Evaluating specialist autism teams' provision of care and support for autistic adults without learning disabilities: the SHAPE mixed-methods study

The call to develop an evidence base on Specialist Autism Teams

The Autism Act29 and NICE guidance’s31 recommendation that localities have a ‘Specialist Autism Team’ has tasked commissioners and practitioners with developing a new type of provision in the absence of any robust evidence about what it should look like in terms of its organisation, service structure, delivery and practice characteristics. The NICE GDG recognised this, and one of its key research recommendations was that as SAT provision emerged and developed this should be evaluated, with particular attention paid to identifying service characteristics associated with positive outcomes (Box 2). This study was developed specifically in response to this call for evidence. To our knowledge, this remains the only study of this sort of provision for autistic adults without LDs. 32

Stage 1: the mapping study

Multiple data sources [survey of Autism Leads across England, searches of Clinical Commissioning Groups (CCGs) and local authority (LA) websites and published reports] were used to identify services in England that, potentially, fulfilled the NICE guideline31 description of a SAT in terms of functions and staffing. All identified services were subject to a two-stage screening process, with additional data collected directly from the services potentially fulfilling the NICE criteria after the first stage of screening. Data on services identified as SATs were subject to structured content analytical techniques to describe them, to identify service characteristics that distinguished them and to test whether or not they could be organised into a typology of SAT service models. Purposive sampling techniques were used to identify SATs to act as research sites for stage 2. Stage 1 took place in late 2014 and 2015.

Stage 2: the evaluation of Specialist Autism Teams

Stage 2 comprised a mixed-methods observational study of two cohorts [the SAT cohort and the Diagnosis-Only (DO) cohort] and a nested qualitative study of the views and experiences of senior members of SATs.

The SAT cohort comprised users of SATs, who were recruited at the time of their first full assessment appointment. Individuals in this cohort included those referred to the SAT who were already diagnosed with autism [the ‘Support-Only’ (SO) group] and those referred for diagnostic assessment and ongoing support [the ‘Diagnosis and Support’ (D&S) group].

Three of the research sites also provided a regional or national diagnostic assessment service for individuals living outside its CCG/LA boundaries via block contracts with neighbouring CCGs or on a case-by-case basis. The DO cohort comprised individuals who accessed the diagnostic assessment service at these sites via this pathway. Thus, these individuals did not receive any post-diagnosis support from the SAT.

The SAT and DO cohorts were followed up for 12 months from the point of the first full assessment appointment (T0). Outcomes and resource use data were collected at 3 (T1), 6 (T2) and 12 months (T3) for the whole cohort. T3 was our primary outcome time point. For those recruited early to the study, it was decided that outcome data would also be collected at 18 (T4) and 24 (T5) months, as this would provide initial data on longer-term outcomes.

A purposively selected subsample of service users from both cohorts was recruited to take part in a semistructured, in-depth interview about their experiences as service users, perceived outcomes and views on factors (service-level and individual-level characteristics) that affected outcomes. Where the interviewee gave permission, a family member (e.g. a parent or partner) nominated by the interviewee was also invited to take part in an interview.

The nested qualitative study of senior SAT practitioners used individual interviews and focus group discussions to collect data on their views and experiences of setting up, managing and delivering a SAT; factors affecting outcomes; and ensuring sustainable developments and improvements in the care and support for autistic adults without LDs.

Finally, for the economic evaluation element, SAT service leads were asked to provide relevant financial information.

Identification of services potentially fulfilling Specialist Autism Team criteria

An overview of the process by which SATs were identified is shown in Figure 2. Data collection instruments are available in Report Supplementary Material 1.

FIGURE 2.

The process of identifying SATs.

A survey of Autism leads across England, web searches and reviews of documentary evidence identified services that, potentially, fulfilled the NICE description of a SAT in terms of functions and staffing. Information gathered on identified services (n = 96) was independently scrutinised by at least two members of the research team. It soon became apparent that the predominant population served by potential SATs were autistic adults without LDs. In response, one of the functions of SATs set out in the NICE guidance31 – ‘care and interventions for adults with autism living in specialist residential accommodation’ – was not used as an inclusion criterion (NICE 2012 Autism spectrum disorder in adults: diagnosis and management. Available from www.nice.org.uk/guidance/cg142 All rights reserved. Subject to Notice of rights. NICE guidance is prepared for the National Health Service in England. All NICE guidance is subject to regular review and may be updated or withdrawn. NICE accepts no responsibility for the use of its content in this product/publication). Where insufficient information had been identified to allow a screening decision, services were taken forward to next stage of data collection. Twenty-eight services were taken through to this stage. The main reasons why services were lost at this screening stage were that they provided (1) diagnostic assessment service only or (2) social care provision with no integrated pathway to/from a diagnostic assessment service.

Structured telephone interviews with service leads of ‘potential SATs’ (n = 28) gathered further data. These interviews were conducted from late 2014 to mid-2015. Interviews, lasting 50–75 minutes, were audio-recorded and a detailed summary was subsequently generated. The topics covered included:

• commissioning and funding arrangements

• the population served and eligibility criteria

• the structure of the service

• the service/team composition

• the diagnostic assessment process

• the approach to meeting health, social care and other needs (e.g. deliver interventions, refer on and/or ‘up-skill’ other services)

• the wider service context (local availability of other specialist autism provision, including third sector).

The interviewees were asked to supply any relevant publicly available documentary evidence (e.g. annual reports/audits, service commissioning briefs and invitations to tender for services) that was not already collected. Where interviews were not achieved (n = 6/28), further extensive efforts were made to gather publicly available documentary evidence.

Using the data gathered, a detailed ‘service description’ of each potential SAT was created and organised under the following high-level headings:

• service history and overview

• staffing, skill mix and location

• structure of the service and commissioning and funding arrangements

• eligibility criteria and referral

• services/interventions offered

• ways of working

• the care pathway

• discharge and caseload.

Before the final screen and informed by an initial analysis of the data, further adjustments to the inclusion criteria were made. First, minor deviations in skill mix from the NICE guidance31 were not used as exclusion criteria. Second, any degree of intensity of ‘support to families/partners and carers’ was acceptable.

‘Service descriptions’ were independently scrutinised by at least two members of the research team. Where necessary, follow-up telephone calls/e-mails with services were carried out to gather additional information. Final decisions regarding whether or not a service (or configuration of services) was classified as a SAT were made in the context of a review of evidence and discussion involving two or more members of the research team.

Data analysis

We had proposed to use cluster analysis to analyse the data and support the generation of a typology of models of service delivery. However, a first look at the data made it apparent that this was neither feasible nor appropriate. First, there were just 18 SATs in our sample. Second, it was clear that these were complex and highly idiosyncratic services and there were no patterns in the co-occurrence of certain features or characteristics. Third, and related to the previous point, no relevant existing evidence was available that could inform selecting certain service characteristics/organisational features to prioritise in the development of a typology.

Service descriptions were, therefore, subject to structured content analysis. 40 Qualitative data were interrogated for descriptive evidence on service characteristics and explanations given for service characteristics or ways of working, etc. Data were also extracted into Microsoft Excel^® (Microsoft Corporation, Redmond, WA, USA) spreadsheets to facilitate comparison between SATs and the identification of any consistent clustering of service characteristics or features. Analytical writing, with iterations shared and commented on by all members of the research team, supported data analysis and conclusion drawing. We also carried out a brief descriptive analysis of all relevant quantitative and qualitative data that were collected to generate high-level information about specialist autism provision in localities without a SAT.

The number of Specialist Autism Teams identified and their broad characteristics

Eighteen localities in England were identified as having a SAT (based on the revised inclusion criteria reported in Methods).

A number of factors influenced both the original ‘design’ of services and the changes to service features/characteristics over time. External influencers were the funding available, the service specifications set out in commissioning briefs and the nature and extent of multiagency working. These were, to some extent, interdependent. Internal influences were personal clinical opinion and cumulative clinical experience acquired through running a SAT.

The majority of SATs came into existence from 2009, with only two existing prior to that date (Figure 3). For those more recently established, the Autism Act,29 Autism Strategy30 and 2012 NICE guidance31 were identified as providing the impetus or justification for the development of the SAT.

FIGURE 3.

The number of SATs in existence by year (missing data, n = 2).

The loss of Specialist Autism Teams

A very small number of services were identified that, previously, would have been regarded as fulfilling the criteria for a SAT. Owing to a reduction in NHS funding and/or a loss of LA involvement, services had constricted to being a diagnostic service.

Population served

Entry into the service was commonly via a diagnostic assessment. The majority of SATs (n = 16/18) also accepted referrals of adults without LDs who were already diagnosed with autism. The proportion of ‘already diagnosed’ referrals within these services varied from < 10% to around half. Only one-quarter of SATs accepted self-referrals. All SATs operated an eligibility criterion of an IQ of > 70. The explanation for this selective approach was a perceived gap in support for autistic adults without LDs, and a belief that there were significant differences in the types of provision and expertise needed for autistic adults with LDs compared with those without LDs.

Organisational features

Staffing and skill mix

The size of the team (in terms of whole-time equivalents) did not necessarily reflect the size of a locality’s population. Constraints in funding were reported. The NICE guidance31 recommended a multidisciplinary team, with a range of professions represented. All SATs were multidisciplinary, but considerable variation in approaches to staffing were observed.

Clinical psychology was the only profession represented in all SATs, with diagnostic assessment the dominant aspect of that role. Within each SAT, the proportion of staffing resource assigned to clinical psychology ranged from < 15% to 50%. Differences in the time requirements of diagnostic assessment protocols and whether or not the SAT delivered specific mental health interventions, rather than referring to another service, appeared to determine this. Typically, if psychiatry featured in the staff team it represented a small proportion of staffing resource. The exception was one service in which the diagnostic assessment was led by psychiatry and not clinical psychology.

Around two-thirds of SATs had a (mental health or social care) social work post. Many SATs also had generic posts in which a set of competencies and autism expertise, rather than a particular professional qualification, were required. Speech and language therapists (SLTs), social workers, mental health nurses and/or occupational therapists occupied these posts. The whole-time equivalent of total staffing resource allocated to generic posts ranged between 25% and 35%. Specific SLT and occupational therapist posts were unusual, although in some SATs had relatively high whole-time equivalents.

Around two-thirds of SATs also employed staff who did not hold a professional qualification. These were typically assistant psychology posts, but other ‘support’ posts were also used. The roles that they assumed included initial/screening assessments, supporting diagnostic assessments and co-facilitating group-delivered interventions. Support workers were more likely to be involved in meeting social/social care needs, such as being involved in running ‘drop-in’ sessions/support groups and providing individuals with community-based ‘low intensity’ support. A few SATs also employed ‘employment support workers’. The proportion of staff resource assigned to ‘support worker’ posts ranged between 20% and 50%.

The diagnostic assessment

Specialist Autism Teams differed in their diagnostic assessment protocols and each was unique. Protocols varied in terms of the:

• use of published diagnostic tools and/or clinical interview protocols [e.g. Diagnostic Interview for Social and Communication Disorders (DISCO)]41

• approaches involving informants for the developmental history

• number of sessions (from one to around four)

• number of professionals involved (between one and three)

• decision-making process

• process by which the outcome of the assessment was shared with the client.

There was wide variation in reported rates of diagnosis between SATs, ranging from < 50% to > 80%. SAT professionals believed that this variation could be attributed to a number of factors, including referrals to longer-established services potentially being ‘harder’ to diagnose (i.e. present more subtly), between-clinician differences and differences in diagnostic assessment protocols.

Psychoeducational support regarding diagnosis

All SATs offered a psychoeducation intervention after diagnosis. As an approach, such interventions integrate psychotherapeutic and educational elements. Their objective is to develop understanding and acceptance of autism, address information needs and support the development of adaptive strategies to manage everyday life. The content of psychoeducation interventions was broadly similar across SATs. Some SATs used a multisession, group-delivered intervention, others used two or more individually delivered sessions and a few offered flexibility regarding mode of delivery, which was based on the individual’s needs.

Needs assessment and ‘care planning’

All SATs conducted a comprehensive needs assessment (covering mental health, social care, employment, housing and sensory needs). This took place either within the diagnostic assessment process or when ‘already diagnosed’ referrals entered the service. This resulted in a ‘care plan’ that incorporated the ‘offer’ from the SAT in terms of interventions and support, and any planned onward referrals or signposting. Services varied in the extent to which the care plan was co-produced with the service user.

Types of care provided by Specialist Autism Teams

The interventions being delivered by SATs could be organised as two ‘levels’ of care, both of which were included in the care plan:

• supporting self-management – interventions that increase knowledge and understanding of autism, improve or develop coping/problem-solving skills and self-efficacy, and develop informal support networks

• managing or addressing specific mental health and/or social needs in which the specialist nature or severity of needs and/or the individual’s capacity to self-manage mean that professional support/intervention is required.

Most SATs did not manage forensic cases or individuals with significant or ‘high-risk’ mental illnesses. If involved, they typically assumed an advisory/consultancy role.

Supporting self-management

A range of interventions to support self-management were reported (Table 1). Information provision and psychoeducation were provided by all SATs, and almost all offered informal support groups. Provision of other self-management interventions was idiosyncratic.

Specialist Autism Teams differed according to the priority given to offering interventions that supported self-management. Commissioning arrangements, clinical opinion and/or availability of autism-specific voluntary sector groups/services in their locality accounted for this. A small number of SATs were distinctive in the relative high priority and investment that was given to this aspect of their service. Others reported that they had plans to expand this aspect of their work. Although psychoeducation was delivered soon after diagnosis, other self-management interventions were not confined to a specific time point in the care pathway. Practitioner judgement (particularly in terms of clients’ readiness) and, in the case of rolling programmes of group-delivered interventions, the availability of an intervention influenced when an individual might access such interventions. In some SATs, self-management interventions were explicitly used as a way of ‘stepping down’ care.

Management of specific mental health and/or social needs

Where identified mental health, social care, employment, housing and sensory needs were sufficiently severe to require direct therapeutic intervention from a suitability qualified professional, there were substantial differences in the ways that this was approached:

• One-to-one work – as well as direct work with the individual, this could also include contact with other agencies/organisations (e.g. employer, landlord or LA housing department) in an advocacy role.

• Manualised, group-delivered interventions.

• ‘Supported referral’ to another service. By ‘supported referral’, we mean that SAT staff support the engagement of an individual with the service (e.g. attending assessments, supporting an individual to complete application for benefits, co-working with the service during assessment and care planning). Services/agencies that SATs referred to included:

  • community mental health services for psychological well-being interventions

  • LAs for assessment of eligibility for statutory social care provision

  • secondary adult mental health services (more severe mental health difficulties)

  • specialist employment support services (statutory and third sector)

  • welfare/benefits services.

Specialist Autism Teams differed in the types of need that were managed within the team and those that were routinely managed through a ‘supported referral’. This variously depended on commissioning arrangements, the perceived suitability of mainstream services, an individual’s ability to engage or cope with a mainstream intervention, the skills/competencies of the team and, in terms of accessing statutory assessment of social care need, the nature of the involvement of the LA in the SAT.

In some SATs, management of employment, welfare and/or housing needs occurred only when significant mental health needs were also present. Where this was not the case, signposting (e.g. providing information about sources of support and contact details for agencies) was used.

For mental health needs, a small minority of SATs reported that it was highly unusual for them to undertake direct work. More common was a time-limited intervention (e.g. cognitive–behaviour therapy for anxiety). Sometimes this preceded a referral to a mainstream psychological well-being service. In one SAT, mental health interventions were spot purchased, as the CCG commissioned the diagnostic assessment only. In terms of social needs (i.e. social care, employment, housing and welfare needs), commissioning arrangements and the skill mix of the team determined whether a SAT was directly involved or used a supported referral to address a need. Finally, SATs varied considerably in the extent of resources directed to specialist sensory processing interventions; this reflected differences in clinical opinion regarding their effectiveness.

Management and oversight of the care plan

There were two broad approaches to overseeing implementation of the care plan: managed care and episodic involvement.

In the majority of sites (n = 14), a named member of the team held responsibility for co-ordinating and overseeing implementation of the care plan; we refer to this as ‘managed care’. In some SATs, this individual was also presented to the service user as their ‘named contact’ while they were in the service. In 12 SATs, there was no predefined duration for an individual to be in the service, but there was an aspiration to achieve discharge (or for the client to use drop-in type provision only) for the majority of clients within at least 1 year. However, in two SATs all referrals were eligible to receive up to a maximum number of sessions (11 or 12 sessions). In one of these SATs, there was no time limit by which these sessions had to be delivered. Both of these SATs used ‘named contacts’, and one-to-one work was a core feature of both.

A second approach was ‘episodic involvement’. Here, the individual is placed on waiting list(s) for each intervention identified in the care plan, receiving each intervention when it becomes available, should they choose to. Two SATs adopted this model. There is no review prior to discharge, rather the individual is regarded as no longer ‘in the service’ once the last intervention has been delivered/offered.

Type of discharge

The majority of SATs operated closed discharge. Two SATs used an open discharge system in which individuals could re-refer themselves to the service within the first 12 months post discharge. A further two SATs used stepped discharge, offering monthly contact from the service for the first 6 months post discharge.

Changes in delivery models and practice

Many, and, particularly, the longer established, SATs reported ways in which their service had changed or evolved. These were driven by one of more of the following factors:

• unprecedented levels of demand for the service caused by unanticipated numbers of referrals and/or high levels of unmet need

• changes in commissioning arrangements

• reductions in funding

• observing existing practices (e.g. open-ended involvement) were creating a dependency on the service

• cumulative clinical experience of working with adults with autism.

Changes implemented included introducing triaging of referrals in terms of level of need, shifting from individual to group-delivered interventions, the introduction of, or increased investment in, preventative and low-intensity support in terms of social inclusion and self-management.

Advice and training to mainstream services and professionals

One of the functions of SATs stipulated in government strategy and clinical guidance is to upskill other professionals and services in their locality. All SATs were delivering on this, although the resource and priority allocated to this varied according to whether or not such activities were included in service specification and the staff views on the suitability of mainstream services/interventions for autistic adults.

Some delivery models were fundamentally based on upskilling and co-working with other services to deliver care and support to autistic adults. Here, clinical leads believed that this was the only sustainable way to meet demand for specialist autism provision. Aside from this, SATs reported upskilling a wide range of professionals/services, including mental health learning disability (LD) teams, adult social care mental health and LD teams, general practitioners (GPs), police, prison service, employers and local industry. Box 3 summarises the types of upskilling work that SATs undertook.

Access to the Specialist Autism Team by the wider community of autistic adults

To make themselves available for low-level support and advice to the wider population of autistic adults without LDs living in their locality, a small number of SATs offered an open drop-in service. However, service leads reported that it was highly unusual for someone not previously known to them to attend or, indeed, this had never occurred. One SAT ran a programme of open workshops/seminar on various topics related to autism.

Support to family members and supporters of autistic adults

Supporting family members is the final identified function of SATs. This aspect of provision was not prioritised and SATs undertook limited or no direct work in this area. Where it was provided, the types of support offered included:

• provision of written information

• responding to simple requests for advice (raised at drop-in or via telephone calls to the service)

• leading informal ‘family member’ support group meetings

• enabling and hosting a peer-led support group

• extending an existing drop-in services for use by family members

• organising and hosting occasional social events for autistic adults and their families

• at diagnosis, offering the opportunity to attend an individually delivered post-diagnosis psychoeducational intervention with the individual being diagnosed.

Many SATs regarded local third-sector groups and peer-led networks as an important source of support for family members. Where this was the case, SATs signposted and promoted them. There were instances of joint work with these organisations (e.g. support groups and social events). Some SATs, however, reported such partnerships were not available in their locality.

Specialist autism provision for adults with autism/no learning disabilities in localities without Specialist Autism Teams

In localities that did not have a SAT service, one or both of the following types of provision were observed.

Sociodemographic and population characteristics

Sites varied in the size of the population they served and their geographical size. Most were localities representing a single CCG and LA. They represented a range of deprivation and urban/rural characteristics (see Appendix 1, Table 24).

Organisational characteristics

Four sites were neurodevelopmental services and the remainder were autism-specific services (Table 2). Two (sites D and H) were multiteam services, with separate teams delivering the diagnostic assessment and ongoing support functions; these teams were not co-located. One multiservice SAT (site D) was commissioned entirely by the local CCG. In the other (site H), the diagnostic assessment service was commissioned by the CCG and the ongoing support service by the LA. Close joint-working arrangements ensured continuity of care between the services.

The different patterns of commissioning and funding identified in the mapping study (see Chapter 2) were represented in the research sites. Among the single-service SATs, three out of the seven had no LA involvement. In another (site A), the LA contributed to the funding of the drop-in service. One (site J), however, was jointly commissioned (with relatively equivalent levels of funding) by the LA and NHS. In two other sites (sites F and IA), the CCG was lead commissioner, with the LA seconding social work posts. However, in site F, which had three LAs within its boundaries, LAs varied in whether or not they invested in the service.

Thus, the range of organisational characteristics observed across all SATs identified in the mapping study (see Chapter 2) were represented in the sites recruiting to the study.

Service lead and skill mix

Seven research sites were clinical psychology led and the remaining two sites were nurse led (site CA) and psychiatry led (site J) (see Appendix 1, Table 25). The only profession represented across all sites was clinical psychology. In the majority of services (n = 7), the staff team included four or more professional disciplines (e.g. psychiatry, clinical psychology, mental health nursing, speech and language therapy, and occupational therapy) or roles (e.g. autism clinical specialist and specialist autism support worker). The remaining two services both had clinical psychologists and autism clinical specialists/support workers, with the latter working across a range of needs. Sites varied in the relative resource allocated to staff with the same professional qualification. However, as reported in Chapter 2 (and also discussed in Chapter 5), care should be taken when interpreting this given that services reported, on occasion, prioritising autism expertise and a generic skill set over discipline-specific expertise. Overall, research sites represented the different patterns of staffing and skill mix observed in the mapping study (see Chapter 2).

Eligibility and referral pathways

The research sites represented both open and closed referral processes observed in the mapping study (see Chapter 2). Four out of the nine sites operated an open referral process, including self-referrals (Table 3). The majority (n = 6) accepted referrals of those already diagnosed. A further two accepted such referrals, but only for those on their complex care pathway. This represented a very small minority of their caseload. Only one service triaged referrals at the intake assessment stage, prioritising referrals in terms of severity of mental health symptoms or social need.

Diagnostic assessment processes

The majority of research sites were using a standardised diagnostic assessment tool (Table 4). The number of sessions used to complete the diagnostic assessment process ranged from one to four or more. Rates of diagnosis ranged from 36% to 90%. Where the assessment was completed in a single session, this tended to be a half-day appointment. Practice varied in terms of the number of staff involved and when service users learnt the outcome of the assessment. The majority of services conducted a single feedback appointment, after which service users were offered a psychoeducational intervention. In one site (site D), the mental health and social needs assessments were split between the two teams delivering the service. This range in approach to diagnostic assessment in the research sites was expected given that findings from the mapping study indicated services were idiosyncratic in their approach and practice. Work soon to be completed by Newcastle University45 on diagnostic assessment practices in England provides further analysis of this issue.

Delivery of the care plan: key features

A range of delivery and practice characteristics were represented in the research sites (Table 5). The three approaches to the management and oversight of the care plan identified in the mapping study (see Chapter 2) and approaches to addressing specific presenting needs (direct work vs. supported referral) were represented. The range of intensity of involvement with supported referrals that was reported by the mapping study was not fully represented by the research sites. Unfortunately, the service that had invested the most in supporting non-specialist services to deliver care and interventions had to withdraw from being a research site.

In terms of group-delivered interventions, each service had developed its own; none was a published, manualised intervention. With respect to communication/social skills interventions, in some services this was led by a SLT, in others this was not the case.

The research sites also represented the three types of discharge observed in the mapping study (see Chapter 2), and the use or non-use, of drop-in provision. Unfortunately, the service in which drop-in provision was (perhaps) the most developed had to withdraw from acting as a research site.

Provision for carers

The research sites varied in provision for carers (see Appendix 1, Table 26). This ranged from signposting to receiving care and support alongside the family member, where appropriate (site CA). Three services (sites D, E and H) described their provision as being limited to a psychoeducational intervention post diagnosis. Another service (site IA) provided limited access to a more general support group-type provision. A couple of services noted that take-up of carers support was higher among parents than other family members. All but one service (site J) reported that there were active local autism carers groups to which they were routinely signposted. This sort of provision was not available in site J’s locality. This range of provision for carers and, overall, its limited nature is representative of the wider findings from the mapping study (see Chapter 2).

Training and consultancy

Training and consultancy was a core element of the range of services provided by the SATs acting as research sites. All but one site were commissioned to routinely deliver autism awareness training and/or more specialist training in their trust and, often, to other statutory services (see Appendix 1, Table 27). Less common, reported by only three sites, were autism awareness activities in the local community among the public. All sites reported providing advice to services/professionals in their locality on a case-by-case basis, although one site reported that this was unusual. Such input was not restricted to NHS or LA services. One site (site F) also offered an ‘advisory clinic’ (two appointments available per week) whereby individual professionals or whole teams could consult with the SAT. Again, this was used by a range of statutory agencies. In addition to these services and activities, one site (site C) had developed e-learning packages for its trust and LA.

Increasing numbers of referrals

A very significant concern for all services was the number of referrals. All services reported a year-on-year increase. In addition, all reported an increase in the proportion of referrals who had complex needs. Critically, none had received an equivalent increase in funding. Indeed, a minority had experienced a constriction in available resources (e.g. loss of funding for posts, post being frozen, withdrawal or reduction of LA involvement):

We’ve constantly historically doubled over [the 4 years existed] . . . And pretty much the same amount of money.

SAT1

. . . the number of referrals are constantly increasing. We thought 7 years ago that we’d have a mass input and then it would slow down. Unfortunately it hasn’t, and we’ve got no more staff than we had to start with.

SAT2

As well as the increasing rates of autism diagnoses in children, services believed that there were three main reasons for this situation. First, mainstream/generic services could be unwilling to work with this population, even with supervision from a SAT. Second, other services were referring to the service as means of managing their own caseload. Third, the absence of any other non-LD, specialist autism provision in the locality.

Issues with service throughput

Services operating a more open-ended care pathway were, unsurprisingly, more likely to identify issues with service throughput. More generally, a reluctance of mainstream services (e.g. CMHT and IAPT) to accept referrals and the absence or loss of low-level community support services, such as third-sector services, peer-led groups/networks and LA provision, were identified as adversely affecting throughput:

. . . the [third-sector organisation] withdrew everything . . . virtually all their volunteering services and all that sorta stuff, which was a big loss . . . and unfortunately the cutbacks in terms of the voluntary sector and local authority and all of those sorts of thing (means) virtually all support has gone.

SAT2

Services acting as care co-ordinators for those with complex needs reported the additional difficulty of being unable to discharge these service users because of the complexity of their needs or the absence of another specialist autism service on which to refer. This further compounded the issue of long waiting lists:

We’re only supposed to care co-ordinate eight people. We’ve got 11 people at the moment and a lot waiting . . . massive pressure of people coming through who are very, very complex, that do need specialist care co-ordination but we can’t do it. And it’s a real area of stress for us trying to find out where those people can go. It’s very difficult to ‘review and move on’ the people that we have got because their needs remain constant and they don’t get better . . . so it’s really difficult, that.

SAT4

Increasingly constrained resources

Increasing numbers of referrals and growing caseloads within the context of unchanged levels of resource meant that all services reported an increase in wait times, both at referral to the service and in the delivery of the interventions set out in the care plan. Inadequate financial resources were attributed to both commissioners’ demands and within-trust cost-improvement programmes:

They [commissioners] are putting a lot of pressure on us to change our practice and looking at really limiting what interventions we’re going to be able to do.

SAT3

All services had changed their service offer or aspects of practice to manage these pressures. Staff described the strain and sense of conflict experienced when they felt that the quality of care and service users’ outcomes were being compromised:

. . . we’re just getting bigger and bigger waiting lists . . . and how do you deal with that? Do you sacrifice to some extent the quality of your assessment and try and just do it as best, you know, when your commissioners are just saying, ‘well the NICE guidelines are only guidelines, the quality standards are just guidelines’.

SAT5

Cos that’s what we’re talking about, trying to change our referral process, our assessment process and still maintain quality, because people aren’t giving us any more money.

SAT4

Concerns about the negative impacts of changes in practice/provision on the quality of provision and user outcomes particularly centred on the diagnosis and assessment process (e.g. in the clinic rather than at home and conducting the assessment in a single session), wait times for interventions once diagnosed and the intensity of support that could be provided. Examples given here include increasing the use of group-delivered interventions, which may not suit all service users, and reducing the intensity or duration of one-to-one work:

There are some cases [undergoing diagnostic assessment] that we talk about an awful lot and deliberate about a lot. And people are different at home and different when they come to the office, different in, you know, different environments.

People do need time to process their diagnosis . . . but they shouldn’t be waiting months inbetween getting the diagnosis and getting the intervention.

We [now only] offer eight (one-to-one) sessions. That’s only very recently – we did have a much more ongoing approach but, cos obviously, the amount of referrals [mean] we’ve got to limit that . . .

SAT6

We’re doing more groups now to try and free up some of our time to do more of the one-to-one stuff. Having clinics as opposed to going out and doing that kind of more bespoke stuff, even though it’s not preferable, that’s kind of helped us, managed to keep our heads above water.

SAT5

Resource constraints also meant that some services found that their community-based provision was delivered in less-than-adequate venues, which was particularly an issue for those with sensory sensitivities:

. . . the financial squeeze means we’ve got no money to pay for anything, so you’re trying scrambling about trying to find some free rooms somewhere, and usually they’ll end up in some horrible old NHS community building which isn’t great. It’s that kind of practical thing . . . And you know if you don’t get it right then it won’t work.

SAT2

At the same time, services noted that constrained resources had specifically driven, or were driving, what was regarded as positive service innovation. Examples included introducing a weekly drop-in service for individuals on waiting lists for interventions and setting up a ‘supervision and consultation’ service to support mental health locality teams. We return to the issue of managing demand and ensuring sustainability in a later section of this chapter (see Ensuring sustainable improvements).

The impact of wider resource constraints across statutory services

Services also noted the impact of the wider issue of (increasingly) constrained resources across statutory services. They partly attributed the perceived reluctance of other services to accept referrals as part of a wider strategy by these services to manage their demand:

Everyone is under massive amounts of pressure with the resources they have. We see a pattern . . . We try and refer to adult social care, they come back to us because they don’t have a learning disability, or we try and signpost them to the community mental health team [and we get the response] . . . ‘no, they’ve autism, they’re not eligible’. So we just keep finding these barriers for this client group.

Our clients do go to the commissioner for mental health to request funding [for specialist psychological interventions]. It’s very hit and miss as to whether they get it unfortunately.

SAT6

Alternatively, when referrals were accepted, long waiting times were a strong possibility:

. . . we do have a referral pathway for Community Care Assessments . . . if you’ve got 2 years to wait to get one. So it’s not, that’s not great.

SAT5

The impact of the commissioning cycle

A minority of services described the impact of an annual or bi-annual commissioning cycle. It was regarded as affecting recruitment and as a significant barrier to strategic planning. Delays in commissioning decisions had led to one service being commissioned on a month-to-month basis for a period of time:

. . . 'cause of the commissioning recruitment’s been a massive problem, because we can’t offer people 3- to 4-week posts . . . no one has a job like that; and also, forward planning.

SAT8

Core service features

Those with very holistic multidisciplinary teams consistently identified this as a very positive aspect of their service, including the contribution of all disciplines to the diagnostic and needs assessment process. Depth of autism expertise was another feature that services highlighted as key to accurate diagnoses and needs assessments. Many services spoke very positively about specific group-delivered psychoeducational and skills development interventions.

Specific features

Further specific examples of service provision regarded as working well, as identified by senior practitioners, included:

• involving ex-service users in developing information resources, acting as peer mentors in psychoeducational interventions, volunteering at drop-in services and running autism awareness/training programmes

• commissioning a third-sector provider of specialist employment support to deliver work/employment interventions

• offering a drop-in service for those waiting to receive one-to-one sessions

• implementing stepped discharge arrangements; here, examples included after formal discharge, service users being able to telephone into the service for a 6-month period and providing access to a ‘drop-in’ type service

• in localities with strong user- or carer-led groups, collaborative working with these groups to support access to some form of ongoing support and autism-specific network.

Motivated and committed staff team

Finally, a number of senior practitioners noted that, despite the challenges, staff were highly motivated and committed to the service. One service lead gave evidence of this in the way staff were prepared to assume additional roles to enable the service to function.

Greater emphasis on ‘consultation and supervision’ function

Service leads believed that a change of emphasis or attention to acting as autism consultants and offering supervision to mainstream services was one of the key solutions to ensuring that adults with autism (without LDs) accessed services in a timely way. This might involve specific joint working or supervision of a specific case, supporting services with adjustments to their services and interventions (e.g. IAPT and Community Care Assessments), and more general autism awareness and education across the workforce.

Although recognising that integrating a consultative function within their delivery model was key, a note of caution was sounded in terms of the potential risks it posed for inappropriate and potentially harmful care and misdiagnosis:

. . . and you get people coming on a day’s training thinking they’re expert. That’s a danger, especially with diagnosis sometimes.

SAT9

For SATs to achieve a greater emphasis on consultative support to mainstream services, senior practitioners stressed that shifts in understanding and attitude were required within SATs, among professionals in other services and at the trust level.

Supporting self-management and minimising dependency

Service leads agreed that the overall approach of SATs should be to minimise service users’ dependency on the service and develop their self-management abilities. This was an aspect of provision that had often already seen significant changes in some services.

A post-diagnosis psychoeducation intervention, covering information about autism and living with autism, was regarded as the critical starting point to supporting self-management:

. . . enable people to get to this point where they’re very aware of their condition, they’re very aware of what’s going to be difficult, and what then to do about it. And also people need to take that responsibility too, about making those choices about [for example] who you’re going to disclose to . . .

All services also offered other interventions, typically group delivered, that related to acquiring skills and understanding which enabled or supported self-management (e.g. social skills, coping skills and anxiety management).

Group delivery was, typically, regarded as a positive decision rather than being regarded as a solution to very limited resources. Indeed, a lack of adequate, protected time for practitioners to prepare for, deliver and review group-delivered interventions affected the quality of the intervention and the outcomes achieved. Identified benefits included the opportunity to hear and learn from others’ experiences, and using ex-service users as co-facilitators or speakers. In addition, the opportunity for self-development associated with simply going along to a group was stressed. A potential additional benefit was that groups (or some members) sometimes chose to continue to meet informally after the intervention was completed:

. . . we’ve got some self-sustaining groups that go on from there [group-delivered intervention].

SAT7

However, although advocating group-delivered interventions, practitioners made it clear that this required investing time in preparing and supporting individuals to attend:

Lots of people with autism would say, ‘I can’t do a group’, but actually they can, it’s just they’re very anxious and it’s something that’s really scary to them. So, a lot of preparation work with them helps them go into that group . . . And making sure facilitators in that group are people who can hold that group, which is a skill in itself when you’ve got lots of different people . . .

SAT7

In addition to specific interventions, senior practitioners described more general ways of working with service users that supported self-management and minimised the risk of individuals becoming dependent on the service. Often these had been developed in response to observing undue dependency and/or a realisation that previous approaches to care were unsustainable:

[we give the message from the start that] . . . we cannot scoop you up and fix you, that’s the message, because everybody comes and families come thinking you’re going to scoop them up nicely, fix them, and then give them back, and we’re not able to do that . . . Because it’s not fixable.

SAT2

Yes, I’m an information giver, I’m somebody who can enable, but they don’t look to me to solve anything they want.

We made a decision several years ago now about things that we wouldn’t do . . . and it might sound really simple, but it’s things like not filling in forms with people. So we spend probably as much time not filling in forms as we would if we actually filled it in for them. So it’s about that emotional resilience to show them that they, they can do it. So it’s kind of short-term pain for long-term gain really.

SAT2

Working with local peer-led networks

Services with strong local peer-led groups (including those for carers) consistently identified them as a core element to a sustainable framework of long-term support for autistic adults without LDs and their families. Connecting individuals to these groups was seen by some service leads as a key function of their service and a whole-team awareness of potential support groups/networks was important:

It’s just about using what’s out there and being really knowledgeable as a team about what is out there.

It was noted that information provision may not be sufficient and that more proactive support may be required to enable service users to engage with such provision:

I think it’s about signposting to whatever services there are out there, to get them to be engaged with those services . . . Linking them into local groups, getting them into them. Getting the carers into anything you can get them into so that the carer feels they’ve got somewhere to go to as well.

We’ve got to use what we’ve got and it’s a matter of getting people into them. Making them, helping them to feel that they’ve got something outside of your service.

Practitioners suggested that the geographical characteristics of a locality (i.e. rural vs. urban) may affect the number and range of peer-led networks and other third-sector organisations with which SATs can partner.

Drop-in services and other low-level support

A number of services used drop-in services or telephone contact as a means of providing ongoing, low-level support. For some, access to such support was time limited and was the way in which stepped discharged was managed:

So we have this step-down service now where they can still link in with the service for 6 months after, perhaps have a telephone clinic each month; and then they tend to kind of just go off and obviously if they need to come back they can come back, but that seems to have worked really well.

SAT5

Others, however, offered service users the open facility to telephone the service for one-off contact and four services provided a weekly drop-in that was open to ex-service users. All those providing this sort of low-level, ongoing support were strong believers in its value. Specifically, it was identified as serving to pre-empt problems or crises, nurture independence and self-management skills and reduce the sense of isolation:

I’ve got somebody who rings me twice per year who’s been discharged for 4 years but they still they still ring twice per year. But that’s what keeps them going, they don’t go into crisis and end, end up, using loads and loads of services. And it’s generally only a quick phone call about probably nothing but it’s just a, a check in that we’re still there and they’re still there and it’s a 5-minute call and they’re gone; and, and I think that, you can’t replace that really, that’s, that’s really invaluable.

SAT5

. . . it’s the ‘there when you’re needed’ level of support. It might be a 5-minute conversation it might take half an hour. But just holding their hand through some difficulties until they learn to do it themselves. I think it’s really valuable and I have increasingly less contact with people as they learn to manage these situations themselves. They generalise a lot better from that than from didactic teaching. People can turn up for half an hour and then we won’t see them again for 6 months, and then they’ll come again. But they know it’s there.

SAT6

With (our drop-in service), the idea of that is to help people network and then move on to back into mainstream.

SAT5

Light touch, access when you need it. Drop-in services are great . . . a little bit can, can go a long way, can’t it?

SAT9

There was a high level of interest among those not currently offering such services. However, some spoke of the difficulties of persuading commissioners that provisions such as this should be included in their service specification.

Functioning and complexity of need

As already noted, practitioners reported a great range in functioning and complexity of need among those referred to their services. These were felt to affect the type of impact that a SAT could expect to achieve. For some, maintaining existing health and functioning or preventing further decline are appropriate outcomes. For others, positive change, sometimes quite specific to a particular aspect of their lives, can be hoped for or expected:

Some of it’s about current functioning. So, like for our complex clients, it’s still improved outcomes, but it’s not a lot of change. And then we’ve got people who are, have jobs, they have families and that kind of thing. They have more of an idea about what they want to do [achieve from using the service].

Ownership of the referral

Practitioners identified three types of referrals: those highly engaged and positively wanting an autism diagnosis, those with mixed feelings and anxiety about the process and passive participants. The last group typically comprised those in which parents have driven the referral:

There’s a massive range, isn’t there? There’s some people that are incredibly motivated and really want a diagnosis, and some people that are a bit ambivalent, and some people that other people are saying: ‘Oh, I think this is an issue for you’.

Reaction to the diagnosis

A strong and consistent theme within practitioners’ discussions was the way that service users’ reactions to their diagnosis affected longer-term outcomes. For some, this reaction was closely related to ownership, or not, of the referral. Thus, in some cases, the diagnosis fulfils a (potentially long term) desire to make sense of their lives. Here, seeking a referral or self-referral may not be precipitated by a particular difficulty or crisis, but a readiness or desire to make sense of themselves and their lives:

There are individuals where the diagnosis can give them sort of a clarity. It can give them an understanding of what it, why they’ve had these difficulties. Those people that are kind of, almost on a sort of quest of understanding why . . . It’s a positive thing in a way cos it kind of makes sense.

It’s about acceptance of the diagnosis when you’ve always felt different and lost and, you know, having an explanation of how that can help you understand yourself and move with things.

Importantly, psychoeducation was regarded as playing a critical role in supporting positive outcomes from being diagnosed.

Engagement with the interventions offered by the service

Practitioners believed that engagement with the service following diagnosis was affected by ownership of the referral and reaction to the diagnosis (as described above). Crucially, they may affect engagement with psychoeducation interventions, which was the first intervention typically offered. Lack of engagement with this was regarded as having the potential to stymie ongoing engagement with the service and/or the potential for positive impact of subsequent interventions:

Outcomes would be improved for people who either will or can engage in [psychoeducation interventions]. But it’s whether they have that ability, or whether they have severe anxiety, to be able to do that.

Mental health difficulties, such as anxiety and depression, were also implicated. In general, practitioners referred to service users having different levels of motivation in terms of wanting or feeling able to make changes in their lives:

. . . its’ about the level of motivation to want to do something and change something.

Family support

Senior practitioners agreed that family members’ responses to the diagnosis can differ widely. For example, a diagnosis of autism may improve one couple’s relationship, as it provides an explanation for why a partner behaves in the way that they do. For another couple, however, an individual may find the realisation that certain things about their partner cannot be changed is a catalyst for ending their relationship. Therefore, overall, families were regarded as having the potential to facilitate or hinder positive outcomes. This was attributed partly to existing family relationships and dynamics, but also to family members’ ability and willingness to understand the additional challenges faced by autistic people, and what they can do to accommodate specific needs:

It can be a really, really positive thing if they have a supportive family . . . then having that diagnosis will possibly enable some more empathy or understanding around the difficulties and can improve things like family dynamics at home.

Across these discussions, the value of family members receiving psychoeducation related to the diagnosis of autism was highlighted.

The intersection of individual need and service characteristics

Another key factor that was thought to affect user outcomes was the ‘fit’ between the support and the interventions provided by a SAT and individuals’ needs. This was not necessarily about severity or complexity of need. In some instances, for example, it concerned whether or not services were able to offer low-level support on an ad hoc basis for those individuals who, for the majority of the time, managed their everyday lives independent of formal support:

[There are some] who don’t need very much support but need to be able to access informal support, so a telephone call or a ‘cuppa tea’ kind of chat as and when they need it as opposed to something formal, eight sessions kind of thing. And if that can be made available to them, they seem to have very, very good outcomes.

However, practitioners also noted that the types of needs presented by service users, and their ability to identify specific and realistic aspects of their lives that they wanted to change, could affect whether or not substantive changes in outcomes were observed. They explained that some needs or difficulties are more amenable to change or address, or that it is reasonably straightforward to equip individuals with self-management skills:

People who have sensory issues, so where you can actually [help them with] coping strategies for dealing with those things, you know, practical things that can help on a day-to-day basis.

The people that are goal-driven but also have realistic expectations. They’ve gone into the service with a clear idea of what their chosen outcome is and then work with someone towards it.

The quality of the diagnostic assessment process

There was strong consensus that the quality of the diagnostic and needs assessment process was important in terms of supporting positive outcomes. All participants believed that their service provided this, and this service characteristic was stressed as distinguishing SATs from other diagnostic pathways.

Skill mix with autism expertise

Skill mix of the team, specifically its multidisciplinary nature, was consistently regarded as one of the most important service characteristics in terms of supporting positive outcomes. Practitioners noted that some key skills are shared across different professions and it may be the case that it is the skill set, rather than the professional qualification, that is the key issue. Integral to this was autism expertise. Indeed, some noted that, on occasion, when considering the skill mix in a team, this expertise was prioritised over a particular professional being represented on the team. Practitioners from one service described the value of having someone with direct experience of autism within the team:

The skill mix of the team is vitally important.

We all bring different things to the team from our professional backgrounds so it’s useful to have a mix of professions with experience . . . [of supporting autistic people without LDs].

We need skill mix, but it isn’t so much your profession, it’s about your skills and knowledge of autism.

Alternative modes of intervention delivery

Another service characteristic seen to have the potential to impact outcomes was services’ ability to be flexible in the way that they provided care and support to accommodate service users' needs and preferences. Two components were identified: the overall delivery model (time-/session-limited vs. open-ended contact) and the mode of delivery (i.e. individual vs. group). The first quotation below comes from a senior practitioner working across two SATs, one mainly using group-delivered interventions and the other using individual work alongside groups:

You can see the difference being able to offer groups and individual work alongside, or just individual work. It makes quite a big difference.

There’s so many positives of both approaches [time/session-limited vs. open-ended].

There are so many people that identify as our services as ‘I’ve been using you for years, and I don’t use you very often but I’ve got that if I need it’. Whereas other people respond so much better to ‘I’ve have these sessions and I go and then I’m never going to see you again’. I just think they’re both equally as important.

Services should be able to be flexible around the individual’s needs.

Having a ‘named contact’

Service users having a ‘named contact’ in the service was regarded as, potentially, impacting outcomes. It was also noted that it was very relevant to service user experience:

A key worker is really important for people. Like a contact, somebody that you can form a relationship with and contact.

We have a named contact primarily because of feedback from the service users. They [said they] don’t like being passed from pillar to post [because] they might move through different professionals in the team. So one person holds responsibility for that service user while they’re moving through. Then, if they need to ring, they leave a message specifically for that person.

Not all services had a formal ‘named contact arrangement’. However, it was reported that this could happen informally on an ad hoc basis:

. . . they will tend to ‘take’ to one of the members of staff and call them.

Local authority involvement

Some practitioners regarded LA involvement in the SAT as one of the more important factors potentially affecting service user outcomes. Specifically, it was seen as supporting a holistic approach and direct access to community care assessments:

. . . in terms of local authority involvement, we all felt it was really important and crucial to a holistic autism service with social work involvement.

Features of service organisation and delivery with less influence on outcomes

Service characteristics identified as being less likely to affect individual outcomes were if a service was autism-only or a neurodevelopmental service, and whether the service was delivered by a single team or two/three teams, each delivering specific functions.

The sample

Over half the T3 sample (138/260) completed both questions, as follows:

• SAT cohort, n = 106/164 (64.60%)

  • D&S group, n = 74/133 (55.65%)

  • SO group, n = 32/75 (42.70%)

• DO cohort, n = 32/52 (61.50%).

Appendix 4, Table 30, provides an overview of the sample characteristics.

Overall impression of the service

In total, 60.4% of respondents reported that the service had a positive impact on their lives, 33.8% reported little or no impact and 5.8% reported a negative impact. However, examination of the qualitative data suggests that a positive impact rating does not necessarily indicate a singularly positive experience and outcome. Rather, a more mixed experience was possible, with some describing the support that they had received as insufficient (Table 6). This table also displays the difference between cohorts. Thus, although 60.0% of SAT D&S group respondents and 56.3% of DO respondents reported a positive impact, this figure was much lower (28.1%) for SAT SO group respondents.

In explaining the reason for a non-positive rating, a range of unmet needs (e.g. mental health difficulties, family understanding of autism and access to social care) and/or inadequacy of support were described. DO cohort respondents typically wrote more generically about insufficiencies, often stating simply that they needed more follow-up support after diagnosis than had been offered, both for themselves and for their family. The following sections report further findings from the content analysis that shed light on why respondents from different cohorts may have had different experiences.

The accounts of positive impact

Appendix 5, Table 31, summarises results of the content analysis of respondents’ accounts of the positive impact(s) that the SAT had on their lives. It shows that positive impacts were wide ranging in nature, with 18 different types of impact identified. These were grouped into seven broad categories:

1. understanding and acceptance of diagnosis and self

2. improved mental health and coping

3. help with employment and education

4. access to other services

5. improved social skills, relationships and networks

6. contact with supportive practitioners

7. reduced sense of isolation.

The number and range of positive impacts experienced varied between cohorts (see Appendix 5, Table 31). All types of positive impact were reported by at least one D&S group respondent. However, the range of positive impacts experienced by SO group and DO cohort respondents was more limited (10/18 and 7/18, respectively). It is notable that ‘contact with supportive practitioners’ was reported only by SAT cohort respondents. This impact category includes a number of different aspects of supportive practice, including being made to feel valued, feeling understood, being treated with respect and staff being experts in autism, easy to contact and responsive to individual need. The follow extract illustrates the significance of such an experience:

Being understood and being treated with such respect and compassion was a hugely beneficial experience after not having been very understanding or compassionate with myself.

SU37, SAT cohort: SO group

The most frequently reported type of positive impact, which was observed across the three cohorts, was ‘greater understanding of self’. Examination of the qualitative data linked to this code provides some explanation as to what this concept means and why it was valued. First, it referred both to their current lives and to their past experiences. Second, ‘greater understanding of self’ had allowed respondents to be more accepting and forgiving of themselves. They described both specific effects or changes (e.g. worrying less about why they behaved as they did, ceasing to label themselves as ‘weird’, changing job/career to one more suited to them) and more global positive impacts, such as increased confidence and self-esteem and improved well-being and/or mental health:

I no longer worry or analyse my thoughts and actions like I used to. I feel much more at ease and finally feel that I understand myself as a person. Although I still struggle with anxiety, I am grateful that I have been given answers. The fact that I am now aware I have Asperger’s is great, it makes sense and provided me with what I have needed all my life – answers and closure.

SU39, SAT cohort: D&S group

I can now learn to live with who I am and what I am capable of/and what I am not. I don’t feel I have to push myself as much to fit in (which caused me the most stress). I still have other problems with moods and sensory issues. I am learning more about this condition all the time.

SU40, SAT cohort: D&S group

It is worth noting here that although respondents across all cohorts reported ‘increased understanding of self’, only respondents in the D&S and SO groups reported that they had benefited from ‘help with accepting and/or seeing the strengths of their diagnosis’. Comments such as the following were made:

It’s given clarity in my life. I know the reason why I’m the way I am. And thanks to [name of service] I look at my condition in a positive manner.

SU41, SAT cohort: D&S group

Given the connection respondents made between greater understanding and acceptance of self, it perhaps is not surprising that improved mental health (including self-esteem) was also one of the most frequently reported positive impacts. Indeed, scrutiny of responses related to improved self-esteem or mental health revealed that such changes were most frequently linked to improved self-understanding. However, for a minority, improved mental health was attributed to other interventions provided by the service (e.g. sessions with a psychologist, joining hobby groups and receiving additional support at university):

I feel strongly that if I had not been sent to the [name of service] team when I had I would not be in such a positive place as I am today. I will always have difficulties with mental health but I now have coping mechanisms and a support system in place which I would never have had without this team. Thank you.

SU42, SAT cohort: D&S group

The accounts of little or no impact

Table 7 provides an overview of respondents’ reasons for why the services that they used had little or no impact on their life. This group represented 22 out of 75 D&S respondents, 18 out of 32 SO group respondents and 7 out of 32 DO group respondents. Across all groups, the most frequent reason for little or no impact (27/47) was that no support had been received.

As would be expected, among DO cohort respondents, little or no impact was solely attributed to not receiving any input or support from the service after the diagnostic assessment process was complete. Notions of isolation were evident and this could offset any positive impacts of being diagnosed. For example:

They diagnosed and left me. The diagnosis was positive as it helped me claim PIP [Personal Independence Payment]. Being left alone and questioning my entire life due to my diagnosis was a negative, left entirely to my own devices.

SU43, DO cohort

These respondents also described the types of issues that they continued to struggle with or the types of support that they wanted. These included coming to terms with the diagnosis, managing mental health problems, helping with relationship difficulties, finding out about/connecting with support groups/services for autistic people and managing difficulties with employment/workplace.

Among the SAT cohort, not receiving input or support from a service was attributed to either the service not being in touch since the diagnostic or needs assessment, or the respondent still being on a waiting list for a group intervention(s). A sense of frustration was observed in many responses:

After my first two appointments in which I was assessed and then told the result, I have not had any further contact with [name of service] other than when I inquired when I would receive the report of my assessment, which I did not get until almost 12 months after my diagnosis. Since my diagnosis I have not been given any information about help/support that is available.

SU44, SAT cohort: D&S group

They have totally failed to put on group sessions. I’ve not heard from them in months.

SU45, SAT cohort: D&S group

The second most frequently reported reason among the SAT cohort for a service having no impact was that the support received did not (fully) address the individual’s needs. This was either because the desired support had been unavailable or not offered (e.g. employment support, explaining the diagnosis to relatives, etc.) or because a respondent was discharged from a service before they felt ready:

It’s a bit like having the carpet pulled from under your feet. It would only have an ongoing positive impact if the support was ongoing. It’s a bit like trying to help a homeless person by giving him exactly six nights of accommodation and then nothing.

SU46, SAT cohort: D&S group

A third reason that SAT cohort respondents rated the service they used as having little or no impact was that the support/interventions offered or provided were regarded as unsuitable. Where further detail was provided, it typically concerned being offered group-delivered (as opposed to a one-to-one) intervention(s). Finally, a small proportion of SAT cohort respondents reported declining any/further input because of other demands on their time or because they felt too old to change how they lived their lives.

The accounts of negative impact

Just 8 out of 139 respondents rated the service that they had used as having a negative impact on their life. In the majority of cases (6/8) this was because the respondent had not been offered any post-diagnosis support.

Needs predominantly met

Eight of our interviewees, across six SATs, reported their needs to be fully or predominantly met by the service. This included individuals not previously diagnosed and individuals already diagnosed with autism. Some individuals had very substantial and wide-ranging difficulties at the point of referral (e.g. homelessness, redundancy, depression and suicidal ideation) and described the considerable impacts that receiving support from a SAT had on their lives:

When I was diagnosed I was not far off being incapable of work and now I’m 6 months into a new job and coping incredibly well, despite massive upheaval and significant change. And I’m doing that at the same time as trying to move house and coping with the rug being pulled out from under my identity in terms of my diagnosis. There is no universe in which I would be coping half as well if it wasn’t for the support I’ve received.

SU35, 33 years

I felt they went above and beyond, and I am truly grateful. I wanted to complete this interview to express that, as I can’t thank them enough.

SU22, 37 years

Others had sought referral to the SAT for a diagnostic assessment and did not identify themselves as having any additional support needs. In these instances, the diagnosis often enabled individuals to self-initiate any changes or resolve problems in their lives, and they had not required support from the SAT to achieve this:

The main thing I really wanted was the diagnosis for work. I had support that I needed from ex-wife and friends so I wasn’t really looking for any sort of ongoing support from the service.

SU18, 62 years

Met and unmet needs

Thirteen interviewees, across seven SATs, were allocated to the ‘mixed experience’ group, with some of their needs remaining unmet or only partially addressed. Although half regarded themselves as still ‘in the service’, no interviewees were actively engaged when we interviewed them. Unmet or partially met needs were attributed to ‘ineffective’ interventions and limitations of the SAT in terms of format, duration, intensity, scope and flexibility of the support available: we return to these issues in detail in Service characteristics: impacts on outcomes. Reports of refusing, not using or dropping out of the interventions/support offered were much more common in this group than in the ‘needs met’ group. The cause of limitations to the support received was most frequently attributed to funding constraints within the NHS and participants expressing sympathy for SAT practitioners. Nevertheless, a number of people expressed disappointment at the help they received.

Needs predominantly unmet

Seven interviewees, across four SATs, were allocated to the unmet needs group. All reported that, aside from the diagnostic assessment, the support they had received had not helped. All had been discharged or had disengaged from the service. The unmet needs that were reported were wide ranging and, in some cases, quite significant or dehabilitating. They were consistently attributed to limitations in the support offered by the SAT and/or the support/interventions being delivered in a way that they could not access (e.g. group-delivered interventions). We return to these issues in detail in Service characteristics: impacts on outcomes.

Some presented the experience of being discharged and/or being offered intervention/support as inadequate or inaccessible, with language that indicated a sense of rejection and a perceived lack of compassion on the part of the service; for example, ‘thrown out’ (SU8) and ‘ticked off [waiting list]’ (SU17), as well as:

The service is like, is almost like detached, it’s like doesn’t even wanna try and help. . . . I feel like I’m not getting across, like I can’t get across how bad it is kinda thing.

SU47, 20 years

Another common theme in this group’s accounts was a frustration that needs had been articulated or identified during assessment, but not then addressed:

It felt like there was a disconnect between the person doing the diagnostics and the follow-up. Identifying an issue is isn’t the same as providing solutions or help with the issue.

SU12, 28 years

A noteworthy difference in experience between the ‘unmet needs’ group and the ‘needs met’ and ‘mixed experience’ groups was that the ‘unmet needs’ group were more likely to describe waiting times for the diagnostic assessment as being problematic, with some having waited > 2 years. It is possible that these waiting times increased the severity of personal difficulties or the hope vested in how the SAT would impact their lives, which, in turn, heightened or coloured their emotional response to their experiences of using a SAT. Despite this sense of frustration and disappointment, as with the ‘mixed experience’ group, interviewees also referred to the high level of demand for services and lack of funding as an explanation as to why they had received so little support.

Scope of and access to psychoeducation about autism

A range of emotional reactions and responses to an autism diagnosis were described. Positive (e.g. relief) and negative (e.g. anger, distress or grief) emotional responses were reported, with some individuals experiencing both, simultaneously and/or over time. All SATs offered an extended psychoeducation intervention (i.e. an intervention seeking to increase understanding of autism, self-understanding and self-management skills) via either group or one-to-one sessions. However, not all SATs offered one-to-one work for those unable to attend a group (see Availability of alternatives to group-delivered interventions).

Individuals in the ‘needs met’ and ‘mixed experience’ groups typically spoke very highly of the psychoeducation that they had received:

It ended a lifetime’s worth of feeling inadequate and feeling worthless.

SU36, 21 years

Aspects of psychoeducation identified as being particularly helpful included content that gave insight into how autism may affect them, including the positive aspects of the condition; sharing experiences, including speakers with positive stories of living with autism (group delivery only); and learning coping techniques for dealing with difficulties commonly experienced by autistic people (e.g. sensory overload and anxiety). A dominant theme in these individuals’ accounts was the impact on the way that they viewed themselves: feeling more accepting or forgiving of themselves. Some also spoke of feeling less need to mask autistic behaviours and/or seeing the positive side of the condition:

It [the diagnosis] could have been life-changing in a way that left me stranded and in fact it has been life-changing in a way that’s given me enormous support and enormous hope, and a new capacity for thinking about myself and the world around me.

SU35, 33 years

By contrast, those in the unmet needs group had either not accessed this intervention (owing to long waiting lists or feeling unable to attend a group) or dropped out early because they found it unhelpful. Being unable to access (good) psychoeducation appeared to be a key influence in whether or not individuals had been able to resolve any negative emotional reactions to the diagnosis.

The service’s overall model of care delivery

All of the individuals in the ‘met needs’ group with support needs extending beyond the diagnostic assessment had attended SATs that provided individualised managed care that was overseen and co-ordinated by a single member of the SAT, who also acted as the ‘named contact’ for the service user. For individuals experiencing a mental health crisis, being able to contact their ‘named contact’ outside appointment times was consistently regarded as extremely valuable. As well as responding to needs that the individual was aware of on referral or were identified in the assessment process, this model of care was also able to respond to needs emerging during the time that the individual was in the service. Furthermore, in adopting this model of care, these services typically had the skills and expertise within the team to respond directly to a wide range of support needs. Where required, they supported referrals to other services/agencies:

It has been individualised to me, I’ve felt seen and heard and respected during the entire thing, but the breadth, but also that the breadth of professionals involved has meant that I’ve had several different appropriate touch points, each of which has had a measurable significant impact like, and it, it’s been, yeah, it’s been fantastic. Considering, considering what a shitty, complicated weird situation it has been it has been, the, the service has been brilliant.

SU35, 33 years

She does things that I wouldn’t have ever known about, having the one to ones . . . I think they’ve exceeded what I expected with it, they’ve explained it [the diagnosis] to me so I understand it, and try to put coping methods in place. That’s exceeded what I expected.

SU15, 47 years

There was some evidence that limiting the duration of individualised managed care (e.g. the SAT being able to offer only a fixed number of sessions) could mean that some needs remained unmet at discharge. This was observed among individuals in the ‘mixed experiences’ group attending such services, especially those with complex and long-standing difficulties. However, even in SATs not working to a fixed number of sessions, a few individuals in the ‘mixed experience’ group believed that they were discharged before they were ready to leave:

I was hoping to stay a client. But they thought that I had all the support I needed, that’s why, but I don’t feel that, because they thought I was finished but I wasn’t, and I didn’t get a chance to explain that to them.

SU9, 18 years

By contrast, interviewees whose SATs did not work within an individualised model of care, offering instead a limited range of interventions and referring elsewhere to address other needs, consistently typically reported (at least some) unmet needs. A later section, Approach to managing referrals to other services, further explores practices around referring to other agencies or sources of support, and how this may be implicated in needs remaining unmet.

Availability of alternatives to group-delivered interventions

Many SATs used groups to deliver one or more interventions. Typically, this was a positive choice, with group delivery perceived as the best approach to achieve the specific objectives of that intervention (see Chapter 4). However, not all SATs offered one-to-one sessions if an individual felt unable to attend a group-delivered intervention. The lack of an alternative to group delivery of an intervention was a frequently cited reason for a need remaining unmet. In these cases, the individual had either refused or dropped out of the intervention early on. In both situations, no alternative (i.e. one-to-one session) was offered.

The most common reason for refusing or dropping out of a group-delivered intervention centred on social anxiety and/or a lack of confidence about being in a group. For some, past experiences of finding groups emotionally draining or distressing contributed to this decision. Other reasons, less frequently described, included an unwillingness to spend time with and/or be seen with other autistic people, particularly if they were perceived to be more severely affected; fears that hearing others’ experiences may be upsetting; and not believing that sharing experiences would be of use or value. An experience of a poorly facilitated group (e.g. not keeping to time/agenda and feeling that distressed participants were not well supported) was another reason for drop out.

Two interviewees spoke positively about the support that was given to attend a group-delivered intervention. Valued practices included pre-meetings with the group facilitator to hear more about what would happen; having the chance to view or be reassured about the suitability of the venue in terms of lighting, etc.; and allowing and supporting participation in group discussions via written, as opposed to verbal, contributions. Similar or complementary practices were suggested by interviewees who had not taken up a group-delivered intervention, but felt that, with support, it might have been possible. These interviewees also suggested offering a mentor or buddy to accompany individuals.

Timeliness of group-delivered interventions

Some of the interviewees who attended SATs using group-delivered interventions spoke of long gaps between the offer of support being made and the relevant group starting. This increased the risk of needs remaining unmet in a number of ways: individuals’ willingness to attend the group diminished or the intervention was regarded as no longer relevant or seen as coming too late to be useful. Some interviewees described how the presenting need increased in severity during the waiting period, which made attending the group more problematic. Finally, and in addition, a lack of communication from the SAT while waiting for a group to start (this could be several months) was, for some, a source of further anxiety with concerns that the SAT had forgotten about them.

Approach to managing referrals to other services

The ways in which individuals were referred to other agencies or sources of support to address needs that could not be met by the SAT also appeared to influence whether or not needs were met. We use the term ‘supported referrals’ to describe instances where SATs assisted with completing application forms (e.g. benefits); arranged and/or accompanied individuals to appointments with other agencies or organisations (e.g. employers, housing officers and social services); and/or acted as an advocate in meetings with other agencies. These sorts of practices were consistently reported by individuals in the ‘needs met’ and many of those in the ‘mixed experience’ group:

She’s very like to the point, very like, this is what she needs, this is what she’s like entitled to type of thing. I do think they like fight your corner for you, not that you should have to fight your corner, but like I do think it’s quite like a driven service.

SU7, 27 years

Experiences such as these contrasted with those who had been ‘signposted’ (i.e. provided with information and contact details for other agencies or other potential sources of support) with regard to an identified need. There was very little evidence that signposting was actively used or successful; as a result, needs remained unmet. For example, two interviewees spoke of being given the contact details for social or housing services. One had made no attempt to contact the service, the other had attempted to visit the service, but it had been closed and he had not returned.

The Specialist Autism Team cohort

This cohort comprised users of SATs who were recruited at the time of their first full assessment appointment. Individuals in this cohort included those referred to the SAT but already diagnosed with autism (the SO group), and those referred for diagnostic assessment and ongoing support (the D&S group).

The Diagnosis-Only cohort

Three of the SATs that acted as research sites also provided a regional or national diagnostic assessment service for individuals living outside its CCG/LA boundaries via block contracts with neighbouring CCGs or on a case-by-case basis. The DO cohort comprised individuals who accessed one of these regional diagnostic assessment services; thus, these individuals did not receive any post-diagnosis support from the SAT.

Setting

Nine SATs (referred to as ‘research sites’) that were broadly representative of the range of service characteristics and ways of working observed in current SAT provision in England (see Chapter 3).

Study participants

The SAT cohort inclusion criteria for participants were that they were:

• diagnosed with autism by the SAT or already had a confirmed diagnosis when referred

• able to give informed consent, as judged by SAT practitioner in the research sites.

The DO cohort inclusion criteria for participants were that they were:

• living outside the geographical commissioning boundaries of the SAT to which the individual has been referred

• referred to a SATs regional/national diagnostic assessment service

• able to give informed consent, as judged by SAT practitioner in the research sites.

Variables: standardised outcome measures

Outcomes were captured immediately prior to, at or immediately after the first full (diagnostic/needs) assessment appointment (T0), and at the following follow-up time points: 3 months (T1), 6 months (T2) and 12 months (T3) after T0. For those recruited earlier in the study, 18- (T4) and 24-month (T5) follow-ups were also possible. The following suite of standardised outcome measures and categorical outcome indicators was used (further details are provided in Appendix 7).

Study recruitment and retention

Recruitment and T0 (baseline) data collection took place within the following time window: from no longer than 3 weeks prior to the start of the full assessment (that is, diagnostic and needs assessment or, for those already diagnosed, needs assessment) to no later than 7 days after (see Appendix 8 for an account of the recruitment and data collection process). Figure 4 summarises recruitment and retention to the study. The recruitment rate was 57.20% (424/741). Over one-quarter (n = 114; 26.88%) of participants became ineligible for the study after the diagnostic assessment did not diagnose them with autism (n = 103) or the individual withdrew from the diagnostic assessment process (n = 11). Retention to the study at T3 (primary follow-up time point) was as follows: SAT D&S group, 133 out of 164 participants (81.1%); SAT SO group, 75 out of 88 participants (85.2%); and DO cohort, 52 out of 56 participants (92.8%).

FIGURE 4.

Flow chart: recruitment and retention to the study (not all participants complete each wave).

Analytical plan

The statistical software Stata^® 14.2 (StataCorp, College Station, TX, USA) was used.

Sociodemographic characteristics

Outcomes at baseline

Scores on standardised outcome measures at baseline are presented in Appendix 10, Tables 33 and 34. The SO group WHOQOL-BREF psychological domain mean score was significantly higher than that of the D&S group and DO cohort. No other differences between groups in baseline outcome scores were observed.

Discharge status at T3

At T3 (12-month follow-up), the SAT cohort study participants were asked to report if they had been discharged (and date of discharge) or were still in the service (Figure 5). The quality of data was poor; 13.5% of responses were coded as missing data (highly unusual in this study and perhaps indicating uncertainty) and a further 20.1% were coded as ‘study participant uncertain’ (based on information provided under the ‘other’ response option). Just under 40% reported that they were still using the service and over one-quarter (26.9%) had been discharged. From the data we have, it would appear that the D&S group were more likely to have been discharged at T3.

FIGURE 5.

Status at T3 (12-month follow-up) in (a) the D&S group and (b) the SO group [n (%)].

When the study was designed, and in consultation with service leads, we had expected the majority of study participants to be discharged by T3.

Concerns worked on: Specialist Autism Team cohort

At the 12-month follow-up (T3), we used a checklist to ask about the concerns that participants had worked on with the SAT (Figures 6 and 7). There was no difference between the D&S and the SO groups in the total number of concerns worked on. However, there were some differences in the type of concern. For example, a smaller proportion of the SO group reported working on understanding of autism, living with autism and family members’ understanding of autism. By contrast, compared with the D&S group, a greater proportion of the SO group had worked on employment, financial and social network/relationship issues. Across both, around one-third reported that they had worked on managing anxiety and/or other emotional difficulties.

FIGURE 6.

Concerns worked on with the SAT: D&S vs. SO groups.

FIGURE 7.

Distribution of D&S vs. SO groups in terms of number of concerns worked on.

Mental health

Table 8 describes changes in mental health outcomes for the D&S and SO groups between T0 and T3. For both of the groups, differences in WHOQOL-BREF psychological domain and GHQ-12 mean scores between T3 and T0 were not significant. For the D&S group, the proportion of the sample scoring below the GHQ-12 clinical threshold (i.e. sample mean score at T0) increased between T0 and T3, with this increase being statistically significant (p < 0.01). This was not the case for the SO group.

Quality of life

For the D&S group, no statistically significant changes in scores from T0 to T3 were observed in measures of health-related quality of life (EQ-5D-5L) and the other WHOQOL-BREF domains (social, physical and environmental) (Table 9). This was also the case for the SO group, except for WHOQOL-BREF social domain scores, for which a statistically significant deterioration was observed (p < 0.05), representing a small effect.

Perception of social networks

In terms of perceptions of social networks (measured using the ‘belonging’ subscale of the ISEL-SF), changes in score between T0 and T3 were not significant for either the D&S group or the SO group (see Table 9).

Daily living: occupations and activities

Five categorical indicators of daily living were used in the study:

1. perceived ability to manage the usual activities of daily living, captured using the relevant item of ED-5D-5L

2. perceived sufficiency of information needed for daily living, assessed using the relevant item of WHOQOL-BREF

3. change in employment status of those in paid work, job-seeking or unable to work because of illness or disability at T0

4. satisfaction with capacity for work, assessed using the relevant item of WHOQOL-BREF

5. satisfaction with how they spent free time, captured using an item designed for this study.

Access to autism-specific networks and support

Longer-term outcomes

For individuals recruited early to the study, the opportunity was taken to collect longer-term follow-up data at 18 (T4) and 25 (T5) months (D&S group: T4, n = 94/133, T5, n = 62/133; SO group: T4, n = 56/88, T5, n = 42/88). Tables reporting our descriptive analyses are presented in Appendix 12. We offer a brief overview of findings here.

Characteristics with no evidence of association with T3 mental health outcomes

Regression analysis explored, one at a time, the association between individual and service characteristics and mental health outcomes (WHOQOL-BREF psychological domain, GHQ-12). Individual characteristics for which no evidence was found of an association between the characteristic and the T3 mental health outcomes were:

• diagnostic status at referral (i.e. D&S group vs. SO group)

• functioning at referral (EQ-5D-5L usual activities domain)

• contact with autism-specific groups/communities at T3.

The status in service (i.e. discharged vs. still in service) was significantly associated with T3 mental health outcomes. However, it was not taken forward to modelling owing to the large numbers of missing data (42/180, 23%).

Service characteristics for which we found no evidence of an association between the characteristic and the T3 mental health outcomes were:

• LA involvement in commissioning/funding the service (taken to indicate social care expertise and easier access to LA services)

• single versus multiteam service structure

• autism versus neurodevelopmental service

• mode of delivering psychoeducation post diagnosis

• whether or not they routinely do one-to-one work for (non-complex) mental health problems.

Characteristics with evidence of association with T3 mental health outcomes

Tables 16 and 17 present outputs from the multiple regression modelling for T3 for the WHOQOL-BREF psychological domain and GHQ-12, respectively. Varying degrees of strength for evidence of an association (almost always in the same direction) were observed for all individual and service characteristics with one or both of the mental health outcomes.

Changes in mental health around the point of diagnosis

A descriptive analysis of mean GHQ-12 scores at T0, T2 (post-diagnosis time point) and T3 (12-month follow-up) was carried out, and results for the D&S group and DO cohort were compared (Figure 8).

FIGURE 8.

Mean GHQ-12 scores at baseline, post diagnosis and the 12-month follow-up in (a) DO cohort and (b) D&S group.

The overall trajectory of the mean GHQ-12 score for the D&S group was in a positive direction. This was not the case for the DO cohort, with evidence of a deterioration between T0 and T2.

Outcomes at the 12-month follow-up

There were no statistically significant changes in the DO cohort mental health outcomes at T3 compared with T0 (Table 18).

Similarly, no statistically significant changes in outcomes were observed in terms of quality of life (EQ-5D-5L, WHOQOL-BREF social, physical and environment domains) and perceptions of social networks (ISEL-SF belonging subscale) (see Appendix 13, Table 55). Likewise, no changes were observed on indicators of daily living or regarding access to autism-specific networks (see Appendix 13, Tables 56 and 57).

Comparing T3 mental health outcomes for the Diagnosis-Only cohort and Diagnosis and Support group

We then used ANCOVA to compare the T3 mental health outcomes of the DO cohort with those of the D&S group. This was repeated, restricting the D&S group to those individuals who used SATs that also provided a regional diagnostic service and recruited to the DO cohort. Outputs are presented in Appendix 14 (Table 58: WHOQOL-BREF psychological domain; Table 59: GHQ-12). No significant differences in mental health between the two groups were observed at T3.

Family members’ accounts

Family members’ accounts of their partner’s or son’s/daughter’s reaction to learning their diagnosis fell quite evenly into one of three types. Some believed it had been a relief to their family member. One parent recalled being surprised at the level of their son’s relief. This had made them realise that they had been unaware of how much the diagnosis meant to their child:

He was thrilled to know why he’d been feeling the way he had for so long, and we didn’t know he was suffering like that.

F3

Others described their relative’s reaction as neutral or ambivalent. Interestingly, these observations did not always tally with their partner’s or child’s account where, as well as providing a new understanding of themselves, feelings of shock, depression and anxiety were described.

A further group reported a very negative reaction, with one describing high levels of distress:

The diagnosis completely floored her. She was beside herself and because she suffers with other mental health issues she literally says, ‘Well that’s it, my life’s over, I may as well not even be here’.

F9

Another spoke of how, although she felt very positive about her son getting a diagnosis because it ‘opened doors all over the place’ (F4), her son’s response had been very different. She further explained that he still (approximately 12 months after diagnosis) got ‘depressed’ if he thought about it.

Diagnosis-Only cohort interviewees

Diagnosis-Only cohort interviewees consistently described the follow-up session as including a discussion of the written report of the diagnostic assessment; being provided with information leaflets about autism; being encouraged to seek further information from the internet; and being signposted to relevant services.

Two elements of this session were regarded as valuable and helpful: explanations as to which aspects of their difficulties were linked to autism (and which were not) and advice on living with autism (e.g. encouragement to develop daily routines and techniques for avoiding ‘meltdowns’).

However, shortcomings and concerns with regard to other aspects of the feedback session were described. First, some reported that information leaflets were age inappropriate (e.g. targeted at young people) or included negative suggestions about what autism meant for people. Second, some who had followed the advice to use the internet as a source of information described coming across information and statistics that were highly negative about the implications of being autistic. This rendered interviewees feeling worried for their future.

Finally, signposting to other services (e.g. social services and employment advisors) was typically regarded as ineffectual and even demoralising. All of those advised to contact other services had not done so, pointing out that the very fact that they were autistic made this difficult, even if they were ‘high-functioning’:

They [service] say your diagnosis opens doors. What doors? I find that quite difficult, because somebody, for me, has to open the door else I can’t get through it.

SU33

Thus, a clear and strong theme in the DO cohort interviewees’ accounts was the insufficiency of the input that they had received post diagnosis. Indeed, for some, there was a sense of an experience of abandonment in their accounts:

What I got was I got a diagnosis and then they gave me a few leaflets and said ‘There you go, good luck, you won’t be seeing us again, take care, bye.’. And that was really frustrating.

SU27

You shouldn’t be diagnosed and then left to your own devices. There was disappointingly little . . . I got something from them [the diagnosis]. I suppose I should be grateful for that. Some people get nothing at all.

SU10

Among those interviewees who described being shocked or confused by their diagnosis, the limited nature of follow-up support seemed to be particularly problematic, exacerbating negative emotional responses. For example, one participant, who had not expected to be diagnosed, said that discharge left her feeling like she had been in a ‘hit and run accident’ (SU30). Another participant said diagnosis because it ‘opened doors all over the place’, her son’s response had been very different that if it had not been for the support offered by people outside the service he would have felt like he had been ‘hung out to dry’ (SU27).

Specialist Autism Team cohort interviewees

The SAT cohort interviewees who attended the psychoeducation intervention offered were, overall, very positive about this intervention:

It ended a lifetimes worth of feeling inadequate and feeling worthless.

SU36

It was valuable beyond words. It [the diagnosis] could have been life-changing in a way that left me stranded and in fact it has been life-changing in a way that’s given me enormous support and enormous hope, and a new capacity for thinking about myself and the world around me.

SU35

Aspects of the intervention that were identified as particularly helpful included:

• content that gave insight into how autism may affect them, including the positive aspects of the condition

• content on dealing with disclosure of diagnosis

• learning coping techniques for dealing with difficulties commonly experienced by autistic adults (e.g. sensory overload, anxiety and dealing with social situations)

• being directed to other reliable sources of information

• listening to and/or sharing experiences (group-delivery only)

• speakers with positive stories of living with autism (group-delivery only).

A dominant theme in these individuals’ accounts was the impact on how they viewed themselves. They spoke about feeling more accepting or forgiving of themselves. Some also described feeling less need to mask their autistic behaviours and/or the positive aspects of being autistic:

It helped me gain a deeper understanding of my own struggles and why I have them, but it also emphasised that having autism has given me a lot of strengths which is not something I had considered before.

SU36

However, a small proportion of individuals had not accessed any psychoeducation input. For the majority, this was because the SAT that they used offered only a group-delivered intervention that they were unable or unwilling to attend. In addition, one interviewee had waited 1 year to receive psychoeducational input. He described enduring months of sleep difficulties, regularly staying up into the early hours to reread the lengthy assessment report (over 150 pages) in an attempt to understand the diagnosis and its potential impact over the course of his life.

Positive impacts

All participants said that being diagnosed with autism had some positive impacts, although the nature and extent of this varied considerably.

The most commonly reported positive impact was an increased understanding of self, including why they found certain environments or situations difficult. Some believed that this understanding had reduced their anxiety and/or led to improvements in self-esteem. Others reported that it had enabled them to develop strategies to help them cope better with life (e.g. developing a better morning routine and making greater use of calendars and lists) and taking more care of themselves (e.g. allowing themselves more ‘downtime’, avoiding ‘meltdown’ triggers and ending difficult relationships). Having the diagnosis was also said to have helped family and friends to better understand their behaviour and the things that they found difficult or challenging. In some cases, this resulted in improved relationships.

A second way that being diagnosed with autism had achieved positive impacts in interviewees’ lives was that it had reduced interviewees’ sense of isolation. Comfort was drawn from simply knowing that there were other people who had similar life experiences to their own. Very occasionally, interviewees reported joining a local autism group or had friends disclose that they were also autistic.

A third mechanism by which being diagnosed had a positive impact was the way it enabled access to practical assistance, such as welfare benefits (Personal Independence Payments, Employment Support Allowance, Disabled Students Allowance), university/college/workplace assessments, adjustments and support.

Fourth, being diagnosed caused improvements in the quality of care and support received from other services, with practitioners using diagnostic assessment reports to help direct care or respond more appropriately to an individual’s needs. Finally, and occasionally, it had triggered other family members to undergo an autism diagnostic assessment.

Negative impacts

Set against these positive experiences, some interviewees reported long-standing or unresolved difficulties related to the diagnosis. The characteristics of this group are consistent. It comprised all interviewees from the DO cohort and those from the SAT cohort (3/20) who had received very limited or no psychoeducational input because either they had chosen not to or they had been unable to access the psychoeducational intervention offered.

Among those in our sample from the DO cohort, more than half said that they were still struggling to come to terms with and understand their diagnosis and the implications it may have for their lives:

. . . still quite overwhelming at times . . . the diagnosis in what it actually means. It’s very much a positive thing. However there isn’t a day that goes by that the A word doesn’t flit through my thoughts. It’s everywhere, life-changing.

SU29

I just don’t understand it and I don’t get the condition, I don’t, and as much as I try and read about it, it just doesn’t make any sense to me.

SU32

I want to understand what happens now. So it’s kind of like I’ve been given a diagnosis and a pile of paperwork and it’s like, well, what does this actually mean? Is this gonna hold me back at work, is it gonna hold me back in university, do I need to compensate for the way that I act socially? I don’t really know, I don’t really know what it means.

SU27

The vast majority of DO cohort interviewees said that they wanted further help to understand and come to terms with their diagnosis. Some described how they were still actively searching for information and sources of support. Even those who were initially relieved to be diagnosed and continued to view the diagnosis positively spoke of struggling or feeling overwhelmed by it at times. As a result, they were keen to point out the importance of high-quality psychoeducation and wider post-diagnostic support:

It’s a dedicated assessment centre so I believe it does well what it does. However . . . diagnosis is a traumatic experience, life-changing. Some will handle it better than others. I guess I’m one of the lucky ones who is quite positive about the diagnosis and life in general and have found ways to self-manage quite well . . . others are not.

SU29

A few of the DO cohort interviewees believed that receiving the diagnosis had caused a significant detrimental effect on their mental health:

I spent a year thinking I needed to die because of what someone had told me, because there’s nothing for you if you’re autistic, because you can’t change the way you are.

SU33

Even a year down the line I’m still seeing it as kind of a death sentence like, right, that’s it, my life’s over, I don’t know how to deal with this condition and everything, and you end up looking at everything differently, and it triggers loads of stuff off that you might not necessarily have thought of before, and I think that can then get you quite stressed and quite anxious.

SU32

In all of these cases, interviewees attributed the severity of their mental health difficulties to the fact that psychoeducation and other post-diagnostic support had not been available to them. Two interviewees explained that they had eventually sought help from a ‘therapist’ (unclear if this was an NHS referral or private practitioner) when their feelings of anxiety and depression and suicidal thoughts spiralled to an unmanageable level.

Family members’ accounts: impacts on the individual

Family members of individuals from both cohorts described positive impacts of their family member receiving a diagnosis. These included their relative seeming more ‘comfortable’ with themselves, and that it had enabled access to practical support related to participating in everyday life and supporting achievement of things they had hoped for (e.g. travelling overseas).

Family members of SAT cohort interviewees also typically regarded the diagnosis as opening access to additional care and support (including that directly provided by or accessed through the SAT). This had brought a range of, sometimes very significant, benefits. These included improvements in mental health, greater independence, an improved ability to manage everyday life and assistance with disclosure of their diagnosis to work colleagues:

Since he’s had the assessment he’s not talked of killing himself . . . he’s not actually said, since then, that, I might as well hang myself or I wish I could just cut my head off. He hasn’t said anything horrible like that, since this has been going on with these, you know, weekly meetings, etc. So they must be doing something, mustn’t they?

F4

By contrast, some family members of DO cohort interviewees saw no positives coming from the diagnosis, per se. Furthermore, when the diagnosis did not then lead to additional support (and potentially positive impacts for the individual), this could be a source of frustration:

It was time consuming to do and where did it get us? Although they’ve said you now have an official diagnosis, it’s like so what? Because it doesn’t help, it doesn’t do anything for her, she can’t access any services.

F7

One parent (DO cohort) firmly believed that the diagnosis had had a negative impact on her child’s life:

She’ll say . . . ‘I can’t help it. I’ve got Asperger’s. I’m on the autistic spectrum. There’s nothing I can do about it.’ It’s not changed her behaviour as such because she’s always had meltdowns, but it’s as if she thinks it is her lot now and how life has to be.

F9

Family members’ accounts: impacts on themselves

Parents who were closely involved in their child’s life also reported impacts on their own lives. For the couple of other family members we interviewed, impacts were minor or much less significant.

Across both cohorts, parents reported that the diagnosis, in itself, had been sufficient to have a positive impact on family relationships. This was for two reasons. First, having some understanding of the condition meant that situations that might trigger upset or arguments could be avoided. Second, it prevented family members misinterpreting behaviour as deliberately difficult.

However, where individuals had reacted negatively to the diagnosis, support from the service appeared to play a major role in determining whether or not and how this affected parents:

She constantly tells us ‘you’re rubbish, you’re a bad parent, you don’t do this, you don’t support me, you’re not getting me any help, you’re not getting me this, you’re not getting me that’. You do feel . . . well yeah, I am a rubbish person.

F9

This contrasts sharply with another family’s experience in which the individual received a one-to-one psychoeducation intervention. This parent reported a marked improvement in her son’s mental health, which, in turn, led to significant improvements in their own levels of stress.

Finally, few family members said that they wanted help for themselves. Instead, they explained that by helping the individual, services would also be helping the family. A minority, however, did express a desire for support. Suggestions included, for emotional support, having access to a parent/carer support group and information about other possible sources of support.

Setting

The setting was the same as that for the outcomes evaluation (see Chapter 8).

Study participants

Inclusion criteria for the SAT cohort and characteristics of sample members are described in Chapter 8.

Costs of Specialist Autism Teams

The SAT costs (measured in Great British pounds) were calculated using financial information (overall annual budget for 2017–18) obtained by the research team from each SAT or its managing organisation, combined with data on the total number of clients supported in each site (also obtained from SATs themselves), to give an overall average cost per study participant for each SAT. It was not possible to carry out microcosting of activities within the SATs because of the considerable time that any such exercise would have taken, nor did those services hold information in a way that would allow separate costing of their various diagnostic and post-diagnostic activities. The advantage of top-down costing is that it is easier to operationalise than microcosting, considerably less time-consuming for the researcher and considerably less intrusive for services. An exploration of cost variations partially compensated for this disadvantage by including regressors that reflect key service and individual components in the analyses.

Some, but not all, SATs provided us with high-level budgetary information. Where financial information was missing, costs were imputed, taking into account differences between SATs in the proportion of D&S and SO service users on caseloads, as this potentially has an effect on average per-person costs. This was estimated from the proportions in our study sample. When we later analysed cost variations and cost-outcome links, we used both SAT costs and (other) service costs.

Service utilisation and associated costs

Service utilisation patterns (health, social care and other statutory sectors) for individuals were collected with an adapted version of the CSRI that was created specifically for this study. It was completed at each time point, covering a retrospective period of 4 weeks.

Services provided within the SAT were assumed to be covered by the SAT budget. However, contacts with professionals and services reported in the CSRI are likely to represent some double-counting, as we know generally that study participants can find it hard to attribute professionals to specific roles and/or services. We, therefore, conducted additional analyses to test how sensitive key findings were to the possibility of some degree of double-counting. We did this, for example, in the cost variations analyses by setting each of the largest cost components to zero (to represent no contact with that particular service outside the SAT) and examining whether or not it affected comparisons between different types of SATs/SAT characteristics.

The CSRI also collected information on employment status (allowing calculation of productivity costs associated with days off because of sickness absence) and privately borne costs. It was not possible to collect data from family members or others on the amount or nature of unpaid care and support. If a participant was categorised as being in full-time paid employment, the number of days taken off work in the last 3 months was multiplied by the number of hours worked per day (assumed to be 7) multiplied by the minimum wage (£7.83 for adults). If an individual worked part time, the number of hours worked per week was divided by 5 working days to obtain the number of hours per day. This was multiplied by the minimum hourly rate at that time of £7.83. If someone was self-employed, the costs for lost productivity were treated in the same way as for a full-time worker. For other work status responses, such as trainee, voluntary worker, job-seeking, not working owing to sickness/disability, student, maternity/paternity leave, carer and retired, we did not attach monetary values to them and cost was assumed to be zero.

The focus of the economic evaluation, as set out in the prespecified analysis plan, was costs between baseline and 12 months that were computed from data collected for 4-week retrospective periods at 3 months (T1), 6 months (T2) and 12 months (T3), and interpolated between time points to get a full 52-week costing. Sample sizes at 18 and 24 months were insufficient to conduct robust analyses, and analysis beyond 12 months was not part of the original design for the study. Unit costs utilised in our analyses are given in Table 19.

Statistical tests

Independent t-tests compared means for the annual health and social care costs and societal costs by different cost components. Bootstrapping was used (bias corrected) owing to skewed costs.

Specialist Autism Team costs

Cost information was obtained from four of the nine SATs. One SAT (multiservice delivery model) provided partial information. Of the remainder, two declined to provide cost information, whereas the other two agreed to provide cost information but did not deliver it, despite reminders. For the SATs with cost information, average costs per client were estimated based on the method described in Costs of Specialist Autism Teams. This costing did not take account of other functions/services included in commissioning arrangements (e.g. training/support to staff in mainstream services, public awareness raising and provision of low-level support post discharge). The average cost per client for the four SATs from which we could obtain budget data and information on the total number of service users per year were £360, £768, £781 and £2951. Differences between SATs may be linked to operational scale (for example, the lowest-cost SAT supported more than four times as many autistic people than the highest-cost SAT), location (influencing some input prices), characteristics of people being supported and the range of services delivered (e.g. some SATs did not accept referrals from those already diagnosed and the extent of direct work). We can examine the last two potential sources of variation with data collected in this study. For SATs that did not provide financial information, we imputed from these observed cost data (see Costs of Specialist Autism Teams).

Comprehensive support costs

Service utilisation data and costs for services provided by SATs (captured by SAT budgets) were summed to give a subtotal for the health and social care costs. In addition, costs were calculated for employment-related and other ‘formal’ sectors (police officer contacts) (see Appendix 15, Tables 60 and 61).

Significant differences in costs between D&S and SO groups over the 12-month period were found in relation to accessing three professions/interventions: psychologist (p = 0.05), nurse (p = 0.05) and fixed-session, group-delivered interventions (p = 0.05) (see Appendix 15, Table 62). These are key features of SAT provision. Sensitivity analyses to address the possibility that there was double-counting between the SAT budgets and the service utilisation reported by study participants were, therefore, conducted.

Aggregating costs across the different service components revealed that the D&S group had service-related costs totalling £2546 over the 12-month period, compared with £1699 for the SO group. Societal costs (again excluding SAT costs for the moment) summed to £2733 for the D&S group and £1931 for the SO group (see Appendix 15, Table 60).

Aggregated costs for each 4-week period of data collection by time point and group (D&S vs. SO) are shown in Table 20. At baseline, the total health and social costs were £125 for the D&S group (n = 164) and £151 for the SO group (n = 88). Costs increased over time, probably as a result of SAT involvement, although comparisons between time points should be made cautiously because of sample attrition: individuals for whom costs were available are not identical between time points. There were no significant differences in costs between the D&S and SO groups at any individual time point.

Cost variations

As noted earlier, analyses of cost variations took each measure of cost over the 12-month period between T0 and T3 as a dependent variable, which was regressed on individual characteristics at baseline and SAT characteristics, adjusting for clustering by site. Generalised linear modelling with gamma distribution and identity link was used for these analyses as the best-fitting model with the lowest Akaike information criterion (AIC) values among other generalised linear modelling model specifications. As noted above, two series of regression analyses were conducted: health and social care costs, and societal costs.

Cost–outcome links

Examination of the cost–outcome links focused on the primary outcome: WHOQOL-BREF psychological domain. Seventeen cost components (over the 12-month period to T3) were included (GP, psychologist, occupational therapist, nurse, SLT, social worker, support worker, group activities lasting a fixed number of sessions, support group, social group, outpatient care, accident and emergency, walk-in centre, day case, inpatient care, police officer and private appointment with other therapists), along with referral group, age, gender, WHOQOL-BREF psychological domain at T0 and baseline societal costs.

The regression equation is presented in Appendix 15, Table 64. Some individual and SAT characteristics were found to be associated with inter-individual differences in outcome (broadly consistent with the analyses reported in Chapter 8), but there was no significant association between the WHOQOL-BREF psychological domain at 12 months and the total annual societal costs. Significant predictors were baseline WHOQOL-BREF psychological domain scores and GP costs. In other words, differences in GP costs and WHOQOL-BREF psychological domain scores at baseline were driving some differences in the WHOQOL-BREF psychological domain scores at T3.

Cost-effectiveness analysis

The cost-effectiveness of different service characteristics was examined by looking at the estimated coefficients on service characteristic indicators in the cost and outcome regressions. The cost regressions (one for health and social care costs and one for societal costs) have been reported in Cost variations. In this section we report the outcome regressions, first for WHOQOL-BREF psychological domain and then for QALYs. Again, we used generalised linear modelling and found that a Gaussian family distribution and identity link had the best fit (lowest AIC values).

In the first regression (WHOQOL-BREF psychological domain, health and social care perspective) (see Appendix 15, Table 65), three of the service characteristics indicators were significantly associated with T3 outcome variations:

• Team structure – individuals in SATs with multiservice teams had, on average, a 5-point higher score on the WHOQOL-BREF psychological domain than individuals in SATs with a single team structure, taking into account other covariates.

• Autism versus neurodevelopmental service – individuals in SATs with a neurodevelopmental service had, on average, a 10-point lower score on the WHOQOL-BREF psychological domain than individuals in SATs with an autism-only service.

• One-to-one work – individuals in SATs that routinely provided one-to-one work regarding (non-complex) mental health problems had, on average, a 17-point higher score on the WHOQOL-BREF psychological domain than individuals in SATs with a single team structure.

In the second regression (WHOQOL-BREF psychological domain), this time from a societal perspective (see Appendix 15, Table 66), four of the service characteristics indicators were significantly associated with T3 outcome variations:

• Local authority involvement – individuals in SATs with joint LA and CCG arrangements had, on average, a 9-point lower score on the WHOQOL-BREF psychological domain than individuals in SATs with just CCG arrangements, taking into account other covariates.

• Team structure – individuals in SATs with multiservice teams had, on average, a 4-point higher score on the WHOQOL-BREF psychological domain than individuals in SATs with a single team structure.

• Autism versus neurodevelopmental – individuals in SATs with a neurodevelopmental service had, on average, a 10-point lower score on the WHOQOL-BREF psychological domain than individuals in SATs with an autism-only service.

• One-to-one work – individuals in SATs that routinely provided one-to-one work regarding (non-complex) mental health problems had, on average, a 20-point higher score on the WHOQOL-BREF psychological domain indicator than individuals in SATs in which this was not routinely offered.

The third and fourth regressions took QALYs over the 12-month period as a dependent variable and analysed variations from, first, a health and social care perspective and, second, a societal perspective (in the adjustment for baseline cost). We also adjusted for baseline utility score (computed from EQ-5D); outputs are presented in Appendix 15, Tables 67 and 68.

In the first regression, three service characteristics were significantly associated with QALY variations:

• Local authority involvement – individuals in SATs with joint LA and CCG arrangements had, on average, a 0.683 higher QALY score than individuals in SATs with CCG arrangement only, taking into account other covariates.

• Autism versus neurodevelopmental service – individuals in SATs with a neurodevelopmental service had, on average, a 0.182 lower QALY score than individuals in SATs with an autism-only service.

• One-to-one work – individuals in SATs that routinely provided one-to-one work regarding (non-complex) mental health problems had, on average, a 0.395 lower QALY score than individuals in SATs with a single team structure.

From a societal perspective, the same three service characteristics were significantly associated with QALY variations over the 12-month period:

• Local authority involvement – individuals in SATs with joint LA and CCG arrangements had, on average, a 0.585 higher QALY score than individuals in SATs with just CCG arrangements, taking into account other covariates.

• Autism versus neurodevelopmental service – individuals in SATs with a neurodevelopmental service had, on average, a 0.194 lower QALY score than individuals in SATs with an autism-only service.

• One-to-one work – individuals in SATs that routinely provided one-to-one work regarding (non-complex) mental health problems had, on average, a 0.381 lower QALY score than individuals in SATs with a single team structure.

Bringing these analyses together, the findings in relation to the effects of service characteristics on costs, self-reported psychological quality of life and QALYs can be summarised as follows (Table 23):

• Local authority involvement – from a health and social care perspective, there are no cost or psychological quality-of-life differences between SATs with joint LA/CCG arrangements or SATs that are only CCG-led. From a societal perspective, there is no cost difference for joint LA/CCG SATs, but psychological quality of life is slightly worse (9 points on a scale that runs from 0 to 100 points). However, LA involvement was associated with higher QALY scores over the 12-month period. Overall, the cost-effectiveness of joint LA/CCG arrangements rests on the credibility of the QALY results.

• Team structure – psychological quality-of-life outcomes are slightly better for SATs with multiservice teams than for single team structures; however, the difference is only 4 or 5 points on the 100-point scale. Health and social care costs do not vary with this service arrangement, but societal costs are slightly lower for multiservice SATs. QALYs do not vary with team structure. Overall, these findings do not suggest major differences from a cost-effectiveness standpoint between team structures.

• Autism versus neurodevelopmental service – psychological quality-of-life outcomes are 10 points lower for individuals in SATs with a neurodevelopmental service rather than an autism service, which is marked on a scale running from 0 to 100 points. QALYs were also lower. Health and social care costs are slightly lower for neurodevelopmental service SATs and there is no difference in societal costs. Overall, these findings point to a cost-effectiveness advantage for autism-only services.

• Psychoeducation – there was only one significant difference between SATs whose dominant mode of delivery of psychoeducation post diagnosis was one-to-one delivery compared with SATs in which there was group delivery: societal costs were slightly higher for the former. Sensitivity analyses suggest that one-to-one delivery of psychoeducation might be slightly less costly (health and social care costs only) if our measurement of psychologist costs included some double-counting. Overall, however, there is no strong cost-effectiveness case for either one-to-one or group delivery.

• Skill mix – in the case of skill mix, there are no observable differences in psychological quality of life according to the richness of skill mix in the SAT being used, nor any differences in QALYs; however, individuals supported by SATs that had four or more professional disciplines had substantially higher costs than individuals in SATs that had only two or three disciplines. On cost-effectiveness grounds, and in terms of psychological quality-of-life and QALY outcomes, arrangements in which fewer professions are included appear to be preferred.

• One-to-one work – individuals using SATs that routinely provide one-to-one work for people with (non-complex) mental health problems have substantially better psychological quality-of-life outcomes than individuals using SATs not using one-to-one approaches either at no higher cost (health and social care perspective) or at reduced costs (societal perspective). On the other hand, QALYs were lower with one-to-one work. Sensitivity analyses suggest that one-to-one work might be slightly more costly (health and social care costs only) if both the psychologist and the nurse had been double-counted, although the difference was not great. Overall, given that psychological quality of life was the primary outcome and that the validity of QALYs generated from the EQ-5D has not been established for autistic people, we conclude that one-to-one work by SATs is a cost-effective way to deliver support for people experiencing mental health problems, as indicated by changes in self-reported psychological quality of life.

• Delivery of care plan – individuals in SATs with a managed approach to delivering care plans had significantly higher costs, yet psychological quality-of-life outcomes and QALYs were unaffected. Overall, we would conclude that an episodic approach appears to be more cost-effective than a managed approach with respect to this particular outcome.

Strengths

To the best of our knowledge, this is the first study to investigate SAT provision in England – a model of care consistently set out in government strategy since 2011 and recommended by NICE. The study has identified and described the approaches taken by localities to set up such provision, and explored practitioners’ experiences of implementing and delivering SATs. In the main study phase, and using a mixed-methods approach, we evaluated such provision, seeking to investigate and understand SATs from the perspective of practitioners, service users and family members, as well as tracking service users’ outcomes and their use of a wide range of specialist and generic services. A key objective of this chapter is to synthesise and integrate findings from these various elements of the study. Hopefully we have achieved this and, in doing so, have demonstrated the benefits afforded by the mixed-methods design.

Overall, the services recruited to the study as research sites represented well the range of approaches taken to SAT provision. Retention to the study and data completeness were excellent. All outcome measures performed well, including those not previously used with this population. We believe that this indicates that the outcomes evaluated and standardised measures chosen, in consultation with our PAG, were appropriate and meaningful. In addition, we worked with the PAG on the design and layout of the study questionnaire booklet. Thus, the hard copy was printed on pale-yellow paper with a dark-blue print, the layout was uncluttered and, where required, shading was used to distinguish between rows/items. Furthermore, at the 12-month follow-up, our primary time point and where highest retention was observed (> 85%), we included a question about the specific impact of the SAT on participants’ lives, with the option of using a free-text response format to describe experiences. Over half of the study participants took the opportunity to briefly share their experiences (reported in Chapter 6). As well as generating data, we believe that this also supported retention. Levels of retention also lend support to other strategies used. These included hard and online versions of questionnaires, the use of personalised/non-automated text message and e-mail alerts immediately prior to a data collection time point, and a three-stage reminder process using both electronic (e-mail, text message) formats and letters. At the same time, it is vital to acknowledge the commitment of study participants to this research.

In England, the care and support of autistic people remains a policy priority and there is a long-term commitment to improve provision. 33,54 Thus, the findings from this study (that commenced in 2014) remain highly relevant. Furthermore, although its focus has been on a model of care and service delivery in England – given the limitations in provision and service development in other countries, and a desire on their part to address this issue – the findings have a wider use and application. 27

The study was designed and carried out in the absence of a broader evidence base on autistic adults without LDs. There was, for example, a very limited literature to draw on to inform the selection of outcome measures when the study was being designed in 2013. Since then, however, there has been a burgeoning awareness of the need for a robust evidence base to inform the care and support of autistic adults without LDs. 27,37,55 Although the last couple of years have seen the publication of studies on outcome measurement or service design and delivery,27,37,55 the evidence base remains very limited. Thus, this study comprises a significant contribution not only with respect to models of providing care and support to autistic adults without LDs, but also methodically; for example, learning about recruitment and retention, data collection and outcomes measurement. With respect to outcomes measurement, members of the research team (EM and BB) have conducted an evaluation of the psychometric properties of the GHQ-12 using Rasch analysis (available from the authors).

Limitations

There are a number of limitations that are important to detail given their implications for the extent to which conclusions can be drawn.

Stage 1 sought to identify services that fulfilled NICE’s description of multidisciplinary, community-based provision for autistic adults: referred to by NICE as ‘Specialist Autism Teams’. We did not identify any service that fully aligned with NICE’s descriptive criteria of SAT provision. The key deviation was that all of the services identified worked exclusively with autistic adults without LDs, although some provided consultancy to other services with regard to all autistic adults. For this reason, this is not a study limitation but rather a ‘limitation’ of implementation of the SAT provision model. However, it does mean that the findings and conclusions drawn cannot inform service development for autistic adults with LDs.

Other variations in the integrity of NICE’s vision for SATs were also observed between SATs. For example, differences in the provision for carers, engagement in upskilling practitioners in mainstream services, extent of multidisciplinarity (and, by implication, holistic approach to care and support) and the provision or not of longer-term, low-intensity support. We have explored why services differed in the extent to which they have or could fully implement SAT provision, and this has generated useful and important evidence.

Evaluating the impact of different structural arrangements and approaches to care and service delivery on service user outcomes and resource use was a core study objective. A key finding from stage 1 was that each SAT was idiosyncratic. This meant that we could not cluster SATs into different ‘types’ or service models and then go on to evaluate and compare the different models using one or two exemplar services per model. Thus, as specified in the study protocol, we moved on to investigate how service characteristics (e.g. structural features, delivery approaches and ways of working) may affect user outcomes and resource use.

It is important here to note that in the quantitative component of this investigation, for each characteristic, the research sites clustered together differently. This was exacerbated by needing to have nine research sites to fulfil the sample size requirements within the study timeline. (The study had more research sites than originally planned owing to withdrawal of a large SAT early in study set-up, slower than expected throughput and, hence, recruitment in some sites, and significant delays in recruitment opening in two sites because of delays in the recommissioning processes.) These issues can, to some extent, be managed in the analytical process. However, it remains that we need to be careful in the interpretation of some findings and the weight given to them.

Finally, for the economic evaluation, data regarding SAT funding, budgets and caseloads were secured from fewer than half of the research sites. Perceived vulnerabilities with respect to funding and commissioning arrangements may have inhibited sharing of budgetary and costs information with the research team. Imputation was used to derive estimated costs for each site; however, the wide variation between sites mean that findings should be treated with caution. Taking a top-down approach to understanding service costs meant that a breakdown of staff costs/resource by the different SAT functions (e.g. diagnostic assessment, mental health interventions, drop-in provision and work with other groups, such as training and consultancy) was not possible. The very different approaches taken by SATs to, for example, the diagnostic assessment process (as reported in Chapter 3, Diagnostic assessment processes) illustrate the limitations of the top-down approach. They also serve to demonstrate the need for caution when interpreting findings. Certainly, to carry out microcostings would have required a substantial additional research resource. However, on reflection and in hindsight, given the novelty of the delivery model and the lack of existing research, this may have been a worthwhile investment.

In terms of the qualitative elements, one site was under-represented in the sample of service users interviewed for the study. This was because, unexpectedly, it closed recruitment quite early in the study timeline and there were, therefore, limited numbers of study participants eligible for interview (i.e. had used a service ≈12 months) when these took place towards the end of the study. However, this site was well represented in the sample who provided written accounts of their experiences within the T3 study questionnaire (reported in Chapter 6). In addition, although we used a sampling frame to ensure representation of a range of characteristics and experiences, the sample recruited to interview was self-selecting, although we note that > 70% of those invited were interviewed.

Furthermore, we did not recruit as many family members to the qualitative evaluation as had been planned. Given the topic of the interview, it was essential that the autistic adults chose whether or not a family member was also invited to take part in the study. Fourteen (out of 38) agreed to this and, of these, nine family members were recruited. This means that the conclusions we can draw regarding family members’ views of the impacts that using a SAT had on their relative (e.g. child, partner) are limited. In addition, our understanding of family members’ needs, and the actual or potential role SATs played in meeting those needs, is partial. It is important to note here that many of the interviews we did conduct with family members revealed significant concerns and difficulties, both for themselves and the autistic person. This accords with findings from existing research that has looked at family members of autistic adults. 24,56

Once the study was under way, a small additional element was introduced that offered an exploratory comparison of the outcomes and experiences of those on a diagnostic pathway provided by a SAT (D&S group of the SAT cohort) with those diagnosed by a regional/national diagnostic assessment service (DO cohort) in which no post-diagnosis support was available. It is important to highlight the initial nature of these findings. The quantitative evaluation was underpowered and the number recruited to our qualitative study was limited. Furthermore, we do not know if and how the quality of the diagnostic assessment process from which the DO cohort was recruited compares with other diagnostic services in England. The particular challenges associated with discriminating autism from co-occurring mental health problems is well documented,57 and indicates the need for a high level of expertise. Senior practitioners in our research sites believed that they offered high-quality diagnostic assessments (including at least one feedback session). However, they noted that this was not necessarily the case for other diagnostic assessment pathways operating across the country. Thus, these findings cannot be generalised and more research is certainly required in this area.

Finally, all sites reported higher than expected levels of demand on their service that had not been matched by an increase in resources. In addition, in two of our sites, there were extended periods during which a key post (e.g. specialist practitioner, clinical psychology) was unfilled. The consequence for all services was longer wait times between intake and full assessment and, potentially, longer wait times for interventions set out in the care plan. Findings from our qualitative data collected from service users (see Chapters 6 and 7) indicate that this may affect both the service user experience and the impact.

Supporting sustainability

An important component of this nested study concerned distilling learning and opinions about whether or not and how SAT provision should further develop. All services were convinced of the importance and value of SATs having a core role in the care and support of autistic adults without LDs, both those needing/newly diagnosed and those diagnosed as children or earlier in adulthood. However, to ensure that there are sustainable improvements in support for autistic adults without LDs, senior practitioners believed that some shifts in the role of SATs, and the emphasis and priority given to certain functions by such services, was required. They were as follows:

• while maintaining the availability to provide specialist autism interventions and support, placing greater emphasis and resource on consultation and supervision of practitioners working in mainstream/generic services

• ensuring practice and interventions offered supported self-management rather than fostered dependency

• developing and investing in the low-intensity support available post discharge; ideally, but not only, this would incorporate collaborative working with local peer-led networks.

Such suggestions are not to say that senior practitioners believed that the vision of SATs set out by NICE needed modification. The guideline31 clearly specifies SATs as being responsible for providing and/or co-ordinating care and support (our emphases), and the Autism Strategy30 identified autistic adults without LDs as being likely to particularly benefit from preventative support. Rather, they believed that service specifications may need revising. Based on their experiences to date, senior practitioners emphasised that such changes require ‘buy-in’ and commitment from commissioners. They noted, however, that it had proved very challenging, or indeed impossible, to secure funding for, for example, drop-in services or other preventative-type provision. The absence of an evidence base on the impacts and effectiveness of low-intensity support was a barrier to making a case for incorporating such provision in the SAT offer.

With respect to increasing the emphasis given to the consultative role, it is important to note that senior practitioners were very clear that it would not remove the need and demand for specialist autism interventions provided by practitioners with extensive expertise in autism, and for those with complex needs. They also noted the potential risks associated with under-trained professionals assuming a level of autism-expertise or competence.

Furthermore, adopting a consultative/supervision model was dependent on mainstream services being willing and allowed to work in this way. Although ‘consultative’ models of health-care delivery have been implemented in other fields of health care, it is not a familiar approach in community adult mental health. The recently published NHS Long Term Plan,54 however, points in that direction, noting the requirement for health-care providers to ‘make reasonable adjustments’ so that autistic people can access and use their services.

Supporting self-management, making available long-term low-intensity support from SATs (e.g. telephone ‘clinics’ and drop-in services) and supporting autistic adults to connect with peer-led groups and communities all point to building resilience and preventing future difficulties, even crises. We noted earlier the challenge of making the case for investing in such provision, although some modelling work published by the National Audit Office demonstrated that such provision for autistic adults without LDs may be cost-saving, or at least cost neutral. 28 Wider evidence indicates a potential benefit of supporting self-management and for its integration within the care of people with long-term conditions. 66,67 However, robust evidence of their impact and effectiveness for autistic adults without LDs is required. 36,68

However, and particularly with respect to the notion of promoting community-led and peer-led support, it is important to also draw attention to findings from stage 2 concerning the membership and contact with autism-specific third-sector organisations and peer-led groups and communities (see Chapter 8). We found that less than one-quarter of study participants were members of an autism-specific voluntary organisation, with a similar proportion having had any contact with such an organisation in the 4 weeks prior to each data collection time point. Although an increase in contacts with such organisations was observed at the 12-month follow-up, and appeared to be sustained into the longer term, it is important to remember that, for the majority, no change was observed. This finding highlights that some people may be unable, even with support, to use peer support. Equally, and as we report in Chapter 8, others may be disinterested or unwilling to do so.

Domains of impact

Chapter 6 reported findings from our analysis of free-text data collected at T3 on the ways that SATs had affected study participants’ lives (or not). Over half of the T3 respondents responded to this question. The impacts described were wide ranging and serve to illustrate the range of needs that an autistic adult without LDs may have. We organised them into seven broad categories: understanding of autism, acceptance of self, improved mental health, reduced sense of isolation, improved relationships and social networks, help with employment and education, and supporting access to other services. In addition, the positive impact of contact with supportive and autism-expert practitioners was reported. Together, these provide evidence that some degree of holistic care was being achieved by the SATs we evaluated. Our data on service users’ reports of the number and range of concerns that they had worked on with their SAT (see Chapter 8) corroborate this, as do findings from our in-depth semistructured interviews with service users (see Chapter 7).

Thus, in these interviews, a similarly wide range of needs was described. In addition, it was clear that interviewees varied considerably in the severity and complexity of their needs. For example, some of those referred via the diagnostic assessment pathway (and including post-diagnostic psychoeducation) identified no further needs for which they required the support of the SAT. Others, however, had multiple and/or long-standing needs and difficulties. In terms of perceived outcomes of using a SAT, we heard a range of experiences. Some described their needs as being fully or predominantly met and this included those who, on referral to the service, had significant difficulties. Others had a more mixed experience, with some needs being met and others remaining unmet. The accounts of a third group indicated that their needs remained predominantly unmet. In a subsequent section (see Factors affecting outcomes of using a Specialist Autism Team) we report our findings as to why there were these different experiences.

Outcomes

Overall, and in line with other studies,24,69 on all domains of the WHOQOL-BREF mean scores were lower for the sample members than for UK norms. 70,71 With respect to the GHQ-12, compared with other studies that have used this measure, a greater proportion of our sample were scoring above the population mean clinical threshold (> 80% vs. 40%). 72,73 Using EQ-5D to generate utility scores, the sample had markedly worse generic health-related quality of life than that for population norms (e.g. 0.696 for those aged 23–34 years compared with an England norm of 0.919). 74

A number of changes in outcomes were observed between baseline and 12-month follow-up (T3) (see Chapter 8). For the D&S group, a statistically significant improvement in the proportion of study participants scoring below the GHQ-12 clinical threshold was observed. Although the WHOQOL-BREF psychological domain mean score also improved, this was not a statistically significant change. The picture for the SO group is different, with no statistically significant improvements in mental health observed.

With respect to our secondary outcome measures (EQ-5D-5L, WHOQOL-BREF social, physical and environmental domains, ISEL-SF belonging subscale), for the D&S group, changes in the mean score between T0 and T3 were slight and non-significant. The same pattern of findings was observed for the SO group, except for the WHOQOL-BREF social domain, for which a statistically significant (p < 0.05) deterioration was observed at T3.

We also investigated changes in outcomes using a set of categorical indicators that assessed two broad domains: daytime occupation/usual activities and access to autism networks and support. For the D&S group, a statistically significant increase (p < 0.05) in the proportion of participants reporting no/slight problems with managing daily living was observed. However, there was no evidence of a positive change in terms of perceived availability of information needed for everyday living, employment status, satisfaction with capacity for work and satisfaction with leisure time. For the SO group any changes were non-significant.

In terms of access to autism networks and support, for the D&S group there was no significant change in the levels of membership of autism-specific local/regional/national groups and/or online-only communities. Indeed, at T3, less than one-fifth were members of such an organisation/community. However, there was a significant increase (p < 0.01) in the proportion of the sample who reported some contact with such an organisation/community in the 4 weeks prior to T3. That said, it is important to note that the majority (73.9%) had no contact. For the SO group, membership levels also remained low at T3 (16.1%). Unlike the D&S group, no significant change in levels of contact was observed.

To summarise, for the D&S group, there was some evidence of improvement in mental health, ability to manage everyday living and use of autism-specific third-sector/peer-led organisations at T3. Findings for the SO group differed, with the statistically significant change observed only being a deterioration in social quality of life (as measured by WHOQOL-BREF’s Social domain). We would note that the size of the SO group is relatively small and, therefore, detecting changes in outcomes may be compromised.

There are a number of possible explanations for the differences observed between the D&S and the SO groups. Thus, the groups differed in their sociodemographic characteristics. Overall, the SO group was younger and, therefore, more likely to have been diagnosed in childhood. This also meant that they were more likely to be students and to be living with their parents. Thus, the potential support networks of the two groups may differ and this may have affected outcomes. In addition, at baseline, the SO group reported a better mental health quality of life. This may indicate, for at least some, that needs were less severe or less pervasive in their impact.

An alternative or additional explanation is that some participants in the SO group may have unresolved difficulties regarding their autism diagnosis that meant that they were unable to (fully) benefit from the interventions provided by the SAT. SATs do not routinely offer a psychoeducational intervention to those referred to their service who are already diagnosed (i.e. the SO group). However, recent studies report that young adults diagnosed as children may hold incorrect beliefs about autism. 75,76 Furthermore, evidence is starting to emerge on the potential role of psychoeducation in preventing poor mental health outcomes because it seeks to support understanding, personal acceptance and a positive view of autism. 77,78

It may also be the case that the expectations and needs of the D&S group and the SO group differed. In our interviews with service users, those sampled from the SO group interviewees reported a wide range of, sometimes quite specific, needs associated with managing everyday life and social relationships. By contrast, and aligning with other research,79,80 the expectations of the D&S group were strongly centred on the diagnostic assessment, particularly as a means to offering sense or validation of their lives. That is not to say, however, that some also had quite significant mental health and/or social needs. However, these needs tended to be expressed in more global terms. Therefore, it is possible that the ‘lack of fit’ between the users’ needs (and expectations) and the care and support SATs offered was greater for the SO group than for the D&S group. Other findings lend some support to this argument. First, there was a difference between groups in the proportion reporting that the SAT did not work on any of their concerns (12.3% vs. 6.8%) (see Chapter 8). In addition, the SO group was more likely to report little, no or a negative impact of using a SAT, with the lack of positive impact attributed to no, or unsuitable or insufficient, support being offered (see Chapter 6). Finally, and not possible for us to investigate, we do not know whether there are routine differences in the quality or intensity of support offered by SATs to these two groups.

Factors affecting outcomes of using a Specialist Autism Team

A key objective of the study was to generate evidence on what a SAT should ‘look like’ in terms of its characteristics and ways of working. Our quantitative and qualitative data both contributed to addressing this objective.

Implications

The notion of a SAT was developed by the GDG convened by NICE to develop guidance on the diagnosis and management of autism in adults. This was a novel service model and the GDG acknowledged that the research evidence available to support and inform the way that this recommendation was operationalised and implemented was very limited. It called for research that would help to define those aspects of service organisation and delivery that best support positive outcomes. This was the primary objective of this study.

This study has shown that it is possible to set up such a service. However, very few localities have been able to achieve this. Findings support the notion of SATs and have demonstrated the benefits and positive impacts they can achieve; however, they need to be sufficiently resourced. Importantly, our findings suggest that different organisational structures are possible, including different approaches to integrating health and social care into a single provision (e.g. jointly commissioned vs. separate teams with joint-working arrangements).

The National Institute for Health and Care Excellence called for evidence on what a SAT should look like in terms of staffing and the interventions that it should provide. Autism-specific expertise, interventions and adjustments to service provision were identified as critical to the achievement of positive changes in people’s lives. With respect to staffing, overall, our findings support a diversity of professions (or rich skill mix) within the service and a holistic, individualised approach to care and support. In terms of interventions, we believe that the findings make a strong case for extended post-diagnosis psychoeducation, and interventions and practice that nurture self-management with respect to mental health, managing day-to-day life and strengthening social networks. This should include the provision of low-intensity, drop-in-type advice and support after discharge. There is some limited evidence indicating that it may be cost-effective for SATs to deliver one-to-one mental health interventions in service, rather than supporting referrals to, for example, IAPT.

The great majority of those referred to SATs have mental health needs and addressing these is fundamental to achieving positive changes in other life domains. Some SATs operated on a model whereby non-complex mental health problems (or those not directly arising from autism diagnosis) were managed by referring on to generic community mental health services. However, such referrals (i.e. handover, assessment and intervention) need to be supported by the SAT, with that support being available to both the individual and the receiving service. To make this work, commissioning arrangements need to include sufficient resource for SATs to be able to do this and for mainstream services to be required (and sufficiently resourced) to accept such referrals. The same argument applies to other statutory services.

Findings from this study do not support the use of signposting (i.e. simply providing information about another statutory or third-sector service, or community support group/network). Given that autistic adults may encounter additional barriers to initiating an approach to a new organisation, any service working with autistic people should consider reviewing its use of signposting. In addition, services should not assume that peer-led/autism-community organisations and networks, or support groups/networks run by autism charities, are widely used or replace specialist autism services providing or being available for advice and support on a longer-term basis to autistic adults living in their locality.

Finally, and aligning with existing evidence, findings from this study raise significant concerns about the insufficiency or lack of post-diagnosis psychoeducation provided by services that offer only diagnostic assessment.

Design

Single, in-depth, semistructured interviews with a subsample of study participants and, with their permission, a member of their family were carried out. Interviews took place at around the 12-month follow-up time point. In the original design (based on anticipating the identification of distinct types of service model and evaluating one or two exemplar services per model), it was proposed that we would interview 40 study participants and 20 family members. These remained our target sample sizes.

We purposively sampled to represent the following characteristics:

• research site

• the different groups – D&S and SO groups within the SAT cohort, and the DO cohort

• within each group

  • perceived impact of service at the 12-month follow-up (using impact rating question in T3 questionnaire with the following response options: positive impact; little or no impact or negative impact)

  • age

  • gender.

Recruitment took place over a 4-month period and the profile of the recruited sample was reviewed on an ongoing basis against the sampling framework. Service users were offered a range of ways to participate in an interview: face to face, telephone and instant messenger. Telephone interviews were used with family members. Interviews took place between March and August 2018.