M-CHAT Autism Screening Interpretation
The M-CHAT-R/F uses a two-stage scoring system: an initial score of ≥3 triggers a follow-up interview, and a score of ≥2 after follow-up indicates significant ASD risk requiring immediate referral for diagnostic evaluation and early intervention. 1, 2
Scoring Algorithm and Risk Stratification
The M-CHAT consists of 23 parent-reported yes/no questions that assess communication skills, joint attention, repetitive movements, and pretend play. 1 The interpretation follows a three-level risk algorithm:
Low Risk (0-2 points)
- Continue routine developmental surveillance at all well-child visits 3
- No immediate action required 2
Medium Risk (3-6 points)
- Administer the structured follow-up interview immediately 1, 2
- If ≥2 items remain positive after follow-up: refer immediately for comprehensive diagnostic evaluation by a developmental specialist 2
- If <2 items after follow-up: continue enhanced surveillance 2
High Risk (≥7 points)
- Skip the follow-up interview entirely 2
- Refer immediately for comprehensive diagnostic evaluation and early intervention services 1, 2
- Do not wait for diagnosis to initiate early intervention 2
Critical Discriminatory Items
Six specific questions have the highest discriminatory power between children with and without ASD: 4
- Interest in other children (social relatedness) 2, 4
- Imitation (social relatedness) 4
- Protodeclarative pointing (joint attention) 4
- Gaze monitoring/response to pointing (joint attention) 4
- Bringing objects to show parents 4
- Response to name 4
Failing any 2 of these 6 critical items warrants immediate concern and follow-up. 5, 4
Age-Specific Performance Characteristics
The M-CHAT performs significantly better in older toddlers:
- Ages 16-23 months: PPV = 0.28 (higher false-positive rate) 6, 2
- Ages 24-30 months: PPV = 0.61 (more reliable) 6, 2
- Optimal screening ages: 18 and 24 months per AAP recommendations 2, 3
Despite lower specificity for autism at 18 months, the PPV for any diagnosable developmental disorder remains high (0.98), making early screening valuable even with increased false positives. 6
Statistical Performance with Follow-Up Interview
When the two-stage process (initial screening + follow-up interview) is properly implemented: 7
- Sensitivity: 0.44-0.93 depending on population 2
- Specificity: ~0.99 in general population 2
- PPV for ASD: 0.47-0.65 in low-risk samples 6, 1, 7
- PPV for any developmental disorder: 0.94-0.98 6, 7
Common Pitfalls and Caveats
False positives occur frequently due to: 6, 2
- Overlapping behaviors between typically developing toddlers and early ASD (repetitive behaviors like turning lights on/off, insistence on routines) 6
- Developmental concerns that may resolve spontaneously 6, 2
- Younger screening age (16-23 months) 6, 2
Critical implementation errors to avoid:
- Skipping the follow-up interview for medium-risk scores (3-6 points) - this significantly reduces PPV 1, 7
- Delaying referral while "watching and waiting" for high-risk scores (≥7) 2
- Screening before 18 months, which increases false-negative rates 6
High-Risk Population Considerations
For siblings of children with ASD (14-18% risk): 2, 3
- Screen at minimum at both 18 AND 24 months 2, 3
- Maintain intensified surveillance between and after screening visits 3
- Lower threshold for referral given elevated baseline risk 2
Clinical Action Based on Results
The follow-up interview is not optional for medium-risk scores - it improves PPV from 0.06 (M-CHAT alone) to 0.47-0.65 (M-CHAT-R/F). 7, 8 Primary care pediatricians can reliably administer the follow-up interview during the same visit with 86.6% agreement with autism specialists. 8
All screen-positive children after follow-up warrant referral - 94.6% will have ASD or another developmental delay requiring intervention. 7