The adult outcome of childhood quasi‐autism arising following extreme institutional deprivation

Background Rutter and colleagues' seminal observation that extended early life exposure to extreme institutional deprivation can result in what he termed quasi‐autism (QA), informed both our understanding of the effects of adversity on development and the nature of autism. Here we provide the first detailed analysis of the adult outcomes of the group of institutionally deprived‐then‐adopted children identified as displaying QA. Methods Twenty‐six adult adoptees identified with QA in childhood (Childhood QA+) were compared to 75 adoptees who experienced extended institutional deprivation (>6 months) but no QA (Childhood QA−), and 116 adoptees exposed to Low/No institutional deprivation. The outcomes were child‐to‐adult developmental trajectories of neuro‐developmental symptoms (autism, attention‐deficit/hyperactivity disorder (ADHD), disinhibited social engagement (DSE) and cognitive impairment), adult functioning, life satisfaction and mental health. Results Childhood QA+ was associated with elevated and persistent trajectories of broad‐based autism‐related difficulties, ADHD and DSE symptoms and low IQ, as well as adult mental health difficulties and functional impairment, including high rates of low educational attainment and unemployment. Life satisfaction and self‐esteem were unaffected. Autism‐related communication problems, in particular, predicted negative adult outcomes. Childhood QA+ was still associated with poor outcomes even when ADHD, DSE and IQ were controlled. Conclusions Early and time‐limited institutional deprivation has a critical impact on adult functioning, in part via its association with an early established and persistent variant of autism, especially related to communication difficulties. Apparent similarities and differences to non‐deprivation related autism are discussed.


Introduction
Autism is a childhood-onset neurodevelopmental condition characterised by difficulties with social interactions, communications and repetitive and stereotyped behaviours (American Psychiatric Association, 2013).It has a strong genetic component according to twin (Tick, Bolton, Happé, Rutter, & Rijsdijk, 2016) and molecular genetic studies (Glessner, Connolly, & Hakonarson, 2014).Preconception, prenatal and perinatal environmental factors may also be important (Gardener, Spiegelman, & Buka, 2011;Kerin et al., 2018;Kolevzon, Gross, & Reichenberg, 2007;Ornoy, Weinstein-Fudim, & Ergaz, 2015;Wu et al., 2017).Common, post-natal social environmental factors, for instance those affecting the quality of the parent-child relationship, have been ruled out as a primary cause of autism (Rimland, 1985), although complex and subtle interactions with genetic factors may occur (Mandy & Lai, 2016).Complicating the picture, Rutter et al. (1999) reported exceptionally high rates of marked autistic features in children adopted into UK families after they spent extended periods (up to 43 months) in brutally depriving Romanian institutions of the 1980s.This raised the possibility that exposure to extreme and extended negative social experiences, early in life, could lead to autism.
Based on a combination of clinical observations and systematic assessments using validated instruments (Rutter et al., 1999(Rutter et al., , 2007)), 20 children were deemed to meet full diagnostic criteria while an additional eight met clinical thresholds on screening questionnaires (overall 16% of the sample).All had experienced more than 6 months of deprivation.Initial reports suggested that childhood deprivationrelated autism differed from idiopathic autism in several waysincluding an unusual degree of social approach, a relative lack of communication problems, a high degree of intense circumscribed interests and/or abnormal preoccupations, as well as a weaker association with male sex and less persistence between ages 4 and 6 years.Rutter coined the term 'quasi-autism' (QA) to describe this presentation (Rutter et al., 1999).More generally, QA formed part of a highly distinctive, yet heterogeneous and only partially overlapping, set of what were termed deprivation-specific problems (DSPs) (Rutter, Sonuga-Barke, & Castle, 2010) -attention-deficit/ hyperactivity disorder (ADHD), symptoms of disinhibited social engagement disorder (DSE) and cognitive impairment (i.e.IQ < 80).
Follow-up studies into adolescence (Kreppner et al., 2010;Rutter et al., 2007) demonstrated the strong persistence of all DSPs symptoms from age 6 years, including autism symptoms (Kreppner et al., 2010;Rutter et al., 2007) despite extended periods of good quality adoptive care for nearly all adoptees.More recently, the same pattern of symptom persistence was observed into early adulthood for autism and other DSPs (Sonuga-Barke et al., 2017).In terms of autism symptoms, these recent analyses left many scientific questions of clinical significance unanswered.First, it reported only a single measure of aggregated symptoms across the different autism domains (communication, social reciprocal interactions, and repetitive and stereotyped behaviours) and, therefore, was unable to ask questions about the deprivation-related autism symptom profile or which aspects of it drove developmental outcomes.Second, it focused on comparing groups of deprived and nondeprived individuals as a whole and so was unable to conclude anything about what distinguished adoptees who had experienced extended deprivation with and without previously identified clinically significant autistic symptoms (i.e.QA).Third, there was not enough space to give a sufficiently finegrained analysis of the impact of deprivation-related autism on different aspects of adult functioning to help inform clinical decision-making.The current paper aims to address these outstanding issues to establish the developmental significance and clinical status of the different components of childhoodidentified QA vis-à-vis other deprivation specific problems, on the one hand, and idiopathic autism on the other.
To do this, we compared young adults who had experienced extended institutional deprivation (>6 months) and were identified in childhood as having QA (Childhood QA+) with those who experienced the same level of deprivation but were not identified as having QA (Childhood QA−) with a combined group of Romanian adoptees who experienced less than 6 months of institutional deprivation and UK adoptees who experienced no deprivation, both without QA (Low/No Deprivation).These group designations have been used in many ERA analyses previously because: (a) there have been equally low levels of DSPs in non-deprived UK adoptees and Romanian adoptees who experienced <6 months deprivation for all analyses at all follow-up ages, and (b) there was a striking step increase in DSPs rates for those who experienced more than 6 months of deprivation with little further increase in risk associated with each additional month of deprivation (see Golm et al., 2020;Kennedy et al., 2016Kennedy et al., , 2017;;Sonuga-Barke et al., 2017).We addressed the following research questions:

Participants
One hundred and sixty-five Romanian adoptees and 52 comparisons UK adoptees with no history of institutional deprivation, and their families, entered the ERA study in the mid-1990s.Romanian adoptees typically entered institutions in the first few months of life and experienced between 1 and 43 months of institutional deprivation.For the current analysis, as in previous studies, Romanian adoptees with <6 months of institutional deprivation were combined with a group of UK adoptees, also adopted before 6 months, into a Low/No Deprivation group.This is because prior analyses (e.g.Kreppner et al., 2007;Sonuga-Barke et al., 2017) Lord, 2006).A further group of eight children met clinical cutoffs on the Social Communication Questionnaire only (i.e.SCQ total score > 14; Rutter, Bailey, & Lord, 2003).Only 6 of these were included in the analyses as 2 were excluded because of the presence of specific biological risk factors (e.g.foetal alcohol syndrome).There were no significant differences between the 20 individuals who met the ADOS/ADI criteria and the six who just met the SCQ thresholds (see Appendix S1 -Table S1). ERA

Background characteristics
Adoptees' sex, date of birth and duration of institutional deprivation was obtained from Romanian records taken at the time of entry into the United Kingdom.Social-economic status (SES) was based on adoptive parents' occupation at the age 15 follow-up (Rutter & Sonuga-Barke, 2010).Families were divided into high (i.e.skilled, managerial/technical and professional occupations) and low (i.e.manual and unskilled occupations) status (General Registrar Office, 1971).Parental marital status derived from young adult reports was coded marriage intact = 1, divorced/separated/widowed = 0.

Deprivation-specific problems assessed at all followups
Autism symptoms.The parent reported SCQ (Rutter et al., 2003) was used to assess autism symptoms across the three core domainssocial reciprocity, communication, and repetitive and stereotyped behaviours.For our analysis we selected five items from each domain which were developmentally appropriate across all assessment waves.These can be seen in Appendix S2 -Table S2 and S3 (see Sonuga-Barke et al., 2017 for the rationale for this approach).Items were rated 0 for 'absent' or 1 for 'present'.
ADHD symptoms.In order to ensure developmental consistency across assessment waves, hyperactivity, sustained attention and distractibility were measured using equivalent items from the Revised Rutter Scale (Elander & Rutter, 1996) at ages 6 and 11, the Strengths and Difficulties Questionnaire (SDQ; Goodman, 1997) at age 15, and the Conners Comprehensive Behaviour Rating Scale (CBRS; Conners, Pitkanen, & Rzepa, 2011) in young adulthood.For the Revised Rutter Scale and the SDQ, a symptom was deemed endorsed when a rating of 2 ('certainly applies') was made (0-2 scales).The equivalent rating ('often/very often', rating of 2 or 3) was found in the CBRS (0-3 scale; see Sonuga-Barke et al., 2017 for rationale).
DSE.At all ages this was based on researcher ratings of parents' responses to three interview questions.These tapped the constructs of being too friendly, showing inappropriate intrusiveness and being unaware of social boundaries.A rating of 'definite evidence of disinhibition' represented a positive endorsement.

IQ. The McCarthy Scales of Children's Abilities General
Cognitive Index (McCarthy, 1972) was used at age 6, the short form of the Wechsler Intelligence Scale for Children (WISC; Wechsler, 1974; i.e. block design, object assembly, vocabulary and similarities) at ages 11 and 15 years.The block design and vocabulary subscales from the short-form Wechsler Abbreviated Scale of Intelligence (WASI) (Wechsler, 1999) were assessed in adulthood.

Mental health and wellbeing in adulthood
Psychopathology.Generalised anxiety, major depres- sion, oppositional defiant disorder, manic episode, social phobia, conduct disorder and obsessive-compulsive disorder symptoms were assessed through parental report by means of the CBRS (Conners et al., 2011) using T-scores on a 0-3 scale.Parental report was chosen over self-report because more parents than adoptees completed these measures.However, parent-and adoptee self-report of emotional problems were similarly elevated in young-adulthood in the High Deprivation group (see Sonuga-Barke et al., 2017).T-scores falling below 60 suggest the absence of concerns.
Mental health service use.Young-adult adoptees reported mental health service use since mid-adolescence during an interview.A score of 1 (0-1 scale) indicated at least a two-session contact with a mental health professional that resulted in either psychiatric diagnosis or the conclusion that there was a clinically significant disorder.
Self-esteem.The Rosenberg self-esteem scale (Rosenberg, 1965), containing 10 self-rated items measuring positive and negative feelings ranging from 'strongly agree' (4) to 'strongly disagree' (1).Total scores between 15 and 25 are within the normal range.

Education, employment and daily functioning in adulthood
Education.

Academic achievement. Parents reported Maths and
English difficulties using the CBRS (Conners et al., 2011).Tscores falling below 60 suggested no concerns.
Employment.Adoptees reported their current employment status (Angold et al., 1999).A score of 0 denoted not "in employment, formally enrolled in full-time education, or training".
Daily functioning.The RAPFA interview (Hill et al., 2008) with the young people assessed day-to-day functioning in relation to finances, household tasks, and daily routines: 0 denoted 'manage independently' and 1 'unable to manage or manage with support'.

Relationships measured in adulthood
Romantic relationships.Adoptees reported whether they had ever been in a romantic relationship ( Offspring.Adoptees reported if they had had a child.A score of 0 (0-1 scale) indicated 'never had a child' (Angold et al., 1999).
Relationship with parents.

Statistical analysis
To answer questions 1a (Does the Childhood QA+ group display persisting symptoms in all three autism core domains (social reciprocity, communication, and repetitive and stereotyped behaviours) across adolescence through to adulthood?) and 1b (Does the Childhood QA+ group display elevated trajectories of other DSPs into adulthood i.e.ADHD, DSE and cognitive impairment), the intercept and linear and quadratic slopes of growth trajectories across ages 6, 11, 15 years and youngadulthood were estimated using Latent Growth Models (LGM), which were subsequently included as dependent variables in one-way ANOVAs with group (Low/No Deprivation vs. Childhood QA− vs. Childhood QA+) as the independent variable.Post-hoc testing was performed, with Bonferroni correction, to determine the specificity of significant group effects.A crosssectional comparison of deprivation specific problems at young-adult follow-up with group as the independent variable was also performed.To answer question 1c (Does the Childhood QA+ group experience poor mental health, impaired daily functioning, reduced life satisfaction and self-esteem in adulthood), we conducted group comparisons using ANOVA for continuous outcomes and χ 2 for binary outcomes.We used a similar approach to address the remaining questions.Both involved using Multiple Linear Regression.For question 2 (Is the association between childhood QA and adult functioning and mental health independent of the effects of the other deprivation-specific neurodevelopmental problems -DSE, ADHD and cognitive impairment, expressed from childhood to adolescence?), we included mean childhood to adolescence ADHD, DSE, and IQ scores and QA membership (Childhood QA− and Childhood QA+) as predictors, and adult mental health and functional impairment as dependent variables.For question 3 (Which autism symptom domain (social reciprocity, communication, and/or repetitive and stereotyped behaviours) is most strongly associated with the negative impact of QA on adult outcomes (i.e.mental health and functional impairment)?),we included mean childhood-to-adolescence SCQ domain scores as predictors, and adult mental health and functional impairment as dependent variables.Because we were specifically interested in the predictive effects related to deprivation, we restricted our analyses to those who experienced more than 6 months of institutional deprivation.
Prior to these regression analyses, two sets of principal component analysis (PCA) reduced the number of dependent variables measuring mental health problems and functional impairment into two separate individual scores for mental health problems and functional impairment, respectively.The PCA analysis for the mental health factor included parent-rated CBRS T-scores for major depressive disorder, manic episodes, generalised anxiety disorder, social phobia and obsessivecompulsive disorder (Bartlett's test of sphericity: χ 2 (10) = 265.924,p < .001;Kaiser-Meyer-Olkin (KMO) = .82;Total variance explained = 74.21%).The second PCA for the functional impairment factor included self-reported RAPFA scores for daily functioning skills, household tasks, finances management, living independently, education, and employment status (Bartlett's test: χ 2 (15) = 220.646,p < .001;KMO = .81;Total variance explained = 56.85%).For both regression models, single predictor variables (SCQ scores, ADHD, DSE and IQ) were created by averaging outcomes at ages 6, 11 and 15 years.All tests, except LGM, were run using SPSS version 26.For LGM we used MPlus version 8.7.

Results
In adulthood, Childhood QA+ and Childhood QA− did not differ in terms of sex ratio, age, length of time in institutions, SES or whether parents are still living together.Both groups had more girls, were older and had spent more time in institutions than the Low/No Deprivation group (see Table 1).
Is QA associated with a persisting pattern of symptoms?
The intercept value was greater for Childhood QA+ than Childhood QA− and Low/No Deprivation groups with Childhood QA+ being associated with significantly higher levels of symptoms on all autism subscales, ADHD, DSE and lower IQ (Figures 1 and  2; Table 2).The groups also differed in terms of the slopes with a relative increase in communication difficulties and a relative decrease in repetitive and stereotyped behaviours for the Childhood QA+ group compared to the other two groups.An examination of the slope values suggested that both Childhood QA+ and Childhood QA− groups saw a relative worsening of DSE compared to the Low/No Deprivation group.Models for all variables failed to converge when the quadratic term was added to them.In young adulthood the difference between the groups was still significant for communication, repetitive and stereotyped behaviours, ADHD, DSE and lower IQ (Appendix S3 -Table S4).
Is QA associated with more negative adult outcomes?
Childhood QA+ was associated with impaired functioning and psychopathology (see Table 1).Eight times as many Childhood QA+ adults were registered disabled than their Childhood QA− counterparts.They were twice as likely never to have lived independently and reported higher functional impairment across all domains (i.e.routines, household tasks and finances).They had significantly lower levels of educational attainment and substantially higher rates of unemployment.They were less likely to have had a romantic relationship (see Table 1).Individuals in the Childhood QA+ group were twice as likely to have used mental health services as their Childhood QA− counterparts.They displayed elevated scores across psychopathology categories compared to Childhood QA− (Table 3).Interestingly, the groups did not differ in the quality    -Table S6. of their relationships with family or in life satisfaction and self-esteem (see Table 1).

C o g n i t i v e i m p a i r m e n t
Is QA associated with worse outcomes after controlling for other deprivation specific problems?
Which autism sub-domains drive adult problems?

Discussion
Rutter's discovery that a considerable proportion of children exposed to extreme, global institutional deprivation showed increased rates of autism symptoms raised challenging questions about both the causes of autism and the impact of early maltreatment (Rutter et al., 2007).We present the first detailed and comprehensive longitudinal analysis of the adult outcomes of these young people.There were a number of notable findings.First, there is strong continuity of autism symptoms into adulthooddespite the adoptees being raised in well-functioning families for over 20 years.This was also seen for deprivationrelated ADHD and DSE (Kennedy et al., 2016(Kennedy et al., , 2017)).It is consistent with the notion that exposure to institutional deprivation during early development can produce deep-seated neuro-developmental alterations, which, after a period of rapid catch-up in the years soon after adoption, are largely insensitive to later environmental enrichment.Second, QA manifests across all three autism domainsalthough there are subtle differences in developmental trajectories through to adulthood: Communication problems worsened somewhat, while repetitive and stereotyped behaviours, although persistent, improved to a certain degree.Worsening communication problems is especially notable since, according to Rutter et al. (1999), a relative lack of communication problems was a distinctive feature of QA in the early years.Regarding repetitive and stereotyped behaviours, as previously mentioned, QA had a distinctive pattern limited largely to intense circumscribed interests and/or abnormal preoccupations (Rutter et al., 1999).Although the adapted repetitive and stereotyped behaviours scale used here did include some ritualistic and stereotype behaviours, these were not sufficient for a detailed analysis of this issue.Previous longitudinal studies of the adult outcomes of idiopathic autism suggest that, where symptom improvements occur, it is across all domains, including communication problems (Howlin, Moss, Savage, & Rutter, 2013), highlighting a possible difference between deprivation and non-deprivationrelated autism.Finally, interestingly, elevation of autism symptoms was not limited to the Childhood QA+ group: those in the Childhood QA− group also had elevated levels of autism symptoms compared to the Low/No Deprivation group.This supports the notion that QA is more a continuum than a category and we should consider sub-threshold symptoms when looking at clinical profiles.
Third, there was substantial overlap between QA and the other deprivation-specific problems (ADHD, DSE and cognitive impairment) in adulthood similar to that seen in adolescence (see Kreppner et al., 2010).While ADHD (Panagiotidi, Overton, & Stafford, 2019) and cognitive impairment (Howlin, Savage, Moss, Tempier, & Rutter, 2014) commonly coexist with autism in non-deprived adult samples, DSE appears particularly characteristic of adult QA.This conforms to Rutter's initial observation.Some studies have also observed a combination of DSE and autism in children with a history of abuse and/ or neglect (Mayes, Calhoun, Waschbusch, & Baweja, 2017;McCullough, Stedmon, & Dallos, 2014;Pritchett, Pritchett, Marshall, Davidson, & Minnis, 2013;Sadiq et al., 2012).The persistent co-occurrence of DSE and autism, as well as ADHD, in adults with a history of childhood maltreatment raises questions about the causes driving this overlap and the possibility that together they may constitute the core of what might be called neurodevelopmental deprivation 'syndrome'.
Fourth, QA is associated with a profound set of challenges in adulthood as manifested by the high levels of mental health problems and functional impairments observedeffects not explained by overlapping deprivation-specific problems.These findings highlight the functional impact of QA and the associated clinical needs of this group.Recent studies of the adult outcome of individuals with a childhood diagnosis of idiopathic autism show a similar picture with poor outcomes in employment and education (Toft et al., 2021), independent living (Billstedt, Gillberg, & Gillberg, 2011) and mental health (Lever & Geurts, 2016).A direct comparison with a cohort of non-deprived autistic individuals is needed to calibrate the relative severity of these outcomes.
Lastly, through all their life challenges, relationships with parents remained on average good, The adult outcome of quasi-autism following institutional deprivation although the ability to form romantic relationships seemed to be negatively impacted.The QA adoptees' sense of life satisfaction and self-worth was strong.This contrasts with the findings from studies of adults with idiopathic autism whose outcomes are typically negatively impacted (Nguyen, Ownsworth, Nicol, & Zimmerman, 2020;Schmidt et al., 2015;Soares et al., 2021).Possible explanations for these differences requiring further exploration include special qualities of ERA families including parenting (Castle, Beckett, Rutter, & Sonuga-Barke, 2010) and resilience (Shmotkin, 2005) and/or cohort differences in terms of IQ, given it has been suggested that autistic individuals with higher IQ have better awareness of their limitations and deficits (Huang et al., 2017).
The high rates and strong persistence of autism in the ERA study appear to support a causal role for institutional deprivation in QA.Other explanations of the statistical link though need to be considered.It is possible for instance that there were extremely high rates of idiopathic autism risk in the families placing their children in the institutions.However, although poverty-related pre-and perinatal autism risk factors were likely common in these families, based on current estimates of effect sizes of such risks in the literature, they could not possibly account for the rates of QA seen in ERA (Mayes et al., 2017).Finally, the distinctive features of the QA group would suggest that this was not typical autism.It is also possible that families selectively placed children most at risk for autism in the institutions.However, this is highly improbable given that placements were typically made within the first months of life, where signs of early autism would be extremely difficult to spot.
Despite the methodological strengths of the ERA study there are some limitations.First, there was no comprehensive clinical assessment of autism in adulthood and so diagnostic status could not be established.Second, in order to ensure comparability across ages the assessment of autism symptoms using the SCQ relied on a restricted range of items within each domain.This meant that the assessment used at a particular age may have missed important developmentally relevant characteristics of QA.Third, there was considerable attrition between the adolescence and adult waves which reduced statistical power.However, there was sufficient power to show QA-specific differences across most outcomes where effect sizes were moderate.The possibility that we missed some less pronounced but still important difference cannot be ruled out.
Our findings have clinical implications.First, since deprivation-related autism is associated with poor outcomes in adulthood, it is imperative that autistic adoptees who have suffered from early-in-life deprivation receive adequate specialist support during the adolescence-to-adulthood transition (Anderson, Newlove-Delgado, & Ford, 2022).Second, measuring exposure to early deprivation in children with autism during clinical assessment seems important given the differences between QA and idiopathic autism.Third, we should reflect on whether QA still stands as the right term to refer to autism symptoms following early deprivation.Using the term might lead some clinicians to see the associated problems as idiosyncratic and transitory, limiting adoptees access to evidence-based interventions for autism (Mayes, Breaux, Calhoun, & Whitmore, 2019).Fourth, the question of whether other forms of maltreatment can create risk for QA needs further consideration.QA has been observed in other cohorts of children exposed to less severe institutional deprivation (i.e.Bucharest Early Intervention Project; Levin, Fox, Zeanah, & Nelson, 2015), and in children adopted after early disrupted care and maltreatment (Green, Leadbitter, Kay, & Sharma, 2016).
In summary, QA is persistent and complex and impacts the adult lives of the individual it effects in profound ways.Clinical services should not consider this form of autism as transitory or benign.
Adoptees completed theInventory of Parent and Peer Attachment (IPPA;Armsden & Greenberg, 1987): a 25-item self-report questionnaire that taps into the construct of degree of mutual trust (10 items), quality of communication (9 items) and extent of anger and alienation (6 items).The IPPA uses a 5-point Likert scale (1 = almost never true to 5 = almost always true).

Figure 1
Figure 1 Developmental trajectories for autism symptoms.Error bars represent standard error of the mean.SCQ, Social Communication Questionnaire; YA, young adulthood (average age 23 years)

Ó
2023 The Authors.Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for Child and Adolescent Mental Health.doi:10.1111/jcpp.13767 1. Compared to the Low/No Deprivation and theChildhood QA− groups, does the Childhood QA+ e. mental health and functional impairment)?
study retention was high up to age 15 years (data available from 90% of the sample) but decreased into adulthood Ó 2023 The Authors.Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for Child and Adolescent Mental Health.

Table 3
Comparison of young adult adoptees' mental health and academic problems by group Ó 2023 The Authors.Journal of Child Psychology and Psychiatry published by John Wiley & Sons Ltd on behalf of Association for Child and Adolescent Mental Health.