The microsegmentation of the autism spectrum: research project

3 Scoping Exercise

3.1 The first phase of the study comprised the carrying out of a scoping exercise involving the identification of preliminary questions, the identification of key issues from the current literature, the exploration of different models to serve as a basis for economic analysis and the preparations required for a fieldwork study.

3.2 A considerable initial focus of this phase was the issue of how to approach the key question of prevalence and establish robust findings applicable to Scotland. The outcomes of this work in terms of a systematic review and meta-analysis of world prevalence of ASD in order to establish a more accurate foundation for economic analysis are covered in Chapter 4.

3.3 In addition to the above, and on the basis of searching and reviewing a very extensive literature, the scoping exercise identified the following preliminary questions. 1 What does research evidence tell us about outcomes and life trajectories in ASD? 2 What are the main co-occurring conditions of ASD, other associated features of the ASD profile and any other factors relevant to outcomes or acting as moderators of outcome? 3 How do the various outcomes and life trajectories in ASD translate into economic implications? 4 How do these economic implications map on to the population of Scotland? 5 What is the relationship between outcome and type of intervention received?

3.4 Preliminary investigation of these questions highlighted not only the complexity of this subject but also the extent to which the research literature, despite its vast magnitude, does not at the present time provide clear answers to almost any of the preliminary questions in our investigation except in broad and general terms.

3.5 Regarding Question 1, ‘What does the literature tell us about outcomes and life trajectories in ASD’, there is now a significant body of outcomes studies for the autism spectrum. The quality of these studies has improved in terms of relevance and validity over the passage of a considerable period of time. ASD in itself is still a ‘young science’ in terms of its research history. Autism was first described by Kanner (1943) and simultaneously by Asperger in a dissertation lodged in the same year (Asperger, 1943) and re-published as a journal article the following year (Asperger, 1944/1991). There was a limited amount of systematic research until the 1970s, and it was not until 1979 that the wider presentation of autism as a spectrum disorder based on the triad was proposed (Wing & Gould, 1979). Recognition of Asperger’s Syndrome as a separate diagnosis within the spectrum may be attributed first to Wing (1981), with entry into the international classification systems not taking place until the 1990s (American Psychiatric Association, 1994; World Health Organization, 1992).

3.6 The fact that the history of autism is relatively brief has raised three related issues regarding the development of a robust and detailed evidence base for outcomes. First, the earliest time at which outcome studies of the whole spectrum could be conducted has only been in the last few years. Knowledge of outcome is dependent on follow-up of a sample of people who have already had a reliable diagnosis, and diagnoses that covered the full spectrum were not available in the early period of the history of autism. Second, it was not only longitudinal studies that had to wait for the opportunity of reliable outcome investigations being conducted but also cross-sectional studies. The early interest in autism was focussed almost entirely on child populations, and it was many years before even the issue of separating childhood autism from adult schizophrenia was settled. Thus, the main academic journal in the field of autism, later to become the Journal of Autism and Developmental Disorders, remained the Journal of Autism and Childhood Schizophrenia until 1979. Also, the main organization for autism in the Scottish voluntary sector, Scottish Autism, remained the Scottish Society for Autistic Children until 1998, and a similar change was seen in national societies for autism elsewhere in the world. The overwhelming bulk of research studies were therefore studies of children, with little early work on adults with autism.

3.7 These two issues in themselves restricted the number of outcome studies available at an earlier stage for ASD. The third related issue defined the nature of such outcome studies as did become available. Restricted diagnostic criteria, the way criteria are applied, the availability of diagnostic services, the level of diagnostic expertise available and various other factors have determined that the earlier the date at which the diagnosis was given the more severe the sample is likely to be. Thus, many of the available outcome studies, and especially those that involved follow up over a lengthy period, have illustrated the levels of outcome expected in individuals with more severe conditions. This is reflected in some of the outcomes reported below.

3.8 Kanner (1971) published a 28-year follow-up of his original sample of 11 children, which, together with Asperger’s small sample from the same period, constitutes the earliest known sample of individuals with autism. Although Kanner’s children might have been viewed as being most likely a high functioning sample – he believed they were ‘all unquestionably endowed with good cognitive potentialities’ and they all came from ‘highly intelligent families’ (Kanner, 1943, pp.247-248) – their outcomes were on the whole poor. Five were in institutional care, another was mute, one of those who developed seizures died aged 28 and of two little or nothing was known. Only two would have met outcome criteria normally viewed as good; they lived at home with their parents and had regular employment.

3.9 In general, the early outcome studies bring the disabling aspects of autism to the fore, since they highlight the extent to which adult outcomes have been on the whole poor, with generally only a small minority experiencing independence in areas such as living arrangements, employment or relationships, and with life trajectories marked by high needs for care and service provision (Gillberg, 1990; Lotter, 1978; Nordin & Gillberg, 1998).

3.10 Billstedt, Gillberg and Gillberg (2005) reported on a 13-22 year follow up of a sample of 120 individuals diagnosed during or prior to the 1980s. In common with most individuals diagnosed at that period, the sample comprised on the whole the more severe cases, with 82% having intellectual disability. The criteria they used for assessing outcomes used five categories - Good: (a) being employed or in higher education or vocational training and, (b) if over the age of 23, living independently; if 22 years or younger, having two or more friends or being in a steady relationship; Fair: either (a) or (b) above; Restricted but acceptable: neither (a) nor (b) above, and in addition not meeting criteria for a major ‘psychiatric’ disorder other than ASD; Poor: severely disabled, with no independent social progress; Very poor: ‘unable to lead any kind of independent existence’. Their results were: Good 0%; Fair 9%; Restricted/acceptable 13%; Poor 21%; Very poor 57%.

3.11 Steinhausen, Mohr Jensen and Lauritsen (2016) published a systematic review and meta-analysis of the long-term overall outcome of ASD in adolescence and adulthood, reviewing 15 studies covering 828 individuals, using the Overall Social Outcome ( OSO) ratings developed from Rutter’s original outcome criteria (Howlin, Goode, Hutton, & Rutter, 2004; Howlin, Mawhood, & Rutter, 2000; Rutter, Greenfeld, & Lockyer, 1967). The OSO rating of an individual was arrived at by summing up points on various developmental domains: independent living (0–5 points), friendship (0–3 points), and occupational domains (0–3 points), leading to a composite score classified from ‘very good’ (0–2 points) to ‘very poor’ (11 points). In practical terms these equate to – Very Good: high level of independence; Good: generally in work but requiring some degree of support in daily living; Fair: some degree of independence, and although requiring support and supervision not needing specialist residential provision; Poor: requiring special residential supervision and high level of support; Very Poor: needing high-level hospital care.

3.12 In order to align outcome criteria as closely as possible across studies using different ratings, the ‘very good’ and ‘good’ categories were combined as ‘good’, while the ‘poor’ and ‘very poor’ categories were combined as ‘poor’. Overall results in these terms were as follows. An estimated 19.7% (95% CI: 14.2–26.6) had a good outcome, 31.1% (95% CI: 23.2–40.4%) a fair outcome, and 47.7% (95% CI: 36.6–59.0) a poor outcome. However, there was very considerable heterogeneity in the samples reviewed, as indicated by the wide confidence intervals.

3.13 The studies reviewed by Steinhausen et al. (2016) covered a wide range in terms both of time of study and the nature of the sample. The analysis above did not control for differences within the samples analysed. Earlier studies have tended to focus only on those with the childhood autism diagnosis, while later studies have included the broadened concept of the autistic spectrum. When the above results were reanalysed to take account of studies of childhood autism only and studies with the wider spectrum, they showed very significantly different proportions across the criteria, with more good outcomes and fewer poor outcomes for the wider spectrum. In general, later studies have reflected the inclusion of Asperger’s Syndrome and a larger proportion of higher functioning cases in their samples. In addition, they have often also reflected an improved availability of autism interventions. These studies overall have shown a trend towards more favourable outcomes than would have been expected with more severe cases.

3.14 Some recent studies have focussed on a sector of the ASD population who later ‘lose their diagnosis’ or whose outcomes are such that they cease to be autism service users. This is discussed further in Chapter 5. However, for almost all practical purposes autism may be viewed as being a lifelong neurodevelopmental disorder of a pervasive nature, with diagnosis being made not only on the basis of a cluster of features being observed but also within the concept that the condition causes functional impairment to the individual. This has been, and continues to be, the underlying rationale for clinical diagnosis, and it is stated succinctly in the Beta Draft for ICD-11 in relation to the overall category of autism spectrum disorder: ‘Deficits are sufficiently severe to cause impairment in personal, family, social, educational, occupational or other important areas of functioning and are usually a pervasive feature of the individual’s functioning observable in all settings’ (World Health Organization, 2016). However, since diagnosis is determined on the basis of meeting behavioural criteria at the required clinical threshold, and since the level at which an individual presents may vary across the lifespan, it is logical to conclude that in a certain number of ‘threshold’ cases such variation may include no longer falling within diagnostic levels.

3.15 Nevertheless, it is difficult to draw firm conclusions regarding the extent to which individuals may lose their ASD diagnosis, and there is evidence that one of the issues obfuscating this area is misdiagnosis and overdiagnosis (Blumberg et al., 2015). There is also the question of whether there is mis-reporting of diagnosis. It was clear from our review of the prevalence studies that the term ASD is used in different ways, and all studies were evaluated in terms of what was meant by the diagnoses cited and the method by which these diagnoses were ascertained ( Chapter 4). It remains necessary to emphasise that the overall picture, both for childhood autism alone and for the wider spectrum, still highlights ASD as a disorder which is accompanied by high and enduring levels of need, with only about a quarter of the Steinhausen at al. (2016) sample falling in the ‘good’ category, and more than another quarter in the ‘poor’ category.

3.16 In addition to considering outcomes for the individuals on the autism spectrum themselves, a number of studies have also focussed on outcomes for families and carers of those with ASD, and these are of relevance to this study in terms of their implications for economic impact.

3.17 In terms of outcomes for families, a number of these have been well documented since the 1970s. Historical findings in relation to parents include the following: mothers of children with autism suffer more stress than mothers of children with Down’s Syndrome (Holroyd & McArthur, 1976; Sanders & Morgan, 1997); one third of mothers of children with autism suffer from depression and marital relationships are often adversely affected (DeMyer, 1979); the chronicity of the disorder can leave parents exhausted, pessimistic and at risk of burnout (DeMyer & Goldberg, 1983); and families suffer economic impact and financial worries (Bristol & Schopler, 1983).

3.18 In relation to siblings, many findings are also well documented: in comparison with siblings of adults with Down’s Syndrome, siblings of adults with autism were only half as likely to be married and had substantially lower household incomes (Seltzer, Krauss, Orsmond, & Vestal, 2000), had more impaired sibling relationships in childhood (Knott, Lewis, & Williams, 1995) and had poorer emotional and behavioural adjustment (Rodrigue, Geftken & Morgan,1993).

3.19 These early findings have been confirmed in more recent research on the families of those with ASD. Safe, Joosten and Molineaux (2012) found that mothers of children with autism have poorer health and wellbeing compared with mothers of children with other disabilities or typically developing children. In terms of siblings, a mixture of both negative and positive outcomes has been reported. For example, while there is an elevated risk of social and behavioural adjustment problems, with feelings of inequality, lack of attention from parents, lack of privacy, embarrassment with peers and worries about the future, positive features have included the development of increased levels of care and compassion and greater understanding and experience of difference and of atypical development (Orsmond & Seltzer, 2007; Petalas, Hastings, Nash, Reilly, & Dowey, 2012).

3.20 Regarding Question 2, ‘What are the main co-occurring conditions of ASD, other associated features of the ASD profile and any other factors relevant to outcomes or acting as moderators of outcome?’, the progress of outcome research, the changing nature of the samples studied resulting from the broadening of the concept of the autism spectrum, and the studies of economic impact have highlighted with considerable consistency some of the key factors that moderate life trajectories in autism and levels of care and services likely to be required.

3.21 The most important statement to be made is that the single most significant determinant of outcome variance in ASD is IQ, and in particular the presence or absence of intellectual disability. This has been demonstrated in a large number of outcome, economic and other studies and is discussed in detail in Chapter 5. By way of summary, individuals with IQ below 50 have the poorest outcomes, those in the IQ range 50 to 70, while still having poor outcomes overall, show comparative improvement and those in the IQ range 70+, that is, those without intellectual disability, have the best outcomes.

3.22 Nevertheless, while outcomes are significantly better for higher IQ ranges, all studies have highlighted the fact that ASD adult outcomes still present very significant challenges even for the high functioning groups. This was demonstrated by Howlin (2000) in a review of six follow-up studies for Asperger’s Syndrome. A composite rating of outcome, based on social interactions, level of independence and occupational status, indicated that just over a quarter could be described as having a ‘good’ or ‘very good’ outcome. Most of these had some friends and either had a job or were undergoing training. Even if still living at home they had a relatively high level of independence, being largely responsible for their own finances, buying their own clothes or taking independent holidays. Thirty-seven percent continued to be moderately dependent on their families or other carers for support, and few in this group had any close friendships. The remainder were highly dependent, with 33% living in special residential units and two individuals in long-term hospital care. For these higher functioning individuals, overall outcomes were very variable. A Swedish study by Engström, Ekström and Emilsson (2003) followed up a group of adults with Asperger’s Syndrome or high functioning autism. While the majority were living independently, all but one were unemployed, none were married, none had children, only a few had some kind of partner and most needed a high level of public or private support. Overall adjustment was rated as good for 12%, fair for 75% and poor for 12%.

3.23 In summary, measured intellectual ability is the main determinant of outcome in autism.. In addition, those with higher IQ show the greatest increases in skills over time (Beadle-Brown et al., 2000; Beadle-Brown, Murphy, & Wing, 2006). This has major implications for economic impact and the level of service provision required. At the same time it is recognised that even those with the higher levels of functioning, whose outcomes are considerably more favourable, still have very prominent and enduring needs.

3.24 While the central role of intellectual ability is relatively clear from the outcome studies, the significance of several other variables in the ASD profile has also been highlighted, of which the most important are language function and severity of autistic symptoms. Both of these, together with intellectual ability, have been historically identified as early predictors for autism outcome (DeMyer et al., 1973; Lockyer & Rutter, 1969, 1970). In addition, there has also been consideration of how outcome relates to which diagnosis within the spectrum an individual receives, the focus being on whether it is the childhood autism or the Asperger diagnosis (Cederland, Hagberg, Billstedt, Gillberg, & Gillberg, 2008).

3.25 There is difficulty in establishing from the literature the independence of these variables as contributory factors to adult outcomes because of the ways in which they overlap. There is a significant extent to which language function and severity of autistic symptoms serve as a proxy for intellectual level, and similarly the diagnosis received overlaps with both of these, as those who receive the Asperger diagnosis have been defined in the international classifications as having no clinically significant delay in language function or intellectual development, and a principal reason for their later average age of diagnosis is that their symptoms are generally less severe and more subtle than in autism (Howlin & Asgharian, 1999). Thus, both directly and indirectly diagnostic category may often also serve as a proxy for intellectual status.

3.26 However, the matter is more complex, and it is necessary to take due account of language function and severity of autistic symptoms as features which overlap significantly with intellectual status but which are not comprehended within it. That is, they make a separate and independent contribution to outcome variance. The question of how that relates to the current ICD-10 and previous DSM- IV diagnostic categories of Asperger’s Syndrome/Asperger’s Disorder and Childhood Autism/Autistic Disorder respectively also requires consideration. We have discussed in Chapter 9 our view that the Asperger diagnosis, in its distinction from the autism diagnosis, is comprehended within the three factors of intellectual function, language function and symptom severity, both in terms of the diagnostic criteria and in terms of actual diagnostic practice.

3.27 Regarding language, failure to develop useful speech function is a factor associated with intellectual disability, usually at a more severe rather than at a mild level, while speech development that is very delayed is usually part of a broader picture of developmental delay, with intellectual function being low or falling within intellectual disability range (Ando, Yoshimura, & Wakabayashi, 1980; Lotter, 1974; Matson & Horovitz, 2010; Seltzer, Shattuck, Abbeduto, & Greenberg, 2004). However, language makes a contribution to outcome independent of IQ. Of the 44 children in the Howlin et al. (2004) study who had IQ 70+, seven were rated as having very good adult outcomes. All of these had developed some speech by age five years. Twenty were rated as having very poor outcomes, and only 13 of these had some speech by five years.

3.28 Regarding severity of autistic symptoms, this too shows overlap with intellectual ability. However, symptom severity also makes an independent contribution to outcome variance. In their large-sample twin study, Hoekstra, Happé, Baron-Cohen and Ronald (2009) found that extreme autistic traits were only modestly related to intellectual disability and that the association was mainly driven by language and communication difficulties rather than by the other criterial features of autism, namely, the social impairments and repetitive behaviours.

3.29 The question of co-occurring conditions of ASD and other associated features of the ASD profile and the role of these as moderators of outcome or determinants of service needs is a complex one, and one on which the existing literature provides few consistent indicators. Co-occurring conditions may serve as moderators of outcome, with some, such as epilepsy, often being present from the start as part of the basic, underlying profile; but they may themselves be outcome variables, with some co-occurring conditions, such as depressive disorder, often developing in the adolescent or adult years. What is certain is that ASD across the lifespan is associated with a large number of co-occurring conditions, together with a range of other associated features which may not reach thresholds for diagnosis as a clinical condition. These co-occurring conditions and associated features are complex in terms of assessing what they contribute to service needs and therefore to economic impacts, and we have considered them as ‘additive risk factors’, an approach discussed later in Chapter 9.

3.30 A wide range of specific co-occurring conditions have been associated with ASD. However, estimates of their prevalence within ASD vary extensively. In a number of cases, the co-occurrence is of much greater significance for the co-occurring condition than for ASD. For example, it has been estimated that around 50% of people with Fragile X Syndrome have autism, but only a small proportion of those with autism have Fragile X Syndrome (see Abbeduto, McDuffie, & Thurman, 2014; Bailey, Raspa, Olmsted, & Holiday, 2008). Similarly, in the case of tuberous sclerosis, about 40-45% are estimated to meet criteria for ASD, but with some 1-4% of people with ASD having tuberous sclerosis (Smalley, 1998). In most cases, however, the co-occurring condition has higher prevalence in autism than autism has in the co-occurring condition.

3.31 Other than intellectual disability, which has been considered separately, specific syndromes reflecting intellectual disability such as Down’s Syndrome, and conditions reflecting the sensory and coordination difficulties listed along with the diagnostic criteria for ASD, the following are commonly reported as co-occurring conditions: epilepsy, attention deficit hyperactivity disorder ( ADHD), schizophrenia, obsessive compulsive disorder ( OCD), Tourette’s Syndrome and anxiety and depressive disorders. In addition, at a practical level a wide range of other co-occurring difficulties are noted, such as sleep problems, challenging behaviour, eliminatory disorders and gender identity issues.

3.32 Some of the issues associated with assessing the nature and extent of co-occurring conditions in ASD may be illustrated in relation to the question of personality disorders, since a key difficulty arises in distinguishing between what is essentially part of an autism spectrum profile and what is additional to it or separate from it. This has been well summarised by Gillberg (2002) with specific reference to Asperger’s Syndrome:

‘Personality disorders are often diagnosed in individuals who have had autism spectrum disorders since early childhood… Such diagnoses are not symptomatically inappropriate: the patients actually meet criteria for many of these disorders (perhaps particularly obsessive-compulsive, schizoid, narcissistic, paranoid, schizotypal, avoidant and borderline personality disorder). The question is whether it adds anything to the understanding of the person with Asperger’s Syndrome to say that he/she also has this or that personality disorder’. (p.56)

3.33 Leyfer et al. (2006) assessed co-occurring conditions in a sample of children with ASD aged 5-17 years. They reported that 37% met diagnostic criteria for OCD, 31% for ADHD and 10% for major depression. They were unable to report reliable figures for anxiety. They also reported a very high occurrence of specific phobias, many of these being directly related to sensory stimuli. This was a small sample study and its figures cannot therefore be generalised to the wider ASD population. However, the study was useful both in identifying common co-occurring conditions of ASD and also in highlighting the fact that some individuals have more than one co-occurring condition. Seventy two percent of the children in the sample had at least one disorder in addition to ASD, and some had as many as six. Similarly, Simonoff et al. (2008), in a sample of 112 children aged 10-14 years found that 28% met ADHD criteria. Overall, 70% had at least one co-occurring condition and 41% had two or more. Lower figures for OCD have been found in studies using large samples. Van Steensel, Bogels and Perrin (2011) reported 17.4% for OCD in a meta-analysis of 31 studies, with a pooled sample of 2,121 in young people under age 18 years (mean age range 4-16 years).

3.34 A retrospective prevalence study of co-occurring conditions in over 14,000 individuals under age 35 with ASD across three general hospitals and one paediatric hospital was conducted by Kohane et al. (2012), using electronic health records for a US population. In addition to a range of medical problems not directly associated with mental and behavioural disorders, they found that approximately 20% of their sample had epilepsy and, for those aged 18 years and over, 9% had schizophrenia. As a hospital population, their sample was not representative. However, it pointed clearly to the over-representation of these disorders in ASD.

3.35 Several studies have highlighted epilepsy as a risk factor in relation to outcomes in ASD. In a sample of 75 adults with intellectual disability, Smith and Matson (2010) found that those with both ASD and epilepsy were significantly more impaired than the control group (intellectual disability only) or those with only autism or only epilepsy as a co-occurring condition.

3.36 Tourette’s Syndrome is reported as occurring about three times more frequently in ASD than in the general population, with a prevalence estimated at 6.5% (Baron-Cohen, Scahill, Izaguirre, Hornsey, & Robertson, 1999) compared with 2-3% in the general population (Mason, Banerjee, Eapen, Zeitlin, & Robertson, 1998).

3.37 A variety of associated features and difficulties are reported as having higher prevalence in autism than in the general population. These can present significant issues for management and support. The most prevalent are sleep disturbances, anxiety, depression and challenging behaviour, but it is not possible to make accurate estimates of prevalence as these may occur both at diagnostic levels and at sub-clinical threshold levels. Other issues reported as having elevated prevalence are eliminatory disorders (enuresis 11%, encopresis 6.6% at age 10-14 years, Simonoff et al., 2008) and gender identity issues. Strang et al. (2014) found a prevalence of 5.4% for parent-reported ‘gender variance’ issues in a sample of 147 children with ASD compared with less than 1% in a comparison group, but it is not always possible to establish the extent to which these issues represent gender dysphoria as normally understood as opposed to reflecting aspects of autistic thinking.

3.38 Sleep disturbances are commonly reported as affecting the majority of children with ASD. However, the prevalence of such disturbances is also high in the general population. Couturier et al. (2005) reported a figure of 78% for a group of children with pervasive developmental disorders compared with 26% in a comparison group. Severity was also greater for those in the PDD group. However, this was a small sample with a low return rate from the comparison group, and the figures for both groups may be overestimates. In a much larger and more rigorous study, Krakowiak, Goodlin-Jones, Hertz-Picciotto, Croen and Hansen (2008) investigated sleep disturbances in 529 children aged two to five years across three groups: ASD (n = 303), developmental delay (n = 163) and typically developing (n = 63). In the ASD group 53% had sleep disturbances compared with 46% for developmental delay and 32% for the typically developing group.

3.39 Prevalence estimates of both anxiety and depression in ASD vary considerably, with assessment criteria often being very different or poorly defined. Nevertheless, there is agreement that they are much more prevalent in ASD than in the general population. For anxiety, there is the additional issue that some anxiety disorders, in particular social anxiety disorder, may be construed within the nature of autism in itself.

3.40 In a review of 40 studies of anxiety in children and adolescents with ASD, White, Oswald, Ollendick and Scahill (2009) reported that ‘between 11% and 84% of children with ASD experience some degree of impairing anxiety’, although the lower of these figures is not representative of the overall picture in the study cited (Lecavalier, 2006 – 22% on parental report, 11% on teacher report). Gadow, Devincent, Pomeroy and Azizian (2005), carried out full clinical assessments with a large sample of 301 children aged 6-12 years (autism 103, Asperger’s Syndrome 80, PDD- NOS 118) using parent and teacher reports for a wide range of disorders. Those above the cut-off for generalised anxiety disorder were: boys – 25.2% parent report, 23.3% teacher report, girls – 19.5% parent report, 25.6% teacher report; for separation anxiety (parent report only) the figures were 6.7% for boys and 7.1% for girls. It is not clear whether some of the sample were counted in both categories. Comparative figures for a typically developing group were all low, with a range of 1.5-4.1% across these categories. It was reported both by teachers and parents that those with Asperger’s Syndrome had more severe anxiety than others with ASD. Others have also reported high levels of anxiety in Asperger’s Syndrome and high functioning autism (Gillot, Furniss, & Walter, 2001; Green, Gilchrist, Burton, & Cox, 2000).

3.41 Prevalence estimates for depression in ASD also cover a very wide range (Lainhart, 1999; Stewart, Barnard, Pearson, Hasan, & O’Brien, 2006). Gadow et al. (2005) reported on prevalence of depression in their 6-12 year old sample. Combined figures for major depression and dysthymia were: boys – 18.2% parent report, 10.8% teacher report, girls – 9.5% parent report, 4.7% teacher report. Estimates of depression prevalence vary considerably in the general population, but a direct comparison for typically developing children in the same age group was made: boys – 1.6% parent report, 1.4% teacher report, girls – 0.0% parent report, 1.0% teacher report.

3.42 Depression is reported with higher frequency in adolescents and adults than in young children. Gotham, Unruh and Lord (2015) reported a prevalence of 20% in their sample of 50 high functioning individuals with ASD in the 16-31 age range. Those with high functioning autism or Asperger’s Syndrome are viewed as being particularly vulnerable. De-la-Iglesia and Olivar (2015) studied risk factors for depression in this group, focussing on studies of children and young people, and concluded that the factors that present the greatest specific risk include higher cognitive functioning, self-awareness of deficit and capacity for introspection.

3.43 Challenging behaviour is very common in ASD, especially in the childhood years, but again no precise estimates of prevalence can be made owing to the breadth of definitions used and also distinguishing between behavioural disturbance on the one hand as a co-occurring condition and on the other as a reflection of autism itself. Simonoff et al. (2008) reported 30% for oppositional defiant disorder or conduct order combined in their sample of 10-14 year olds.

3.44 Regarding Question 3, ‘How do the various outcomes and life trajectories in ASD translate into economic implications?’, the issue of ASD and its correlates in relation to economic impacts provides the principal rationale on which this study has been conducted. This question is therefore reflected throughout the report.

3.45 The primacy of IQ in relation to outcome and overall life trajectory gives it a central place in relation to economic studies of ASD, and this has been demonstrated in the earlier studies. Järbrink & Knapp (2001) estimated average lifetime costs of autism as being more than three times greater for individuals with autism and intellectual disability compared with those who had no intellectual disability. The revised estimates in their subsequent study (Knapp, Romeo & Beecham, 2009) indicated about one and a half times the cost where intellectual disability was present, a ratio also found in later UK calculations (Buescher, Cidav, Knapp, & Mandell, 2014). The central importance of intellectual disability to this study in economic and other terms determined its position as a major part of our investigation, and it is covered in detail in Chapter 5.

3.46 Regarding Question 4, ‘How do these economic implications map on to the population of Scotland?’, the Scottish context is considered in Chapter 6 in terms of prevalence and intellectual ability. This question also has had significant implications for the nature of any fieldwork exercise. The lack of clear answers in the literature to almost any key question which this study required to address indicated that the fieldwork must go beyond providing illustrative life trajectories based on known factors, but must provide more detailed data at every level on which life trajectories could be assessed in economic terms. This survey and an overview of the descriptive statistics arising from it are covered in detail in Chapter 7.

3.47 Regarding Question 5 ‘What is the relationship between outcome and type of intervention received?’, this question raises many particularly difficult issues, some of them the subject of work undertaken by groups tasked with taking forward various of the recommendations in the Scottish Autism Strategy. Of the recommendations referred to in Chapter 2, those most relevant to issues of interventions were: Recommendation 7: ‘It is recommended that the ASD Reference Group commissions research to examine and compare the outcomes in relation to quality of life for those who are supported by autism service providers and individuals who access generic provision and that relevant findings are used to inform revised guidance for commissioners of services for people with ASD’; Recommendation 10: ‘It is recommended that agencies and services develop a menu of interventions including advice, therapeutic interventions and counselling for children, young people and adults with an ASD, that are appropriate and flexible to individual need. This menu should identify advice and support that is immediately available, and set out the referral and assessment process for all other services and interventions’; and Recommendation 11: ‘It is recommended that consideration is given to the specific supports needed for the more able individuals with ASD.’

3.48 This question is potentially of crucial importance not only with regard to the relationships between intervention and outcome but also in its relation to the economic burden of autism and any possible cost-benefit analysis of implementing effective interventions. We reviewed a large body of intervention literature and we have returned to this question in Chapter 10 in considering microsegmentation and future research and provision for ASD in Scotland.