When I was training to be a clinical psychologist, telling parents that their child had autism was a regular part of my job. Now that I’m a parent, I understand better the pained expression that came over their faces as they contemplated this news. Among the many questions taking shape in their minds, I can imagine the one looming largest: Could their child ever be like other children?

A recent study, published in February in theJournal of Child Psychology and Psychiatry, suggests that for some people, the answer is yes. The researchers found that some individuals who had been diagnosed with autism as young children no longer had symptoms—such as difficulty interacting and communicating with others, rigid adherence to rituals and routines, and repetitive movements of their bodies and objects—when they were older.

This finding is not the first to suggest that some young adults with autism lose their symptoms. A 2008 literature review reported that 3 to 25 percent of affected people eventually recover. But the recent study was especially rigorous. The researchers recruited three groups of eight- to 21-year old subjects: 34 of them had apparently recovered from autism, 44 had high-functioning autism, and another 34 were control subjects with no developmental issues.

An expert diagnostician thoroughly reviewed the early records of all recovered participants to confirm that they truly had autism when they were younger, and she correctly rejected 24 reports from kids with nonautism diagnoses (such as language disorders) that had been slipped in as foils, verifying that her diagnostic technique was sound. These measures made researchers confident that the now typically functioning children had not initially been misdiagnosed. The team also set a relatively high bar for recovery: participants not only had to be free of autism symptoms, as indicated by a battery of tests—they also had to have typically developing friends and be fully included in regular education classrooms.

The study’s findings were encouraging: those who recovered were on a par with the typically developing individuals and better than the group with high-functioning autism in their social and communication skills and in their ability to go about daily life, such as taking care of themselves and doing housework. These findings, according to some experts, represent a watershed moment in autism research, “clearly demonstrating the possibility of leaving behind the symptoms of [autism] and emerging into a state of healthy functioning,” writes University of California, Davis, psychologist Sally Ozonoff, who was not involved in the study.

Not Autistic, Not Quite Typical

As exciting as these results might be, I believe we should interpret them with caution. For one thing, about 20 percent of those in the recovered group still showed mild difficulties with eye contact, gestures and facial expressions. According to the study authors, “these difficulties were judged … not to have an autistic quality.” But those doing the judging were not “blind,” as researchers are in rigorous clinical trials—they knew if a participant was typically developing, had autism or had supposedly recovered.

It is easy to imagine how this knowledge could have colored their interpretations of participants’ behaviors. Even if those in the recovered group did actually lose their symptoms, “the majority still have serious issues,” including depression, anxiety and inattention, notes Elizabeth Kelley, a psychologist at Queen’s University in Ontario and one of the study’s authors.

Kelley points out another limitation of the study: it looked back at recovery after it had happened. The retrospective design cannot reveal what proportion of kids will shed their diagnosis or why. Parents often try a variety of interventions, including behavioral treatment, speech and occupational therapy, and medication—and they do not always keep detailed records. Until researchers report on the outcomes of children they starting following very early in life, “we have no idea why some people recover,” Kelley says.

Fortunately, scientists are starting to do those studies. For example, Catherine Lord, director of the Center for Autism and the Developing Brain at Weill Cornell Medical College, has been following a group of about 100 people with autism from the time they were diagnosed at age two through their early 20s. Study participants completed a large battery of tests every few years as children and again at age 18, and parents have been filling out questionnaires every year.

Like Kelley and her colleagues, Lord has found that a handful of participants lose their autism symptoms. Moreover, she says, “their eye contact, gestures, the way they hold their body, the way they talk about their friends”—behaviors that have long been thought to be difficult to improve on—are indistinguishable from those of typically developing adults. They are also functioning well in daily life, holding down part-time jobs while attending college. The researchers fittingly refer to this group as having a “very positive outcome.” A more sizable group is considered “more able” than the remaining adults in the sample—they have no cognitive impairment and are generally doing well academically, although they still have clear autism symptoms. A paper presenting these results is currently under consideration at a peer-reviewed journal.

Who Recovers and How?

Although Lord’s study has not resolved the question of why some people with autism improve radically, it has found early signs that may help identify those who will. For example, at a very young age, individuals in the “very positive outcome” group had rapid gains in verbal skills and decreases in restricted and repetitive behaviors, such as flapping their hands and lining up toys. Parents whose children show these early improvements have reason to be hopeful about prognosis.

Interestingly, neither Lord nor Kelley found that those with the best outcomes received more behavioral treatment than the others, as one might expect. This finding does not mean that behavioral treatment is ineffective for autism—in fact, many studies suggest the opposite is true. Researchers simply do not yet know how the amount and type of treatment relate to prognosis. For example, applied behavioral analysis, which focuses on using reinforcement to help children learn and attend to another person, could be especially effective for some youngsters, whereas social skills training, which emphasizes capabilities such as holding a conversation and turn taking, might help others. Similarly, some children might require much more intensive intervention to make gains. The unpredictable effect of treatment could be related to underlying genetic differences; autism most likely has many variations, rather than a single genetic cause.

Until there are more definitive answers, Kelley says, “parents should do as much as they can for their individual child,” within their means. She firmly cautions parents against bankrupting themselves or running themselves emotionally ragged trying to get help for their child, because there may be factors leading to better outcomes that are completely outside parental control. “It may be a genetic variant—who knows?” Kelley remarks.

As exciting as these positive outcomes are, we do not yet know if they will last. As people with autism begin to face the challenges of adulthood, old symptoms might recur and new ones could surface. Studies of adults with autism have consistently found that many struggle to live independently, get and keep a job, and form friendships and romantic relationships. Lord and her colleagues want to keep following study participants to see if those in the “very positive outcome” group continue to flourish.

Of course, people with autism can thrive even if they don’t recover. Lord recalls an adult with autism who put it well: “If I’m independent, have a good job and have relationships with other people, why am I not as good as someone else?” Researchers and clinicians alike should keep working to understand how all people who have autism can achieve the best possible outcome: a happy, fulfilling life.