Impact of the DSM-5 Attention-Deficit/Hyperactivity Disorder (ADHD) Age of Onset Criterion in the U.S. Adolescent Population

Abstract

Objective

The present study aims to compare the prevalence and clinical correlates of DSM-IV versus DSM-5 – defined attention-deficit/hyperactivity disorder (ADHD) and subtypes in a nationally representative sample of U.S. youth based on the age-of-onset criterion.

Method

The sample includes 1,894 participants 12–15 years of age from cross-sectional National Health and Nutrition Examination Survey (NHANES) surveys conducted from 2001 to 2004. Data on DSM-IV and DSM-5 criteria for ADHD were derived from administration of the parental ADHD module of the National Institute of Mental Health (NIMH) Diagnostic Interview Schedule for Children, Version IV (DISC-IV).

Results

Extension of the age-of-onset criterion from 7 to 12 years led to an increase in the prevalence rate of ADHD from 7.38% (DSM-IV) to 10.84% (DSM-5). Youth with later age of onset did not differ from those with earlier age of onset in terms of severity and patterns of comorbidity. However, the later age-of-onset group was more likely from lower income and ethnic minority families.

Conclusion

The comparability of the clinical significance of the early and later age-of-onset groups supports the DSM-5 extension of the age-of-onset criterion in ADHD.

INTRODUCTION

Attention-deficit/hyperactivity disorder (ADHD) is one of the most common psychiatric disorders of childhood with major impact on a child’s academic and social development.¹ However, this diagnosis has been controversial because of its high prevalence and popular conception that it is over-diagnosed, particularly in the U.S.² There has been substantial research that has evaluated the key components of the diagnostic criteria of ADHD including the symptom criteria,³ subtypes,⁴ severity,⁵ role of impairment,⁶ and contextual manifestations⁷ as well as the validity of informants.⁸ Recent studies have demonstrated that ADHD might be better characterized by severity dimensions rather than a categorical diagnosis.⁹ Likewise, the criterion onset under age 7 in the DSM^10,11 that appears to have been arbitrarily determined^12–14 was not supported by empirical evidence from the DSM-IV field trials.¹⁵

After three decades and further field trials, a change in the age criterion has finally been made. In accordance with the conclusions of a comprehensive review and meta-analysis of evidence regarding the age-of-onset criterion for ADHD,¹⁶ the DSM-5 has now increased the age of onset of ADHD symptoms to age 12.¹⁷ The increase in age of onset was supported by the high false negative rate of case detection attributable to over-estimation of age of onset in retrospective studies.¹⁶ For example, a substantial proportion of adults with full criteria for lifetime ADHD with impairment that continued throughout adult life failed to meet criteria of prior DSM editions because of these adults’ retrospectively reported onset after age 7.^13,18 Concern regarding the impact of this change on increasing the already high prevalence of this disorder has been somewhat ameliorated by an increase of only 0.1% in a prospective British twin study¹⁹ and other studies.^14,15 However, the earlier age-of-onset criterion has been shown to differentially exclude youth with the inattentive subtype. Willoughby et al.²⁰ found that of those who did report an onset after age 7 years, 26% had the inattentive subtype compared to 8% with the hyperactive-impulsive subtype, and 13% with the combined subtype.

Two nationally representative samples of U.S. children and adolescents, the National Health and Nutrition Examination Survey (NHANES, surveying 8–15 year olds) and the National Comorbidity Survey Replication-Adolescent Supplement (NCS-A, 13–18 year olds), have provided data on the prevalence of ADHD in the U.S. based on DSM-IV criteria. The 12-month prevalence rates of ADHD were 8.7% in the NHANES, and 7.2% in the NCS-A, with rates of 4.4% for Inattentive, 2.0% for Hyperactive/Impulsive, and 2.2% for the Combined subtypes in the latter study.²¹ However, these rates were based on the early-age-of-onset criterion, and did not present rates for later age of onset.

The goal of this study is to compare the prevalence of ADHD and its subtypes in the NHANES study based on the respective age-of-onset cut-offs of ADHD symptoms of 7 for DSM-IV and 12 for DSM-5 for age of onset among youth aged 12 and older who have passed through the period of risk of ADHD by current diagnostic criteria. Second, we compare the clinical and treatment correlates of youth with onsets under age 7 and ages 7–12 to examine the clinical validity of the expansion of the age-of-onset criteria.

METHOD

Subjects

The NHANES is a large, nationally representative, probability sample survey of the health status of non-institutionalized, U.S. civilians conducted by the National Center for Health Statistics.²² A total of 3,042 children 8–15 years of age were evaluated with psychiatric interview modules, including ADHD, at the Mobile Examination Center for the 2001–2004 NHANES. This study focused on 1,894 participants who were 12 to 15 years old in order to reflect the portion of the sample who had aged out of the “at-risk” time period for development of ADHD as outlined in the DSM-5 criteria.

Assessment of ADHD and Other Psychiatric Disorders

The Diagnostic Interview Schedule for Children-IV (DISC-IV), a structured diagnostic interview, was used to identify psychiatric disorders in children and adolescents. It was designed for trained lay interviewers to use in epidemiological studies and has been shown to have high validity and test-retest reliability.²³ The DISC-IV includes youth and caregiver interview modules. The caregiver was most often a parent who was most knowledgeable about the participating child (hereafter referred to as parent-informant). In the study, diagnostic information such as symptoms, pervasiveness of symptoms, age of onset, impairment, and service and medication utilization during the 12 months prior to the survey was collected from parent-informants and/or youth informants depending on the psychiatric disorder modules: generalized anxiety disorder (GAD) and panic disorder were based solely on youth reports, ADHD and conduct disorder (CD) were based solely on parent reports, and major depressive disorder (MDD) and dysthymia was based on positive endorsement of symptoms by either the parent or youth.

A computerized algorithm was applied to generate ADHD and other DSM-IV diagnosis variables using diagnostic information obtained from the DISC-IV based on parent reports on the participating child. Two sets of ADHD diagnosis variables were made by applying the DSM-IV and DSM-5 criteria. Children and adolescents met criteria for ADHD in the past 12 months if parents endorsed the following (that is, answered “yes” when asked if the patient had experienced certain symptoms): criterion A - six symptoms of either inattention and/or hyperactivity-impulsivity; criterion B - symptoms causing impairment present before age 7 years (DSM-IV) or symptoms present before age 12 years (DSM-5); criterion C - some impairment in at least two settings – at school or work and at home; and criterion D - clinically significant impairment in social, academic or occupational functioning. The diagnostic hierarchy rule (criterion E) was not operationalized. Criterion B (age of onset) was the primary feature differentiating the DSM-IV and DSM-5 diagnoses of ADHD. Children and adolescents were classified as ADHD inattentive subtype if they endorsed predominantly inattention symptoms; or as ADHD hyperactive/impulsive subtype if they endorsed predominantly hyperactivity/impulsivity symptoms; or as ADHD combined subtype if they met symptom criterion A for both inattention and hyperactivity/impulsivity. The question regarding the age of onset for ADHD was the following: “How old was [he/she] the first time [he/she] started to have trouble paying attention or concentrating?” or “How old was [he/she] the first time [he/she] started to be overactive?” If age was not known, parents were probed “What grade was [he/she] in?” The younger age value was used for ADHD combined subtype if the age of onset for inattention and hyperactivity were different.

Clinical Characteristics

Several clinical characteristics of ADHD were also evaluated. A “severe case” of ADHD was defined as DISC impairment level C (i.e., at least one severe rating was endorsed among six impairment questions regarding family, school or peer relationships). The “ADHD-specific treatment” variable was based on parent reports in the DISC ADHD interview module, which asked whether their child had seen someone in an inpatient or outpatient setting for symptoms that included overactivity, hyperactivity, or having trouble paying attention. The “mental health treatment” variable was based on the response to the question in the NHANES section of Hospital Utilization and Access to Care, which asked whether the participating child had seen or talked to a mental health professional such as a psychologist, psychiatrist, psychiatric nurse, or clinical social worker about his/her health in the past 12 months. Parents were also asked about any medication use in the past 12 months for overactivity, hyperactivity, or inattention. Comorbidity patterns for this study were combined into three major groups, which included anxiety, depression, and conduct disorders. The anxiety group included both GAD and panic disorder. The depression group included both MDD and dysthymia.

Socio-demographic Correlates

The socio-demographic variables included age, sex, race/ethnicity and family income. Youth race/ethnicity was separated into the following categories by caregivers: Non-Hispanic White, Non-Hispanic Black, Mexican American, Other Hispanic, and Other Race (including multiracial). Family income was based on the poverty index ratio (PIR), which is the ratio of household income to family-size-specific poverty threshold level. The PIR was categorized into three groups with a value of 1 representing the lowest income and a value of 3 or higher representing the highest income.

Analysis

All statistical analyses were completed with the SUDAAN software package (version 10) (The Research Triangle Institute, Research Triangle Park, NC) using the Taylor series linearization method to take into account the complex survey design. Cross-tabulations were used to estimate prevalence of ADHD and its subtypes by socio-demographic characteristics and clinical features as well as prevalence of mental comorbidity of ADHD and its subtypes. Means were used for continuous clinical features such as the number of inattentive or hyperactive symptoms. The group difference between earlier and later onset was compared statistically using the design-adjusted Wald Chi-square test to examine demographic correlates, clinical features, mental comorbidity of ADHD and its subtypes, controlling for sex, age, race/ethnicity, and PIR. Multivariate logistic regression analysis was performed on the ADHD sample to compare the demographic and clinical characteristics of the two age-of-onset thresholds. Significance was based on two-sided tests evaluated at the level of 0.05.

RESULTS

Demographic Characteristics

As shown earlier, 7.38% of children met 12-month criteria for ADHD with an age of onset prior to 7 years.^21,24 Expanding the required inattentive and hyperactive/impulsive symptoms from those occurring in children prior to age 7 years to those occurring in children prior to age 12 resulted in a 47% increase in prevalence (10.84%). Table 1 presents the twelve-month prevalence rates by demographic characteristics for these two groups. ADHD was more common in males (14.10%) than in females (7.57%), with equivalent sex ratios for those with an age of onset prior to age 7 as compared to those with onsets between the ages of 7 and 12.

Table 1.

Prevalence of 12-Month Attention-Deficit/Hyperactivity Disorder (ADHD) by Demographic Characteristics of Adolescents 12–15 Years Old (National Health and Nutrition Examination Survey [NHANES] 2001–2004)

Demographics	Unweighted N	DSM-5 ADHD			Proportion of Cases with Onset 7 to <12 years	Test statistic (Onset <7 years vs. 7 to <12 years)
		Total (n=163)	Onset <7 years (n=99)	Onset 7 to <12 years (n=64)
		% (SE)	% (SE)	% (SE)	% (SE)	Wald χ2 (df) [p-value]
TOTAL	1,894	10.84 (1.25)	7.38 (0.96)	3.46 (0.58)	31.95 (3.96)	-
SEX
Male	925	14.10 (1.62)	9.89 (1.15)	4.21 (0.87)	29.86 (4.39)	0.3 (1) [.567]
Female	966	7.57 (1.36)	4.86 (1.06)	2.71 (0.87)	35.86 (8.98
AGE
12	466	15.53 (2.81)	10.82 (2.29)	4.70 (1.17)	30.29 (6.05)	0.3 (3) [.804]
13	500	12.13 (2.42)	8.37 (2.11)	3.76 (1.11)	31.02 (8.18)
14	487	9.66 (1.62)	6.74 (1.31)	2.91 (0.90)	30.16 (7.52)
15	441	5.53 (1.38)	3.15 (1.05)	2.38 (0.92)	43.08 (12.59)
RACE/ETHNICITY a
Mexican American	589	6.12 (1.09)	3.26 (0.71)	2.86 (0.54)	46.79 (5.04)	3.3 (3) [.035]
Other Hispanic	74	7.11 (2.83)	4.39 (1.05)	2.71 (2.41)	38.19 (20.46)
Non-Hispanic White	550	12.65 (1.98)	8.89 (1.55)	3.75 (0.80)	29.69 (4.66)
Non-Hispanic Black	622	7.69 (1.33)	3.76 (0.75)	3.93 (0.92)	51.07 (6.73)
FAMILY INCOME b
PIR >=3	498	11.66 (2.72)	9.57 (2.57)	2.09 (0.78)	17.89 (6.61)	4.5 (2) [.020]
PIR 1–2	787	11.73 (1.97)	8.79 (1.61)	2.94 (0.97)	25.08 (6.75)
PIR <1 (poor)	540	10.00 (1.71)	5.63 (1.17)	4.37 (0.99)	43.68 (6.55)

Note: Df = degrees of freedom; PIR = Poverty index ratio; SE = standard error.

^aOther race is not shown

^bUnknown family income is not shown

Prevalence rates for the overall sample were significantly higher in non-Hispanic whites (12.65%) than in Mexican Americans (6.12%), other Hispanics (7.11%), or non-Hispanic blacks (7.69%). There were greater rates of early-onset ADHD in non-Hispanic whites than in other minority groups (p=0.0008), whereas there were no significant differences in race/ethnicity for the later-onset groups. Family income was similar in the overall sample (p=0.1883) and the later onset group (p=0.3339). However, in the earlier onset group, children from lower income families had the lowest prevalence of ADHD (5.63%), while children from the higher income families had greater prevalence rates than those from the lowest income group (9.57%, 5.63%, respectively, p=0.001). In contrast to the early onset group, there was no significant difference in family income in the later age of onset group.

Comparison of the two ADHD age-of-onset subgroups within the demographic strata indicated that those with later age of onset were less likely to be non-Hispanic white (p=0.035) and more likely to be from low-income families (p=0.020). Although not statistically significant, there was also a trend for the later-age-of-onset group to be female and older.

Clinical Characteristics

Figure 1 compares the subtypes of ADHD (inattentive, hyperactive/impulsive, and combined type) by age of onset and sex. The inattentive subtype was the most prevalent subtype in both age-of-onset groups. Inclusion of males with an onset between ages 7 and 12 yielded an increase in the prevalence rate of the inattention subtype of 3.1%. When combined with the early onset males, the total prevalence rate of the inattention subtype was 8.5%. The increased age-of-onset criterion led to an increase in prevalence by 1.8% in females. The male-to-female prevalence ratio for inattention was 2.12 for the earlier onset group compared to 1.67 for the later onset group (p=0.897). There were no differences in the sex distribution of the early- and later-onset groups.

Figure 1.

Prevalence of attention-deficit/hyperactivity disorder (ADHD) and subtypes by sex for adolescents 12–15 years old (National Health and Nutrition Examination Survey [NHANES] 2001–2004)

Application of the later-age-of-onset criterion led to prevalence rates of the hyperactive/impulsive subtype of 0.8% for males and 0.6% for females. When combined with rates of 1.7% for males and 0.9% for females for early onset, the total rates of this subtype were 2.5% for males and 1.5% for females. The male-to-female prevalence ratio for hyperactivity/impulsivity was 1.82 for the earlier-onset group compared to 1.42 for the later-onset group (p=0.436).

The later-age-of-onset criterion had less impact on rates of the combined subtype, with prevalence rates of 0.3% in both males and females. Taken together with early onset, total rates for males are 3.1% and females are 1.7%. The male-to-female prevalence ratio for the combined subtype was 2.04 for the earlier-onset group compared to 1.07 for the later-onset group (p=0.268).

Table 2 compares clinical characteristics and comorbidity patterns between earlier- and later-onset groups by ADHD subtypes. In general, there was no significant difference between the number of inattentive or hyperactive symptoms between the earlier- and later-onset groups for the inattentive and hyperactive/impulsive subtypes. For the combined subtype, the earlier-onset group exhibited significantly more inattentive and hyperactive symptoms when compared to the later-onset group (9.3 vs. 7.9 for inattention, p=0.004; 8.7 vs. 7.2 for hyperactivity/impulsivity, p=0.001). Likewise, there was no difference between the earlier- and later-onset ADHD groups in the proportion of cases of ADHD with severe impairment.

Table 2.

Clinical Characteristics of Attention-Deficit/Hyperactivity Disorder (ADHD) and Subtype for Adolescents 12–15 Years Old (National Health and Nutrition Examination Survey [NHANES] 2001–2004)

Clinical Features	ADHD, Any Type			ADHD Subtypes
				Inattentive Type			Hyperactive/Impulsive Type			Combined Type
	Onset <7 years	Onset 7 to <12 years	p -value	Onset <7 years	Onset 7 to <12 years	p - value	Onset <7 years	Onset 7 to <12 years	p -value	Onset <7 years	Onset 7 to <12 years	p -value
Number (n)	99	64		44	42		18	18		37	4
Inattention Symptom Count, Mean (SE)	7.9 (0.3)	7.8 (0.4)	0.422	8.6 (0.2)	8.7 (0.3)	0.721	3.9 (0.9)	4.4 (0.8)	0.720	9.3 (0.2)	7.9 (0.4)	0.004
Hyperactivity Symptom Count, Mean (SE)	5.4 (0.5)	4.8 (0.5)	0.250	2.8 (0.4)	3.6 (0.5)	0.097	7.7 (0.3)	7.8 (0.1)	0.685	8.7 (0.3)	7.2 (0.3)	0.001
Severe Cases/Higher Impairment, % (SE)	66.8 (6.1)	75.6 (6.5)	0.323	72.1 (8.1)	81.6 (7.9)	0.415	30.2 (14.2)	60.2 (20.8)	0.403	79.5 (7.8)	63.3 (29.4)	0.304
12-month Comorbidity, % (SE)
Anxiety Disorder (Panic/GAD)	0.3 (0.3)	0.0 (0.0)	-	0.0 (0.0)	0.0 (0.0)	-	1.8 (1.8)	0.0 (0.0)	-	0.0 (0.0)	0.0 (0.0)	-
CD	12.1 (2.8)	15.9 (5.7)	0.653	6.7 (4.7)	20.7 (7.5)	0.319	0.0 (0.0)	6.5 (4.7)	-	29.9 (6.9)	0.0 (0.0)	-
Depression (MDD, Dysthymia)	1.8 (1.5)	0.0 (0.0)	-	2.8 (2.7)	0.0 (0.0)	-	1.8 (1.8)	0.0 (0.0)	-	0.0 (0.0)	0.0 (0.0)	-
Treatment and Medication, % (SE)
ADHD-specific treatment	53.5 (6.3)	46.0 (6.0)	0.445	52.5 (7.4)	39.5 (8.8)	0.240	39.7 (13.2)	64.0 (10.6)	0.212	64.0 (11.8)	56.1 (30.0)	0.281
Mental health treatment	56.1 (6.3)	50.2 (5.9)	0.532	56.2 (7.8)	45.4 (8.8)	0.292	41.4 (13.9)	64.0 (10.6)	0.210	64.8 (11.4)	56.1 (30.0)	0.308
Use of medication past year	54.8 (5.9)	34.7 (8.0)	0.178	54.8 (9.6)	35.8 (9.0)	0.403	39.7 (13.2)	42.9 (21.4)	0.661	64.1 (11.8)	8.0 (8.8)	0.008

Note: p-values based on Wald χ2 tests, adjusted for sex, age, race, and poverty index ratio (PIR). CD = conduct disorder; GAD = generalized anxiety disorder; MDD = major depressive disorder; SE = standard error.

Table 2 also compares the proportions of youth with comorbid disorders including major depression, anxiety, and conduct disorders by subtype for each age-of-onset group. In the overall sample, 12.1% of those ADHD cases with an earlier onset and 15.9% of those with a later onset had comorbid CD (p=0.657). In the overall sample with earlier onset ADHD, 1.8% had comorbid depression and 0.3% had comorbid anxiety whereas in the later-onset group, there were no cases of comorbid depression or anxiety in the last 12 months. Although there was no significant difference in treatment by ADHD subtypes among youth with early versus late onset, there was a striking difference in the proportion that used medication in the past year in the combined subtype, particularly in males. The proportion who received medication in the earlier-onset group was 64.1% while only 8% of the later-onset group received medication (p=0.019). Of note, when performing some subgroup comparisons, especially in the ADHD combined subtype, sample sizes became smaller, which may lead to a lack of sufficient power to detect the existent differences.

Table 3 presents the results of a multivariate logistic regression analysis that compared the two ADHD age-of-onset groups adjusting for other demographic and clinical characteristics. The results indicate that the later age-of-onset group was more likely to include youth with lower family income (PIR<1) (OR=3.23, 95% CI 1.11–9.39). However, there was no significant interaction between sex and ADHD subtype in the later-onset group.

Table 3.

Correlates of Attention-Deficit/Hyperactivity Disorder (ADHD) Cases with Later Onset (n=163)

	Adjusted Odds Ratio, 95% Confidence Interval		Test Statistic
	aOR	95% CI	Wald χ2 [p]
SEX
Male	Reference		0.2 [0.649]
Female	1.32	0.38 –4.58
AGE
12	Reference		0.4 [0.941]
13	1.27	0.43 –3.80
14	1.07	0.35 –3.24
15	1.33	0.37 –4.78
RACE/ETHNICITY
Non-Hispanic White	Reference		3.2 [0.074]
Others	2.43	0.88 –6.72
FAMILY INCOME
PIR >=1	Reference		5.0 [0.025]
PIR <1 (poor)	3.23	1.11 –9.39
ADHD SUBTYPE
Inattention	Reference		4.8 [0.092]
Hyperactivity	1.32	0.42 –4.17
Combined	0.22	0.04 –1.26
SEVERE/HIGHER IMPAIRMENT CASE
Yes	1.93	0.60 –6.25	1.3 [0.253]
No	Reference
COMORBID WITH OTHER DISORDERS
Yes	1.07	0.24 –4.74	0.01 [0.931]
No	Reference

Note: Multivariate logistic regression model ran within n=163 ADHD cases. The outcome variable was being a later-onset case (1) versus being an earlier-onset case (0). Demographic and clinical characteristic variables listed in the table were included in the model simultaneously. aOR = adjusted odds ratio; chisq = chi square; df = degrees of freedom; PIR = poverty index ratio.

DISCUSSION

The results of this study provide the first comparison of DSM-IV vs. DSM-5 criteria for age of onset of ADHD in a large, nationally representative, population-based sample of U.S. children and adolescents. The finding of comparable severity, functional impairment, patterns of comorbidity and treatment among the additional 3.46% of children who met all of the ADHD criteria except age of onset under 7 to those with early onset supports the recent age-of-onset change in the DSM-5. Although relaxation of the age-of-onset criterion in the DSM-5 generated an increase in all subtypes of ADHD, the greatest increase was found for the inattentive subtype. This confirms findings from previous studies that the age of onset of those with predominantly inattentive symptoms have later age of onset than those with primarily hyperactive symptoms.^{15, 20} Therefore, extension of the age-of-onset criterion will lead to the recognition of a larger proportion of children with impairment associated with ADHD.¹⁴ Later recognition of inattention symptoms may represent a later stage in development or increased challenge with academic training.

Consistent with population-based research on the sex ratio of DSM-IV ADHD, we found that males had significantly greater rates of ADHD prevalence than females across all subtypes with the DSM-5 criteria. Although not statistically significant, male-to-female ratios remained relatively constant (ratios ranged from 1.1 to 2.1) with slightly lower sex ratios in the later-onset group, indicating increased recognition of female cases under the new criteria.²⁵ This trend suggests that the later age at onset may reflect improved recognition of the disorder in females. Finding potential explanations for these differences, including biologic explanations for differential developmental manifestations of ADHD in males and females, is an important area for future research.^26,27

It is particularly noteworthy that the new criteria identify more children from lower-income families and ethnic minorities who would not have been identified by the DSM-IV criteria. Previous literature has shown that low family income is associated with an increased likelihood of ADHD,^28,29 even though youth from higher-income families are more likely to receive an ADHD diagnosis.³⁰ Furthermore, previous studies have shown that black and Hispanic children are less likely to be diagnosed with ADHD.^31,32 Reasons for these disparities may include a lack of cultural competence, access to health care, or patient knowledge of ADHD symptoms and treatment.³¹ By extending the age criteria, our results suggest that youth from lower income families and ethnic minority groups have an increased opportunity for diagnosis and treatment. However, the clinical implication of this observation warrants a further investigation.

We confirmed the strong comorbid association between ADHD and CD reported in several previous studies.^33–36 Although there was a stronger association between ADHD and CD in the early-onset group with the combined type,³⁶ CD was also strongly associated with the later-onset ADHD inattentive type. Given that comorbidity can impact treatment and the academic, social and family functioning of children with ADHD, clinicians should be vigilant in recognizing and treating both disorders.³⁶

The most compelling support for the DSM-5 age-of-onset change emerged from the finding that, despite equal clinical severity, the later-onset group with either the inattentive or combined type was less likely to receive general mental health and ADHD-specific treatment. However, the opposite was true for those with the hyperactive/impulsive type in whom children with earlier onset were slightly less likely to receive treatment than those with later onset. One possible explanation may be the perception that younger children are naturally more energetic and impulsive, which may delay a parent in searching for medical intervention.³⁷

The findings of this study should be interpreted in the context of several limitations. First, the diagnosis of ADHD was based on parent report only. Although several studies have supported the use of parent rather than child reports based on the poor validity of child self-report for ADHD symptoms,^8,38 our study did not include teacher reports, which provide an important perspective on attention and hyperactivity symptoms.³⁸ However, because older children tend to have multiple teachers, the lack of inclusion of teacher reports may be a less serious limitation in middle and high school youth. Second, because of the cross-sectional design, ADHD symptoms and age of onset are based on retrospective report that may be subject to recall bias. This may be particularly relevant to recall of age at onset, although previous studies have demonstrated “excellent” one-year maternal recall accuracy for ADHD symptoms in children.³⁹ Nevertheless, prospective data are required to establish the validity of these findings.

Despite these limitations, we present the first systematic study of a nationally representative sample of U.S. youth to assess the later age-of-onset criterion introduced in the DSM-5. In support of systematic aggregation of earlier studies¹⁶ and a UK cohort of children followed to age 12,¹⁹ our findings provide evidence to support the extension of the age-of-onset criterion in ADHD. Without this extension, nearly 3.5% of children who meet all symptom, duration and impairment criteria for ADHD would go unrecognized. Those children with an onset between ages 7 and 12 exhibit similar levels of severity to those with an onset prior to age 7, indicating that enhanced recognition and treatment of this group of children are imperative. However, perhaps most important, the imposition of any arbitrary age-of-onset cut-offs should be further examined to provide insight into the development of ADHD across the lifespan.

CLINICAL GUIDANCE.

• ▪Extending the ADHD age-of-onset criterion from 7 to 12 years led to an increase in the 12-month prevalence rate of ADHD from 7.38% to 10.84% based on a nationally representative sample of U.S. youth.
• ▪Children whose onset was between the ages of 7 and 12 years were not systematically different in terms of severity and comorbidity when compared to those children whose onset occurred prior to age 7 years.
• ▪Children whose onset was between the ages of 7 and 12 years were more likely to come from lower-income and ethnic minority families.