Abstract
ADHD occurs together with learning disabilities (LDs) approximately 30% of the time (Du Paul & Stoner, 2003), with subtypes of LD, including math LD (Barbaresi, Katusic, Colligan, Weaver, & Jacobsen, 2007), reading disability (Yoshimasu et al., 2010), and written language disorder (Yoshimasu et al., 2011). Children with comorbid ADHD/LD have more severe cognitive deficits and educational dysfunction, with increased grade retention and placement in special education classes (Faraone, Biederman, Monuteaux, Doyle, & Seidman, 2001). In addition, they may be at increased risk for social and emotional adjustment difficulties such as poor self-esteem, addictive behaviors, and suicidality (Brook & Boaz, 2005). To maximize the long-term prognosis in this high-risk group of children, multimodal interventions that include educational services, behavioral strategies, and medication are indicated.
Stimulant medication has substantial empirical evidence supporting its efficacy for ADHD; however, a significant number of children who might benefit from such medication remain untreated (Hoagwood, Kelleher, Fell, & Comer, 2000; Zima et al., 2010). The causes of unmet needs among specific subgroups such as those with comorbid ADHD/LD have been relatively unexplored; however, one contributing factor may be provider perceptions (Bussing et al., 2012). Indeed, some psychiatrists may prescribe less often to this group because of concerns of decreased efficacy and side effects (Bramble, 2007). In addition, physicians may be reluctant to treat ADHD with stimulant medication if they ascribe a child’s problems to an etiology other than ADHD (Grizenko, Bhat, Schwartz, Ter-Stepanian, & Joober, 2006). This could certainly occur with comorbid ADHD/LD because difficulties with school performance may be attributed to the presence of LD rather than to ADHD. Such provider perceptions could have a significant impact on the treatment options provided to patients and their families, and ultimately on child outcomes.
Unfortunately, data on the extent to which comorbid LD affects the safety and efficacy of stimulant medication in children with ADHD are limited, which may contribute to negative provider perceptions. Most large trials investigating the use of medication for ADHD either fail to assess for LD (McGough et al., 2006; Molina et al., 2009; Schulz et al., 2010), explicitly exclude students not functioning at age-appropriate levels (Biederman et al., 2007; Wigal et al., 2005), or use samples so small that even moderately strong effects might be missed because of lack of statistical power (Shafritz, Marchione, Gore, Shaywitz, & Shaywitz, 2004). The few studies that have examined stimulant efficacy in children with comorbid LD yielded somewhat inconsistent and limited results. Ackerman, Dykman, Holcomb, and McCray (1982) demonstrated that a small group of boys (n = 11) with “hyperkinesis and reading disability” was significantly less hyperkinetic after 3 weeks of flexibly dosed methylphenidate than after 3 weeks of placebo, and the magnitude of improvement did not differ from that of 27 boys with “solely hyperkinesis.” Comparing the extent to which 4 weeks of treatment with low- or high-dose methylphenidate improved psychomotor speed, Dykman and Ackerman (1991) found no differences between boys with ADHD with or without reading disability. In contrast, Grizenko and colleagues (2006) found that fewer children with ADHD/LD responded well to 2 weeks of fixed-dose methylphenidate than children with ADHD per team consensus. These studies have a number of weaknesses that limit interpretation of their findings for clinical application: (a) limited variability in gender, race, and ethnicity of studied samples, (b) brief duration of trials, (c) lack of flexibility or clear explanation of criteria for titrating dosage, (d) lack of adequate sample size, and (e) relative lack of easily interpretable behavioral outcome data from relevant standardized measures. Given these limitations, as well as the number of children who may have ADHD comorbid with one or more LDs, further exploration of the extent to which comorbid LD may impact efficacy, safety, and tolerability of stimulants is merited.
The present study examined the effects of Osmotic Release Oral System (OROS) methylphenidate in the treatment of children with ADHD with and without comorbid LD. We examined the two most commonly identified forms of LD, reading disability and math disability (Seidman, Biederman, Monuteaux, Doyle, & Faraone, 2001). Within the context of these disorders, we were most interested in the following questions: (a) Does methylphenidate effectively and safely treat symptoms of ADHD in children with and without comorbid LD? (b) Do core ADHD symptoms, as rated by parents and trained observers, vary in their response according to the presence or absence of LD? (c) Do cognitive deficits associated with ADHD (i.e., visual working memory and response time variability) vary in their response according to the presence or absence of LD? and (d) Do academically related difficulties associated with either ADHD or LD (i.e., reading fluency, reading comprehension, and handwriting) vary in their response according to the presence or absence of LD?
Method
Study Design and Participants
Participants were combined from two identical, double-blind, placebo-controlled, crossover, multisite laboratory school studies (Murray et al., 2011; Wigal et al., 2011) that evaluated the efficacy of OROS methylphenidate HCl extended release tablets (Concerta® CII, McNeil Pediatrics, Division of Ortho-McNeil Janssen Pharmaceuticals, Inc., Titusville, NJ) in treating ADHD. Children aged 9 to 12 years who achieved a Wechsler (1999) Abbreviated Scale of Intelligence Full-Scale IQ (4-subtest form; WASI FSIQ-4) score >80 were eligible for study participation. Children included in the study attended a public or private school and could read and understand English. Children with a history of or current diagnosis of a medical, neurological, or psychiatric disorder that could compromise their ability to adhere to the study requirements, or a history of failed response to a stimulant product (i.e., at the maximum approved dose) were not eligible for participation.
Identification of ADHD and LD
Children were diagnosed with ADHD as defined by the Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.; DSM-IV-TR; American Psychiatric Association, 2000) with diagnosis supported by the Schedule for Affective Disorders and Schizophrenia for School-Age Children–Present and Lifetime version (K-SADS-PL; Kaufman et al., 1997). Children were required to have total or subscale scores ≥90th percentile for age and gender on the parent-rated ADHD Rating Scale, 4th Edition (ADHD RS-IV; Du Paul, Power, & Anastopoulos, 1998) at screening.
Following work similar in concept to the approach taken by other investigations (Ackerman et al., 1982; Grizenko et al., 2006), LD was defined according to a low-achievement criterion, whereby scores 1 to 2 standard deviations less than expectations for age (between the 2nd and 16th percentiles) were categorized as indicating a mild to moderate LD. Participants with severe LDs (defined as scores >2 SD less than the mean score for age) were excluded from participation in the study.
Reading disability was assessed with measures of phonologic awareness (Elision subtest of the Comprehensive Test of Phonological Processing; Wagner, Torgeson, & Rashotte, 1999) and reading fluency (Fluency score from the Gray Oral Reading Test, 4th edition [GORT-4]; Wiederholt & Bryant, 2001). As phonologic awareness has been described as “the best predictor of children’s future reading ability” (Eden & Vaidya, 2008) and because reading fluency has proven to be a powerful predictor of reading achievement sensitive to impairment for reasons other than impaired phonologic awareness (Meisinger, Bloom, & Hynd, 2010). Math LD was assessed with the Numerical Operations subtest of the Wechsler Individual Achievement Test®, 2nd edition–Abbreviated (WIAT-II-A; Wechsler, 2001).
Procedures
The study procedures have been described in detail by Murray et al. (2011) and Wigal et al. (2011). Parents or legal guardians provided written consent for their children to participate in the study. Children also provided assent. Participants receiving medication to treat ADHD at study entry completed a washout period for up to 6 days before completing baseline assessments at a baseline visit. Doses were then titrated to an individualized dose (i.e., final dose established during titration) over a period of up to 6 weeks. Children initiated treatment with open-label OROS methylphenidate 18 mg per day (morning dose), with an increase in their daily dose by 18 mg every 3 to 7 days until meeting the following treatment goals: parent/guardian ADHD RS-IV total and subscale scores ≤75th percentile for age and gender and Clinical Global Impression–Improvement (CGI-I) rating of “very much improved” or “much improved.” One dose decrease by 18 mg was allowed if required for tolerability. Participants reaching the maximum dose of 54 mg per day and requiring a dose decrease continued to the assessment period if their ADHD RS-IV total and subscale scores were less than the 85th percentile for age and gender with a “very much improved” or “much improved” CGI-I rating.
Following dose titration, children meeting criteria for improvement in symptoms entered the assessment period. In this 6-week period, participants received their open-label individualized dose of OROS methylphenidate each day except on 2 days when they participated in full-day laboratory school sessions (separated by at least 7 days on sequential Saturdays). Each child was randomized to one of two treatment sequences: individualized OROS methylphenidate dose on the first laboratory school day with crossover to placebo for the second laboratory school day or the reverse. The final study visit included assessment of ADHD symptoms within 2 weeks of the last of two laboratory school days.
Outcome Measures
Parents/guardians rated attention, impulsivity, and hyperactivity symptoms with the ADHD RS-IV (Du Paul et al., 1998) at baseline and the final study visit. The children’s regular classroom teachers assessed the same symptoms with the school version of the ADHD RS-IV at baseline and final study visit.
Trained classroom observers assessed behavior using the Swanson, Kotkin, Alger, M-Flynn, and Pelham Scale (SKAMP; Wigal, Gupta, Guinta, & Swanson, 1998). The SKAMP was measured several times throughout the laboratory school day; the 4-hr postdose evaluation was prespecified as the primary assessment of interest. Response time variability was measured with the reaction time variability score of the Test of Variables of Attention (TOVA; Leark, Greenberg, & Kindschi, 2007). The Bedard, Jain, Johnson, and Tannock (2007) version of the Finger Windows subtest from the Wide Range Assessment of Memory and Learning (Adams & Sheslow, 1990) was used to measure visual working memory. Both of these measures were collected at approximately 4 hr after medication was administered on laboratory school days. Measures of reaction time variability and visual working memory were chosen because these abilities have been found to be relevant to attention and academic function in ADHD and are sensitive to the effects of stimulant treatment (Barnett et al., 2001; Bedard, Martinussen, Ickowicz, & Tannock, 2004; Epstein et al., 2006; Tannock, Ickowicz, & Schachar, 1995).
Measures of academically related skills included the numbers of math problems attempted and correctly answered on the Permanent Product Math Test (PERMP-Attempted and PERMP-Correct; Wigal & Wigal, 2006), the reading fluency score of the Dynamic Indicators of Basic Early Literacy Skills (DIBELS; Good & Kaminski, 2002), the Silent Reading Comprehension score of the Gray Silent Reading Test (GSRT; Wiederholt & Blalock, 2000), and the total score from the Test of Handwriting Skill–Revised (THS-R; Milone, 2007). The PERMP was measured several times throughout the laboratory school day; the 4-hr postdose assessment was prespecified as the primary assessment.
Statistical Analysis
All randomized participants who received at least one dose of study medication, had any efficacy data in a randomized sequence, and could be classified by LD status were included in the analysis. Because the trials were identical in design, data from both studies were combined for the post hoc analyses of LD subgroups reported here.
Comparisons of demographic and baseline characteristics in the groups with and without LD were analyzed with Fisher exact tests for categorical variables and Student t tests for continuous variables.
To determine whether ADHD symptoms improved in children with and without LD from baseline to final visit, ADHD RS-IV scores were analyzed using a general linear, mixed-model approach with independent (fixed) variables of visit (baseline vs. final), treatment, period, and sequence.
For variables measured once during the laboratory school day, a general linear, mixed-model approach was used that included terms for independent (fixed) variables of treatment, period, and sequence. For variables measured at repeated times throughout the laboratory school day (PERMP and SKAMP), a repeated measures mixed model was used with appropriate contrast statements. Each model used the observed variable value at 4-hr postdose as the dependent variable and included terms for treatment, time, treatment-by-time interaction, period, sequence, study, and baseline score, assuming a first-order autoregressive (AR [1]) correlation structure among repeated observations. Participants were random effects nested within sequence.
The studies were neither prospectively designed nor powered to examine LD subgroup analyses. Given the exploratory, post hoc nature of these analyses, correction for multiplicity was not performed.
Safety analyses included all participants who received ≥1 dose of study medication and who could be classified by LD status. Percentages of participants with adverse events during open-label and double-blind treatment with OROS methylphenidate were analyzed descriptively. Differences between percentages of participants experiencing adverse events in each LD group were analyzed using Fisher’s exact test.
Results
Of the 167 participants enrolled in the studies receiving at least 1 dose of study medication, 139 completed the open-label dose-adjustment period, met criteria for treatment response, and were randomized to one of the two sequences of treatment for the laboratory school days. Reasons for discontinuation have been detailed (Murray et al., 2011; Wigal et al., 2011). Of the 139 participants, 4 did not receive academic skills testing at baseline and were excluded from further analyses. Of the remaining 135 participants, 46 (34.1%) were identified as having LD. Among the participants with LD, 31 (67.4%) had reading LD, 9 (19.6%) had math LD, and 6 (13.0%) had both reading and math LDs. The remaining 89 randomized participants had neither reading LD nor math LD.
Safety analyses included 161 participants who received ≥1 dose of study medication and could be classified by LD status. Six participants were excluded because they could not be classified by LD status (4 did not receive academic testing at baseline and 2 had WASI FSIQ-4 <80).
Demographic and baseline characteristics of randomized participants with and without comorbid LD are presented in Table 1. A majority of participants were male, White or Caucasian, non-Hispanic, and had Combined type ADHD. Female and Black or African American participants were well represented. A lesser percentage of White/Caucasian participants (24.1%) than Black/African American participants (46.2%) were identified as having comorbid LD (p < .01). There were no significant differences between the LD groups in terms of age (p = .6793), gender (p = .8454), or ADHD subtype (Inattentive vs. Combined and Hyperactive, p = 1.0000). However, according to parent/guardian ADHD RS-IV ratings, the group without comorbid LD was significantly more inattentive (p < .001) and somewhat more hyperactive and impulsive (p = .0529) at baseline than was the group with comorbid LD. Baseline IQ and achievement scores are presented in Table 2.
Demographic and Baseline Characteristics of Randomized Sample by LD Group.
Note: LD = learning disability; ADHD RS-IV = ADHD Rating Scale, 4th edition.
IQ and Achievement Scores at Baseline by LD Group.
Note: LD = learning disability; WASI FSIQ-4 = Wechsler Abbreviated Scale of Intelligence Full-Scale Intelligence Quotient (4-subtest form); CTOPP = Comprehensive Test of Phonological Processing; GORT-4 = Grey Oral Reading Test, 4th edition; WIAT-II-A = Wechsler Individual Achievement Test, 2nd edition–Abbreviated.
Standard scores have a mean of 100 and a standard deviation of 15.
Scaled scores have a mean of 10 and a standard deviation of 3.
The percentage of ADHD/LD participants achieving the individualized dose titration criteria (79.3%; 46 of 58 participants) did not differ significantly from the percentage of participants without LD meeting this criteria (86.4%; 89 of 103 participants), p = .27. A majority of participants in both groups had a final dose of 54 mg. The distribution of final doses of OROS methylphenidate varied by comorbid LD status, with somewhat higher doses in the group with comorbid LD (Table 3). Fisher’s exact test was used to determine whether these group differences in dose were significant. The analyses revealed that children with LD were more likely than children without LD to have a final dose of 54 mg (p = .0282), and children with LD were less likely than children without LD to have a final dose of 36 mg (p = .0026); thus, doses were higher for the group with comorbid LD. The percentage of children with 18 mg as a final dose did not differ significantly between groups (p = .3173).
Number and Percentage of Patients Who Received Each Final Individualized Dose by LD Group.
Note: LD = learning disability.
p value from Fisher’s exact test for difference between LD groups within each dose level.
There were no observed differences in treatment-emergent adverse events (p = .8553). Adverse events (AEs) reported in ≥5% of participants in either subgroup are presented in Table 4. The types of adverse events were similar to patterns previously reported in other studies of stimulants in children with ADHD. Two participants discontinued because of AEs during the dose-adjustment period. No serious AEs or deaths were reported.
Adverse Events Reported in ≥5% of Participants Who Received ≥1 Dose of Study Medication by LD Group.
Note: LD = learning disability. Includes events for all participants who received ≥1 dose of Osmotic Release Oral System methylphenidate and could be classified according to criteria for LD; six participants could not be classified by LD criteria.
ADHD symptoms, rated by parents/guardians and teachers (ADHD RS-IV), improved significantly during OROS methylphenidate treatment from baseline to final study visit, regardless of the presence of comorbid LD (Table 5; all p < .0001).
Baseline and Final Parent/Guardian and Teacher ADHD RS-IV Scores by LD Group.
Note: ADHD RS-IV = ADHD Rating Scale, 4th edition; LD = learning disability.
p values from general linear mixed model for comparison of baseline and final visit scores.
Least square (LS) mean SKAMP scores rated children treated with OROS methylphenidate as more attentive (without LD, placebo 11.0 [SE = .50] vs. OROS methylphenidate 5.8 [SE = .50]; with LD, placebo 10.7 [SE = .68] vs. OROS methylphenidate 6.5 [SE = .68]; both p < .0001) and better behaved (without LD, placebo 8.7 [SE = .50] vs. OROS methylphenidate 2.7 [SE = .49]; with LD, placebo 8.2 [SE = .71] vs. OROS methylphenidate 3.6 [SE = .70]; both p < .0001) on laboratory school days, regardless of the presence of comorbid LD. The pattern of results was consistent across informants for behaviors indicating inattention, hyperactivity/impulsivity, or the combination.
On cognitive and academically related skills (Table 6), overall, mean scores were lesser for children with LD than for children without LD whether they received placebo or OROS methylphenidate. Mean scores were greater overall when children received OROS methylphenidate than when they received placebo for both LD groups.
Results of Cognitive and Academically Related Testing on Laboratory School Days by LD Group and Treatment Received.
Note: LD = learning disability; OROS MPH = Osmotic Release Oral System methylphenidate; TOVA = Test of Variables of Attention; RT = reaction time; WRAML = Wide Range Assessment of Memory and Learning; GSRT = Gray Silent Reading Test; THS-R = Test of Handwriting Skill–Revised; DIBELS = Dynamic Indicators of Basic Early Literacy Skills; PERMP = Permanent Product Measure of Performance; LS = least squares.
p values from general linear mixed model for comparison of placebo and OROS methylphenidate. For PERMP, p values were determined by using appropriate 4-hr time point contrast statement in a repeated measures mixed model with actual value as the dependent variable and terms for treatment, time, treatment by time interaction, period, sequence, study, and baseline score respectively, and assuming an AR(1) correlation structure among the repeated observations. Greater scores indicate better performance.
Children who received OROS methylphenidate had less variability in reaction time (i.e., greater TOVA scores) and could recall longer sequences in reverse order (Finger Windows Backwards) than those who received placebo, regardless of the presence of comorbid LD. Children with LD were also better at recalling forward sequences (Finger Windows Forwards) when they received OROS methylphenidate than when they received placebo (p = .0022); however, for children without LD, the difference between treatments was not statistically significant (p = .0752).
Children read more fluently on the DIBELS (both groups p < .01) and wrote more legibly on the THS-R (both groups p < .0001) when they received OROS methylphenidate than when they received placebo. In contrast, mean silent reading comprehension (GSRT) improved significantly during OROS methylphenidate treatment only in children without comorbid LD (p = .0143), though the pattern of results was in the same direction in the children with LD (p = .0998). On the PERMP test (at 4 hr), children performed more quickly (p < .0001) without sacrificing accuracy when they received OROS methylphenidate than when they received placebo, regardless of the presence of LD.
Discussion
Children with ADHD and comorbid LD are at increased risk for negative educational outcomes. However, providers may have concerns regarding efficacy or side effects of stimulant medications when treating ADHD in children with comorbid LD. The present study addressed these questions empirically, contributing to the literature with analysis of a relatively large, comprehensive, clinical trial–based data set. The reported results were obtained in a rigorously controlled laboratory school setting as well as in the naturalistic school and home environments. These data provide a demographically diverse sample whose performance was measured with an array of validated behavioral, cognitive, and academically related measures.
The results of this analysis support the use of OROS methylphenidate as a safe and effective treatment for children with ADHD regardless of presence of the two most common forms of LD (reading disability and math disability). According to parental reports of inattention and impulsivity, 79% and 86% of children with and without comorbid LD, respectively, responded positively to treatment of their ADHD symptoms. Teachers and trained observers of these children rated their inattention and impulsivity as significantly improved as well, whether in the tightly controlled laboratory school setting or in the more naturalistic setting of the child’s regular classroom. The proportion of children with comorbid LD responding to treatment in this study exceeds that reported by Grizenko et al. (2006); however, the extent to which results are a function of individualized flexible dosing, a longer treatment period, or more explicitly defined outcome measures is difficult to determine.
The examination of core cognitive deficits associated with ADHD (visual working memory and variability in response time) also revealed treatment-associated gains. In particular, variability in reaction time improved significantly regardless of presence of comorbid LD, as did performance on the more demanding of the visual working memory measures. This is the first demonstration that children with explicitly defined comorbid LD also benefit significantly from OROS methylphenidate.
Regardless of the presence of comorbid LD, children with ADHD in this sample performed mathematical calculations more quickly without sacrificing accuracy, read text more fluently, and wrote more legibly when they received OROS methylphenidate than when they received placebo. Silent reading comprehension improved significantly in children without comorbid LD, and this trend was also evident in children with LD.
Perhaps the most salient of the differences between the groups with and without comorbid LD in this study was the individualized dose. Although the individualized dose for the majority of children in both groups was 54 mg, greater doses on average (within the 18- to 54-mg range) were determined for children with comorbid LD, despite ADHD RS-IV parent/guardian ratings at baseline suggesting that the group with comorbid LD was significantly less inattentive, less hyperactive, and less impulsive than children in the group without comorbid LD. This pattern is consistent with data presented in a previous trial using a flexible titration schedule (Ackerman et al., 1982). The finding by Grizenko et al. (2006) that fewer children with reading disability responded to their fixed doses suggests that children with comorbid reading disability may have responded better to greater dosages. Given our relatively small sample size in these LD subgroups, however, such a hypothesis should be examined in a larger study that is prospectively designed to address such questions.
Another difference between the groups with and without comorbid LD in this study was the varying racial and ethnic mix. The proportion of African American children in the comorbid LD group (39%) exceeded that found in the LD absent group (24%). A similar trend was found in a sample evaluated by Dykman and Ackerman (1991). The potential impact of this difference in ethnicity of patient sample is difficult to determine, as variables often influencing ethnic or racial differences in cognitive and academic performances (Burchinal, Roberts, Zeisel, & Rowley, 2008) were not collected. Previous investigations have found that efficacy, safety, and tolerability of OROS methylphenidate in African American children are similar to those of other groups (Starr & Kemner, 2005). Thus, the authors suggest that differences in racial composition did not contribute significantly to differences between LD groups in final individualized doses. The current sample, however, was not of sufficient size to investigate potential interactions between comorbid LD status and race or ethnicity.
Finally, the groups in this study also differed in terms of IQ; children without comorbid LD had greater mean IQ scores than those with LD. This finding is ubiquitous in studies examining psychometric differences between children with ADHD with and without comorbid LD (Dennis et al., 2009). In the current analysis, the authors chose a priori not to control for IQ, as IQ is a poor predictor of severity of reading difficulties, response to intervention, and performance on specific measures underlying reading success. These outcomes appear to be better predicted by fundamental skills underlying reading achievement (i.e., phonological awareness and rapid automatized naming). In previous studies of LD, correcting for general intellectual ability (IQ) has led to counterintuitive results inconsistent with both clinical observation and psychometric theory (Dennis et al., 2009).
The ability to draw more broad conclusions from these data is constrained by limitations of the data set. The response rate of the community teachers was relatively low; consequently, it is not possible to determine bias. Although the sample size from this study compares favorably to the available literature, the ability to comment on math disability, in particular, and on various interactions (e.g., ADHD subtype by LD subtype or ethnicity by ADHD subtype) remains limited. Future investigations with larger, more demographically diverse samples will shed additional light on the extent to which LD subtypes may impact the safety, efficacy, and dosing of medication in different subgroups.
For children with ADHD/LD whose educational challenges may be greater than those seen in children with ADHD alone, stimulant treatment should be considered by practitioners as part of a comprehensive treatment plan. The results of this study demonstrate that OROS methylphenidate may be used to safely and effectively treat symptoms of ADHD in children with or without comorbid LD. In addition, treatment may impact working memory, handwriting, and academically related skills that pose challenges to children with ADHD.
Footnotes
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: David Williamson and H. Lynn Starr are full-time salaried employees of Janssen Scientific Affairs, LLC; CV Damaraju and Steve Ascher are full-time salaried employees of Johnson & Johnson Pharmaceutical Research and Development, L.L.C. For all authors, compensation includes common stock in Johnson & Johnson. Compensation as an employee is not influenced by publication of this report, which describes a study conducted by other contributors.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by funding from Janssen Scientific Affairs, LLC, Raritan, NJ. Dr. Murray has received research funding from National Institute of Mental Health (NIMH), the Department of Education, Ortho-McNeil Janssen Scientific Affairs, and Lilly USA. She has also received salary support from Otsuka Pharmaceutical Development & Commercialization, Inc. (OPDC) within the past 2 years. Ellen Stoltzfus, PhD, of JK Associates, Inc., provided editorial assistance that was funded by Janssen Scientific Affairs, LLC.