In one of the earliest studies of post-childhood outcome for adults with ASD (termed “infantile psychosis” in the original publication), Rutter et al. (1967) examined adolescent outcomes for 63 children who had been identified between 1950 and 1958 through clinical and educational programs at the Maudsley Hospital Children’s Department. About one-fifth of the sample experienced a developmental regression in early childhood. The sample included children at all levels of intellectual ability, with 43 % having severe intellectual disability, and 29 % having IQs in the near-normal or normal ranges. Like other studies from this period, half of the participants were institutionalized at the time of the outcome assessments. The investigators noted prognosis for these individuals was poor as only 17 % could be “said to be well adjusted.” One person was described as having “normal” adult functioning, and eight more were doing relatively well in regard to achieving some independence in adulthood. Outcomes were Poor or Very Poor for 61 % of the sample. Unlike children today whose access to public education is protected by legislation, 21 of the individuals in this sample never attended any school at all, and less than half of the children had as much as 2 years of formal schooling.

Lotter (1974) followed 32 individuals who were identified through an epidemiologic survey in Middlesex, England when they were 8–10 years old. The mean IQ for these individuals in childhood was 71, with a range of 55–90. Eight years later, one person had passed away, and two were lost to follow-up. Sixty-two percent of those remaining were described as requiring “extensive care and supervision.” Outcomes were rated as Good for 14 % and Poor or Very Poor for 60 %. Half were institutionalized. Five of the individuals in the Very Poor outcome category had been excluded from any educational services as they grew up, and Lotter reported that individuals’ placements over time, whether in schools or institutions, were associated with their outcome status.

A decade passed before another outcome study was published (Rumsey et al., 1985). This retrospective study concerned 14 men with a mean age of 28 who were recruited through a national search for eligible adult males with DSM-III autistic disorder in the US. Individuals with seizure disorders, identifiable causal medical conditions, or who were unable to discontinue current medications were excluded. The men were studied intensively in an inpatient setting over 5 days. Nine were “unusually highly functioning” with nonverbal IQ scores above 80 and well-developed language abilities. Stereotyped movements were observed on the unit in 86 % of the participants. Basic achievement was commensurate with measured IQ scores, but measures of adaptive functioning were far below expected, given the men’s IQs, which is a robust finding from subsequent ASD research. Only one man was in a state institution.

Gillberg and Steffenburg (1987) studied outcome for a population-based sample of 23 people with ASD in late adolescence or early adulthood. As children, one-third obtained IQ scores in the mildly intellectually impaired range, and 26 % achieved scores in the normal or near-normal ranges. One-third had communicative speech at age 6. One person achieved a Good outcome and 44 % had Poor or Very Poor outcomes. Childhood IQ and use of communicative speech at age 6 were useful predictors of outcome status.

Kobayashi et al. (1992) conducted a follow-up investigation of 201 adolescents and adults identified with ASD in childhood through clinical services in Japan. Four of the people had died. The mean age for the remaining 197 young adults was 21. About one-fourth of the sample had an IQ score of 70 or better at age 6, and about 20 % were able to speak without echolalia at that age. Forty percent of the sample began school in a general education class, but only 27 % remained in general education at the age of 12. Outcome adjustment for roughly one-fourth of the sample was Good or better, and was Poor or Very Poor for 46 %. Childhood IQ was the only strong predictor of outcome in this investigation. Although there were similarities between the sample in this study and others reported, the outcome for these participants was better than in previous studies. The authors provided some possible explanations including sociodemographic factors in Japan, advances in public education standards for people with disabilities, intensive intervention histories, and a high proportion of people with ASD and average-range IQ scores at baseline.

In 1996, Ballaban-Gil, Rapin, Tuchman, and Shinnar reported on their study of outcome in 54 adolescents and 45 adults with ASD after an 18-year period. The follow-up procedure consisted of a 30-min telephone interview with caregivers for the members of the sample. Three people had died at the time of the follow-up. Two were described as no longer having any social deficits. Behavioral difficulties were reported for 69 % of the sample, and 40 % of the adults were prescribed medications aimed at controlling behavior. Thirteen adults were described as high-functioning.

Howlin, Mawhood, and Rutter (2000) compared outcomes for 19 men diagnosed in childhood with autism. Participants were 7–8 years old at the time of the childhood assessment and were identified through their involvement in hospitals or special school programs in the community. They were 23 years, 9 months old on average at the time of the adult assessment. Roughly three-fourths of the men continued to exhibit severe social difficulties in adulthood. Only one-fourth was rated as exhibiting minimal or no “autistic-type behaviors.” Just over half of the men relied on others to schedule and organize leisure activities for them, and one-third was described as having no or very limited interests or leisure activities. Three-fourths of the sample experienced a Poor or Very Poor outcome and 16 % experienced a Good outcome or better. Analyses of childhood variables that were associated with adult functioning indicated that early language skills for these men were highly related to social functioning in adulthood.

In 2003, Engstrom, Ekstrom, and Emilsson reported on the outcome for 42 Swedish adults identified through public health service records with Asperger’s disorder or so-called high-functioning autism. Eligible participants were over 18 years of age and had adult IQ scores of 70 or higher. Participants included 24 men and 18 women ranging in age from 18 to 49 years. Ten individuals had autism diagnoses, and 32 had Asperger’s disorder diagnoses. The participants with Asperger’s disorder had been diagnosed at a mean age of 28 compared to age 13 for the group with autism. From among the full sample, 16 were selected as a representative subsample that included people from both diagnostic groups, genders, and a range of ages. Participants with ASD and care providers responded to in-depth interviews designed to obtain a picture of the adult situation for these men and women. Only one adult required no support from others. Almost all participants were supported through pensions from the state. While virtually all of the sample required a high level of support, the authors noted this was a selected sample of adults with ASD who had been identified through public service records as having an ASD.

Howlin et al. (2004) studied adult outcome for 68 people with ASD who also had a childhood nonverbal IQ score of 50 or better. The mean age at the initial evaluation was 7 and at follow-up was 29 years. Average nonverbal IQ scores in adulthood were slightly below childhood scores. Almost all of the subjects were known to have attended compulsory schooling; however, only 22 % left school having achieved formal qualifications. At the time of the follow-up investigation, 23 people were employed. Eight worked in regular, independent jobs; one was self-employed as an artist but was unable to earn a living wage; and 14 worked in sheltered or supported employment. Twenty-seven people were occupied in general work/leisure programs at day centers for adults with disabilities. Outcome adjustment ratings for the sample included Good or Very Good outcomes for one-fourth of the sample and Poor or Very Poor outcomes for roughly 60 %. Childhood IQ was a useful predictor of adult adjustment in that those with childhood nonverbal IQ scores of 70 or more functioned more independently than those with scores below 70. A score of 100 or better did not increase the likelihood that a person would do well in adulthood. For those capable of completing a childhood verbal IQ measure, the combination of verbal and nonverbal IQ scores in childhood provided a more precise indication of outcome classification, with scores above 70 in both domains yielding the greatest likelihood of a Fair outcome or better. Language level at age 5 was useful in predicting overall outcome and residential status, but none of the other outcome variables studied had predictive utility.

Billstedt et al. (2005) followed up with members of three population-based studies of autism in Sweden, for a total sample of 120, after a period of 13–22 years. Participants ranged in age from 17 to 40 years and included 84 men and 36 women. Six participants had died at the time of the follow-up assessment, and only six participants from the original sample declined to participate. Outcomes were rated as Poor or Very Poor for 78 %. No participant was rated as having an outcome better than Fair. Only four adults were independent, and these individuals were described as “leading fairly isolated lives.” Sixty-two of 73 participants examined for current diagnostic status continued to meet criteria for autistic disorder, and one person no longer met diagnostic criteria for any ASD. An overall downward shift in measured intellectual ability was reported. In childhood, 46 % of the sample had scores in the severe range of intellectual disability compared to 71 % at follow-up.

Eaves and Ho (2008) followed 48 individuals with ASD from childhood to adulthood in Canada. Eight of the participants had a childhood IQ score above 70. All participants received special education support in childhood, and 30 % engaged in some kind of post-secondary educational program. Overall outcome adjustment ratings were that 21 % had Good or Very Good outcomes and 46 % had Poor outcomes. Almost 80 % received a government disability pension and used the services of social workers. Also in 2008, Cederlund et al. released their study of outcome for 70 adults with autism and 70 adults with Asperger’s disorder. Twenty-seven percent of this sample obtained an outcome categorization of Good, and only two people fell within the Poor category. There were no participants with Very Poor outcome ratings.

Our group (Farley et al., 2009) studied 41 adults who had been identified through a population-based study of ASD in Utah in the 1980s. All of these individuals had historical IQ scores of 70 or greater. Mean age at the first assessment was 7 years and in adulthood was 32 years. Outcome adjustment was better for this sample than previous samples. No systematically collected prognostic factors could be found to explain the difference, but anecdotal information suggested that the relative advantages experienced by the sample we studied over others could be related to the social supports experienced by most of the sample who were members of the Church of Jesus Christ of Latter-Day Saints (LDS Church) in Utah. LDS Church members tend to have large families and organize their religious communities according to the geographical location of their residences, so that in areas that contain a high density of LDS Church members children attend school and church activities with their neighbors. The members of individual congregations, therefore, tend to grow up having frequent interactions with the same community of individuals, with numerous inter-familial relationships due to family size. Participants in our sample routinely reported having found work, friendships, and roles in social groups through their relationships with other members of their church group.

A Canadian sample of 80 adults was identified through clinical diagnostic records, and outcomes were examined through chart reviews (Marriage et al., 2009). The investigators examined outcomes for 45 adults who were diagnosed before 18 years of age compared to 35 adults who were diagnosed in adulthood. As is often the case, individuals with comorbid intellectual disability experienced limited outcomes in adulthood. For those with normal range IQ scores, no differences were detected when comparing those diagnosed before 18 versus those diagnosed later. The investigators looked at a subgroup of individuals with normal range IQ scores who were over 25 years of age. Dividing this subgroup further based on whether they were diagnosed before or after age 18, they found that group diagnosed in adulthood had achieved more in terms of employment and independent living status. The authors suggested that the differences between these groups may be attributable to the fact that these subgroups had a 10-year age difference (average age of 29 for those diagnosed in childhood compared to an average age of 39 for those diagnosed as adults) that may have accounted for the relative increase in achievement levels for those diagnosed as adults. The authors speculated that perhaps having an additional 10 years to develop and achieve normative goals of adulthood offered an advantage to one group.

In a recent study of 20 US adults first examined under 4 years of age, Gillespie-Lynch et al. (2012) analyzed outcomes at an average age of 26 years. This was the fourth data collection point from this sample, with others occurring at average ages of 11 and 18 years. On average, participants had an average mental age of 2 years when they were almost 4 years old and 8 years when they were 18 years old. There was a trend toward reduction in ASD symptoms and improvement in adaptive functioning scores over time. Outcomes were rated as Very Good or Good for 30 % and Poor for 50 %. Early childhood language ratings and IQ scores predicted adaptive functioning in adulthood for this sample. A unique strength of this study was the nature of systematic data collection in early childhood that included specific metrics on the use of joint attention communication strategies. Initiation of joint attention, a voluntary communicative behavior, was not associated with adult variables, but response to joint attention, an involuntary communicative behavior, predicted adult social skills, ASD symptoms, and nonverbal communicative behavior.