Rationale:
Autism spectrum disorder (ASD) is often underdiagnosed in rural Vietnam due to limited awareness, screening, and intervention, particularly in low-income areas like Ca Mau. This study applied a community-based approach to improve early ASD care. Key objectives included:
(1) implementing local screening and intervention;
(2) evaluating M-CHAT-R/F diagnostic accuracy;
(3) estimating ASD prevalence and severity;
(4) assessing PECS-based intervention outcomes; and
(5) identifying perinatal risk factors in rural children.
Methods:
This study employed a two-part design: (1) a cross-sectional descriptive component for community-based ASD screening, diagnosis, and prevalence estimation; and (2) a quasi-experimental component to evaluate early intervention effectiveness. Conducted from January to September 2022, the study involved 3,639 children aged 18–36 months across 117 preschools in Ca Mau province.
Preschool teachers and healthcare workers were trained on ASD signs and screening procedures. Teachers administered the 20-item M-CHAT-R; children scoring 3–20 were referred for follow-up using M-CHAT-R/F by trained medical staff. Those with follow-up scores ≥2 were considered screen-positive and assessed for ASD using DSM-5. Severity was classified with the Childhood Autism Rating Scale (CARS) to estimate prevalence and severity distribution.
For the intervention, 40 children diagnosed with ASD participated in a PECS-based program. CARS scores were measured at baseline, and at 3, 6, and 12 months post-intervention to track progress.
To evaluate the diagnostic accuracy of M-CHAT-R/F, sensitivity and specificity were calculated using DSM-5 as the reference. Logistic regression identified perinatal risk factors associated with ASD, including prolonged labor, medical intervention at birth, and birth asphyxia.
https://drive.google.com/file/d/1KktTZm4uJ2sjcav1MfP9yHrs42Tk1mya/view?usp=sharing
Results:
Initial M-CHAT-R screening identified 6.5% (235/3,639) of children as at-risk. Follow-up M-CHAT-R/F interviews confirmed 203 as screen-positive. ASD prevalence was 2.0% (75/3,639): 36% mild, 41.3% moderate, 22.7% severe. Key perinatal risk factors included medical intervention at birth (OR=2.44), prolonged labor (OR=2.58), and birth asphyxia (OR=7.17). M-CHAT-R/F showed high diagnostic accuracy (sensitivity 98.67%, specificity 95.48%). After 6 months of PECS-based intervention, CARS scores significantly declined in both mild-moderate (35.13→30.93) and severe groups (43.0→35.4), indicating reduced symptom severity (p < 0.01).