M-CHAT-R/F Documentation: Score Interpretation & Sample Note

Not yet: it is the band that requires the second stage, and treating it as positive is the instrument's most damaging misuse. In the 2014 validation of 16,071 children, applying the questionnaire threshold alone produced 700 false positives, against 116 after the completed two-stage algorithm, and the structured Follow-Up resolved 63.2 percent of first-stage positives to screen-negative while raising the positive predictive value for autism from 13.8 to 47.5 percent. The defensible chart entry for an unresolved medium score is "initial score in medium range; M-CHAT-R/F incomplete pending Follow-Up," with the interview scheduled; when the EHR's screening module has no Follow-Up field, say that too, because an un-auditable second stage is a workflow defect worth naming. A medium score becomes a result only when the Follow-Up says so.

Any appropriately trained member of the team: the authors state the instrument requires no special certification, and the Follow-Up can be conducted by office staff, an early-intervention agency, a specialist, or a licensed electronic implementation. It is a structured branching interview, one flowchart per item, administered only for the items flagged on the questionnaire; the interviewer walks through standardized probes and concrete examples until each item resolves to pass or fail, and improvised general conversation does not count as the second stage. Practical honesty about time: the authors state the questionnaire scores in under two minutes but publish no minutes figure for the Follow-Up itself; the 5-to-10-minute estimates in circulation come from secondary sources. Chart who administered it, in what language, and with what interpreter support, because those facts govern the result's interpretability.

Two referral tracks at once, plus hearing, with the chart showing dates and clocks. A positive screen (after Follow-Up, or on a high initial score) triggers referral for a comprehensive diagnostic evaluation and, simultaneously, referral to early intervention or developmental services, because intervention eligibility does not require a diagnosis and diagnostic waitlists commonly run months to a year. In the US, Part C law puts a seven-day clock on referral after identification and a 45-day clock on the program's evaluation and initial IFSP after referral, and parents may self-refer; audiology joins the chain whenever communication concerns exist, since hearing loss is the reversible mimic. The interim plan belongs in writing: developmental services on eligibility, speech-language evaluation, parent coaching, waitlist confirmation, and a dated primary care follow-up. What a positive screen never does is diagnose: roughly half of final positives are ultimately diagnosed with autism, and about nine in ten have some developmental concern worth evaluating either way.

No, and the chart should behave accordingly. Validation sensitivity around 0.85 fell to 33 to 39 percent in large real-world networks, misses concentrate at the 18-month screen and among girls, and longitudinal data show some children screen negative at 18 months and positive at 24, which is why the authors and the AAP direct rescreening at 24 months after any earlier screen and why surveillance continues at every visit regardless of results. A negative result with a live parent or clinician concern is not a closed question: current guidance says concern warrants evaluation whatever the score. Defensible chart language after an early negative: screen negative after Follow-Up, rescreen at the 24-month visit scheduled, parent advised to return sooner for loss of skills or new social-communication concerns.

Yes to both, with the nuance stated. The validation base is principally ages 16 to 30 months, and the authors' current guidance permits use through 48 months based on subsequent studies, while cautioning that older children with clear concerns may be better served by direct developmental or diagnostic evaluation than by another screening pass. Verbal ability changes neither rule: speaking in sentences does not invalidate a positive screen, autism occurs across the language spectrum, and a positive result at 30 months in a talkative child still warrants the two-track referral, interpreted within the full developmental picture. The defensible phrasing for the page and the chart alike: validated principally for 16 to 30 months, usable to 48 under current author guidance, with evidence less uniform outside the original window and concern always sufficient grounds for referral.

No, and the difference is worth charting. The American Academy of Pediatrics recommends universal autism-specific screening at 18 and 24 months (reaffirmed October 2025), while the US Preventive Services Task Force's standing position is that evidence is insufficient for universal screening of children without raised concerns, an assessment currently under active update, and US family medicine's national body sides with the task force. The Canadian Paediatric Society recommends surveillance for all children and reserves standardized autism screening for identified risk or concern, so a Canadian administration should record its indication. Australia's national autism guideline governs assessment after concern arises rather than mandating screening, and Victoria and Tasmania run clinician-observation surveillance (SACS-R) through child health nurses instead of parent-report screening. The practical rule: state the frame your screen sits in, universal schedule or concern-triggered, and let the referral chain that follows a positive result look the same everywhere.

Internal EHR yes, website no, and the boundary is the authors' license. The instrument is free for clinical, research, and educational use provided it is unmodified (title, items, order, instructions, response format intact) and carries the copyright line, and the authors expressly permit a practice to build the unmodified questionnaire into its own internal EHR at no charge. Distribution is the licensed regime: an EHR vendor, assessment platform, telehealth product, app, or any website presentation of the items, free or paid, public or password-protected, requires a licensing agreement through the authors' current inquiry process, so a practice whose vendor displays the items should confirm the vendor holds the license. Authorized translations (more than 65 languages) must include the Follow-Up, and stale reposted PDFs with legacy wording should be replaced with the current form from the authors' site. What every implementation must preserve is the two-stage recordability: a build that stores only the questionnaire cannot produce an auditable screen.

Yes. Give it the facts (age, respondent and language, first-stage total and band, Follow-Up status with its score and administrator, observations, referrals and dates) and it drafts the full entry: both stages separated, the final classification named, screen-level counseling language, the two-track referral chain with the Part C clocks, and the rescreen or interim plan, ready for your review. It can also cross-check a finished note for a medium score standing as a final result, a missing Follow-Up score, legacy scoring language, or a negative screen with no rescreen plan, and it can draft the informed-refusal entry that keeps the classification intact. BastionGPT is HIPAA-compliant with a signed BAA on every plan, and your data is never used to train models.