Timing of the Diagnosis of Attention-Deficit/Hyperactivity Disorder and Autism Spectrum Disorder
Abstract
BACKGROUND AND OBJECTIVE: Symptoms of inattention, hyperactivity, and impulsivity are core features of attention-deficit/hyperactivity disorder (ADHD). However, children with autism spectrum disorder (ASD) often present with similar symptoms and may receive a diagnosis of ADHD first. We investigated the relationship between the timing of ADHD diagnosis in children with ASD and the age at ASD diagnosis.
METHODS: Data were drawn from the 2011–2012 National Survey of Children's Health, which asked parents to provide the age(s) at which their child received a diagnosis of ADHD and/or ASD. Using weighted prevalence estimates, we examined the association between a previous diagnosis of ADHD and the age at ASD diagnosis, while controlling for factors known to influence the timing of ASD diagnosis.
RESULTS: Our study consisted of 1496 children with a current diagnosis of ASD as reported by parents of children ages 2 to 17 years. Approximately 20% of these children had initially been diagnosed with ADHD. Children diagnosed with ADHD before ASD were diagnosed with ASD ∼3 years (95% confidence interval 2.3–3.5) after children in whom ADHD was diagnosed at the same time or after ASD. The children with ADHD diagnosed first were nearly 30 times more likely to receive their ASD diagnosis after age 6 (95% confidence interval 11.2–77.8). The delay in ASD diagnosis was consistent across childhood and independent of ASD severity.
CONCLUSION: To avoid potential delays in ASD diagnosis, clinicians should consider ASD in young children presenting with ADHD symptoms.
- AAP —
- American Academy of Pediatrics
- ADHD —
- attention-deficit/hyperactivity disorder
- ASD —
- autism spectrum disorder
- CI —
- confidence interval
- FPL —
- Federal Poverty Level
- NSCH —
- National Survey of Children’s Health
What’s Known on This Subject:
Many studies have suggested that autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD) are commonly co-occuring neurodevelopmental conditions.
What This Study Adds:
In children with co-occurring ASD and ADHD, an initial ADHD diagnosis may be associated with delayed ASD diagnosis and a higher likelihood of ASD diagnosis older than 6 years of age. Clinicians should consider ASD when evaluating young children presenting with ADHD symptoms.
Although autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD) are unique neurobiological conditions, they have been shown to share genetic factors1–3 and neural pathways.4–6 They also exhibit many overlapping phenotypic features, including attentional and social vulnerabilities7–9 and externalizing symptoms, such as impulsivity and hyperactivity.10,11 In addition, trend analyses using data from the 2011–2012 National Survey of Children's Health (NSCH) reported dramatic increases in the prevalence of both ASD and ADHD. The most recent estimates of parent-reported diagnosis by a health care provider were 2% for ASD in children 6 to 17 years old12 and 11% for ADHD in children 4 to 17 years old.13
ASD can be reliably diagnosed in children as young as 24 months14,15; however, the median age at first ASD diagnosis remains older than 4 years.16 Pringle et al17 found that 40% of 6- to 17-year-old children with special health care needs and ASD were aged 6 years and older when first identified as having ASD. This is well beyond the age when signs and symptoms should clearly be notable and when behavioral therapies appear to be most effective.18–20
Numerous studies have identified family and child factors associated with disparities in the diagnosis of ASD.21 Minority background, lower level of parent education, and disadvantaged socioeconomic circumstances have all been correlated with increased age of diagnosis.22,23 Levy et al24 identified an association between later diagnosis of ASD and the presence of co-occurring developmental, psychiatric, and neurologic disorders. Similarly, 51% of children in Iceland diagnosed with ASD after age 6 had a previous non-ASD developmental diagnosis.25 A more recent study in Israel focused on children who received a diagnosis of ASD at or after age 6 despite having undergone a comprehensive multidisciplinary assessment before age 6. Almost half of the sample had demonstrated ADHD features at the initial assessment. The authors suggest several reasons for why ASD diagnoses may have been missed at younger ages, including social and pragmatic deficits that gradually emerge as social demands increase with age.26
A larger population-based cohort in Nova Scotia, Canada, used administrative health databases to obtain the age at ASD diagnosis for 884 children.27 In adjusted models, the 3 features related to age of diagnosis were maternal age, county of residence, and co-occurring ADHD. An ASD diagnosis occurred 1.29 years later (95% confidence interval [CI] 0.93–1.64) if a child had a comorbid ADHD diagnosis compared with a child without ADHD. Interestingly, in almost 60% of these cases, an ADHD diagnosis had been documented some time before the ASD diagnosis.
Although these studies suggest that symptoms of ADHD may overshadow or mask the symptoms of ASD, none have examined the relationship between the age at ADHD diagnosis and the age at ASD diagnosis. We hypothesized that a diagnosis of ADHD before ASD would be associated with delayed ASD diagnosis compared with children who received an ADHD diagnosis at the same time or after an ASD diagnosis or children with ASD only.
Methods
Data for this study were drawn from the 2011–2012 NSCH public-use data file. The NSCH is directed and funded by the Maternal and Child Health Bureau of the Health Resources and Services Administration and is administered by the National Center for Health Statistics. The 2011–2012 NSCH used a multistage cluster design based on a random digit–dialed sample of households with children younger than 18 years of age selected from each of the 50 states and the District of Columbia. The respondent was a parent or guardian in the household who was knowledgeable about the child’s health. The overall response rate for 2011–2012 was 23%. A total of 95 677 interviews were completed from February 2011 through June 2012.12 Information about parent-reported ASD diagnosis was obtained for 85 556 children ages 2 to 17 years. For more information about NSCH, including its sample design, data collection procedures, and questionnaire content, visit http://www.cdc.gov/nchs/slaits/nsch.htm.
Study Sample
Children classified as having ASD were those for whom a parent provided valid responses to the following questions: “Has a doctor or other health care provider ever told you that [study child] had autism, Asperger disorder, pervasive developmental disorder, or other autism spectrum disorder?” “How old was [study child] when you were first told by a doctor or other health care provider that [he/she] had autism or ASD?” and “Does [study child] currently have autism or ASD?” To obtain the temporal order between ASD and ADHD diagnoses, a parent had to provide a valid response to the question: “How old was [study child] when you were first told by a doctor or other health care provider that [he/she] had ADD or ADHD?” Based on the responses to these questions, we defined a 3-level variable: (1) ADHD before ASD, (2) ADHD same/after ASD, and (3) ASD only. Omissions to these variables, including “refusal” or “don’t know” responses, or no current ASD diagnosis were excluded. We set the lower limit of diagnosis for ASD at 2 years14,15 and for ADHD at 3 years28–30 based on previous literature.
Study Variables
All sociodemographic variables were acquired by parent report and included child gender, race (white, black, other), ethnicity (Hispanic, non-Hispanic), mother’s education (high school diploma or less, more than high school diploma), parent-reported ASD severity (mild, moderate, severe), and household income (≤200% of the federal poverty level [FPL], 201%–400% FPL, >400% FPL). Co-occurring conditions of interest included parent-reported developmental delay, speech problems, and intellectual disability. All study covariates have been significantly associated with variations in the timing of ASD diagnosis in previous studies.21 We also included child’s age at the time the survey was completed to account for any secular trends in diagnostic patterns for ASD or ADHD.
Analytic Methods
All statistical data analyses were performed using SPSS version 22.0 (IBM SPSS Statistics, IBM Corporation, Chicago, IL) with SPSS Complex Samples, which accounted for the complex survey design of NSCH, including clustering of children within households, stratification of households within states, and unequal sampling weights. Using sample weights to calculate prevalence estimates and SEs better represents the noninstitutionalized population of the United States. A .05 level of significance was used in the analyses. In comparing weighted proportions, significance was based on the adjusted F statistic, a variant of the second-order Rao Scott adjusted χ2 statistic used for analyses with sample weights. We used the General Linear Model procedure within the Complex Samples Module to examine the associations between the age at ASD diagnosis and the 3 study groups while controlling for other independent variables. Adjusted odds ratios were obtained from multivariate logistic regression analysis using age at ASD diagnosis dichotomized at 6 years of age as the binary dependent variable. The Boston Children’s Hospital Institutional Review Board exempted this study as nonhuman subject research.
Results
The final sample included 1496 study children with a current ASD diagnosis. For 705 (42.9%) of these children, the parent or caregiver indicated that the child had also received an ADHD diagnosis and provided the age at ADHD diagnosis. Of the children with co-occurring ASD and ADHD, 313 (44.5%) comprised the ADHD before ASD group. The group with ADHD before ASD contained a significantly smaller proportion of children with speech problems than the other 2 subgroups (adjusted F[1.88, 2729] = 4.91, P = .009). Overall, children with ADHD before ASD were ∼4 years older when they received their ASD diagnosis (F[2, 1449] = 51.51, P < .001) and 81% were diagnosed after age 6 (adjusted F[1.79, 2595] = 35.15, P < .001). The 3 groups were similar in the distribution of other demographic characteristics. All of the factors listed in Table 1 were included in our final models because a review of the literature had indicated they were significantly associated with the timing of ASD diagnosis (Table 1).21
Demographic and Clinical Characteristics of Children Stratified by Study Group (n = 1451)
Across study groups, the presence of speech problems was associated with an ASD diagnosis occurring 1.2 years earlier, whereas children with an intellectual disability were diagnosed with ASD 0.6 years earlier. Mild to moderate autism severity was associated with later diagnosis compared with those with severe ASD.
Controlling for all covariates, ADHD before ASD was associated with a 2.9-year delay (β = 2.87 t[1211] = 9.64, P < .001) in the age at ASD diagnosis when compared with children with ASD only. The results of the regression indicated that our full models explained 53% of the variance (R2 = 0.53, F[2, 1210] = 59.51, P < .001). In contrast, the age at ASD diagnosis for the group with ADHD same/after ASD was not significantly different than it was for children with ASD only. We re-ran the same model shifting which group served as the reference to directly compare the 2 groups with comorbid ASD and ADHD. We found that ADHD before ASD was associated with a 3.2-year delay (β = 3.18, t[1211] = 9.86, P < .001) in ASD diagnosis compared with ADHD same/after ASD (Table 2).
Associations for Age at ASD Diagnosis and ASD Diagnosis > 6 y (n = 1212)
Figure 1 illustrates how the gap in age at ASD diagnosis between the 2 groups with ADHD remains fairly constant across early childhood and into adolescence.
Predicted mean age at ASD diagnosis by age at ADHD diagnosis for ADHD before ASD (dark gray line) and ADHD same/after ASD (light gray line).
Looking at the probability of receiving an ASD diagnosis after 6 years of age, children with ADHD before ASD were 16.7 times more likely to be diagnosed with ASD after 6 years of age (t[1211] = 6.38, P < .001) compared with children with ASD only and 29.5 (95% CI 11.2–77.8) times more likely compared with children with ADHD same/after ASD (t[1211] = 6.86, P < .001) (Table 2). Children with mild and moderate autism were also more likely to be diagnosed after 6 years compared with severe presentations. Notably, whereas children with milder cases of ASD were more likely to be diagnosed after 6 years of age, the disparity in age of ASD diagnosis for the ADHD before ASD group was maintained across all severity levels (Fig 2).
Mean age at ASD diagnosis stratified by parent-reported ASD severity (mild, moderate, severe). Adjusted for gender, race, ethnicity, maternal education, household income, and child’s age at survey completion.
Discussion
The American Academy of Pediatrics (AAP) guidelines endorse universal screening for ASD at 18 and 24 months of age.15 The AAP also recommends careful developmental surveillance at every well-child visit, a process that emphasizes eliciting parental concerns, considering family history and observing the child during visits.31 Despite the push for routine screening and surveillance in primary care, our study of children 2 to 17 years old found that, overall, 39.5% were not diagnosed until 6 years or older, similar to the findings of other large surveillance studies16 and population surveys.17
Barkley32 maintained that it was common for children with pervasive developmental disorder, not otherwise specified, to initially receive a diagnosis of ADHD. Jensen et al33 reported that 74% of the children diagnosed with pervasive developmental disorder, not otherwise specified, in their study were originally diagnosed with ADHD. Perry34 estimated that at least half of the patients referred to him for Asperger disorder had previously been diagnosed with ADHD. More recently, Mandell et al35 reported 21% of Medicaid-eligible children with ASD seen in a mental health clinic received a diagnosis of ADHD before ASD. We found ∼20% of children in our sample with ASD had been diagnosed with ADHD before ASD. Although an association between milder cases and a later age of ASD diagnosis was supported by our findings, the delay in diagnosis for the ADHD before ASD group existed regardless of the age at ADHD diagnosis or the severity of the child’s ASD.
Most of the evidence for making an accurate diagnosis of ASD in young children is based on studies conducted with experienced and highly trained staff in research settings. The NSCH reflects what occurs in the population at large. Community-based primary care physicians are often the first to encounter parental concerns about their child’s development,36 yet the variability in presenting symptoms makes it challenging to diagnose ASD in young children.37 Parents are less likely to express specific concerns about social development before their child has been exposed to a consistent peer group,38,39 and the stereotypies and repetitive behaviors that are required for the diagnosis of ASD are often less pronounced at younger ages.40 As a result, the accuracy rate of ASD diagnosis among general pediatricians has remained relatively low.41–43
Conversely, >40% of clinically evaluated children with ASD who are <6 years old have been found to display symptoms of ADHD.26,44 Published research suggests limited adherence to the AAP guidelines for ADHD diagnosis within primary care settings.45 Only 25% of pediatricians reported using the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria, and only 67% used standardized rating scales in making the diagnosis.46 Because the prototypical behaviors and social deficits that characterize ASD may not manifest themselves in the relatively short time frame allotted for routine clinic visits,37 general practitioners may be inclined to attribute maladaptive behaviors to ADHD, the most common neurobehavioral disorder of childhood. This is especially true if they have more experience managing and treating ADHD, if they do not feel comfortable giving a diagnosis of ASD,47 or if they lack access to specialists who can conduct comprehensive neurodevelopmental assessments.48 However, ADHD is rarely an easy diagnosis to make in very young children.49,50 Although many of the children in our study retained their ADHD diagnosis after their ASD diagnosis, 12.6% of those who were diagnosed at younger than 7 years of age reported having only ASD and not ADHD at the time of survey completion. Similarly, Law et al30 re-contacted children who had been diagnosed with ADHD after undergoing a multidisciplinary consultation when they were younger than 7 years; they found that 7 years after the initial evaluation, 11.4% of the children reported having a current ASD diagnosis but no longer carried the original ADHD diagnosis.
Our findings are subject to several limitations. First, the cross-sectional design of the study precludes us from making statements about the causality of our associations. Although the parent-reported timing of ASD and ADHD diagnoses allows us to examine the temporal order of events, we cannot prospectively control or account for other intervening factors. Second, the indicators used here rely on parent-reported diagnoses by a health care provider, which is not only subject to recall bias but lacks the validity of clinical assessment using standardized measures. Without collateral information from medical records or teacher reports, what parents perceive as ADHD may actually be within the normative range of behavior for young children. However, a recent analysis indicated that parent-reported survey data produced similar estimates as those from insurance claims data for ADHD51 and parent-reported diagnosis of autism was also consistent between 2 nationally representative surveys,52 providing evidence of convergent validity. Although NSCH estimates are potentially subject to survey nonresponse bias, analyses of nonresponse bias suggest that differences between respondents and nonrespondents should not have a major impact on the conclusions in this report.12 Finally, information on the severity of ASD at the time of diagnosis and the diagnostic subtype of ASD were not available. Therefore, we attempted to control for this by using current parent-reported ASD severity as a proxy for baseline ASD severity in our models.
Conclusions
We found a substantial number of parents reporting an ADHD diagnosis by a health care provider before an ASD diagnosis. Moreover, these children were significantly older when they received their ASD diagnosis compared with other children diagnosed with ASD. Our study supports the hypothesis that receiving a diagnosis of ADHD before ASD may delay the diagnosis of ASD, and that this delay persists across age and severity of the ASD. Furthermore, if ADHD is diagnosed first, the ASD diagnosis has a higher probability of occurring late (ie, >6 years of age). ASD that goes unrecognized and untreated until the child is older may negatively affect their long-term prognosis.
Our findings have important implications: (1) children with ADHD before ASD may exhibit unique dimensional traits that could bias clinicians toward an ADHD diagnosis, and (2) diagnostic criteria and screening measures for ASD may need to reflect the overlapping symptomatology between ASD and ADHD. However, prospective studies are still needed to characterize the overlapping and distinct phenotypic presentations of individuals with ADHD before ASD. For now, clinicians should consider ASD when evaluating young children presenting with ADHD symptoms.
Footnotes
- Accepted July 6, 2015.
- Address correspondence to Amir Miodovnik, MD, MPH, Division of Developmental Medicine, 1 Autumn St, Boston, MA 02215. E-mail: amir.miodovnik{at}childrens.harvard.edu
Dr Miodovnik conceptualized and designed the study, conducted the analyses, drafted the initial manuscript, and reviewed and revised the manuscript; Dr Harstad reviewed and revised the manuscript; Dr Sideridis critically reviewed the analyses; Dr Huntington reviewed all analyses completed by Dr Miodovnik and critically reviewed and revised the manuscript; and all authors approved the final manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: Dr Harstad has been the recipient of grants from SynapDx. The other authors have indicated they have no potential conflicts of interest to disclose.
References
- Copyright © 2015 by the American Academy of Pediatrics
