Epidemiology of ASD in Preschool-age Children in Japan

Abstract

In recent years, it has been reported that the prevalence of autism spectrum disorder (ASD) is increasing, but there are few research reports in Asia equivalent to those in Europe and the United States. Since large-scale epidemiological studies of neurodevelopmental disorders (NDDs) have not been conducted in Japan, the delay in early detection is conspicuous compared to other countries. Therefore, we started epidemiological studies in a medium-sized city (Hirosaki City) in northern Japan from 2013 to elucidate the prevalence of ASD and have been conducting a 9-year community cohort survey. In 2020, we published an adjusted prevalence of ASD of 3.2% at the age of 5 years, no change in 4-year incidence, and comorbidity of ASD. Since then, we have focused on sleep problems at the age of 5 years and have been studying the estimation of the prevalence of sleep disorders and the relationship with neurological development disorders. In this chapter, in addition to our research results since 2013, we will introduce the screening and support system in the community in Japan.

1. Introduction

In epidemiological studies, the prevalence of autism spectrum disorder (ASD) has changed significantly over the last 20 years. A 2018 study by the Centers for Disease Control and Prevention (CDC) reported a prevalence of 2.30% (1 in 44 children) among children aged 8 years [1]. In 2008, the prevalence was 1 in 88 children [2], so the ADDM network reported that the prevalence of ASD has doubled in 10 years. Diagnostic criteria are American Psychiatric Association’s the Diagnostic and Statistical Manual of Mental Disorders (DSM), Fourth Edition, Text Revision (DSM-IV-TR), 5th ed. (DSM-5), the World Health Organization’s International Statistical Classification of Diseases and Related Health Problems (ICD), Ninth Revision (ICD-9) or Tenth Revision (ICD-10), indicating the importance of following the same criteria [3, 4, 5, 6].

In Asia, a large-scale study by YS Kim et al. in 2011 reported the prevalence of ASD with similar diagnostic criteria, reporting 2.64% in children aged 7–11 years [7]. In Japan, there has been no large-scale epidemiological study by DSM criteria, and in 2020 we reported for the first time that the prevalence of 5-year-old ASD in the community was 3.22% [8].

In recent years, as the heterogeneous nature of brain and sensory functions in neurodevelopmental disorders (NDDs) has been clarified, it has been suggested that ASD symptoms can be improved by early intervention, and ASD has undergone a paradigm shift as a disorder in which neuroplasticity can be expected [9]. In addition, it has been reported that people with neurodevelopmental disorders often have complications such as depression, anxiety disorders, and conduct disorders [10, 11]. Many of these complications occur as secondary disorders of neurodevelopmental disorders (NDDs), and early detection of developmental characteristics and early intervention are considered important to prevent secondary disorders [12]. In this chapter, in addition to our research results since 2013, we will introduce the screening and support system in the community in Japan.

2. Epidemiology of ASD in preschool-age children in Japan

2.1 Infant health checkup system in Japan

In Japan, the Maternal and Child Health Law requires municipalities to conduct health and development checkups for children aged 18 and 36 months [13]. At these checkups, public health nurses and pediatricians assess physical, motor, social, emotional, behavioral, verbal development, and general health (medical and dental). If these assessments are found to be abnormal, the public health nurse and/or pediatrician should refer the patient to a specialist for further evaluation, if necessary, and have the child’s caregiver available for early intervention, such as speech therapy. Providing local resources where possible is encouraged (but not compulsory). According to a Ministry of Health, Labor, and Welfare report published in 2020, the 18-month health checkup rate in Japan was 95.2%, and the 36-month health checkup rate was 94.5% [14].

While the participation rate in infant health checkups is high and infant deaths are sufficiently prevented, the rate of finding prominent neurodevelopmental disorders in health checkups is extremely low at 0.2–1.3%. Therefore, in 2017, the government made recommendations for early detection of neurodevelopmental disorders in infant health checkups [15]. The number of children receiving special support after entering elementary school has tripled from 10 years ago [16]. There is a gap between the high need for support due to the awareness of neurodevelopmental disorders by guardians and caregivers and the low detection rate in the health checkup system.

2.2 The Hirosaki five-year-old children developmental health check-up (HFC) study

The Hirosaki Five-year-old Children Developmental Health Check-up (HFC) study was established in 2013 using a total population sample of 5-year-olds living in Hirosaki City, Japan. The rationale for this study is that, despite the high participation rate in these assessments, the data of children who screened positive for NDD at 18-month or 36-month health checkup or follow-up arose from concerns about the limited. There were also concerns about whether children who screened positive before the age of 5 years were subsequently evaluated and received appropriate services. This study, in collaboration with the University of California, San Francisco (UCSF), used HFC data from 2013 to 2016 to estimate the prevalence and 5-year cumulative incidence of ASD at the age of 5 years. We investigated patterns of comorbidity of NDDs in children with ASD, including hyperactivity disorder (ADHD), developmental coordination disorder (DCD), and intellectual disability (ID) [8].

Five-year-old Children Developmental Health Check-up was conducted annually between January 2013 and December 2016, and four cohorts were created. HFC was conducted in two stages and aimed to detect all children with NDD in the community. The primary screening was a series of questionnaires mailed by Hirosaki City to the parents and kindergarten or nursery school teachers of all 5-year-olds living in the city. The questionnaire used epidemiological information (children’s gender, family composition, parental education, and employment history, etc.) and the following scales: Autism Spectrum Screening Questionnaire (ASSQ), Strengths and Difficulties Questionnaire (SDQ), ADHD Rating Scale-IV (ADHD-RS-IV), Developmental Coordination Disorder Questionnaire (DCDQ), and Parenting Stress Index (PSI) [17]. The parent has completed all of the above questionnaires, and the teacher has completed SDQ. All of the above tools have been translated into Japanese, and their reliability and validity had previously been established [18, 19, 20, 21, 22].

In the second stage of evaluation, a comprehensive evaluation of NDDs in screen-positive children was performed at Hirosaki University Hospital. Caregivers of screen-positive children were invited to participate in a comprehensive evaluation. Figure 1 is a flow chart of the screening and evaluation stages of HFC research.

Figure 1.

Flow chart of the Hirosaki five-year-old developmental Checkup and assessment.

At the comprehensive assessment, developmental history and concerns were collected using items derived from the Diagnostic Interview for Social and Communication Disorders (DISCO) [23]. The DISCO is a semi-structured interview schedule designed to collect information on development and behavior. It can be used to assist in identifying possible diagnostic categories, including ASDs and other developmental disorders affecting social interaction and communication. Cognitive assessments were conducted by psychologists using the Japanese version of the Wechsler Intelligence Scale for Children, 4th edition (WISC-IV) [24] only for children who did not have an ID diagnosis before participating in the present study. Assessment of children’s coordination skills was performed by trained occupational therapists and psychologists using the Movement Assessment Battery for Children, 2nd edition (MABC-2) [25]. MABC-2 is a test of coordination disorders for children aged 3–16 years. It consists of three fine motor measures, two ball skill measures, and three balance skill measures. Screening test scores, parent interviews and child examinations, and other test results were reviewed by multiple professionals, including child psychiatrists. If ASD is diagnosed or suspected, the child is added to the Autism Diagnostic Observation Schedule (ADOS)-2 [26], and they were assured of research reliability. The definitive diagnosis was determined on the basis of findings consistent with screening evaluation and diagnosis. We used the DSM-5 criteria for the diagnosis of ASD, ADHD, and both the DSM-5 and the European Academy of Childhood Disability guidelines [27] for the diagnosis of DCD. Criteria for ID were defined as an IQ <70.

2.3 Prevalence and cumulative incidence of ASDs and the patterns of co-occurring neurodevelopmental disorders in a total population sample of 5-year-old children

Of the 559 children who underwent secondary assessment, 87 children (60 boys and 27 girls) were diagnosed with ASD. The 4-year mean ASD crude prevalence was 1.73% (95% CI 1.37–2.10%), with a 95% CI of 1.37 to 2.10%. Gender crude prevalence estimates for ASD were 2.35% (95% CI 1.76–2.94%) for boys and 1.09% (95% CI 0.68–1.51%) for girls, with the gender ratio of 2.2:1. After statistically adjusting for nonparticipants in comprehensive developmental assessment, the adjusted prevalence of ASD was estimated to be 3.22% (95% CI 2.66–3.76%). Gender-adjusted prevalence of ASD was 4.06% (95% CI 3.20–4.92%) in boys and 2.22% (95% CI 1.57–2.88%) in girls, with the gender ratio of 1.8:1.

The cumulative incidence of ASD by the age of 5 years within this research period (2013–2016) was 1.31% (95% CI 1.00–1.62%), with no significant increase in the 5-year cumulative incidence. The prevalence and 5-year cumulative incidence for each study year are summarized in Table 1.

Table 1.

Crude prevalence, adjusted prevalence, and cumulative incidence up to the age of 5 years of autism spectrum disorders in each survey year.

Of the children with ASD (N = 87), 88.5% (n = 77) had at least one comorbid NDD (ADHD, DCD, ID, and/or borderline intellectual function (BIF)) and 20 children with ASD (23%) had 3 comorbid NDDs. Gender ratio of comorbidities was ADHD 50.6% (boys: girls = 2.4:1), DCD 63.2% (2.1:1), ID 36.8% (1.7:1), and BIF 20.6% (2.6,1) (Figure 2 and Table 2).

Figure 2.

Number of comorbidities of autism spectrum disorder.

Co-occurring NDDs	n	%
None (i.e. ASD alone)	10	11.49
ADHD alone or ADHD and other NDDs	44	50.57
DCD alone or DCD and other NDDs	55	63.22
ID alone or ID and other NDDs	32	36.78
BIF alone or BIF and other NDDs	18	20.69

Table 2.

Comorbid patterns of neurodevelopmental disorders in 87 individuals with ASD.

Only 21 of 87 ASD children had received a diagnosis of ASD prior to this study. Of the 59 children who were assisted by the age of 5 years, 38 had other diagnoses (developmental or language delay). Twenty-eight (32%) had no developmental problems and no remedial intervention by the age of 5 years. Figure 3 shows the problem of undiagnosed and unintervention of ASD.

Figure 3.

The problem of undiagnosed and unintervention of ASD.

Our study revealed that 5-year-old children with ASD have a high incidence of concurrent NDD, suggesting that ASD has a wide range of difficulties in daily life, such as attention and motor control, in addition to social problems. In infant screening, it is necessary to broadly evaluate various characteristics and provide early developmental support to children.

For this reason, children who were diagnosed with some form of NDD at a health checkup were promptly given support and prepared for entering elementary school. Children who have been diagnosed with the disease should visit the Hirosaki University Hospital regularly at least once every 1 to 2 years for examinations and consultations and follow up until the age of 15 years to reconsider the diagnosis and determine the need for treatment. Some children require medication. The effects of these interventions need to be analyzed separately.

2.4 Prevalence of sleep problems in Japanese Preschoolers and children with developmental disabilities

Sleep problems are not only associated with emotions and behaviors but also affect mental and physical health over time. Our previous study reported 80% of 482 children had sleep problems in Japan [28]. Among Chinese urban kindergarten children, also, almost 80% (78.8%) of the children scored above the original Children’s Sleep Habits Questionnaire (CSHQ) cutoff point for global sleep disturbance [29]. However, there has been no report of a larger-scale study and the comorbid rate of neurodevelopmental disorders. The aim of this study was to estimate the prevalence of sleep problems in preschoolers and children with developmental disabilities using the Children’s Sleep Habits Questionnaire (CSHQ), which is widely used in large community-based surveys.

Subjects were 1800 children who participated in 5-year-olds developmental checkup in a city, Japan. Six hundred and nine participants in the secondary checkup were diagnosed whether NDD or not according to DSM-5 criteria. The data include 1421 TDs (boys: girls = 726:695), 118 ASDs (83:35) and 125 ADHDs (79:46), and 136 other DDs (91,45). Caregivers of 5-year-old children completed CSHQ. We compared with CSHQ total and subscale scores in four groups using Kruskal–Wallis’s test and analyzed the relation between z-score of CSHQ total and subscale and diagnosis using a logistic multiple regression analysis (p < 0.05).

Children’s Sleep Habits Questionnaire (CSHQ) consists of nine subscales: Bedtime Resistance, Sleep Onset Delay, Sleep Duration, Sleep Anxiety, Night Waking, Parasomnias, Sleep Disordered Breathing, Daytime Sleepiness, and Sleep/wake patterns [30].

Percentage of children suspected of having sleep problems (CSHQ co > 41) was 80% in TD, 89.0% in ASD, 90.4% in ADHD, and 83.8% in Other DD, respectively (see Figure 4). ASD and ADHD children have significantly higher scores of Total score, Sleep Duration, Sleep Anxiety, Parasomnias, Sleep-Disordered Breathing, and Daytime Sleepiness than TD (see Figure 5). When the CSHQ total score z-score increases by 1 (1SD), the probability of being diagnosed with ASD increases by 1.45 odds.

Figure 4.

Comparison of the probability of sleep disorders by diagnosis in 1800 preschoolers.

Figure 5.

Comparison of CSHQ subscales.

This study showed Japanese preschoolers have high percentage of sleep problems. In addition to comorbid rate of those in ASD were so high that we can predict ASD diagnosis from CSHQ. We must pay attention more that many children have sleep problems, and it would occur some health problems in the future.

2.5 Verification of new screening tools for neurodevelopmental disorders in 5-year-old children

We analyzed multiple questionnaires completed by the parents and teachers of 954 5-year-old primary screening participants in 2013 and the DSM-5 diagnoses of 156 individuals who participated in the secondary examination, then we invented an algorithm to extract the NDD risk group. Children were considered “screen positive” if one of the following criteria (a)–(d) was met:

1. Parent assessment ASSQ score of 19 or higher.

2. PSI scores >75 percentile.

3. Parent-rated ASSQ scores between 9 and 19 and parent-assessed ADHD-RS total scores or at least one of subscale scores were above-defined cutoff.

4. One or more among parent-assessed ASSQ , ADHD-RS, and DCDQ scores were above the cutoff and the teacher-assessed SDQ total or one SDQ subscale scores were above the cutoff.

We validated the algorithm on 965 people in 2014, 1004 people in 2015, 1031 people in 2016, 967 people in 2017, and 1040 people in 2018. The confirmation method is as follows.

First, we compared the rate of high-risk group extraction for her NDD by year under the old and new algorithms (see Figure 6). Table 3 shows old algorithm and new algorithm.

Figure 6.

Comparison the proportion of NDDs high-risk children with old and new algorithm.

Old algorithm	Children who scored at least one of the PSDQ , PASSQ , PADHD-RS, PDCDQ , and K6 cutoffs were considered risk children.
New algorithm	Children who met one or more of the following four criteria were considered risk children.

Table 3.

Old algorithm and new algorithm.

Next, we compared between the proportion of children with special needs who diagnosed with NDD, those who were below the diagnostic criteria but required observation and had no problems by year (see Figure 7). Finally, we calculated the sensitivity and specificity of the new algorithm for extracting children with special needs and observation needs, which were 0.89 and 0.99, respectively.

Figure 7.

Changes in the rate of children who need supports and observation, or no needs.

The new algorithm, created by combining multiple tests, has made it possible to extract risk children more efficiently than the old algorithm. This algorithm was put into a web system in 2019, and from 2020 it was used in developmental health checkups in Hirosaki City [31]. Due to the spread of COVID-19, it was difficult to hold face-to-face health checkups, but with the introduction of the web system, the participation rate of 5-year-old developmental health checkups increased from 85–94%, and the secondary health checkup rate also increased. As a result, more children with NDD are receiving support for adjusting to primary school than before.

3. Conclusions

Epidemiological research in medium-sized cities in northern Japan will soon reach its 10th anniversary. Epidemiological studies are carried out as observational studies over time, repeating the same methods as much as possible. In Japan, where early detection is delayed, intervention support is provided after investigation and diagnosis. It had to be used for individual profit. Although this study reported an analysis of a 4-year survey, a 10-year trajectory and prognostic study of NDD diagnosis in 5-year-olds will begin in the future. We hope that our research will continue and help build a comfortable society for NDD.

Acknowledgments

We would like to express our heartfelt gratitude to all children, their families, teachers, and Hirosaki City staff who participated in HFC study and the Hirosaki University faculty members and researchers (Hirota T, Sakamoto Y, Mikami T, Terui A, Osato A, Koeda S, Mikami M, Kuribayashi M, and Nakamura K) who were involved in the examination, diagnosis, and statistical analysis of health checkups.

This study was supported by the following grants: Grants-in-Aid for Scientific Research (KAKENHI: grant numbers JP16K10239 and JP19K08062), Hirosaki City Commissioned Research Fund, Hirosaki University Institutional Research Grant for Future Innovation, Industry-Academia Joint Development Research Fund, and Hirosaki Institute of Neuroscience in Japan.

Conflict of interest

The authors declare no conflict of interest.