Does the Early Start Denver Model (ESDM) help autistic toddlers?
A well-regarded early approach with real but modest evidence — one of the more hopeful, lower-risk options in this library.
Podcast·Two-voice deep dive
Listen to the discussion
Chapters & transcript
This is an AI-generated discussion of this article, transcribed automatically. The written summary below is the reviewed source — if the two differ, go by the summary.
0:00What is this discussion about?
Host A Back in 2010, an early autism therapy made a promise that honestly seemed almost impossible.
Host B Oh, yeah. The Dawson study.
Host A Right. A study was published claiming that this specific intervention could raise an autistic toddler's IQ by a staggering 17 points.
Host B 17 points is massive.
Host A It really is. I mean, if you're a parent or a caregiver reading that, you're naturally going to clamor for it, right? You would move mountains to get your kid into that exact program.
Host B Oh, absolutely.
Host A But then, nearly a decade later, when scientists tested it again in a much larger independent trial, that massive IQ jump just completely vanished.
Host B And that is the ultimate whiplash for a family.
Host A Yeah.
Host B Because you're trying to make these really crucial decisions during a, you know, very sensitive window of your child's early development.
Host A Right. You're just desperate for answers.
Host B Exactly. You're standing in the center of this absolute hurricane of information and you just expect science to give you a clear binary roadmap. Do this. Don't do that.
Host A But it's never that simple.
Host B No, not at all. As we're going to see, early intervention research is incredibly messy. And the gap between a flashy medical headline and the nuanced reality, well, it's vast.
Host A Okay, let's unpack this. Our mission for today's deep dive is to act as your shortcut to clarity. We're cutting through all that noise to look at the exact therapy behind those studies.
Host B The Early Start Denver Model.
Host A Right, or ESDM for short. And to do this, we're drawing on two really comprehensive internal documents from Spectrum Connect. Which are great resources. They really are. We have an evidence synthesis audit document and a findings report. Now, these are basically demonstration runs to show how their engine reasons, but they synthesize the real foundational science on this topic.
Host B So we aren't just looking at, you know, opinion blogs here.
Host A No, definitely not. We're looking at the raw clinical trials.
Host B Which means we can actually see what the data proves, what it maybe only suggests, and where the massive blind spots are.
1:57What does ESDM actually look like day to day?
Host A Right. But before we get into the spreadsheets and the disappearing IQ points, we kind of have to ground ourselves in what this therapy actually looks like in, say, a living room.
Host B Yeah, context is everything. That's actually why I've adjusted my backdrop today.
Host A Oh, yeah, I see that. You've gone from the usual towering library bookshelves to like a bright, slightly messy toddler playroom.
Host B Yeah. I've got the blocks, a toy kitchen, some sensory bins.
Host A It's the perfect visual anchor. Yeah. Because ESDM doesn't happen in a sterile clinical lab with a kid sitting at a desk, right? It happens on the floor.
Host B Exactly. It's a comprehensive early intervention designed specifically for autistic toddlers. So roughly kids between 12 and 48 months old.
Host A Okay.
Host B And the technical classification for it is an NDBI. That stands for a naturalistic developmental behavioral intervention.
Host A That is a mouthful.
Host B It really is, yeah. But what that essentially means is it merges two historically opposing philosophies. It takes this warm relationship based developmental approach and it layers it with very targeted behavioral principles.
Host A So to understand the mechanics here, let's contrast it with traditional behavioral therapy.
Host B Good idea.
Host A Because traditional therapy is often like, I don't know, building a house from a strict blueprint. The therapist has a rigid plan, sits the kid down, runs drills with flashcards, and rewards them for the correct answer.
Host B Very structured.
Host A Yeah, the child is basically conforming to the blueprint. But this naturalistic approach, ESBM, it feels more to me like sneaking vegetables into a delicious smoothie.
Host B Oh, I like that.
Host A Right, you're using everyday play and natural routines to just kind of seamlessly deliver this intensive behavioral and developmental support, and the kid just thinks they're having a treat.
Host B That is a fantastic analogy. The core tenet is really about following the child's lead.
Host A So what does that actually look like, minute by minute?
Host B OK, let's imagine a parent giving their autistic toddler a bath.
Host A OK, bath time.
Host B And the toddler isn't interested in the bath toys at all. They're just repeatedly slapping the water to watch it splash.
Host A He's splashing over and over.
Host B Right. Now, in a rigid traditional model, a therapist might try to redirect that child to play appropriately with a rubber duck.
Host A Like, no splashing, look at the duck.
Host B Exactly. But in ESDM, the therapist, or even the parent who is actively coached to do this, they join in on the splashing.
Host A Oh, wow. They just splash with them.
Host B They join the splash. The adult gets right in their line of sight and splashes the water with them. And then they pause.
Host A Like to see what the kid does.
Host B Exactly. They wait just for a second to see if the child looks at them to figure out why the fun stopped. And the moment the child makes eye contact or makes a sound, the adult enthusiastically says, splash, and hits the water again.
Host A So they're taking the child's natural obsession in that exact moment and turning it into a behavioral loop.
Host B Yes. They're building a pathway for joint attention, social imitation, language, but the kid just thinks they're having a great time in the bathtub.
Host A That's incredible.
Host B And what's fascinating here is that because it's developmentally appropriate, play-based, and heavily reliant on parent inclusion, the harm profile for ESDM is considered minimal.
Host A Because you aren't fighting their neurodivergence.
Host B Exactly. You're using their natural motivations as the vehicle for learning. It makes it a distinct category from the intensive, Highly structured, traditional ABA. And that's a really vital distinction for you to understand when you're weighing the lower risk profile of this therapy.
5:23How did the first study show such a massive IQ jump?
Host A Okay, so if the therapy is just highly calibrated play, how did it possibly produce a 17-point IQ jump?
Host B That is the million-dollar question.
Host A Right. Let's move from the living room to the raw data. In 2010, Geraldine Dawson and her colleagues published this landmark study in the journal Pediatrics.
Host B It was a single-site randomized controlled trial with 48 toddlers.
Host A And over a two-year period, the toddlers receiving ESDM gained an average of 17.6 IQ standard score points.
Host B While the control group, which was just getting standard community treatment, they only gained 7.0 points.
Host A That's a massive difference.
Host B It really is. The ESDM group also showed significant gains in everyday adaptive behaviors. It was the kind of massive, undeniable effect size that immediately put the therapy on the map as a potential game changer.
Host A But then we hit 2019.
Host B Ah, yes.
Host A The Rogers study gets published in the Journal of the American Academy of Child and Adolescent Psychiatry. And this is a multi-site replication attempt, right?
Host B Right. They tested 118 toddlers across three different universities.
Host A And it was single blind. So the people assessing the kids had no idea who was getting ESDM and who wasn't.
Host B Exactly.
Host A And the result, that giant IQ jump is just gone. The effects were much more modest and primarily seen in language.
Host B Yep.
Host A So wait, did the first study do something wrong, or is this just the messy reality of science getting bigger and more rigorous?
Host B It's the latter.
6:49Why did the effect shrink in the bigger, more rigorous trial?
Host B What happened is a phenomenon called attenuation at scale, and it's actually incredibly common in intervention research.
Host A Walk me through that. What happens when a therapy scales up like that? Why do the numbers just shrink?
Host B Well, think about the conditions of that first 2010 study. It was a single site. It was heavily overseen by the brilliant people who literally invented the therapy.
Host A Oh, so they were super invested.
Host B No, highly invested. The therapists delivering the care were likely the most highly trained, highly motivated practitioners in the world. The conditions were absolutely pristine.
Host A Okay, I see where this is going.
Host B Right. When you move to a multi-site trial in 2019, you're introducing real-world messiness. You have different therapists with varying levels of experience, different community dynamics.
Host A And the blinding you mentioned earlier.
Host B Yes. The assessors were masked, which removes any subconscious bias where an evaluator might inadvertently grade a child more generously just because they know the kid is in the special treatment group.
Host A The messy reality of science leaving the pristine lab and hitting a normal community clinic.
Host B Exactly. And because of the rigorous design of that 2019 trial, you know, the blinding, the larger sample size, the multiple sites, the scientific consensus dictates that we really have to weigh those modest replicated results much more heavily than the initial 2010 outlier.
Host A Makes sense.
8:08What is ESDM actually changing, if not core autism traits?
Host A But if the massive cognitive explosion didn't hold up, what exactly is ESDM changing? Like, what are these kids gaining?
Host B Let's look at the nuance in the numbers.
Host A Yeah, because we shouldn't just look at two studies, right? We need to look at the meta-analyses that pool everything together.
Host B Right. So we have a 2020 meta-analysis by Fuller. It pooled 12 studies involving 640 children. Okay.
Host A Pretty solid sample size.
Host B And it found a moderate overall effect size favoring ESDM with a statistical G value of 0.357. Okay. Then there's a later 2022 meta-analysis that stratified these effects by specific domains, which is where the real story emerges.
Host A I'm going to pause on this term G value really quick because just throwing a decimal at someone doesn't really help them visualize the impact on their actual toddler.
Host B That's a fair point.
Host A You mentioned the 2022 analysis showed a moderate replicated benefit for developmental milestones, specifically cognition at 0.28 and language at 0.29. Right. What does a 0.29 language bump actually look like in real human behavior?
Host B So a G value around 0.3 indicates a moderate, reliable shift. It's not a magical cure where a non-speaking child suddenly starts reciting Shakespeare.
Host A Right, of course not.
Host B What it looks like in the real world is the difference between a child having a 45-minute meltdown because they're thirsty and can't express it versus that same child learning to take their parent's hand, walk them to the refrigerator, and point at a juice box. Wow. It's the acquisition of functional downstream developmental tools.
Host A OK, so it has a real measurable impact on those milestones. But what about the core traits of autism itself?
Host B This raises an important question, because when the researchers looked at the effects on core autism symptoms, things like repetitive behaviors. hand flapping, or intense hyperfixations, the effects were significantly smaller.
Host A How small?
Host B Very small, mixed, and highly inconsistent. The G value hovered around 0.27. And in several restricted analyses, there was absolutely no significant effect on core autism symptomology at all.
Host A Interesting. So ESDM is measuring downstream developmental outcomes way more effectively than it changes the core neurodivergent traits.
Host B Yes.
Host A It's helping the child build skills to interact with the world, but it isn't making them less autistic.
Host B Exactly. It's an intervention that promotes development, not one that attempts to cure an R-type.
Host A Let's ground this in reality for the parents listening, because when you're in the trenches with a toddler, I mean, two days feels like a lifetime.
Host B Oh, absolutely.
Host A Two years is an absolute eternity. The exhaustion is just bone deep.
Host B Yeah.
Host A So if a family commits to this play based intervention to help with those language milestones. How long are they realistically waiting to see that 0.29 shift?
Host B We have to set incredibly realistic expectations here, drawing right on the parameters of the Dawson 2010 study.
Host A OK. What's the timeline?
Host B Because this is a naturalistic play-based intervention, it relies on the slow, steady building of neural pathways over thousands of tiny positive interactions. The developmental tracking for this model is measured over a two-year timeline.
Host A Wow. Two years. So you're not going to see a transformation in three weeks.
Host B Absolutely not. You aren't getting overnight results. You're signing up for a marathon of gentle, persistent support over 24 months. You're waiting for the cumulative effect of a thousand bath time splashes and playtime routines to finally crystallize into a new cognitive skill.
Host A Which means as a parent, you are investing a massive amount of time, energy and likely money into this two year marathon.
Host B A huge investment.
11:45Who funded and ran this research?
Host A Which makes it incredibly crucial to examine exactly who is writing the science you're basing this massive decision on. We have to look at the footnotes.
Host B We do.
Host A Because the Spectrum Connect audit flags something called a developer COI, or conflict of interest.
Host B Right. The provenance of the research is always critical. And when we analyze the evidence base for ESDM, a very clear pattern emerges.
Host A Which is?
Host B The key trials and the major meta analyses were led by the creators of the model themselves.
Host A Oh, wow.
Host B Yeah. Sally Rogers and Geraldine Dawson, the brilliant minds who developed ESDM, are the prominent authors on the 2010 trial, the 2019 trial, and even the 2020 meta analysis by Fuller.
Host A Wait, hold on. You're telling me the people who invented the therapy, who have basically staked their entire careers on it, are the ones running the studies and pooling the data to see if it works? Isn't that just grading your own homework? I mean, how is that allowed in scientific literature?
Host B I completely understand the skepticism. And honestly, it's a valid reaction. But in behavioral and educational interventions, developer involvement is essentially the non drug equivalent of industry funding in pharmaceutical trials.
Host A OK, so it's just kind of how things are done.
Host B It's how these therapies get off the ground. The people who invent them secure the initial grants to test them. It doesn't mean they're falsifying data. You know, these are highly respected academics operating under rigorous peer review.
Host A But it inherently introduces a bias.
Host B It does. A 2016 systematic review looked at 15 studies involving 209 children. It found mostly positive outcomes for ESDM, but it explicitly flagged that about half the studies had quality or rigor weaknesses.
Host A And those were tied to the developers.
Host B Largely tied to this close developer involvement, yes.
Host A Here's where it gets really interesting because it's like a brilliant chef reviewing their own restaurant.
Host B Oh, I see what you mean.
Host A Right. The food might genuinely be great. The chef isn't lying about the ingredients. But because their name is on the door, you still desperately want an independent food critic to walk in, eat the meal, and confirm that it's actually that good.
Host B That is exactly the consensus of the broader scientific community right now. That developer involvement just underscores a desperate need for independent non-developer replications.
Host A We need community effectiveness trials.
Host B Yes. We need to know what happens when a regular therapist in a community clinic in Ohio or London runs this program completely unaffiliated with the creators.
Host A Are there any studies like that happening?
Host B Fortunately, yes. There's a large ongoing cluster randomized trial right now called the PolyMES study, and it's aiming to do exactly this.
Host A That independent data will be the missing puzzle piece. OK, so we've gone from the living room floor through the messy reality of replication and into the footnotes of who is funding the science.
Host B We've covered a lot.
14:31What should you ask before starting ESDM with your child?
Host A But if I'm a parent listening to this and I have an appointment with an early intervention clinic tomorrow morning, how do I actually use this information? I can't just walk in and quote meta-analyses at them.
Host B Right. Knowledge is only powerful when it's applied.
Host A Exactly. What exactly should I be asking?
Host B Based on the data we've synthesized, there are three highly specific, actionable questions you must ask your prospective therapy team. The first is about intensity.
Host A Because the dosage matters.
Host B Exactly. The main trials that demonstrated those moderate gains in language and cognition, they were studying programs that offered about 15 to 20 hours a week of intervention.
Host A Okay, so 15 to 20 hours.
Host B You need to look the provider in the eye and ask, is your program offering the 15 to 20 hours a week that the science is actually based on?
Host A Because if they aren't.
Host B If a clinic is only offering you two hours a week, you cannot realistically expect the outcomes promised by those clinical trials.
Host A That makes total sense. And the second question.
Host B The second is about coaching. We established that ESDM relies heavily on integrating therapy into daily routines, right?
Host A Right, a bath time splashing.
Host B Yes. So ask the therapist, is parent coaching actively included in the natural routines? Are they going to teach you how to run these behavioral loops during mealtime, or are they just working with your child in isolation?
Host A So the parents have to be involved.
Host B The model only works if the parents are empowered to continue the play-based support around the clock.
Host A And the third question ties right back to that grueling 24-month marathon we talked about. Tracking.
Host B Yes. Tracking is essential.
Host A You have to ask them, how exactly are we tracking developmental progress over that two-year timeline?
Host B Yeah.
Host A You need to know what metrics they're using to ensure your child is actually making those moderate gains.
Host B So you aren't just spinning your wheels for two years hoping something happens.
Host A Exactly. Intensity, coaching, and tracking. Asking those three questions just shifts your dynamic from being a passive recipient of medical advice to being an active, informed partner in your child's developmental care.
Host B Which is so incredibly important.
Host A And empowering you to be an active partner is the entire reason we're doing this. Currently, Spectrum Connect is bridging the incredibly complex science of autism therapies to make it more approachable for parents.
Host B They're doing all the heavy lifting of auditing the science.
Host A Exactly. So you don't have to go get a PhD in statistics just to figure out how to help your toddler communicate.
Host B It's a vital service because parsing the difference between an initial miracle study and an attenuated multi-site trial is incredibly difficult without a dedicated guide.
Host A It really is. And we actually have a specific call to action here. If you are a credentialed expert listening to this deep dive right now.
Host B Like an MD or a PhD.
Host A Right. An MD, a PhD, a PsyD or a DO. If you have those credentials and you would like to help us validate our findings and contribute to this mission, email the team at research at SpectrumConnect.us.
Host B It's a phenomenal opportunity.
Host A Again, that's research at SpectrumConnect.us. We'd love your help to ensure the information reaching families is as rigorous and independent as possible.
17:37What are the key takeaways?
Host A So let's bring it all together. We've learned that the Early Start Denver model is a promising, developmentally appropriate, play-based option.
Host B With a very minimal harm profile.
Host A Right, because it doesn't force a child to sit at a desk and drill. It uses their natural interests as a trellis for learning.
Host B We also saw that while the initial studies promised massive IQ leaps, the more rigorous scaled-up trials show that the benefits are real, but they're moderate.
Host A Mostly concentrated in language and cognition.
Host B Yes, and they require a realistic two-year commitment to fully emerge.
Host A And we also learned that while the evidence base is heavily led by the therapy's developers right now, independent trials are on the horizon. But ultimately, the science is just a tool. The choice of how to use it stays with the family and their care team.
Host B The data can guide the blueprint, but you are the one tending the garden.
Host A I love that. So what does this all mean? We started by talking about the desire for a clean roadmap when you get an autism diagnosis, but looking at how ESDM actually works.
Host B How it improves those downstream milestones, like language, but has a very weak effect on core autism symptoms, like repetitive behaviors.
Host A Right. It leaves us with a fascinating final thought to mull over.
Host B It really demands a profound shift in how we view the goal of therapy itself.
Host A Yeah. Because if a therapy successfully gives a child the language to communicate that they're hungry or overwhelmed, but it doesn't stop them from flapping their hands when they're excited, Is that actually a failure?
Host B That's the question.
Host A Or does it suggest that a truly ethical intervention shouldn't even be trying to erase a child's autistic traits in the first place? Maybe it should just give them the developmental tools, like language, to navigate the world on their own terms.
Host B It's the fundamental difference between trying to fix a child and trying to equip a child.
Host A Equipping, not fixing. Thank you for joining us on this deep dive. Keep asking those hard questions, keep trusting your instincts, and we invite you to keep exploring.
Key Takeaways
One of the more hopeful, lower-risk options — ESDM is a play-based early intervention for toddlers that blends developmental and behavioral methods, delivered in natural settings by therapists and parents. It's one of the better-studied, developmentally-appropriate options in this library.
A landmark trial found large gains — an influential 2010 trial found toddlers in ESDM gained substantially more in thinking skills, about 17 IQ points versus about 7 in a comparison group. That result was exciting and widely cited.
A larger, more rigorous trial found more modest gains — a bigger multi-site trial designed specifically to check the first result found more modest gains, mainly in language, and didn't reproduce the large across-the-board IQ jump. Striking early results shrinking in larger trials is common in intervention research.
The effect is stronger on development than on autism traits — pooling the studies, ESDM shows moderate benefits for early cognition and language. Effects on core autism traits, like repetitive behaviors, are smaller and less consistent.
Much of the research was developer-led — and it's relatively low-risk — about half the studies had quality limitations, so independent confirmation matters. On the other hand, ESDM is play-based, includes parents, and is generally viewed as developmentally appropriate for young children.
What this means for you
ESDM is a comprehensive early intervention for toddlers (roughly 12 to 48 months). It combines developmental and relationship-based approaches with behavioral principles, delivered through play in everyday routines by trained therapists and coached parents — one of the “naturalistic developmental” approaches that follows the child's lead rather than relying on highly structured drills. An influential 2010 trial found toddlers in ESDM gained substantially more in thinking skills — about 17 IQ points, versus about 7 in a comparison group — along with gains in everyday adaptive skills. That result was exciting and widely cited.
A bigger 2019 trial across multiple sites, designed specifically to check the first result, found more modest gains — mainly in language — and did not reproduce the large across-the-board jump in IQ. This pattern, where a striking early result shrinks in larger, more independent trials, is common in intervention research, and it's why the fuller picture matters more than any single study. Pooling the studies together, ESDM shows moderate benefits for early development, especially cognition and language; effects on core autism traits, such as repetitive behaviors, are smaller and less consistent. Much of the research was led by the model's developers, and about half the studies had quality limitations — so independent confirmation matters. On the other hand, ESDM is relatively low-risk: it's play-based, developmentally oriented, and includes parents.
What was studied. Studies enrolled toddlers (roughly 12–48 months) and measured developmental outcomes — cognition, language, and adaptive behavior. Intensity varied, at about 15–20 hours a week in the main trials.
Worth asking your early-intervention team. How intensive is it? Is parent coaching included? How is your child's progress tracked?
Where the studies landed
Promising — real but modest evidenceThe two pivotal trials agree ESDM helps — the more rigorous, larger one just found a smaller effect than the first.
5 Points toward it helping
1 Narrower measure, no effect
Tap any tile to read it.
Points toward it helping5
Narrower measure, no effect1
Spectrum Connect reviews published research on interventions parents are exploring for their autistic children — so you can see where the evidence actually stands. No agenda, no selling, no cherry-picking. Just the studies, our method, and what it means for you.
See the research behind this
For anyone who wants to check our work · 6 sources
Where we looked
This run was a scoping search only — done via general web search, not the reproducible Boolean search of record and not the PubMed/Epistemonikos API layer we use on a fully conformant run. That means we can't publish reproducible per-database counts or a formal PRISMA flow for this run. Below is the search string a full conformant pass would run against PubMed, the Cochrane Library, PsycINFO, Embase, ERIC, and CINAHL — we haven't executed it against the database APIs yet.
("autism spectrum disorder"[MeSH] OR autistic[tiab] OR ASD[tiab]) AND ("Early Start Denver"[tiab] OR ESDM[tiab] OR "naturalistic developmental behavioral"[tiab] OR NDBI[tiab] OR "Denver model"[tiab]) AND (toddler*[tiab] OR "early intervention"[tiab] OR efficacy[tiab] OR systematic[sb] OR "meta-analysis"[pt] OR "randomized controlled trial"[pt])
Run on PubMed →
What we did with what we found
What the strongest evidence says
A landmark trial found large IQ gains; a larger, more rigorous multisite trial found more modest gains, mainly in language. Pooled estimates land on a moderate, real benefit for early cognitive and language development.
Effects on repetitive behaviors and other core autism traits are smaller and less consistent than the developmental gains — some analyses find no significant effect at all.
Much of the key evidence — the landmark trial, the main meta-analyses — comes from the model's own developers, and about half of studies in one review had quality limitations. Independent, non-developer confirmation would raise certainty here.
Keep reading
What would you like to do?
Want to talk to a parent who’s been there?
We match you with families whose kid is like yours. Free, takes two minutes.
Heard of something we haven’t covered?Suggest an article
If another parent, a clinic or a video mentioned a therapy you can’t find here, tell us. We’ll read the research and review it next.
Suggest an articleThink we got something wrong?Report a factual error
We publish the whole record so it can be checked, and that only counts if we act on what you find. If a number looks wrong, a study is missing or has been retracted, or we’ve read a finding in a way the evidence doesn’t support, tell us.
You don’t need a research background to file one. “This doesn’t match what our doctor told us” is a useful report. Every one reaches a person: we reply within seven days, and within thirty we have either corrected the page or told you when we will. Substantive reports send the affected steps back through the protocol and need fresh sign-off before anything here changes.
Report a factual errorAre you a clinician or researcher?Review this article
If you have read this and the evidence holds up, you can sign off on it. We send you a personal review link where you confirm your credentials, and our team checks every review before anything is published. Once approved, your name and a link to your profile appear on this article.
Ask to review this