![[SIGN thistle header]](../../../images/page-header-template.jpg)
SIGN Guideline 98: Assessment, diagnosis and clinical interventions for children and young people with autism spectrum disorders
Guideline Index Page | SIGN Methodology
The early detection of children requiring assessment for health problems and developmental disorders is desirable and is the aim of child health screening and surveillance programmes. These programmes are reviewed regularly by the Royal College of Paediatrics and Child Health. The most recent review entitled Health for all Children,18 and commonly referred to as Hall 4, has led to a significant change in the provision of child health surveillance and screening in Scotland.19
Hall 4 states that every child and parent should have access to a universal or core programme of preventative pre-school care, but that formal screening should be confined to the evidence based programmes agreed by the UK National Screening Committee.18 Hall 4 does not recommend formal universal screening for speech and language delay, global developmental delay or autism, but states that staff should elicit and respond to parental concerns as part of child health surveillance. The report emphasises the need for an efficient preliminary assessment, or triage process, to determine which children may need referral for fuller assessment and/or intervention.
Screening has been defined by the UK National Screening Committee as “a public health service in which members of a defined population, who do not necessarily perceive they are at risk of, or are already affected by a disease or its complications, are asked a question or offered a test, to identify those individuals who are more likely to be helped than harmed by further tests or treatment to reduce the risk of a disease or its complications”.20
Any screening test must have a known specificity (analogous to the risk of false positives) and sensitivity (analogous to the risk of false negatives) within the population to which it is being applied. The UK National Screening Committee20 and a systematic review21 have not identified any research into ASD screening instruments that meet the rigorous criteria for a robust population screening test. Evidence level 2+
Population screening for ASD is not recommended. False positive or false negative results from inappropriate use of screening tests may delay correct diagnosis. The decision about the need for referral and further assessment should be made on clinical grounds.
C Population screening for ASD is not recommended.
Child health surveillance takes a broad clinical approach involving partnership between parents, children and health professionals. Child health surveillance can contribute to the early recognition and diagnosis of ASD.22 Surveillance for ASD should follow general developmental surveillance and should be considered by all professionals working with children and young people.
Responding to concerns raised by parents has a role in surveillance, and healthcare professionals should be aware that parental concerns about the absence of normal developmental features are as important as the presence of abnormal features.22-27 Evidence level 3
The recognition of children requiring further assessment for ASD requires a high level of vigilance for features indicative of abnormal development, both at any specific age and as they emerge over a period of time. Two structured instruments are of potential use to help identify young children with possible ASD during child health surveillance.
The Checklist for Autism in Toddlers (CHAT) was designed to identify 18 month old children at risk of ASD. It has been tested in a general population setting and was found to have acceptable specificity, but the sensitivity was too low for it to be used in total population screening.28, 29 Evidence level 4
The modified CHAT (M-CHAT) is a parent report version of the CHAT designed to be used as part of clinician led child health surveillance, with 18-24 month old children.30 Evidence level 4
A preliminary study suggests the M-CHAT is useful but final data on the psychometric properties from the ongoing follow up study are awaited.
These instruments can provide a useful structure for considering relevant clinical features during surveillance by healthcare professionals. Surveillance remains dependent on the use of clinical knowledge and skills to identify unusual patterns of development. Not all children with ASD will be identified during child health surveillance, and parents should be encouraged to return for further assessment, if they remain concerned about the development of their child.
Features which should alert healthcare professionals to the possibility of ASD are shown in Tables 1, 2 and 3.
D As part of the core programme of child health surveillance, healthcare professionals can contribute to the early identification of children requiring further assessment for ASD, and other developmental disorders:
Table 1 General developmental warnings of possible ASD in pre-school children31
Warning signs
Table 2 Warnings of possible ASD in school-age children2
Warning signs
Communication impairments
Social impairments
Impairments of interests, activities and/or behaviours
Other factors
Table 3 Additional warnings of possible ASD in adolescents*
NB difficulties are likely to be more subtle in older individuals or those without learning disability.
Warning signs
General picture
Language, non-verbal skills and social communication
Social problems
Rigidity in thinking and behaviour
* developed by the guideline group based on their knowledge of the evidence base and their clinical experience
The screening of children and young people thought to be at high risk (defined as secondary screening) may be applied, for example, to children referred to services because of developmental delay, emotional and behavioural problems, certain genetic syndromes or to siblings32 of children and young people with a diagnosis of ASD.
Secondary screening is dependent on an awareness that a child is at higher risk of ASD, and the application of sound clinical knowledge and skills. Several structured instruments for use in secondary screening have been examined in a number of studies using relatively small cohorts.30, 33-38 With all these instruments, the findings of the studies have not been replicated outwith the study settings. Evidence level 2+
The use of these instruments can be considered as a supplement to the clinical assessment of at-risk children, and may improve the reliability of the process used to screen for ASD, see annex 4. A single specific instrument cannot be recommended as each one is designed for use within a limited age group, and often focuses on one particular ASD eg Asperger’s syndrome.
![[Good practice point]](../../../images/bwtick.gif){kind=small}
The assessment of children and young people with developmental delay, emotional and behavioural problems, or genetic syndromes should include surveillance for ASD as part of routine practice.
Healthcare professionals should consider informing families that there is a substantial increased risk of ASD in siblings of affected children.
C The use of an appropriate structured instrument may be a useful supplement to the clinical process to identify children and young people at high risk of ASD.
In children under two years old typical ASD behaviours may not be evident. Absence of such behaviours should not rule out the possibility of diagnosis.22 Evidence level 4
The evidence regarding the minimum age at which ASD can be reliably diagnosed is not clear. Findings suggest that:
D ASD should be part of the differential diagnosis for very young (pre-school) children displaying absence of normal developmental features, as typical ASD behaviours may not be obvious in this age group.
Regardless of the findings of any earlier assessments, referral for further diagnosis of an ASD assessment should be considered at any age.
Suggested criteria for alerting features for ASD in older children are given in Tables 2 and 3.
The initial presentation can be to a wide range of professionals in primary care, education or social services. Important information can be gathered at this stage that may suggest the need for specialist assessment. Those involved in carrying out the initial assessment should be aware of the core features of ASD as well as of the wide range of different possible presentations, depending on the child’s level of communication and intellect, personality, gender differences, family and educational supports.
Key areas to explore at this stage include:
If, on the basis of initial assessment, it is suspected that a child or young person may have ASD, they should be referred for specialist assessment.
The aim of specialist assessment is to gather and record information that enables diagnosis and to formulate a multiagency management plan, leading to the development of an appropriate programme of supportive intervention. Such an assessment is necessarily comprehensive and may take place over a period of time.2
A diagnosis of ASD may be seen as a life long ‘label’. For this reason, it is of equal importance that clinicians diagnose, and not diagnose, accurately. Specialist healthcare professionals must ensure that they are sufficiently informed and experienced to confidently diagnose in the majority of cases and that they collaborate, where possible, with relevant multiagency colleagues, so as to achieve diagnostic consensus. Healthcare professionals should also have a low threshold of referral to more specialised colleagues in cases of diagnostic disagreement or subtle presentation.
The process of assessment and diagnosis aims to review functioning in relevant domains, make diagnoses as appropriate and facilitate seamless, multiagency intervention. It should acknowledge that other conditions (for example, specific language impairment in a three year-old, or first onset depression in a 13 year-old) may present in a superficially similar way to ASD and also that there is significant potential for comorbidity.
Although the research evidence is limited, there is support for the use of multidisciplinary or multiagency teams. 42-45
The use of different professional groups in the assessment process is recommended as it may identify different aspects of ASD and aid accurate diagnosis.
Specialist assessment should involve a history-taking element, a clinical observation/assessment element, and the obtaining of wider contextual and functional information.
Specialist assessment should be available for any children and young people who need it. Specialist teams should assess if their service is being used equitably. Apparent inequalities should be investigated and addressed.
The appropriateness of an assessment of mental health needs should be considered for all children and young people with ASD.
History taking (Parent/carer interview)
This is an important component of any ASD assessment. Without it, evidence of ASD-like behaviour cannot be put into context. Use of ASD-specific history-taking instruments can be useful in this process, although healthcare professionals should be mindful of a global perspective on the circumstances of a child or young person, taking into consideration the possibility of comorbidities and the possible differential diagnoses.
A clinical history should include:
A framework for an ASD-specific developmental history is important and a version is available in the NAPC.2 In an older or more able individual, there may be successful compensation for disabilities, and problems may only be evident within a detailed developmental history.46
ASD-specific diagnostic instruments may be used to supplement the process of clinical history taking. There are two theoretical approaches to the diagnostic classification of ASD – the categorical and the dimensional. Categorical systems (such as ICD and DSM) have led to the development of such instruments as the Autism Diagnostic Interview – revised (ADI-R).47, 48 The dimensional concept has led to the development of the Diagnostic Interview for Social and Communication Disorders (DISCO)49 and the Developmental, Dimensional and Diagnostic Interview (3di).50
The Autism Diagnostic Interview – revised (ADI-R) has been shown to be a reliable diagnostic instrument. 47, 48 It should be used with caution in children with a developmental level below the age of two years. It has also been shown to be a valid instrument for diagnosing autism in children of pre-school age.51 Evidence level 2+
The 3di and DISCO allow structured data collection in relation to ASD and other conditions.
The published data on the 3di suggests that it is a reliable and valid ASD diagnostic interview schedule when compared to the ADI-R. 50 Evidence level 2+
The published data on DISCO suggest that it has adequate inter-rater reliability for ICD-10 categories.49, 52 Evidence level 3
D Healthcare professionals involved in specialist assessment should take an ASD-specific diagnostic history.
C ASD-specific history-taking instruments may be considered as a means of improving the reliability of ASD diagnosis.
Clinical observation/assessment (Child/young person assessment/ interview)
The experience of interacting with a child or young person, in order to elicit clinical evidence of ASD that is compatible with ICD-10 or DSM-IV, is a significant professional task, which cannot be undertaken without a substantial amount of clinical experience. Such skills are not exclusive to disciplines. The crucial ingredients are training and experience.
Assessments of children and young people for ASD cannot be rushed. It may not be possible to obtain sufficient evidence in one session and the child/young person may require observation in different settings, eg at school (especially in unstructured activity such as break-time) as well as the clinic.2 Evidence level 4
ASD-specific diagnostic instruments may be used to supplement the process of clinical observation, as part of the diagnostic assessment.
The Childhood Autism Rating Scale (CARS) is an older instrument which encompasses history and observation of spontaneous behaviours relevant to autism.53, 54
The Autism Diagnostic Observation Schedule–Generic (ADOS-G),55 has been shown to be a reliable diagnostic instrument and can be used to supplement clinical history. It provides standard contexts to elicit relevant social and communicative behaviours, rather than relying on what is spontaneously manifested by a child or young person. ADOS-G has an excellent diagnostic validity for autism versus non-ASD conditions, if controlled for expressive language level.55 A study of an earlier version (the ADOS) found that it was also a very specific diagnostic instrument.48 Evidence level 2+
D Healthcare professionals should directly observe and assess the child or young person’s social and communication skills and behaviour.
C Healthcare professionals should consider using ASD-specific observational instruments, as a means of improving the reliability of ASD diagnosis.
Contextual and functional information
Helpful information about a child or young person’s functioning should be available from pre-school or school provision, and additional input can be sought from any other educational or social care professionals involved. Frameworks for information gathering to guide education professionals are available.
This type of information increases understanding as to how a child functions in groups, in unstructured settings, and when performing real life tasks. It may point clinicians towards difficulties that are not evident in one to one observations, or in more structured assessment contexts.
Information about children’s and young people’s functioning outside the clinic setting, should routinely be obtained from as many available sources as is feasible.
Children and young people with ASD vary considerably in their individual strengths and difficulties. More detailed assessment of communication, neuropsychological functioning, motor and sensory skills, and adaptive functioning may be helpful.
By definition, all children and young people with ASD have an impairment in communication which ranges from profound comprehension problems and lack of speech to subtle pragmatic or functional use of language difficulties, such as failure to understand sarcasm or use of metaphor. A wide range of speech and language and communication assessments are available56-58 but there is limited evidence to support the use of one assessment tool over another (see annex 3 for communication, speech and language assessments). Evidence level 3
D All children and young people with ASD should have a comprehensive evaluation of their speech and language and communication skills, which should inform intervention.
Practitioners should note that an individual’s level of comprehension may be at a lower developmental level than that suggested by their expressive language skills.
Children and young people with ASD will have a range of impairments in intellectual, neuropsychological and adaptive skills. A wide range of assessments were included in the search strategy (see annex 3). These are useful for individual profiling but are not diagnostic instruments.59-64 Some impairments, such as “theory of mind”62-64 and executive function60 are not specific to autism, although they may be more severe in children and young people with ASD. The degree of impairment is also influenced by levels of speech and language, communication and verbal mental age. Evidence level 3
Insights from these assessments may promote understanding by care-givers, therapists, education and social work staff in optimally supporting the child and young person with ASD to reach their potential.
“Theory of mind” is not a diagnostic marker for autism but relates to communication and linguistic development. It may be of value as part of an assessment to inform intervention. Verbal mental age should be taken into account to avoid over interpreting deficits in “theory of mind”.
DChildren and young people with ASD should be considered for assessment of intellectual, neuropsychological and adaptive functioning.
There was insufficient evidence to make recommendations about occupational therapy or physiotherapy assessments.
Occupational therapy and physiotherapy assessments should be considered where relevant.
There is a range of potential biomedical investigations that may be appropriate for a child or young person with suspected ASD. These are carried out to aid diagnosis through establishing aetiology, to exclude treatable conditions, to identify comorbid conditions and to establish baseline information prior to starting treatment. The evidence does not support the use of routine magnetic resonance imaging (MRI) brain imaging.71-73 Whilst epilepsy is common in children with ASD,74 there is no indication for an electroencephalogram (EEG) in the absence of other clinical criteria.75
A fifth to a third of pre-school children with ASD have a history of regression in acquired language skills during their second year of life. A total loss of acquired language skills is associated with a high probability of autistic conditions when this occurs in children under the age of three.76 When children undergo language regression over the age of three, they are more likely to experience seizures and the differential diagnosis should include consideration of an acquired epileptic dysphasia/Landau Kleffner dysphasia.76 Other conditions such as Rett disorder may appear superficially similar to ASD77 and other neurodegenerative conditions such as mitochondriopathies may need to be considered and investigated.78 Evidence level 3
Around 10% of children with ASD have an identifiable cause65 such as tuberous sclerosis and genetic investigations may be helpful.66-69 Clinical examination for dysmorphic features and the presence of a learning disability may aid in the decision to investigate further.66,70 For these reasons, medical paediatric history and examination may indicate that further biomedical investigations are warranted. 70 Evidence level 3
D Where clinically relevant, the need for the following should be reviewed for all children and young people with ASD:
There is considerable interest in the role of the immune system and the influence of bowel function in children and young people with ASD. An extensive search was carried out for research in this area, using the terms listed in annex 3. In addition a variety of additional investigations for children and young people diagnosed with ASD were considered (included within the list of investigations given in annex 3). The guideline development group found no research evidence of an acceptable level in support of the clinical use of these investigations, and it is not possible at present to make a recommendation.
Children with ASD can experience the full range of developmental, medical and mental health problems that are experienced by children who do not have ASD. It is as crucial to their development as to any other child’s that all comorbid conditions are appropriately assessed and managed. Clinicians should not assume that any problems are inevitable aspects of an ASD, as many comorbid conditions benefit from careful assessment and management.
Equally, children with ASD that has not been recognised may initially present to clinical services with a separate problem, eg epilepsy, a sleep disorder, or school refusal.79
A case control study found no evidence that children with autism were more likely than children without autism to have had defined gastrointestinal disorders at any time before their diagnosis.80 Parent-reported gastrointestinal symptoms, in particular frequent vomiting and constipation, were more common post-diagnosis in one study. Parents also reported higher rates of food selectivity.81 Evidence level 2+
Children and young people with ASD have higher rates of epilepsy65, 82-84 visual impairment65, 84 and hearing impairment.84, 85 As these associations have been described in the main in children and young people with learning disabilities, the extent of the specific association across the ASD spectrum is uncertain. Evidence level 2+
There are some clinical conditions which seem to occur more frequently in children and young people with ASD, regardless of intellectual ability. Children with ASD experience higher rates of mental ill health and behaviour problems.86, 87 In particular, there is evidence that anxiety and depression88-92 and attention deficit and hyperkinetic disorders (ADHD)93, 94 are more common. Evidence level 2+
Parent-reported sleep problems are more frequent in children and young people with ASD.95-98 Evidence level 2+
There is also evidence that neuromotor problems, such as clumsiness99 and tics94,100 are commonly experienced by children and young people with ASD. Evidence level 2-
Children and young people with ASD display the same attachment behaviours as children who do not have ASD. However, children and young people with ASD are more likely to be insecurely attached, affecting their responsiveness in contact with care-givers.101 Evidence level 2+
Healthcare professionals should recognise that children and young people with ASD may also have medical problems or emotional difficulties/disorders and should have access to the same range of therapeutic interventions as any other child.
C Healthcare professionals should be aware of the need to routinely check for comorbid problems in children and young people with ASD. Where necessary, detailed assessment should be carried out to accurately identify and manage comorbid problems.
Only the evidence for prognostic indicators in childhood was reviewed.
In one small study, early joint attention and imitation skills were found to be predictive of pre-school language levels.102 High IQ and language skills at an early age were also found to predict better eventual outcome in communication and social competence domains,103-107 although social impairments and repetitive behaviours103 may persist. Evidence level 2+
Improvements in adaptive behaviour and decline in atypical features have been reported for adolescents with ASD and a high IQ, with poorer outcomes evident in social impairment and social skills for young people with learning disability.108, 109 Evidence level 3
Around a quarter of young children with ASD are reported to have had regression of skills. Early language regression before three years of age, in children referred for paediatric neurology assessment,76 or those referred for ASD assessment110 has a high probability of being associated with an ASD diagnosis. The majority of children with ASD who are reported to regress have not had normal skills prior to the loss, and most are reported to subsequently regain the lost skills.111 Regression does not appear to be associated with worse prognosis during pre-school years.41,110 Evidence level 2+
There have been no adequate studies of later childhood or adolescent onset regression and it is not clear whether the phenomena are clinically the same.
Guideline Index Page | SIGN Methodology
Scottish Intercollegiate Guidelines Network,
Elliott House, 8-10 Hillside Crescent,
Edinburgh EH7 5EA
Tel. 0131 623 4720 Fax. 0131 623 4503 Web contact duncan.service@nhs.net
Last modified
10/05/10
© SIGN 2001-2009