reaction time variability d-prime $

EXECUTIVE FUNCTIONING IN ADHD

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 46:12, DECEMBER 2007

1599

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

1600 0.29 0.40 0.33 0.31 0.42 0.40

0.37 0.42 0.25 0.29 0.31 0.27 0.39

12.34y 30.14y 15.22y 14.01y 23.00y 20.95y 18.72y 22.78y 8.56y 11.66y 13.70y 10.16y 20.10y 14.54y

WAIS Vocabulary WAIS Block Design WAIS subtests Digit Symbol/ Coding Symbol Search

Reading fluencya Text reading time

Word reading time Verbal fluency NEPSY: category NEPSY: phonological NEPSY: total Trailmaking Testa Trails A Trails Alphabet A

Stop Signal Testa SDRT 0.33

0.35 0.23

15.90y 6.55***

SWAN behavior ratingsa Inattentive Hyperactiveimpulsive Combined General intelligence WAIS Est FSIQ

ANOVA Effect Size F3,403 Cohen`s f

10.86y

5.90*** 11.50y

4.00**

2.65* 2.91*

14.97y

9.68y

10.04y

11.78y

V V

V

5.03**

6.72*** 2.75*

IQ Covary F3,402

NI, WM
NI
NI>WM, RIN, Both

NI>WM, Both NI>WM, RIN, Both

NI, WM, RIN
NI>WM, RIN, Both; RIN>Both NI>WM, RIN, Both; RIN>Both

Dx F2,400

8.33**

19.47y 30.97y

25.88y

16.96y 19.50y

31.21y

19.39y

40.56y

37.71y

3.68*

1.73 4.03*

<1

<1 1.19

6.06**

8.49***

5.09**

8.54***

5.44** 8.18***

8.60***

5.56* 135.92y

13.24***119.02y <1 84.92y

EFD F1,400

NI>WM, RIN, 30.14y Both; RIN>Both NI>WM, Both; RIN>Both 17.29y NI>WM, RIN, Both 19.61y

NI
NI
Pairwise Comparisons

<1

<1 <1

3.67*

3.55* 2.05

<1

<1

<1

1.32

2.43 <1

1.35

<1

<1 1.62

Dx x EFD F2,400

TABLE 2 Behavioral and Cognitive Correlates of Executive Functioning Deficits Type of EFD EFD x Dx

0.14

0.00 0.00

0.14

0.14 0.10

0.00

0.00

0.00

0.10

0.10 0.00

0.10

0.00

0.00 0.10

5.97*

8.96** 22.37y

10.88***

7.45** 9.26**

16.42y

16.46y

12.04***

14.63***

15.39y 11.52***

22.22y

4.11*

5.35* 1.35

0.20

0.23 0.37

0.25

0.20 0.23

0.31

0.31

0.27

0.29

0.31 0.27

0.37

0.14

0.18 0.10

2.46

4.93* 14.58***

2.87

1.89 2.57

11.35***

8.35**

5.33*

5.73*

V V

V

3.04

4.17* <1

IQ Covary F1,162

ADHD T EFD Effect Size ANOVA Effect Size Cohen`s f F1,163 Cohen`s f

LOO ET AL.

J. AM. ACAD. CHILD ADOLE SC. PSYCHIATRY, 46 :12, DECE MBER 2007

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

6.31** 0.00 <1 2.80 12.75*** NI
4.58**

4.22* 8.02** 2.69 0.00 0.00 0.00 <1 <1 <1 5.99** 4.74** 5.79** 4.27* 17.50y 5.69* NI
Conners CPTa Omission errors Commission errors Reaction time variability D-prime

Note: Cohen`s f represents effect size and is interpreted according to the following guidelines: 0.10 = small effect, 0.25 = medium effect, and 0.40 = large effect. EFD = executive function deficit; Dx = diagnosis; ADHD = attention-deficit/hyperactivity disorder; IQ = estimated intelligence; FSIQ = Full Scale IQ; SDRT = SD of Reaction Time; CPT = Continuous Performance Test; NI = No EFD; WM = working memory deficit; RIN = response inhibition deficit; Both = WM and RIN deficit. a Denotes that higher score indicates poor performance. * p e .05; **p e .01; ***p e .001; y p e .0001. Means and SDs are available online via the Article Plus feature.

4.61* 0.20

<1 5.55* <1 0.18 0.23 0.14

Effect Size IQ Covary Cohen`s f F 1,162 1,163

Dx x EFD Effec t Size F 2,400 Cohen`s f ANOVA F Dx F 2,400 ANOVA Effect Size IQ Covary F 3,402 F 3,403 Cohen`s f

Type of EFD

Pairwise Comparisons

EFD F

TABLE 2 Continued

1,400

EFD x Dx

ADHD T EFD

EXECUTIVE FUNCTIONING IN ADHD

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 46:12, DECEMBER 2007

as having any EFD scored worse on nearly all of the cognitive measures that were not used in the principal components analysis than the No EFD group (Table 2). For measures of intelligence, fine-motor functioning, and reading fluency, the EFD groups had significantly lower scores than the No EFD group, regardless of type of EFD (WM, RIN, or Both). Although the Both group obtained the lowest scores on almost all of the measures, the only statistically significant difference from the single impairment groups (RIN and WM) was on the reading fluency measures. The pure RIN and WM impaired groups did not differ from one another on any cognitive measure except response variability (where RIN had increased response variability). The RIN group had lower scores than the No EFD group on almost all of the measures, but significantly higher scores on measures of intelligence, fine-motor processing, response speed, reading fluency, and letter fluency than the Both group. Association of EFD by Diagnostic Status. To examine whether the cognitive and behavioral correlates of EFDs differed across diagnostic group, we analyzed the main effects of EFD and diagnostic status as well as their interaction. Because the cell sizes were too small to analyze according to individual EF type (RIN, WM, and Both), we collapsed the three groups and included them together in the EFD group. Overall, analyses revealed significant main effects of EFD and diagnostic status; however, the interaction effect of EF impairment and diagnosis was not significant for any measure (Table 2). Inspection of mean scores indicates that the scores for the EFD groups were generally lower than those without EFD, regardless of diagnosis. The lack of significant interaction effects indicates that the impact of EFD is similar across the diagnostic groups and that this finding is consistent across various domains of cognitive functioning. Behaviorally, the main effect of EFD on SWAN behaviors is with the inattention dimension (F1,401 = 13.24, p < .001), whereas its association with SWAN Hyperactivity/ Impulsivity was not significant (F1,401 < 1, not significant). Reflecting the trend toward significance of EFD on SWAN Inattentive behaviors, the EFD association with SWAN Combined behaviors trended toward significance (F1,401 = 5.56, p = .02). Presented in Table 2 are the p values for the main effects of EFD impairment and the interaction effect of EFD by diagnosis.

1601

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

LOO ET AL.

Effect of EFD Within the ADHD Group. To assess the behavioral and cognitive correlates of EFD within ADHD, we compared those with EFD (ADHD + EFD) and without EFD (ADHD j EFD). As seen in Table 2, the ADHD j EFD and ADHD + EFD groups showed a trend toward significance on two of the SWAN scales of ADHD behaviors: inattentive behaviors (F1,163 = 5.35, p = .03) and combined (F1,163 = 4.11, p = .04) and the ESs are small (e0.18). This suggests that having EFDs do not significantly affect symptom severity among adolescents with ADHD and that the presence of EF impairment may be difficult to detect using only behavior rating scales. Cognitively, the ADHD + EFD group exhibited worse performance than the ADHD j EFD group on all of the measures; however, these differences were significant only on measures of reading and letter fluency after covarying for differences in estimated IQ. DISCUSSION

The purpose of the present article is to examine cognitive functioning among adolescents with and without ADHD in the NFBC. The data presented herein afford three conclusions. First, the cognitive functioning patterns, both in terms of sex and ADHD status, are largely comparable to other populations and clinical studies of cognition in ADHD in that a host of cognitive performance differences are evident (Biederman et al., 2004; Seidman, 2006). Second, EFDs are significantly more frequent in ADHD. Using a 10th percentile definition of EF impairment in the control sample, 52% of adolescents with ADHD exhibit impairment in WM, RI, or both compared with 18% in controls and 29% in the non-ADHD BD group. Third, EFDs are associated with poorer cognitive performance on a host of cognitive tasks irrespective of diagnostic status. Consistent with previous literature and as hypothesized, adolescents with ADHD exhibit deficits in nearly all of the EF and non-EF processes. The BD group obtained scores that were generally intermediate between the ADHD and control groups on all of the measures, but had deficits similar to the ADHD group in intellectual functioning and EFs such as interference control and WM. This group primarily consists of adolescents with oppositional defiant disorder (ODD)/ conduct disorder (CD) and elevated levels of ADHD

1602

symptoms (Smalley et al., 2007). Because ODD and CD are often comorbid with ADHD, it is of interest that these adolescents are intermediate between ADHD and controls, suggesting that they may represent a less severe form of ADHD. In the present study, adjusting for elevated levels of ADHD symptoms in the BD group did not substantially alter the pattern of diagnostic group differences. Similarly, controlling for comorbid ODD/CD within the ADHD and BD groups was not associated with worse scores on the EF components (data not shown). This suggests that the mid-level EF problems observed in the BD group are not caused by increased ADHD symptomatology or ODD/CD per se and that better characterization of the BD sample is warranted. Our impairment analysis indicates that 52% of ADHD adolescents show impairment in RIN, WM, or Both compared to 29% and 18% of the BD and control groups, respectively. The definition of impairment used here is the same as that used by Nigg et al. (2005) and similar to that used by Biederman et al. (2004), who used a 1.5-SD cutoff, which corresponds with the 93rd percentile. All three studies have remarkably similar findings demonstrating that approximately half of children and adolescents with ADHD have EFD, making EF impairment either an etiologic variable in a subset of individuals with ADHD or an associated feature or common comorbidity seen with ADHD. Despite differences in frequency of EF impairment, there were few cognitive correlates that differed between subjects who were classified as impaired in RIN, WM, or both. A surprising and unexpected result was the substantial role of WM deficits in ADHD as indicated by ESs that were uniformly in the medium size range and a more than threefold increase in pure WM deficits in the ADHD group compared with controls (30% versus 8%, respectively). Although a few associations with individual processes emerged (WM associated with verbal fluency deficits, RIN associated with increased response variability, and Both associated with slower reading and letter fluency), the most consistent finding is between the No EFD group and all of the other groups. This finding suggests that having one or more EFD is associated with more generalized cognitive dysfunction and that the effects of specific processes are not easily dissected. Just as the type of EFD was not associated with different cognitive correlates, the presence of EFDs was

J. AM. ACAD. CHILD ADOLE SC. PSYCHIATRY, 46 :12, DECE MBER 2007

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

EXECUTIVE FUNCTIONING IN ADHD

not differentially related to cognitive performance as a function of diagnosis. The present results suggest that the ADHD + EFD subgroup is distinguishable by lower intelligence and slower reading fluency that may not be readily identifiable by traditional diagnostic probes (behavior ratings). These findings converge with those reported by Biederman et al. (2004), who found that the impact of EFD in ADHD is evident primarily in lower academic achievement rather than in behavioral functioning. Taken together, the data suggest that the ADHD + EFD group is not merely a more severe ADHD group, but rather a potentially unique group in terms of the nature of the mechanistic deficit. Further study to determine whether this subgroup has additional clinical characteristics (e.g., other psychiatric comorbidities or treatment response) or different patterns of etiological factors is warranted. Limitations

The NFBC is a large population sample, and these results may not generalize directly to clinical populations. We used a stringent p value (.01); however, the large sample size afforded increased statistical power to detect small ESs. As a result, the clinical effect of such group differences may be hard to discern. The PCA of WM and inhibition components is exploratory in nature. Component 2, which reflects primarily RIN, had strong loadings from only two tests. Although the components were calculated using component loadings for all tasks, it may be less stable than the component 1 (WM) because there were five tests that loaded strongly on that component. The results of our PCA are consistent, however, with factor analyses of inhibition and WM factors reported by other groups (Friedman et al., 2006). Those authors have also shown that a latent variable approach is more powerful than a manifest variable approach, which should be considered in future research. Finally, we have reported the impact of EFDs on cognitive performance; however, the generalizability of these cognitive constructs to actual daily functioning requires further study. Clinical Implications

Overall, EFDs are a risk factor for poor cognitive functioning across multiple domains irrespective of diagnosis. Approximately half of the adolescents with ADHD have EFDs, the majority of which (~80%) have a WM deficit (alone or in combination with an RIN

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 46:12, DECEMBER 2007

deficit). EF impairment within ADHD is not readily apparent through behavioral probe; however, adolescents with ADHD and EFD are likely to have lower intelligence and slower reading fluency leading to subsequent academic difficulties. Disclosure: Dr. Moilanen is a member of the Lilly Strattera advisory board, Finland. Dr. McCracken is a consultant to Abbott, BristolMyers Squibb, Eli Lilly, Janssen, Johnson & Johnson, McNeil, Novartis, Noven, Pfizer, Shire, and Wyeth; he also receives research support from Abbott, Bristol-Myers Squibb, Eli Lilly, Gliatech, McNeil, Pfizer, and Shire; and is on the speakers` bureaus of Eli Lilly, Janssen, and UCB. Dr. McGough receives grant research support from Eli Lilly, McNeil, New River Pharmaceuticals, Novartis, Shire, and Pfizer, is also a consultant to Eli Lilly, Novartis, and Shire, and serves on the speakers` bureaus of Eli Lilly, McNeil, Novartis, and Shire. The other authors have no financial relationships to disclose.

REFERENCES Barkley RA (1997), Behavioral inhibition, sustained attention, and executive functions: constructing a unifying theory of ADHD. Psychol Bull 121:65Y94 Biederman J, Monuteaux MC, Doyle AE et al. (2004), Impact of executive function deficits and attention-deficit/hyperactivity disorder (ADHD) on academic outcomes in children. J Consult Clin Psychol 72:757Y766 Castellanos FX, Sonuga-Barke EJ, Milham MP, Tannock R (2006), Characterizing cognition in ADHD: beyond executive dysfunction. Trends Cogn Sci 10:117Y123 Cohen J (1988), Statistical Power Analyses for the Behavioral Sciences, 2nd ed. Hillside, NJ: Erlbaum Conners C (2000), CPTII: Conners` Continuous Performance Test. Toronto: Multi-Health Systems Doyle AE, Biederman J, Seidman LJ, Weber W, Faraone SV (2000), Diagnostic efficiency of neuropsychological test scores for discriminating boys with and without attention deficit-hyperactivity disorder. J Consult Clin Psychol 68:477Y488 Fan J, McCandliss BD, Sommer T, Raz A, Posner MI (2002), Testing the efficiency and indpendence of attentional networks. J Cogn Neurosci 14:340Y347 Friedman NP, Miyake A, Corley RP, Young SE, Defries JC, Hewitt JK (2006), Not all executive functions are related to intelligence. Psychol Sci 17:172Y179 Ja¨rvelin MR, Hartikainen-Sorri AL, Rantakallio P (1993), Labour induction policy in hospitals of different levels of specialisation. Br J Obstet Gynaecol 100:310Y315 Korkman M, Kirk U, Kemp S (1998), NEPSY: A Developmental Neuropsychological Assessment. San Antonio, TX: Psychological Corporation, Harcourt Brace Lijffijt M, Kenemans JL, Verbaten MN, van Engeland H (2005), A metaanalytic review of stopping performance in attention-deficit/hyperactivity disorder: deficient inhibitory motor control? J Abnorm Psychol 114: 216Y222 Logan GD, Schachar RJ, Tannock R (1997), Impulsivity and inhibitory control. Psychol Sci 8:60Y64 Nigg JT, Blaskey LG, Huang-Pollock CL, Rappley MD (2002), Neuropsychological executive functions and DSM-IV ADHD subtypes. J Am Acad Child Adolesc Psychiatry 41:59Y66 Nigg JT, Willcutt EG, Doyle AE, Sonuga-Barke EJ (2005), Causal heterogeneity in attention-deficit/hyperactivity disorder: do we need neuropsychologically impaired subtypes? Biol Psychiatry 57:1224Y1230 Pennington BF, Ozonoff S (1996), Executive functions and developmental psychopathology. J Child Psychol Psychiatry 37:51Y87

1603

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

LOO ET AL. Reitan RM (1979), Manual for Administration of Neuropsychological Test Batteries for Adults and Children. Tucson, AZ: Author Schoechlin C, Engel RR (2005), Neuropsychological performance in adult attention-deficit hyperactivity disorder: meta-analysis of empirical data. Arch Clin Neuropsychol 20:727Y744 Seidman LJ (2006), Neuropsychological functioning in people with ADHD across the lifespan. Clin Psychol Rev 26:466Y485 Seidman LJ, Biederman J, Faraone SV, Weber W, Ouellette C (1997), Toward defining a neuropsychology of attention deficit-hyperactivity disorder: performance of children and adolescents from a large clinically referred sample. J Consult Clin Psychol 65:150Y160 Smalley SL, McGough JJ, Moilanen IK et al. (2007), Prevalence and psychiatric comorbidity of attention
adolescent Finnish population. J Am Acad Child Adolesc Psychiatry 46: 1575Y1583 SPSS (2003), SPSS for Windows. Rel 12.1.0.2003 edition. Chicago: SPSS Swanson JM, Schuck S, Mann M et al. (2001), Categorical and dimensional definitions and evaluations of symptoms of ADHD: the SNAP and SWAN ratings scales. Available at: http://adhd.net. Accessed June 6, 2006 Wechsler D (1981), Wechsler Memory Scale-Revised: Manual. New York: The Psychological Corporation Wechsler D (1992), WAIS-R Kasirkirja [in Finnish]. Helsinki: Psykologien Kustannus Willcutt EG, Doyle AE, Nigg JT, Faraone SV, Pennington BF (2005), Validity of the executive function theory of attention-deficit/hyperactivity disorder: a meta-analytic review. Biol Psychiatry 57:1336Y1346

Use of Complementary and Alternative Medicine in a General Pediatric Clinic Dany Jean, MD, Claude Cyr, MD, FRCPC Background: Use of complementary and alternative medical therapies is common and increasing, particularly for children with chronic disease. Objectives: The purpose of this work was to describe the use of complementary and alternative medicine by children and to identify factors that may influence the use of complementary and alternative medicine. Patients and Methods: We conducted a crosssectional descriptive study with children who were visiting a pediatric outpatient clinic. Parent_s satisfaction about primary care was evaluated with the Parent_s Perceptions of Pediatric Primary Care Quality questionnaire. Results: Fifty-four percent of children used Q1 type of complementary and alternative medicine in the previous year. No sociodemographic characteristic difference was found between user and nonuser groups. Children most often used complementary and alternative medicine to treat musculoskeletal problems (27%), psychological problems (24%), or infections (20%). Factors that influenced complementary and alternative medicine use were Bword of mouth[ (36%), Breference by a physician[ (28%), Bpersonal experience by parents[ (28%), and Bno adequate resources in Ftraditional_ medicine[ (21%). Forty-seven percent of complementary and alternative medicine users used prescribed medications simultaneously. Most users (75%) believed that complementary and alternative medicine had no potential adverse effects or interactions with prescribed medication. Only 44% of complementary and alternative users were known as such by their physician. The primary care satisfaction was significantly lower in complementary and alternative users versus nonusers. Parents of complementary and alternative users were less satisfied in the areas of accessibility, knowledge of the patient, and communication. Conclusions: Complementary and alternative medicine was used by 54% of the children in our cohort. Complementary and alternative medicine users were less satisfied with primary care than nonusers. Only 44% of complementary and alternative medicine users were known by their physician. It is important that physicians systematically elicit families’ expectations of treatment and be aware of the range of therapies used by children. Pediatrics 2007;120:e138Ye141.

Malpractice Claims Involving Pediatricians: Epidemiology and Etiology Aaron E. Carroll, MD, MS, Jennifer L. Buddenbaum, MHA Objective: Our goals were to examine malpractice claims data that are specific to the specialty of pediatrics and to provide a better understanding of the effect that malpractice has on this specialty. Methods: The Physician Insurers Association of America is a trade association of medical malpractice insurance companies. The data contained in its data-sharing project represent 25% of the medical malpractice claims in the United States at a given time. Although this database is not universally comprehensive, it does contain information not available in the National Practitioner Data Bank, such as information on claims that are not ultimately paid and specialty of the defendant. We asked the Physician Insurers Association of America to perform a query of its data-sharing project database to find malpractice claims reported between January 1, 1985, and December 31, 2005, in which the defendant_s medical specialty was coded as pediatrics. Comparison data were collected for 27 other specialties recorded in the database. Results: During a 20-year period (1985Y2005), there were 214226 closed claims reported to the Physician Insurers Association of America data-sharing project. Pediatricians account for 2.97% of these claims, making it 10th among the 28 specialties in terms of the number of closed claims. Pediatrics ranks 16th in terms of indemnity payment rate (28.13%), with dentistry ranked highest at 43.35%, followed by obstetrics and gynecology at 35.50%. Indemnity payment refers to settlements or awards made directly to plaintiffs as a result of claimresolution process. Data are presented on changes over time, claim-adjudication status, expenses on claims, the causes of claims, and injuries sustained. Conclusions: Malpractice is a serious issue. Some will read the results of this analysis and draw comfort; others will view the same data with alarm and surprise. Regardless of how one interprets these findings, they are important in truly informing the debate with generalizable facts. Pediatrics 2007;120:10Y17.

1604

J. AM. ACAD. CHILD ADOLE SC. PSYCHIATRY, 46 :12, DECE MBER 2007

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.