Abstract
Screening for autism spectrum disorder is the first step toward early detection and diagnosis, thereby impacting the likelihood of children accessing early intervention and, importantly, improving long-term outcomes. This systematic review aimed to (a) establish a clear baseline of autism spectrum disorder screening tools currently used throughout mainland China and surrounding regions, (b) identify the strengths and limitations of these instruments, and (c) develop specific recommendations regarding screening for autism spectrum disorder throughout Chinese-speaking countries. Databases were searched for recent (2015–2018) articles published in Chinese or English languages. Twenty-two studies (13 Chinese, 9 English) met inclusion criteria; two from Taiwan and the remainder from mainland China. Studies varied greatly in the extent of psychometric analyses and reported autism spectrum disorder prevalence. The majority of diagnoses were based on Diagnostic and Statistical Manual of Mental Disorders (4th ed. (DSM-IV) or 5th ed. (DSM-5)) criteria, although a small number of studies utilized gold-standard diagnostic assessment instruments. It is recommended that a systematic, multi-tiered, screening network be established to improve the identification of autism spectrum disorder in China and surrounding regions. Assessment and diagnosis need to be culturally appropriate, and amenable to low-resource settings. In addition, increased public awareness programs to reduce stigma will be important in improving outcomes for children with autism spectrum disorder.
Autism spectrum disorder (ASD) is a neurodevelopmental condition involving impairments in social interaction and communication skills, and the presence of restricted and repetitive behaviors and interests (American Psychiatric Association, 2013). The first published report of ASD in China was by Tao Kuo-Tai in 1987; since then, estimates on the number of children diagnosed with ASD in China have been increasing. However, due primarily to the lack of a national monitoring system, the true prevalence of ASD in China is unknown. To date, there has been no large-scale epidemiological survey to address this issue (N. Li, Chen, Song, Du, & Zheng, 2011). Several regional studies reported relatively low prevalence rates of ASD in toddlers; for example, only 1.1 to 2.3 cases per 1000 people in Tianjin, Jiangsu, Beijing and Harbin (W. H. Wang et al., 2003; Yu et al., 2010; Zhang, Ji, Li, & Sun, 2004), with a recent meta-analysis showing a pooled prevalence of ASD in China of 39.2 per 10, 000 (F. Wang et al., 2018). A more recent study (Sun et al., 2019) conducted in three Chinese cities found similar prevalence rates to Western countries (i.e. about 1%). This latter study included children from mainstream schools and used internationally agreed-upon screening and diagnostic methods.
Screening for ASD is a crucial first step to early diagnosis. Early detection and diagnosis are both critical for accessing early intervention, which is widely considered critical for improving long-term outcomes (Nah, Young, & Brewer, 2014). In high-resource countries, it is not unusual for primary health care professionals and physicians to conduct routine identification and screening for ASD (Barbaro & Dissanayake, 2010; Carbone, Norlin, & Young, 2016). Where children at increased risk of ASD are identified, formal developmental assessments leading to diagnosis then follow (Steiner, Goldsmith, Snow, & Chawarska, 2012). Despite the lack of formal screening networks in China, early ASD detection has progressed in recent years, with several translated Western instruments being used in different cities and regions. However, there are reported differences in the psychometric properties of these instruments such as internal consistency, test–retest reliability, sensitivity, and specificity (Stewart & Lee, 2017). Furthermore, discrepancies may arise from the lack of established formal training and processes, or lack of knowledge of the early signs of ASD among primary health care professionals. In addition, currently well-recognized diagnostic instruments such as the Autism Diagnostic Interview–Revised (ADI-R; Rutter, Le Couteur, & Lord, 2003) and Autism Diagnostic Observation Schedule, second edition (ADOS-2; Lord et al., 2012) are time-consuming, costly, and require extensive training (Hedley, Young, Juarez Gallegos, & Marcin Salazar, 2010; McEwen et al., 2016). Thus, there is limited use of these instruments in clinical settings throughout China, particularly in rural settings (Stewart & Lee, 2017). Clinical diagnosis is therefore reliant on criteria from the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders(5th ed.; DSM-5; American Psychiatric Association, 2013) and a range of other ad hoc developmental assessments that are not specifically designed to screen for ASD; for example, the Gesell Developmental Schedules (Ball, 1977), or the Sign-Significant Relations, which is a language-based developmental instrument (Jin, Yang, Liu, Huang, & Jin, 2018).
Despite early signs of ASD being present prior to the second year of life, formal identification and diagnosis is often not confirmed until much later, even in Western and other developed countries (Filipek et al., 2000; Planche, Lazartigues, & Lemoniier, 2004). For example, studies from Australia and the United States found that although signs of ASD are present from the child’s first birthday, with many parents having concerns about their child’s development at this age (Barbaro & Dissanayake, 2013; Clifford, Young, & Williamson, 2007; Osterling & Dawson, 1994). However, diagnosis is often delayed, with one Australian study finding an average age of diagnosis from 4 to 6 years (Bent, Dissanayake, & Barbaro, 2015). Delays in diagnosis can occur for various reason, such as variations in levels of service in lower resourced and rural areas, variable presentation of ASD symptomatology, and continuation of pediatric care (Mandell, Novak, & Zubritsky, 2005). A delay in professionals’ response to parents’ early concerns may also contribute to the gap between first concern and age at diagnosis (Zuckerman, Lindly, & Sinche, 2015). Evidence from China of a delay from parental concern to diagnosis was limited to one paper conducted in the Heilongjiang Province (J. Wang et al., 2010), and while it showed a shorter delay (19 months) than in Western Countries, a delay in diagnosis was still evident, with reported average age of diagnosis around 4 years. Thus, compared with Western samples, Chinese parents in the region studied may not become concerned about their child’s development until children are older. This bolsters the argument for the development of universal screening for ASD in Chinese-speaking countries in order to lower the age of diagnosis.
A delay in diagnosis as well as many studies reporting an overall lower prevalence rate compared with Western countries (W. H. Wang et al., 2003; Yu et al., 2010; Zhang et al., 2004) could suggest possible failure to accurately detect ASD, or other potential barriers to diagnosis. Barriers to receiving a diagnosis in China may result from widespread lack of public awareness concerning ASD and other disabilities, and also cultural influences (P. Wang, Michaels, & Day, 2011). For example, it has been suggested that parents may keep their concerns from friends and communities and avoid seeking professional help, which is possibly related to the legacy of the one child policy, or shame and guilt associated with having a child with a disability (J. Wang et al., 2013; P. Wang et al., 2011). Indeed, it is reportedly not uncommon for children with ASD or other disabilities to be locked inside the family home, rarely, if ever, leaving the residence (P. Wang et al., 2011). Furthermore, due to the lack of a national health system and low economic development in many regions, parents often find navigating the public health system to be complex. This includes receiving information concerning ASD, navigating pathways to assessment and diagnosis, and ultimately accessing appropriate treatment (Sun et al., 2013). Finally, in regional and low-resource settings characterized by poor economic development, families may be unable to access public health services due to inadequate numbers of primary care physicians and lack of a comprehensive and strong primary health care system (Sun et al., 2013). Thus, when compared with Western countries, there are many differences regarding the social, economic, and cultural context of China that may impede early detection and diagnosis of ASD.
As an inroad into improving services to children with ASD and their families, it is necessary to gain an understanding of the current state concerning screening and assessment practices and to gain understanding of any cultural considerations or adaptations that have been made or may be required. To the authors’ knowledge, there are currently no systematic reviews of studies examining ASD screening practices in China or surrounding regions. Therefore, the purpose of this systematic review was to (a) identify and report the assessment tools that are being employed in China at the population and clinical levels, (b) examine the psychometric properties of these tools, (c) identify how screening tools are being changed or adapted to address cultural differences in China and its surrounding regions, and (d) describe the implications of these findings for further research and practice.
Method
A systematic review of recently published literature was conducted adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Liberati et al., 2009).
Search strategy
Web of Science, Medline, PsycINFO, ERIC, Scopus, the Chinese National Knowledge Infrastructure database (CNKI), the Chongqing VIP database for Chinese Technical Periodicals, the Wanfang Data and the Chinese Biological Medical Literature databases were searched for published articles from January 2015 to January 2018. Key journals were also monitored after this date, until submission of the manuscript. This led to the identification of two additional studies (C. Li et al., 2018; Sun et al., 2019). Keywords were identified based on our research aims (see Table 1). Studies published in both Chinese and English languages were included in the review.
Search terms by domain.
Terms from each domain were connected with “OR” and between domains with “AND.”
Inclusion and exclusion criteria
To be included in the review, study populations and regions had to be located in China or surrounding regions (i.e. Taiwan, Macau, and Hong Kong), and studies were required to report the use of an ASD screening tool to identify possible ASD cases in the general population, community, or clinical settings. The age of the study population was required to be under 18 years; thus, studies involving adults were excluded. Non-human studies, unpublished reports, reviews, and case reports were also excluded. Identified tools were required to be widely used and with an established evidence base. Self-compiling scales were excluded. Refer to Figure 1 for an overview of the review process.
Search results at each stage of the process.
Definitions
Level 1 screening tool—used at population levels to identify children at risk of ASD
Level 2 screening tool—used to discriminate children at high risk of ASD from typically developing children or children with other developmental disorders
Community sample—refers to participants recruited from the community (e.g. schools)
Clinical sample—refers to participants referred for assessment from clinical settings (e.g. hospitals)
Results
Twenty-two studies met inclusion criteria. Nine were published in English and 13 were in Chinese languages. Study details are summarized in Table 2. Two study samples originated in Taiwan and 20 samples came from mainland China. Participant age ranged from 4 months to 18 years. Eleven different screening tools were identified across the 22 reviewed studies.
Search results.
NR: not reported; ASD: autism spectrum disorder; CM-CHAT: Modified Checklist for Autism in Toddlers; FYI: the First Year Inventory; SCQ: Social Communication Questionnaire; CSBS: infant Communication and Symbolic Behavior Scale Developmental Profile; CABS: Clancy Autism Behavior Scale; SRS: Social Responsiveness Scale; CAST: Childhood Autism Spectrum Test; ABC: Autism Behavior Checklist; ASRS: the Autism Spectrum Rating Scale; T-STAT: Taiwan version of the Screening Tool for Autism in Two-Year-Olds; PPV: positive predictive value; NPV: negative predictive value. ADOS: The Autism Diagnostic Observation Schedule; ADI-R: The Autism Diagnostic Interview–Revised; DSM-IV: Diagnostic and Statistical Manual of Mental Disorders (4th ed.); DSM-5: Diagnostic and Statistical Manual of Mental Disorders (5th ed.); TD: typically developing; DD: developmental disability; IDD: intellectual developmental disorder; AD: autistic disorder; PDD-NOS: pervasive developmental disorder not otherwise specified; SLI: specific language impairment.
Study inclusion required diagnosis to be based on comprehensive diagnostic evaluations.
Data were collected in 2012–2014, researchers re-diagnosed some of the participants using the DSM-5 criteria, and no psychometrics were provided for this subsample of the Study 2 participants.
Instrument type
Ten studies (45.5%) utilized a Level 1, population-based screening tool (Bu, Chen, Cen, Xie, & Zou, 2017; Deng, Chen, Gu, He, & Gao, 2017; J. Huang, Ma, Chen, Xu, & Chen, 2017; Jiang et al., 2015; Jin et al., 2018; Q. Li, 2015; W. Li et al., 2016; F. Wang, Zhou, & Xu, 2017; Xiao & Wang, 2018; Yang et al., 2015), and 10 studies (45.5%) utilized a Level 2 screening tool, to discriminate children at high risk of ASD from typically developing children (29%) or children with other developmental disorders (19%) (Cen et al., 2017; Gong et al., 2015; Guan, Gong, Chen, Zhou, & Li, 2016; C. Huang, Zhu, & Zang, 2016; M. Huang et al., 2016; Liu & Xu, 2015; Sun et al., 2015; Wong et al., 2018; Wu, Chiang, Hou, Chu, & Liu, 2018; Zhou et al., 2015). One study (4.5%) (C. Li et al., 2018) used the CHAT-23 as both a Level 1 screening tool (CHAT-23-A, consisting of the parent report section of the M-CHAT) and Level 2 screening tool (CHAT-23-B, consisting of the observational section of the CHAT). The Level 1 screening tool was applied at 18- to 24-month well-child visits, with children who screened positive referred for Level 2 screening on the same day. Positive cases for ASD were referred for diagnostic assessments which utilized the ADOS-2 and DSM-5. These assessments typically occurred within a week. One study (4.5%) (Sun et al., 2019) used the CAST-Level 2 screening tool in general and clinical populations.
Sample characteristics
A total of 194,468 children aged 4 months to 18 years were included in the reviewed studies. Children were either from low-risk, community-based cohorts (12 studies; n = 188,881 children, aged 4 months–12 years) (Bu et al., 2017; Deng et al., 2017; J. Huang et al., 2017; Jiang et al., 2015; Jin et al., 2018; C. Li et al., 2018; Q. Li, 2015; W. Li et al., 2016; Sun et al., 2019; F. Wang et al., 2017; Xiao & Wang, 2018; Yang et al., 2015) or were from clinically referred samples (8 studies; n = 3921 children, aged 1–18 years) (Cen et al., 2017; Gong et al., 2015; Guan et al., 2016; C. Huang et al., 2016; M. Huang et al., 2016; Wong et al., 2018; Wu et al., 2018; Zhou et al., 2015). Two further studies were drawn from mixed community-based and clinical settings (n = 1666 children, aged 4–10 years) (Liu & Xu, 2015; Sun et al., 2015). None of the studies reviewed involved “high risk” sibling cohorts. Of the studies using low-risk, community-based populations, three studies used screening tools within a local community clinic (J. Huang et al., 2017; W. Li et al., 2016; Xiao & Wang, 2018), and seven used screening tools within local Maternity and Child Health Care Hospitals (Bu et al., 2017; Deng et al., 2017; Jiang et al., 2015; Jin et al., 2018; Q. Li, 2015; F. Wang et al., 2017; Yang et al., 2015). Three studies reported children who had been first identified at risk of ASD in a local Maternity and Child Health Care Hospital setting who were then transferred to a specialized clinical for a final diagnosis, which was completed by a pediatrician (C. Li et al., 2018; Liu & Xu, 2015; Sun et al., 2015). Two studies validated screening tools (CAST and ASRS; Sun et al., 2015; Zhou et al., 2015) in local autism rehabilitation centers, special schools, and primary schools, using clinical samples with existing diagnoses.
Eight studies (36%) screened children under the age of 2 years (Bu et al., 2017; J. Huang et al., 2017; M. Huang et al., 2016; C. Li et al., 2018; Q. Li, 2015; W. Li et al., 2016; F. Wang et al., 2017; Xiao & Wang, 2018), nine studies (41%) screened kindergarten-aged children (3–6 years) (Deng et al., 2017; Gong et al., 2015; Guan et al., 2016; C. Huang et al., 2016; Jiang et al., 2015; Liu & Xu, 2015; Wong et al., 2018; Wu et al., 2018; Yang et al., 2015), two studies (9%) utilized both kindergarten and primary school-aged children (3–12 years) (Cen et al., 2017; Jin et al., 2018), and three studies (13.5%) screened children of primary school age or older (6–18 years) (Sun et al., 2015; Sun et al., 2019; Zhou et al., 2015).
Quality and type of diagnostic assessment
There was much variability among the studies in both the type and quality of the diagnostic assessments completed to confirm children’s diagnostic status following a positive ASD screening result. Among the 22 studies, 7 (32%) studies confirmed ASD based on DSM-IV diagnostic criteria (Cen et al., 2017; Gong et al., 2015; M. Huang et al., 2016; Q. Li, 2015; Liu & Xu, 2015; Yang et al., 2015; Zhou et al., 2015) and 13 (59%) were based on DSM-5 diagnostic criteria (Bu et al., 2017; Deng et al., 2017; Guan et al., 2016; C. Huang et al., 2016; J. Huang et al., 2017; Jiang et al., 2015; Jin et al., 2018; C. Li et al., 2018; W. Li et al., 2016; Sun et al., 2015; F. Wang et al., 2017; Wong et al., 2018; Xiao & Wang, 2018). Two studies reported diagnosing ASD based on both the DSM-IV and DSM-5 criteria (Sun et al., 2019; Wu et al., 2018).
Seven studies (32%) reported using gold-standard assessment tools (i.e. ADOS and ADI-R), with three studies (13%) using the ADOS only (C. Li et al., 2018; Wong et al., 2018; Wu et al., 2018) and two studies (9%) using the ADI-R only (Jiang et al., 2015; Zhou et al., 2015). Only two studies (9%) used both the ADOS and ADI-R (Sun et al., 2015; Sun et al., 2019), which is the most robust method for confirming children’s diagnoses among research samples, particularly in younger children (<3 years) (Barbaro & Dissanayake, 2009).
Psychometric properties
Thirteen studies failed to report psychometric data (e.g. sensitivity, specificity, positive and negative predictive values: PPV; NPV); thus, the efficacy of the instruments used in identifying children with ASD could not be determined for these studies. The remaining nine studies (41%) provided sufficient psychometric data to evaluate the instruments used (Cen et al., 2017; Gong et al., 2015; Guan et al., 2016; C. Li et al., 2018; Liu & Xu, 2015; Sun et al., 2015; Wong et al., 2018; Wu et al., 2018; Zhou et al., 2015). The nine studies providing sufficient data utilized high-risk samples and reported sensitivity estimates ranging from 0.71 to 0.95, and specificity estimates ranging from 0.77 to 0.97, in discriminating children with autistic disorder (AD) and other developmental conditions (developmental delay, attention deficit hyperactivity disorder (ADHD)), and those with broader ASD classifications (e.g. pervasive developmental disorder not otherwise specified (PDD-NOS)) (Cen et al., 2017; Gong et al., 2015; Guan et al., 2016; Liu & Xu, 2015; Sun et al., 2015; Wong et al., 2018; Wu et al., 2018; Zhou et al., 2015).
In one study, C. Li et al. (2018) applied a two-stage screening process with the Chinese CHAT-23, which was followed up with a diagnostic assessment using the ADOS-2 and DSM-5 criteria. Using this method, 437 of 17293 (2.53%) children screened positive for ASD during the first phase (i.e. Level 1 screening) and 110 of 364 (30.22%) children who competed the second phase (i.e. Level 2) were found to screen positive. Overall, PPV for the full model was reported as 11% and 46% for Levels 1 and 2, respectively.
Prevalence
Four studies used comprehensive diagnostic evaluations to establish diagnosis and reported prevalence rates ranging from 1 in 1075 to 1 in 84 (0.09%–1.19%) (Jiang et al., 2015; C. Li et al., 2018; Sun et al., 2015; Sun et al., 2019).
Cultural adaptations and changes
Six studies made cultural adaptations and changes to the screening tool. In one study (Gong et al., 2015), five items of the Chinese revision of the M-CHAT (CM-CHAT) were deleted, although the content and scoring method of the other screening items remained unaltered. However, the cutoff score differed to that of the original scale. One study reported that items of the original cutoff scores on the CM-CHAT were not successful at discriminating children with ASD and DD; they therefore used different items for the cutoff scores, which resulted in higher sensitivity and specificity (Wong et al., 2018). Four studies reported that modifications to the content or cutoff scores were made to screening tools for cultural reasons (Cen et al., 2017; W. Li et al., 2016; Wu et al., 2018; Zhou et al., 2015). Notably, one of these studies (W. Li et al., 2016) failed to report psychometric properties of the modified instrument.
Cultural and administration issues
Among the reviewed studies, all reported that parents or caregivers had difficulty understanding some of the items on the screening tools. While this may be explained in part by educational level, it was also reported that social stigma, parenting styles, and cultural differences could have affected the interpretation of some of the screening tool items. Furthermore, screening instruments were completed by a range of informants including parents and grandparents; in particular, it was found among many studies that grandparents refused to accept that there may be something atypical or different about their grandchild’s development, had little knowledge of child development more generally, or held “old” beliefs about development and disability, with this being prevalent in studies conducted in lower socioeconomic areas (Wong et al., 2018). Consequently, grandparents were reported to often provide incorrect information to health care professionals when filling out the screening tools (Guan et al., 2016; J. Huang et al., 2017; M. Huang et al., 2016; Q. Li, 2015; W. Li et al., 2016; Wong et al., 2018; Xiao & Wang, 2018; Yang et al., 2015). Furthermore, there was a common theme of a lack of, or limited information, awareness and service for parents and caregivers regarding ASD and other developmental disabilities. For health care professionals, it was reported that there were limited opportunities for professional training on ASD, leading to failure to recognize symptoms early, delaying diagnosis and consequently treatment, particularly in rural and remote regions (J. Huang et al., 2017; W. Li et al., 2016; Wong et al., 2018; Xiao & Wang, 2018).
Discussion
The current systematic review identified a growing awareness of ASD, and increasing use of screening instruments, among health professionals throughout China and surrounding regions. Encouragingly, over half of the studies focused on the detection of ASD in younger children (<2 years), suggesting growing awareness that symptoms of ASD present early, and thus can be detected early, which is critical for facilitating early intervention. Specifically, awareness of the presence and prevalence of the disorder in young children is an important first step to improving access to services (Nah et al., 2014). The following sections expand on the main findings from the systematic review.
Sample characteristics
Given the limited amount of literature concerning ASD screening in China, it is encouraging to be able to report that around half (55%) of the studies were conducted in community-based samples. The results from community-based studies may be more generalizable to the broad population than “high risk” sibling or clinical sampling studies (Barbaro & Dissanayake, 2010, 2017; Zwaigenbaum et al., 2007). That is, community-based studies offer the opportunity to assess and identify cases in the general population, rather than selective cases referred to a clinical setting due to pre-existing developmental concerns (Stewart & Lee, 2017). Furthermore, it is suggested that community-based studies are well suited to low-resource environments, particularly rural settings, which is characteristic of much of mainland China (Zhang & Ji, 2005).
Prevalence rates
Four studies using comprehensive diagnostic evaluations reported prevalence rates, with a wide range in reported prevalence estimates. Possible reasons for this wide variance in prevalence estimates include (1) differences in study design and methodology, including reliance on parent, teacher, or caregiver completed questionnaires compared with formal clinical assessments; (2) sampling, for example, diagnoses that were based in community hospitals versus specialized clinical settings, and studies from different regions; and (3) cultural differences across regions. Given that many of the reviewed studies did not include comprehensive diagnostic assessments, future studies that incorporate universal Level 1 screening and comprehensive diagnoses are required in order to effectively calculate prevalence rates in Chinese-speaking countries.
Psychometrics
Overall, the screening instruments used in the various studies were reported to show fair to good sensitivity (0.71 to 0.95) and specificity (0.77 to 0.97) across low-risk (i.e. general population) and high-risk (i.e. clinically referred) samples. Two studies reported different psychometric properties using the same tool (CM-CHAT), one showed higher sensitivity (0.92) and specificity (0.83) in a clinical and general population control sample from a higher socioeconomic setting (Beijing; Gong et al., 2015) and another showed lower sensitivity (0.70) and similar specificity (0.82) in high-risk samples in a lower socioeconomic setting (Chia-Yi; Wong et al., 2018). Furthermore, three studies reported higher sensitivity (0.84 to 0.95) and specificity (0.90 to 0.96) across low- and high-risk, school age samples in higher socioeconomic settings (i.e. Beijing, Guangzhou, Shenzhen) (Cen et al., 2017; Guan et al., 2016; Sun et al., 2015). Thus, psychometric properties may also have been affected by factors related to child age and region where the study was conducted (Stewart & Lee, 2017; Wong et al., 2018). While the psychometric properties for the instruments reported above are encouraging and support their use within the Chinese population, less than half of the reviewed studies reported sufficient data to adequately evaluate the instruments; thus, these findings are limited.
Quality of the diagnostic assessment
Only six studies (28%) used either the ADOS or ADI-R to confirm diagnosis (Jiang et al., 2015; C. Li et al., 2018; Sun et al., 2015; Wong et al., 2018; Wu et al., 2018; Zhou et al., 2015). Gold standard instruments for assessing ASD such as these are not widely used and are typically only found in larger cities such as Beijing, Shanghai, Guangzhou, and in Taiwan. Furthermore, the ADOS-2 (Lord et al., 2012) has only recently become available in China; hence, only one study included this instrument (C. Li et al., 2018). Thus, diagnostic confirmation was mostly reliant on correspondence to DSM criteria. Furthermore, information concerning the qualifications or training of the assessor or diagnostician were generally absent.
Cultural adaptations and administration
Six (27%) studies reported cultural adaptations or changes to screening instruments. Reported adaptations included relatively minor changes to scale items and cutoff scores (CM-CHAT, FYI, SRS, ASRS, T-STAT), to removal of items from the measure. For example, one study reported that five items were deleted due to poor validity or reliability, including items 1 (swing or jump), 3 (climbing), 11 (sensitive to sound) 16 (walking), and 18 (fingers movement) (Gong et al., 2015). Wong et al. (2018) suggested that the need to adjust items and cutoffs in their population was due to parents under-reporting because of the importance they placed on these behaviors, or to avoid social stigma. They also suggested that parents in poorer provinces have reduced time to observe these behaviors in their children. These changes complicate the use of these instruments more broadly, as standardization is lost due to different versions or scoring algorithms being used. China is also vast in area and varied in culture; even within China, different cultural groups have different behavioral norms, expectations, and understanding or perceptions concerning similar concepts. Recognition of ASD features in one culture may also differ from another and it cannot be assumed that features in Northern China are perceived to be the same in the South. It will therefore be important in future to not only ensure instruments are adapted appropriately to Chinese culture, and standardized versions are developed and distributed, but also to ensure that sufficient consideration has been given to variance between cultural sub-groups.
The review also identified differences when comparing populations from lower and higher socioeconomic settings. First, socioeconomic status is likely to impact access to high quality medical services, and the training that health professionals receive in low resource, rural settings. Indeed, several studies reported administration difficulties within these poorer community settings (Deng et al., 2017; W. Li et al., 2016; Wong et al., 2018). Parents or caregivers from lower socioeconomic backgrounds are likely to have less access to information about ASD and other developmental disabilities. In addition, there are likely to be fewer health care professionals with specific training in identifying and screening children with ASD in settings characterized by lower socioeconomic status (Stewart & Lee, 2017; Wong et al., 2018). As has been identified by others, it will be critical to develop an ASD service system and increase social and governmental support for low socioeconomic communities, including rural regions (Sun et al., 2013).
On a related subject, it is reported that there are millions of “left-behind children” who are cared for by their grandparents in rural and remote areas throughout China, due to their parents seeking employment in larger cities (Zhao et al., 2014). It is not clear whether separation from parents might affect the social development of children with ASD, particularly if siblings are absent. Furthermore, there is a traditional belief among older generations that a child with a language delay or other developmental delay in childhood will be highly intelligent or clever in the future, especially if the child is male (Wong et al., 2018; Zhao et al., 2014). This belief may adversely affect the likelihood of the child’s carer seeking professional advice in the presence of developmental delays or other concerns.
Future directions
Taken together, the results of the systematic review suggest that the assessment tools currently employed in China show reasonable psychometric properties for identifying and diagnosing ASD. However, these results are somewhat limited due to psychometrics not being reported in some studies, and difference in item inclusion and cutoff scores in some studies due to cultural differences. Cultural factors were also seen to play a role in the administration and interpretation of items in measures, both between regions and socioeconomic levels. Future research would benefit from establishing administrative and reporting norms, which take into account culture and socioeconomic position.
The results of our review have implications for future ASD screening studies in China and surrounding regions. First, as we did not identify a consistent approach to the screening and diagnosis of ASD in China, there is a need to develop a multi-tiered child healthcare network from screening to diagnosis that is suitably adapted to the Chinese environment (e.g. Filipek et al., 2000; C. Li et al., 2018). Specific stages of the process could include community-based screening, where psychometrically and culturally validated screening tools are used to screen children in the general population. Further assessment of children identified at “high likelihood” of ASD within the Maternal and Child Health Care Hospital network can then be referred with appropriate follow-up, and confirmation of diagnosis made by a specialized, multidisciplinary clinical team (Filipek et al., 2000; C. Li et al., 2018).
Second, public awareness of ASD can substantially improve identification, diagnosis, and intervention (C. Li et al., 2018; Zwaigenbaum et al., 2015). We identified a need for increased public awareness about ASD, and programs aimed at reducing the stigma associated with ASD and other disabilities. In the Chinese population, social, school, and family support can play a significant role in increasing positive outcomes for children with ASD, but this will most likely require government support to break down existing barriers and perceptions concerning autism and disability. Third, health insurance policy should be used in mainland China for autism, and government needs to provide more financial supports for poor families. Finally, there is a need to increase the availability and training in the use of gold-standard diagnostic instruments within specialized clinical settings.
Supplemental Material
AUT871174_Lay_Abstract – Supplemental material for A systematic review of screening tools for the detection of autism spectrum disorder in mainland China and surrounding regions
Supplemental material, AUT871174_Lay_Abstract for A systematic review of screening tools for the detection of autism spectrum disorder in mainland China and surrounding regions by Ji Wang, Darren Hedley, Simon M Bury and Josephine Barbaro in Autism
Footnotes
Author contributions
JW and DH conceived the study. JW and SMB completed the systematic review. DH, JB, JW, and SMB constructed the results based on the review. All authors contributed equally to the writing of the manuscript and reviewed and approved the final submitted version.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
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 the Key Laboratory of Myocardial Ischemia, Harbin Medical University, Ministry of Education, Heilongjiang Province, China (KF201611) and Natural Science Foundation of Heilongjiang Province of China (H2017004).
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
