Some children diagnosed with autism do eventually stop meeting the diagnostic criteria, but this is uncommon and doesn’t mean autism simply disappears with age. Depending on the study, somewhere between 3% and 37% of children diagnosed with autism in early childhood no longer qualify for the diagnosis a few years later. That wide range reflects differences in study size, how children were originally diagnosed, and how strictly researchers defined “no longer autistic.” The reality is more nuanced than the phrase “growing out of it” suggests.
What “Losing” an Autism Diagnosis Looks Like
Researchers use the term “optimal outcome” to describe children who once had a confirmed autism diagnosis but no longer show significant autism traits. To qualify in research studies, these individuals must have been diagnosed before age 5 by a specialist, and they must later score below autism thresholds on standardized assessments. They also need to function in the normal intellectual range, attend regular classrooms without one-on-one support, and demonstrate typical communication and social skills on validated measures.
That sounds like a clean bill of health, but it isn’t quite. The definition of optimal outcome explicitly acknowledges that other difficulties often persist. Weaknesses in executive functioning (things like planning, organizing, and shifting attention) are common. So is a heightened vulnerability to anxiety and depression. The autism-specific behaviors fade, but the person’s neurological profile doesn’t become indistinguishable from someone who was never autistic.
How Common Is It?
The numbers vary significantly. In smaller clinical studies tracking children diagnosed at age two, between 18% and 37% no longer showed signs of autism two years later. A broader review of long-term studies found a more conservative range of 3% to 25%. The higher figures tend to come from studies of very young children, some of whom may have been on the milder end of the spectrum or whose initial diagnosis captured a developmental delay that later resolved. Children diagnosed under more rigorous criteria, or those with more pronounced traits, are far less likely to lose the diagnosis.
Diagnostic stability also increases with age. In one decade-long screening study, all nine children who received an autism diagnosis after comprehensive evaluation at 18 to 48 months retained that diagnosis at every follow-up point. Two additional children who weren’t initially diagnosed later met criteria, suggesting that under-diagnosis may actually be more common than over-diagnosis in the long run.
What Predicts a Change in Diagnosis
Not every child has an equal chance of moving off the spectrum. Research comparing children who lost their diagnosis to those who kept it has identified several early predictors. Children who eventually lost the diagnosis tended to have milder autism symptoms from the start, fewer repetitive behaviors (like lining up objects or rigid routines), and stronger everyday adaptive skills such as dressing, communicating needs, and interacting with family members.
Interestingly, baseline IQ and language ability at the time of initial diagnosis did not significantly differ between the two groups. What did differ were motor skills and overall symptom severity. Children who later lost the diagnosis had stronger motor skills at their first evaluation and notably lower scores on symptom severity scales. A diagnosis of PDD-NOS (a milder, now-retired diagnostic category that fell under the autism umbrella) also predicted a higher likelihood of eventually moving off the spectrum compared to a full autistic disorder diagnosis.
The Role of Early Intervention
Early therapy is widely recommended for children with autism, and many children make meaningful gains. But the relationship between intervention and losing a diagnosis is less straightforward than parents might hope. A large multisite randomized trial comparing different therapy styles and intensities found that neither the type of intervention nor the number of hours per week consistently predicted improvement in autism symptom severity, expressive communication, receptive language, or nonverbal ability. At one of three study sites, higher-intensity treatment did lead to greater improvement in autism severity, but this effect didn’t replicate at the other two sites.
This doesn’t mean therapy is ineffective. Many children develop better communication, stronger social skills, and greater independence through intervention. But the evidence suggests that the children who lose their diagnosis aren’t simply the ones who received more or “better” therapy. Their individual neurological and developmental profiles at the time of diagnosis appear to play a larger role than the specific treatment they received.
What’s Happening in the Brain
Brain imaging research offers some clues about what distinguishes children who move off the spectrum. Studies examining white matter, the wiring that connects different brain regions, found that children who retained their autism diagnosis had more pronounced structural differences in key pathways involved in language, attention, and sensory processing. Children who achieved optimal outcomes showed fewer of these differences, though their brain structure still wasn’t identical to children who were never autistic.
One specific finding stood out: the structural integrity of a pathway connecting the thalamus (a brain relay station that routes sensory information) to the cortex was a significant predictor of whether a child would lose the diagnosis. Children with stronger connectivity in this pathway were more likely to achieve optimal outcomes. This suggests that how efficiently the brain routes and processes incoming information may set the stage for how much autism-related behavior a child shows over time.
Masking Versus Genuine Change
One of the most important questions in this area is whether children who lose their diagnosis have truly changed or have simply learned to hide their differences. Masking, sometimes called camouflaging, is the conscious or unconscious suppression of autistic traits to appear more typical. Autistic adults describe it as exhausting and often harmful, involving constant monitoring of facial expressions, tone of voice, and social responses.
In research settings, optimal outcome participants are evaluated specifically to guard against this possibility. Their social communication is assessed by experienced clinicians reviewing video recordings, not just by parent report or self-assessment. Still, the line between genuinely losing autistic traits and becoming very skilled at compensating for them is difficult to draw with certainty. Some researchers and many autistic adults argue that this distinction may matter less than the question of whether the person is actually thriving.
Long-Term Mental Health After Losing a Diagnosis
Adults who lost their autism diagnosis in childhood don’t have a perfectly smooth road ahead. Research on their psychiatric profiles shows that while active mental health conditions tend to decrease over time (unlike in adults who retain the diagnosis, where co-occurring conditions persist), these individuals still carry elevated rates of ADHD and specific phobias compared to the general population. More recent work has found that as young adults, people who lost the diagnosis have higher rates of depression and anxiety than their neurotypical peers, pointing to a lasting vulnerability even after the core autism traits have resolved.
The lifetime prevalence of psychiatric diagnoses is roughly equivalent between people who lost the diagnosis and those who kept it. The difference is in the trajectory: conditions tend to wane in the group that lost the diagnosis, while they persist in those who remain autistic. But “waning” is not the same as “gone,” and this pattern suggests that a history of autism leaves a lasting imprint on mental health regardless of whether the person still meets diagnostic criteria.
The Neurodiversity Perspective
Not everyone agrees that losing a diagnosis should be framed as a goal. The neurodiversity movement views autism as a natural variation in human cognition, not a disease to be cured. Supporters of this perspective generally oppose interventions aimed at making autistic people “act less autistic” and instead advocate for changing environments to be more accommodating. They draw a clear distinction between treating co-occurring conditions like anxiety, depression, or epilepsy (which most advocates support) and trying to eliminate autistic traits themselves.
This perspective raises a practical concern worth considering. If a child learns to suppress stimming, rehearse social scripts, and mask their natural communication style well enough to lose a formal diagnosis, is that a success story or a different kind of harm? Community surveys find broad support for teaching adaptive skills that help autistic people navigate daily life, but strong opposition to normalization, the idea that the goal of intervention should be to make someone appear non-autistic. Many autistic adults and researchers now advocate for well-being and societal reform as primary goals, with skill-building as a tool in service of the individual’s quality of life rather than their diagnostic status.