The Case Against 'Wait and See'

What is 'wait and see' actually proposing?

It is proposing a bet: that the thing you noticed will resolve on its own, and that nothing of value is lost while you find out. The first half of that bet is sometimes right. The second half is almost never examined, and it is where the argument fails.

Waiting is worth doing when it changes something. So the useful question is not "is my child autistic?" — nobody in the room can answer that today — but a narrower one: what would be done differently in six months that cannot be done now? Run through the honest answers:

• If the plan is nothing until the re-check, waiting has bought a delay and no information.
• If the plan is a referral in six months, then the referral could be made today and the waitlist would be six months shorter by the time it matters.
• If the plan is start speech therapy now and reassess, that is not waiting. That is acting, and it is usually the right call.

Watchful waiting is only watchful if something is actually being watched, written down, and reviewed on a named date. Without those, it is just waiting.

The reason this matters is structural rather than emotional. Evaluation capacity is the bottleneck in autism care, not parental willingness. Autism evaluation wait times are measured in months in much of the country, which means the clock on the thing families are told to postpone starts running the day the referral is written — not the day the concern is confirmed.

The gap between when autism can be found and when it is

There is a well-documented gap between the age autism becomes detectable and the age it is actually detected, and it is large enough to be the central fact of this page. Consensus review of the evidence finds the earliest signs and symptoms of autism are detectable by 24 months, and that this is the basis for acting early rather than waiting 1Ref 1Zwaigenbaum L, Bauman ML, Stone WL, et al. (2015).Early Identification of Autism Spectrum Disorder: Recommendations for Practice and Research.That the earliest signs and symptoms of autism are detectable by 24 months, and the consensus rationale for acting early on a concern rather than waiting. Not used here for any claim about treatment effect size.. The pediatric system is built around that finding: 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 2Ref 2American Academy of Pediatrics (2024).Developmental Surveillance and Screening.That developmental surveillance occurs at every well-child visit, with general developmental screening at 9, 18, and 30 months and autism-specific screening at 18 and 24 months..

Then there is what happens. In CDC surveillance data, the median age of earliest known autism diagnosis was 49 months 3Ref 3Maenner MJ, Warren Z, Williams AR, et al. (CDC ADDM Network) (2023).Prevalence and Characteristics of Autism Spectrum Disorder Among Children Aged 8 Years — Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2020.That the median age of earliest known autism diagnosis was 49 months in CDC ADDM surveillance, establishing the gap between when reliable identification is possible and when it actually happens..

Signs are detectable by 24 months; the median child is identified at 49 months 1Ref 1Zwaigenbaum L, Bauman ML, Stone WL, et al. (2015).Early Identification of Autism Spectrum Disorder: Recommendations for Practice and Research.That the earliest signs and symptoms of autism are detectable by 24 months, and the consensus rationale for acting early on a concern rather than waiting. Not used here for any claim about treatment effect size.3Ref 3Maenner MJ, Warren Z, Williams AR, et al. (CDC ADDM Network) (2023).Prevalence and Characteristics of Autism Spectrum Disorder Among Children Aged 8 Years — Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2020.That the median age of earliest known autism diagnosis was 49 months in CDC ADDM surveillance, establishing the gap between when reliable identification is possible and when it actually happens..

That is roughly two years, and it is worth being precise about what fills it. It is not usually one decisive act of neglect. It is a chain of individually reasonable pauses: a screen not done at a sick visit, a concern raised and softened, a six-month re-check that becomes nine, a referral placed onto a nine-month waitlist, an evaluation rescheduled once. Each link is defensible. The sum is two years.

The implication for a specific worried parent is the opposite of the one they usually draw. If the median child is identified at 49 months, then a parent asking questions about a two- or three-year-old is not overreacting or jumping the gun. They are, statistically, ahead of the curve — and the thing standing between them and an answer is mostly queue, not certainty.

The strongest case for waiting, stated fairly

A page arguing one side owes you the other side at full strength, so here it is: sometimes the child does catch up, and the clinician saying so is not being lazy. The clearest example is late language emergence — a delay in language onset in a child without another diagnosed disability. Some of these late talkers do catch up to their peers, while others remain at risk, and the recommended response is early assessment and periodic monitoring 4Ref 4American Speech-Language-Hearing Association (2024).Late Language Emergence (Practice Portal).That late language emergence is a language-onset delay without other diagnosed disability, that some late talkers catch up while others remain at risk, and that the recommended response is early assessment and periodic monitoring rather than waiting..

Read that carefully, because it is doing two things at once. It concedes the point that pure waiting rests on: yes, some children resolve without intervention. And then it declines to recommend waiting anyway. The recommendation attached to a population where some children genuinely catch up is still assessment and monitoring 4Ref 4American Speech-Language-Hearing Association (2024).Late Language Emergence (Practice Portal).That late language emergence is a language-onset delay without other diagnosed disability, that some late talkers catch up while others remain at risk, and that the recommended response is early assessment and periodic monitoring rather than waiting. — not watching from home.

The other honest points on that side:

• Early evaluations are not infallible. A young child's picture can be genuinely unclear, and a rushed conclusion in an ambiguous case is its own harm.
• Some of what worries parents is normal variation. Development is uneven, and not every plateau is a sign.
• Diagnosis has costs. It is not weightless for a family, and pretending otherwise is dishonest.

None of that argues for doing nothing. It argues for the difference between a plan and a pause. Assessment with periodic monitoring is a plan: it has a schedule, a person, and a record. "Let's see how he's doing at the next visit" is a pause wearing a plan's clothing.

Why the bet is asymmetric

Because the two ways of being wrong do not cost the same, and almost nobody does this arithmetic out loud. Set it out honestly, given that some children in this position will turn out not to be autistic and some will:

The bottom-left cell is the only clean win for waiting, and its prize is modest: you skipped some appointments. The bottom-right cell is the expensive one, and it is where the cost of the whole strategy sits.

Two things sharpen the asymmetry further. First, the evaluation queue means the cost of acting is paid mostly in your time, while the cost of waiting is paid in the child's development — and those are not the same currency. Second, most of what gets started while waiting is useful regardless of the outcome. A child with a language delay who turns out not to be autistic still needed the speech therapy. This is why the eligibility vs diagnosis distinction matters so much: the help is aimed at the difficulty, not at the label.

There is one more failure mode worth naming. Children whose difficulty is masked by a strength — the child with startling vocabulary who cannot converse, or the early reader with almost no functional speech that people describe as hyperlexia — are the ones "wait and see" holds longest, because every observer keeps finding the reassuring half of the picture.

What the evidence on acting early actually shows

It shows something real, and less than the internet claims. The landmark trial is a randomized controlled study of 48 toddlers aged 18 to 30 months, in which the Early Start Denver Model produced significant gains in IQ, adaptive behavior, and diagnostic status compared with community intervention over two years 5Ref 5Dawson G, Rogers S, Munson J, et al. (2010).Randomized, Controlled Trial of an Intervention for Toddlers With Autism: The Early Start Denver Model.That a randomized trial of 48 toddlers aged 18 to 30 months found the Early Start Denver Model produced significant gains in IQ, adaptive behavior, and diagnostic status versus community intervention over two years — cited here as a single-site trial and explicitly not as a promise of individual outcomes.. That is genuine evidence that intensive early developmental-behavioral intervention improves outcomes, and it is the study underneath most of the claims made in this space.

It is also a single-site trial of 48 children 5Ref 5Dawson G, Rogers S, Munson J, et al. (2010).Randomized, Controlled Trial of an Intervention for Toddlers With Autism: The Early Start Denver Model.That a randomized trial of 48 toddlers aged 18 to 30 months found the Early Start Denver Model produced significant gains in IQ, adaptive behavior, and diagnostic status versus community intervention over two years — cited here as a single-site trial and explicitly not as a promise of individual outcomes.. What follows from being careful about that:

• It does not promise your child a specific outcome. A group-level average across 48 children is not a forecast for one.
• It does not say intervention must be intensive or expensive to be worth anything, and it does not rank the available approaches against each other.
• It does not support "recovery" or "a closing window." Neither claim is in the study, and both get sold hard.

The stronger version of the early-action argument does not actually rest on that trial. It rests on the fact that acting early is what the consensus recommends on the basis of when signs become detectable 1Ref 1Zwaigenbaum L, Bauman ML, Stone WL, et al. (2015).Early Identification of Autism Spectrum Disorder: Recommendations for Practice and Research.That the earliest signs and symptoms of autism are detectable by 24 months, and the consensus rationale for acting early on a concern rather than waiting. Not used here for any claim about treatment effect size., and on the plain observation that a child who is identified is a child whose environment can be adjusted — at home, in preschool, by the adults who set the expectations. That does not require an RCT. It requires someone knowing.

Nothing here says a family who waited has ruined anything. Children identified later do well. The argument is about the cost of a delay, not about a door that shuts.

Be skeptical of anyone selling urgency. A page that tells you the window is closing is usually selling something. This one is telling you the queue is long, which is a different and more mundane problem.

You do not have to wait for a diagnosis to start

This is the fact that dissolves most of the dilemma, and it is the one families most often do not know. IDEA Part C provides early intervention services to children from birth through age two who have disabilities or developmental delay, along with their families, delivered in natural environments through a statewide system 6Ref 6U.S. Department of Education (2024).IDEA Early Intervention Program for Infants and Toddlers with Disabilities (Part C).That IDEA Part C provides early intervention services to children birth through age two who have a disability or developmental delay, and their families, in natural environments through a statewide system — establishing that eligibility rests on delay rather than on a completed autism diagnosis.. Read the eligibility category: it is delay. It is not an autism diagnosis.

What that means in practice:

• A family can request an early intervention evaluation directly. It is a statewide system with a defined referral route 6Ref 6U.S. Department of Education (2024).IDEA Early Intervention Program for Infants and Toddlers with Disabilities (Part C).That IDEA Part C provides early intervention services to children birth through age two who have a disability or developmental delay, and their families, in natural environments through a statewide system — establishing that eligibility rests on delay rather than on a completed autism diagnosis., and a pediatrician's blessing is not the entry ticket.
• This runs in parallel with a diagnostic evaluation, not instead of it. Joining the diagnostic waitlist and starting Part C on the same day is the move. Neither cancels the other, and one of them starts far sooner.
• The services are aimed at function. Speech, occupational therapy, and developmental support address what the child is finding hard now.
• Age three is a handoff, not a cliff. Part C runs birth through age two 6Ref 6U.S. Department of Education (2024).IDEA Early Intervention Program for Infants and Toddlers with Disabilities (Part C).That IDEA Part C provides early intervention services to children birth through age two who have a disability or developmental delay, and their families, in natural environments through a statewide system — establishing that eligibility rests on delay rather than on a completed autism diagnosis.; after that, the route is an evaluation request to the school district.

So the sentence "we should wait and see before doing anything" contains a false premise. The waiting was never protecting the child from an unnecessary diagnosis, because the help on offer during the wait does not require one. Starting therapy before diagnosis is the ordinary path, not an exception, and this is exactly why early intervention doesn't diagnose autism — it is a different system answering a different question.

What to say when you are told to wait

The goal of the conversation is not to win it. It is to leave with three things: the concern in the chart, a referral placed, and a named date. A clinician who recommends waiting may be right, and it is still reasonable to ask them to make the reasoning explicit.

Questions that tend to move the visit:

• "What specifically would you expect to see in six months if this is nothing?" This converts a vague pause into a testable prediction, and it gives you both something to check against.
• "Can we do the autism-specific screen today rather than at the next visit?" Autism-specific screening belongs at 18 and 24 months, plus surveillance at every well-child visit 2Ref 2American Academy of Pediatrics (2024).Developmental Surveillance and Screening.That developmental surveillance occurs at every well-child visit, with general developmental screening at 9, 18, and 30 months and autism-specific screening at 18 and 24 months.. If it has not been done at the right age, asking is not pushing.
• "Would you document the concern in the chart today?" A documented concern has a trail. One raised in a hallway has to be raised again from scratch.
• "Can we place the referral now and cancel it if things resolve?" This is the crux. Given the wait, a referral placed today and cancelled in six months costs almost nothing and buys the queue position. A referral placed in six months costs six months. Shortening the wait is mostly a matter of when the clock starts.
• "Can we start speech or OT while we wait?" Function-based help does not depend on the diagnostic answer.
• "Can you put in the early intervention referral as well?" Different system, different clock, eligibility based on delay 6Ref 6U.S. Department of Education (2024).IDEA Early Intervention Program for Infants and Toddlers with Disabilities (Part C).That IDEA Part C provides early intervention services to children birth through age two who have a disability or developmental delay, and their families, in natural environments through a statewide system — establishing that eligibility rests on delay rather than on a completed autism diagnosis..

If the answer to all of that is still to wait, a second opinion is a normal part of medicine and not an act of aggression. Parental concern is one of the more reliable early signals in developmental care — you have a sample size of one, watched for thousands of hours, and it is legitimate for that to count. It is also true that you cannot diagnose your own child from a page like this one, and the point of every step above is to get your observation in front of someone who can.