When Pediatricians Screen for Autism

The schedule, in one place

American pediatric guidance asks for two distinct things on a schedule. General developmental screening — covering language, motor, problem-solving, and social skills — happens at the 9-month, 18-month, and 30-month visits. Autism-specific screening happens at 18 and 24 months. Underneath both, developmental surveillance runs at every well-child visit from birth 1Ref 1American Academy of Pediatrics (2024).Developmental Surveillance and Screening.The AAP-recommended schedule itself: developmental surveillance at every well-child visit, general developmental screening at 9, 18, and 30 months, and autism-specific screening at 18 and 24 months — including that surveillance continues at every visit and that raised concerns warrant screening outside the dated schedule..

The bottom row is the one that gets forgotten, and it is arguably the most important line in the table. The ages are a floor, not a gate. A concern raised by a parent, a teacher, a grandparent, or the clinician is itself a reason to screen, at whatever age it surfaces 1Ref 1American Academy of Pediatrics (2024).Developmental Surveillance and Screening.The AAP-recommended schedule itself: developmental surveillance at every well-child visit, general developmental screening at 9, 18, and 30 months, and autism-specific screening at 18 and 24 months — including that surveillance continues at every visit and that raised concerns warrant screening outside the dated schedule..

18 months is the only visit that carries both screens. If a family is going to attend exactly one developmental appointment in the second year, that is the one.

Surveillance and screening are two different jobs

These get used interchangeably in conversation and they are not the same activity. Developmental monitoring, or surveillance, is the ongoing, longitudinal process — noticing milestones, asking the parent what's new, tracking whether the trajectory bends. Screening is a formal, validated instrument administered at set ages that produces a result 2Ref 2Centers for Disease Control and Prevention (2024).Developmental Monitoring and Screening.The distinction between ongoing developmental monitoring or surveillance — watching milestones over time — and formal developmental screening with a validated instrument at recommended ages..

Surveillance is a conversation. Screening is a measurement. One runs continuously and depends on the relationship; the other happens on specific dates and depends on a form.

The reason both exist is that each fails in a way the other catches. Surveillance is sensitive to everything and specific to nothing — it depends on what the clinician happened to notice in a room where the child was tired, or shy, or unusually charming. Screening is reproducible but blind: it asks its questions and nothing else, on its dates and no others. A system with only screening misses the child whose difficulty does not sit inside those questions. A system with only surveillance misses the child nobody thought to worry about.

Developmental surveillance is the ongoing watching. Developmental screening is a validated instrument on a schedule. Guidance asks for both, because each one catches what the other drops 2Ref 2Centers for Disease Control and Prevention (2024).Developmental Monitoring and Screening.The distinction between ongoing developmental monitoring or surveillance — watching milestones over time — and formal developmental screening with a validated instrument at recommended ages..

This distinction has a practical edge for parents. "We keep an eye on development at every visit" is a description of surveillance, and it is a true and good thing. It is not an answer to the question was a screen done at 18 months. If a family wants to know whether the measurement happened, that has to be asked as a separate question, because from the waiting room the two are indistinguishable.

Why 18 and 24 months, and not earlier or later?

The ages are a compromise between two competing facts. The behavior that autism screening asks about is not reliably present much before 18 months — ask too early and the instrument returns noise. But autism can be identified as early as 18 months, and the whole benefit of finding it is front-loaded into early childhood 3Ref 3Hyman SL, Levy SE, Myers SM; AAP Council on Children With Disabilities, Section on Developmental and Behavioral Pediatrics (2020).Identification, Evaluation, and Management of Children With Autism Spectrum Disorder.That autism can be diagnosed as early as 18 months; the rationale for standardized screening at 18 and 24 months alongside ongoing surveillance; the primary-care role in identification and referral; and heightened attention for younger siblings of autistic children.. So the schedule lands at the earliest point the signal is readable.

The second screen at 24 months exists because one reading is not a trajectory. Two things happen between 18 and 24 months that a single screen cannot capture:

• Some of the relevant development is only just emerging at 18 months. A child can be genuinely ambiguous then and clear six months later — in either direction.
• A subset of children lose skills they previously had. That pattern is invisible to any single point-in-time screen, and it is only detectable by comparing two.

A parent asking about signs of autism at 18 months is asking about the exact hinge the schedule was built around. Which is also why a negative screen at 18 months is not a clearance and was never meant to be read as one — the guideline builds in a second look precisely because the first is provisional 3Ref 3Hyman SL, Levy SE, Myers SM; AAP Council on Children With Disabilities, Section on Developmental and Behavioral Pediatrics (2020).Identification, Evaluation, and Management of Children With Autism Spectrum Disorder.That autism can be diagnosed as early as 18 months; the rationale for standardized screening at 18 and 24 months alongside ongoing surveillance; the primary-care role in identification and referral; and heightened attention for younger siblings of autistic children..

A screen at 18 months that comes back reassuring is one reading of a moving thing. The 24-month screen is not a re-test because something went wrong. It is the design working as intended.

What actually happens in the room

For most families, autism screening looks like a clipboard. A validated parent questionnaire — in American primary care, most often the M-CHAT-R/F — gets handed over at check-in, filled out in the waiting room, scored by staff, and mentioned in one sentence or not at all 4Ref 4Centers for Disease Control and Prevention (2024).Clinical Screening for Autism Spectrum Disorder.That validated instruments such as the M-CHAT-R/F are used for autism screening in primary care, and that a positive screen is not a diagnosis but an indication for referral to a separate comprehensive diagnostic evaluation.. Many parents complete an autism screen without ever knowing they did.

That is not a scandal, but it does explain a common and disorienting experience: a parent who has been worried for months hands over a form, hears "everything looks great," and leaves without ever having had the conversation they came for.

What the process actually consists of:

• A parent-report questionnaire, not an examination of the child. Nobody tests the toddler. The instrument is asking the person who watches him every day.
• Scoring by the practice, usually against a key that produces a result rather than a description.
• For some instruments and some results, a structured follow-up conversation — a second stage that is part of the design and is sometimes skipped when a clinic is busy.
• A referral decision, if the result warrants one.

The screen is only as good as what goes on the form, and the form is answered in a waiting room, in three minutes, by a parent managing a toddler. Answering it honestly rather than optimistically is the single largest thing within a family's control. The instinct to round up — well, he does point sometimes — is human and it degrades the measurement. If an item is a genuine "sometimes, and I'm not sure," that is worth saying out loud to the clinician rather than resolving privately on the page.

Developmental screening more broadly follows the same shape: a validated form, scored, at set ages. Knowing that it is a form and not an exam is what lets a parent see it happening.

Who gets screened outside the schedule

The calendar describes what happens to a child about whom nobody has said anything. It is a net, not a ceiling. Any raised concern is grounds for screening at any age, and guidance is explicit that surveillance continues at every visit rather than switching off at 30 months 1Ref 1American Academy of Pediatrics (2024).Developmental Surveillance and Screening.The AAP-recommended schedule itself: developmental surveillance at every well-child visit, general developmental screening at 9, 18, and 30 months, and autism-specific screening at 18 and 24 months — including that surveillance continues at every visit and that raised concerns warrant screening outside the dated schedule..

Situations where screening is warranted independent of the schedule:

• A parent, teacher, childcare provider, or relative raises a concern. A parent's concern is a recognized reason to act, not an anxiety to be managed.
• A younger sibling of an autistic child. Familial recurrence is a known consideration in identification, and these children warrant closer watching than the routine net provides 3Ref 3Hyman SL, Levy SE, Myers SM; AAP Council on Children With Disabilities, Section on Developmental and Behavioral Pediatrics (2020).Identification, Evaluation, and Management of Children With Autism Spectrum Disorder.That autism can be diagnosed as early as 18 months; the rationale for standardized screening at 18 and 24 months alongside ongoing surveillance; the primary-care role in identification and referral; and heightened attention for younger siblings of autistic children..
• Any loss of skills the child previously had — words, gestures, social engagement, play — which is a reason to act promptly rather than wait for the next dated visit.
• A child arriving late to the system: adopted, from another country's care system, or simply not seen consistently in the first three years.
• A child flagged on general developmental screening at 9 or 30 months, where the concern is developmental but not specifically autism-shaped.

The last one is worth pausing on. The 9-month and 30-month screens are not autism screens, but they are the two dates on the calendar most likely to catch a child whose autism was going to be missed — the 9-month because it is early, the 30-month because it is the last routine one and the demands have risen by then. A concerning general screen at either age is a live reason to look at autism specifically, even though the calendar does not ask for it 1Ref 1American Academy of Pediatrics (2024).Developmental Surveillance and Screening.The AAP-recommended schedule itself: developmental surveillance at every well-child visit, general developmental screening at 9, 18, and 30 months, and autism-specific screening at 18 and 24 months — including that surveillance continues at every visit and that raised concerns warrant screening outside the dated schedule..

Why your practice may do it differently

Practices vary, and a family encountering that variation deserves to know why rather than concluding their clinician is negligent. Some practices screen at 18 and 24 months exactly. Some screen at every visit in that window. Some screen only when something has been raised. The variation is real and it has an evidentiary basis rather than being carelessness.

The U.S. Preventive Services Task Force concluded in 2016 that the evidence was insufficient to weigh the benefits and harms of universal autism screening in children aged 18 to 30 months for whom no concerns had been raised 6Ref 6U.S. Preventive Services Task Force (2016).Autism Spectrum Disorder in Young Children: Screening — Final Recommendation Statement.The 2016 'I' (insufficient evidence) statement on universal screening of children aged 18-30 months in whom no concerns have been raised — used here to explain practice-to-practice variation, and noting that it is not a recommendation against screening and does not apply to children with signs or raised concerns.. This is an "I" statement — a finding that the necessary studies have not been done. It is not a recommendation against screening, and it explicitly does not apply to a child anyone has raised a concern about 6Ref 6U.S. Preventive Services Task Force (2016).Autism Spectrum Disorder in Young Children: Screening — Final Recommendation Statement.The 2016 'I' (insufficient evidence) statement on universal screening of children aged 18-30 months in whom no concerns have been raised — used here to explain practice-to-practice variation, and noting that it is not a recommendation against screening and does not apply to children with signs or raised concerns.. The American Academy of Pediatrics, weighing the same landscape against a different question, continues to recommend screening at 18 and 24 months 1Ref 1American Academy of Pediatrics (2024).Developmental Surveillance and Screening.The AAP-recommended schedule itself: developmental surveillance at every well-child visit, general developmental screening at 9, 18, and 30 months, and autism-specific screening at 18 and 24 months — including that surveillance continues at every visit and that raised concerns warrant screening outside the dated schedule..

American pediatrics overwhelmingly followed the AAP, and the 18/24 schedule is the standard of practice. But two respected bodies genuinely disagree on the universal-screening question, and a practice that has read both is not doing anything indefensible.

Neither body says screening is a bad idea for a child anyone is worried about. The entire disagreement is about children nobody has noticed anything about — a category your worried child is not in.

Which means the practical answer to "my pediatrician doesn't routinely screen" is not to change practices. It is to stop being in the disputed category: state the concern plainly, and the disagreement no longer applies to your child.

The distance between the schedule and what actually happens

The schedule is good. Its execution is where children get lost. A systematic review and meta-analysis of international studies from 2012 to 2019 found the mean age at autism diagnosis was around 60 months — roughly five years old 5Ref 5van 't Hof M, Tisseur C, van Berckelaer-Onnes I, et al. (2021).Age at autism spectrum disorder diagnosis: A systematic review and meta-analysis from 2012 to 2019.The mean age at autism diagnosis of approximately 60 months across international studies from 2012 to 2019, documenting the persistent gap between the age at which identification is possible and the age at which it actually happens. An international meta-analysis rather than a single-country figure.. The guideline aims at 18 months. The world lands, on average, three and a half years later.

That gap is not caused by one failure. It accumulates across a chain in which every link is individually forgivable:

• The visit gets missed, or moved, or the 24-month one quietly merges with the two-year checkup that ran long.
• The screen is handed out but never scored, or scored but never acted on.
• The result is positive and the referral is placed, and the waitlist is fourteen months.
• The follow-up stage of a two-stage instrument gets skipped, and the referral is never triggered.
• A parent raises something, hears "let's watch it," and does not know that watching it was supposed to have a date attached.

Autism can be identified as early as 18 months 3Ref 3Hyman SL, Levy SE, Myers SM; AAP Council on Children With Disabilities, Section on Developmental and Behavioral Pediatrics (2020).Identification, Evaluation, and Management of Children With Autism Spectrum Disorder.That autism can be diagnosed as early as 18 months; the rationale for standardized screening at 18 and 24 months alongside ongoing surveillance; the primary-care role in identification and referral; and heightened attention for younger siblings of autistic children., but the international mean age at diagnosis is about 5 years 5Ref 5van 't Hof M, Tisseur C, van Berckelaer-Onnes I, et al. (2021).Age at autism spectrum disorder diagnosis: A systematic review and meta-analysis from 2012 to 2019.The mean age at autism diagnosis of approximately 60 months across international studies from 2012 to 2019, documenting the persistent gap between the age at which identification is possible and the age at which it actually happens. An international meta-analysis rather than a single-country figure.. The distance between those numbers is a system problem, not a knowledge problem.

None of this is an argument for panic, and it is not an argument that clinicians are failing. It is an argument for a specific, unglamorous kind of parental attention: knowing which dates carry which screen, asking whether the screen was actually done, and knowing that "let's watch and see" is a legitimate clinical plan only when it comes with a return date. Watchful waiting without a date is not a plan. It is a deferral, and it is where most of those three and a half years live.

What happens after a screen, and what to ask

A screen produces one of two outcomes, and both of them are the beginning of something rather than the end. A result that raises no concern means the instrument found nothing at this reading, and surveillance continues. A positive autism screen is an indication for further evaluation — it is not a diagnosis, and it means the child should be referred for a comprehensive assessment 4Ref 4Centers for Disease Control and Prevention (2024).Clinical Screening for Autism Spectrum Disorder.That validated instruments such as the M-CHAT-R/F are used for autism screening in primary care, and that a positive screen is not a diagnosis but an indication for referral to a separate comprehensive diagnostic evaluation..

Diagnosis is a two-step structure by design: screening, then a separate and much longer diagnostic evaluation conducted by a clinician trained to do it 4Ref 4Centers for Disease Control and Prevention (2024).Clinical Screening for Autism Spectrum Disorder.That validated instruments such as the M-CHAT-R/F are used for autism screening in primary care, and that a positive screen is not a diagnosis but an indication for referral to a separate comprehensive diagnostic evaluation.. Screens do not diagnose. Nothing on a clipboard ever will.

Questions worth asking at any well visit in the 18-to-30-month window:

• Was an autism-specific screen done today, or a general developmental one? They are different, and both are on the schedule at 18 months.
• What was the result, and can it go in the chart where I can see it?
• If there was a follow-up stage to the instrument, was it completed?
• If a referral is being placed — to whom, for what, and has an early intervention referral gone in as well, today?
• If the plan is to watch and wait: watch for what, and come back when?

That last pair of questions is the one that changes outcomes. Referrals to evaluation and to early intervention run in parallel rather than in sequence — a child does not need a completed diagnosis to start receiving birth-to-three services, and treating the two as sequential can cost a year of a small child's development.

If a screen comes back positive, the first 90 days are mostly logistics: waitlists, phone calls, paperwork. Choosing an evaluator is a real decision with real trade-offs in wait time and cost. It is a strange and administrative response to a frightening morning, and it is also, honestly, most of the job.