Autism prevalence in China is comparable to Western prevalence

Xiang Sun, Carrie Allison, Liping Wei, Fiona E. Matthews, Bonnie Auyeung, Yu Yu Wu, Siân Griffiths, Jie M. Zhang, Simon Baron‐Cohen, Carol BrayneMolecular 2019 · DOI

Abstract

Recent research into autism spectrum conditions (ASC) in China suggests that previous estimates of low prevalence were likely due to methodological limitations. By employing standardized Western diagnostic tools and including mainstream school populations, this study found a prevalence rate in Jilin City of approximately 1.08%, aligning closely with international estimates. Findings from Shenzhen and Jiamusi also support the conclusion that autism is as common in China as it is in the West, though many cases in mainstream schools remain undiagnosed.

Introduction: Addressing the Prevalence Gap

Historically, reported autism prevalence in China has been significantly lower—often cited around 0.4% or less—compared to the 1-1.5% typically reported in Western nations. This discrepancy raised fundamental questions about whether biological or environmental differences existed between populations, or if the "gap" was simply a measurement error.

This study identifies two primary methodological flaws in earlier Chinese research:

  1. Sampling Bias: Most prior studies focused exclusively on "special schools," neglecting the mainstream school population where many high-functioning children are educated.
  2. Diagnostic Lag: Many studies relied on outdated criteria or lacked contemporary, validated screening and assessment tools like the Autism Diagnostic Observation Schedule (ADOS).

Methodology: The Three-Step Funnel

To establish a more accurate baseline, researchers implemented a rigorous three-step "funnel" approach to identify cases among children aged 6–10.

Fig. 2: Flowchart of the screening and diagnostic process in Jilin City, illustrating the three-step funnel approach.
  1. Screening: Parents completed the Childhood Autism Spectrum Test (CAST), a 37-item questionnaire validated for high sensitivity in Chinese populations.
  2. Clinical Assessment: Children scoring above the CAST threshold (≥15), along with a control group of "borderline" and "low-score" children, underwent clinical evaluation by local child psychiatrists using DSM-IV-TR and DSM-5 criteria.
  3. Research Diagnostic Assessment: To ensure "gold-standard" accuracy, children identified in the second step received intensive assessments using the ADOS and the Autism Diagnostic Interview-Revised (ADI-R), conducted by researchers with international reliability certification.

Geographic Scope and Datasets

The study targeted three cities representing diverse geographic and economic profiles across China to capture a broad demographic range.

Fig. 1: Map showing the locations of the three study sites: Jiamusi, Jilin, and Shenzhen.
  • Jilin City: A total population approach (including both mainstream and special schools) involving 7,258 children.
  • Shenzhen City: A mainstream-only sample of 21,420 children.
  • Jiamusi City: A mainstream-only sample of 16,358 children.

In total, over 45,000 children were screened, making this one of the largest and most methodologically consistent studies of its kind in the region.

Key Results and Findings

The findings strongly suggest that when the same diagnostic standards are applied, autism prevalence in China mirrors that of Western countries.

  • Total Prevalence in Jilin: The study identified a prevalence of 108 per 10,000 (1.08%). This is the most significant figure, as it accounts for the entire school-age population, placing China squarely in line with global estimates.
  • Mainstream Prevalence: In Shenzhen, the mainstream-only prevalence was 42.3 per 10,000, while in Jiamusi it was 19.0 per 10,000.
  • The "Hidden" Population: A critical discovery was the high rate of under-diagnosis. In all three cities, the majority of cases identified in mainstream schools had no prior diagnosis. Under the Sui Ban Jiu Du policy, which encourages children with disabilities to attend regular classes, many children with autism are present in general education but lack formal support.
  • Consistency of Phenotype: By using the same tools (ADOS/ADI-R) used in the West, the authors demonstrated that the "autism phenotype" is remarkably consistent across cultures.

Discussion and Future Directions

While the overall prevalence is comparable to the West, the study noted non-significant but visible variation between cities. This suggests that local factors—such as economic development, parental awareness, and the availability of school psychologists—influence how many cases are identified.

Limitations

The authors acknowledge that while the three cities are diverse, they do not represent the entirety of China, particularly rural or impoverished areas. Additionally, participation was voluntary, which may introduce selection bias. The study also occurred during the transition from DSM-IV to DSM-5, which can complicate long-term comparisons.

Conclusion

This research provides a vital baseline for understanding autism in China. It highlights an urgent need to improve support systems for the newly identified "mainstream" autism population and calls for a more robust national picture by expanding this methodology to more diverse regions.

Summary updated

Abstract

Results from Jilin City, where both mainstream and special school data were available, revealed a similar prevalence of autism in China to the West, at around 1%. Results from Shenzhen and Jiamusi cities, where only mainstream data were available, prevalence is also in line with Western estimates. In all three cities, new cases of autism were identified by the study in mainstream schools, reflecting current under-diagnosis. Non-significant variation across different cities is seen indicating the need to explore potential variation of autism across diverse Chinese regions with large sample sizes to achieve a fully robust national picture.
BibTeX
@inproceedings{sun2019autism,
  author = {Xiang Sun and Carrie Allison and Liping Wei and Fiona E. Matthews and Bonnie Auyeung and Yu Yu Wu and Siân Griffiths and Jie M. Zhang and Simon Baron‐Cohen and Carol Brayne},
  title = {{Autism prevalence in China is comparable to Western prevalence}},
  booktitle = {Molecular},
  year = {2019},
  doi = {10.1186/s13229-018-0246-0},
}
Figure