To determine the impact of SIT on irritability and agitation, as measured by the corresponding sub-scale of the Aberrant Behaviour Checklist (ABC)(13)(14).
i. effectiveness of SIT for additional behaviour problems such as hyperactivity/non-compliance, lethargy/social withdrawal, stereotypic behaviour, inappropriate speech
ii. the impact of SIT on adaptive skills, functioning, and socialisation
iii. sensory processing scores post-intervention (i.e. at six months) as a potential mediator of any association observed between SIT and the primary outcome at 12 months
iv. age, severity of SP difficulties, adaptive behaviour, socialisation and comorbid conditions as potential moderators of any association between SIT and irritability/agitation, adaptive functioning (child) and carer stress
v. the impact of the intervention on carer stress and Quality of Life (QoL)
vi. cost-effectiveness of the intervention, including direct intervention costs, health, social care, education services, carer expenses and lost productivity costs
vii. fidelity, recruitment, acceptability, adherence, adverse effects and contamination in a process evaluation conducted alongside the main trial. An internal pilot with specific progression criteria will assess the feasibility of proposed recruitment and trial retention rates and whether usual care (UC) differs from expected provision of SIT.
A two arm individually randomised (1:1 ratio) effectiveness trial of SIT for children with ASD and SP difficulties in mainstream primary education (aged 4 to 11 years old). The comparator will be Usual Care (UC). Manualised SIT will be delivered in clinics meeting full fidelity criteria (structural equipment elements). The target is to recruit 216 children. Those allocated to the intervention group will continue to receive any care currently being received provided it does not contravene the eligibility criteria.
A process evaluation following Medical Research Council (MRC) guidance(15) will examine contamination, fidelity of intervention delivery, adherence and any adverse effects. It will also include assessment of recruitment, retention, adherence and intervention reach. Therapist and carer interviews will explore: barriers/facilitators, adherence, therapeutic relationship, mechanisms of change, SP deficit, engagement in activities and contamination.
i. have a diagnosis of ASD (as documented on medical and/or educational records), OR have probable/likely ASD (defined as currently being assessed within the local ASD pathway);
ii. aged 4-11 years at the start of the trial;
iii. plan to remain in mainstream primary education until the primary outcome time-point (6-months post-randomisation, and end of intervention for SIT arm);
iv. have definite or probable SP difficulties defined as: (a) a definite dysfunction on at least one sensory dimension (all domains except social participation) and the total score on the Sensory Processing Measure (SPM)(16) or (b) at least a probable dysfunction on two or more sensory dimensions and the total score;
v. provide carer consent/child assent.
Other than the obverse of the inclusion criteria, participants will be excluded if:
i. Currently undergoing or previously undergone SIT
ii. Currently undergoing an intensive, comprehensive Applied Behaviour Analysis-based intervention.
Participants with a range of ASD and sensory symptom severity, as well as functional and cognitive ability will be recruited from Child and Adolescent mental Health Services (CAMHS)/paediatrics, occupational therapy, schools and support/social services. The study will also be advertised on relevant websites (i.e. related charities’ websites) and via social media and trial specific website. It will also be possible for carers to make a self-referral.
Potential participants will have a range of impairments and some may have a degree of intellectual disability. No child will be excluded on this basis, or due to other co-morbid conditions, provided all other inclusion criteria are met and exclusion criteria not met. Informed consent from carers and assent from children will be sought. The child’s school may be asked for feedback on the child’s behaviour and will be asked to complete the Aberrant Behaviour Checklist – Irritability scale (ABC-I) at 6 months.
Video recording: Assessment of fidelity and mentoring require sessions to be video recorded. Permission to video record the Autism Diagnostic Observation Schedule (ADOS)(17) will also be sought. Eligibility to participate in the trial is not however contingent on provision of consent to video record.
Primary outcome measures
The primary outcome, to be measured at baseline, six and 12 months post-randomisation (Fig. 1.) is irritability/agitation as measured by the corresponding Aberrant Behaviour Checklist sub-scale (Community version ABC-I: 15 items(13)(14)). The primary outcome time point is at six months post-randomisation (i.e. immediately post-intervention in the SIT arm). Teacher/teaching assistant ratings of ABC-I (assessed at six months post-randomisation only) will be measured to explore agreement between teacher and carer assessments and as way of measuring carer response bias.
Secondary outcome measures
All secondary outcomes will be measured at baseline, six and 12 months post-randomisation (Fig. 1.).
Other problem behaviours will be measured using the remaining four ABC sub-scales(13)(18)(19).
Adaptive behaviours, socialisation and functional skills change will be assessed using the Vineland Adaptive Behaviour Scales (VABS-II: parent/carer rating version(20)).
Carer stress will be assessed using the Autism Parenting Stress Index (APSI(21)).
Carer Quality of Life will be measured using two measures the EQ5D 5L(22) scale and CarerQol(23).
Scores on the SPM(16) are assessed at screening and six months post-randomisation in order to determine whether any effects of the intervention on the primary outcome at 12 months (if observed) are mediated by severity of SP difficulty post-intervention.
Health Economics measures
Detailed information on staff and non-staff inputs directly associated with the SIT intervention and UC will be recorded for each participant during the intervention period. Data on services and support external to the intervention will be collected at interview for each participant in the study at baseline (covering the previous 6 months), six, and 12 months post-randomisation. The Client Service Receipt Inventory (CSRI)(24) will be adapted for use in this study to not only collect service and support data for the child but also data on health and social care services used by the child’s main carer. This will include carer out-of-pocket expenses and time taken off work to care for their child.
The SPM(16) is included at screening to confirm definite/probable sensory processing difficulties (Fig. 2.). The measure provides norm-referenced standard scores for seven domains: visual, auditory, tactile, proprioceptive, vestibular, praxis, social participation; and a total sensory systems score. Scores then fall into one of three interpretive ranges: Typical, Some Problems or Definite Dysfunction. The ‘home’ version will be administered as a questionnaire to the child’s parent.
For the purposes of the current trial, sensory processing difficulty is defined as either: (a) a definite dysfunction on at least one sensory dimension (defined as all domains except social participation) and the total score or (b) at least a probable dysfunction on two or more sensory dimensions and the total score. Treating therapists will access these scores in order to aid with delivery of the intervention.
To characterise the recruited sample according to ASD symptoms, the ADOS will be included as a baseline measure. ADOS administrators will be trained to research standard and assessments will be video recorded. A sample of these recordings will then be used to ensure consistency in administration and coding between researchers.
Sensory Integration Therapy
Those allocated to the intervention arm will receive 26 one-hour sessions of SIT(25)(26), delivered over 26 weeks: two sessions per week for 10 weeks (intensive phase), followed by two sessions per month for two months, then one telephone session per month for two months (tailoring phase). A detailed assessment (SIT arm only) of sensory processing deficit will be undertaken (Sensory Integration and Praxis Tests: SIPT(27)) along with clinical observations post-randomisation. Following this assessment, the data will be scored to generate a SIPT report and a hypothesis developed as to the nature of the underlying sensory difficulty affecting function. In addition, background history, and the Canadian Occupational Performance Measure (COPM)(28) will be carried out. SIT uses the ‘just right’ challenge – a key principle of the sensory integrative approach – for each child and is therefore able to adjust the therapy to functional ability (as measured at baseline). Carers will be encouraged to observe or actively participate in sessions to facilitate engagement. Between sessions, carers may be given brief written guidelines of specific sensory-motor activities to support their child’s sensory integration. Success of these strategies will be discussed at the following session.
The intervention will be delivered by occupational therapists (typically NHS Band 7) trained in SIT and meeting fidelity criteria in regional clinics which also meet fidelity. Initially clinics will be located in South Wales and Cornwall with the potential for more to be included based on recruitment rates and therapist availability. For the duration of the trial, each intervention therapist will be paired with a mentor - a therapist independent to their clinic who is trained in both SI and mentoring. The mentor will support the therapist in the assessment, interpretation and intervention of the child. This is a critical part of the trial process that will help provide evidence of meeting intervention fidelity criteria. Mentoring sessions will be approximately one hour long and provided ideally fortnightly during the first two months for the first participant during the intervention delivery phase, tapering to once per month or at least once every 6 weeks thereafter. A Facebook group for treating therapists to support each other will also be established. Intervention therapists will provide therapy to participants recruited to the SIT arm only. Those participants receiving any form of usual care (such as provision of sensory strategies and/or face-to-face sessions delivered once per week or less) will be seen by occupational therapists not delivering SIT in the current trial.
After the participant has entered the trial, the therapist must remain free to give alternative treatment to that specified in the protocol, at any stage, if he/she feels it to be in the best interest of the participant, e.g. provision of specialist equipment (such as adapted seating) or advice to teachers on management of the child.
Intervention delivery will be assessed using the Ayres Sensory Integration® Intervention Fidelity Measure(29). The first two video recorded face-face sessions delivered to any participant for each therapist will be assessed purely to address any training issues at the earliest opportunity and to ensure ongoing fidelity can be rated. A sample of recorded sessions in the intensive phase will then be rated for fidelity by an independent SIT-trained therapist. Demonstration of adequate fidelity is defined as scoring 80/100 or above on the fidelity measure(29) across at least 80% of sessions sampled. An ‘effective’ dose for SIT has not yet been established however, based on clinical experience and currently available evidence(7)(9)(10)(11) attending 13 of a possible 20 sessions delivered during the intensive intervention phase (two thirds) is likely to indicate sufficient exposure.
Structural fidelity is assessed according to level of therapist training/qualifications, followed by a score of 85/110 for four areas: safety of the environment, detail and content of therapist-held records including therapist-carer collaboration in relation to goals set during therapy, physical space and equipment, and communication with carers.
We plan to identify suitable additional resources to use the video recording to look into fidelity of delivery in more detail. As part of this, we will include specific items to gauge the impact of non-specific therapist effects, using an adapted version of a tool developed for evaluation of psychosocial interventions for individuals with intellectual disability(30)(31).
Usual Care (UC) will be recorded by carers in diary format. The current standard care pathway is variable across the UK, ranging from minimal contact/no specific treatment targeted at sensory processing, to provision of manualised SIT in some regions. However, within the proposed trial sites, we estimate that usual care will be much less intensive than the 26-week programme detailed above, ranging from some provision of sensory strategies not meeting full fidelity criteria for SIT (and should not occur more frequently than once per week) to no specific treatment. Notes will be kept according to usual policy. Usual care for ASD will also be recorded more generally including any contact with NHS services (e.g. speech therapy, paediatrics and CAMHS).
Usual care for the current trial will be assessed and fully defined following a brief pre-recruitment survey of therapists, and discussions (e.g. as interviews or focus groups) with carers and occupational therapists. The potential for contamination, if participants recruited to the UC arm receive enhanced/additional support from clinicians who are aware of their participation in the trial is acknowledged, thus there will be an examination as to whether the UC received differs in any way from the provision mapped out as a result of the scoping focus groups.
Internal pilot and progression criteria for full trial
An initial internal pilot phase will assess feasibility of recruitment, retention to the intervention and the nature of UC for sensory processing difficulties in the control arm.
Progression criteria are as follows:
1. Recruitment feasibility criteria will be met if at least 70% of those approached meet eligibility criteria for trial entry and at least 50% of those eligible are willing to be randomised. The proposed internal pilot end date is study month 18. Overall recruitment rates will be formally reviewed at this time point.
2. Once approximately 10% of participants have completed the post-intervention/six-month follow-up, carer-completed diaries will be qualitatively assessed to determine whether UC is sufficiently different from the SIT intervention for the full trial to continue. Broadly defined, this criterion will be met provided those in the UC arm do not receive any intervention meeting criteria for full SIT.
3. If dropout at the six-month post-randomisation time point exceeds 20%, the sample size calculation and associated implications for feasibility of recruitment will be re-assessed.
4. To confirm the accuracy of the sample size calculation and other features of the proposed design, an estimate of the following will be obtained: (a) proportion of participants providing primary outcome data; (b) SD of the ABC-I at the primary outcome time point (post-intervention) in both SIT and UC groups; (c) intra-cluster correlation coefficient (ICC) of SIT therapists within participants for the ABC-I at the primary outcome time point (post-intervention, SIT arm only); (d) correlation between baseline and six-month post-randomisation ABC-I. Sample size may be adjusted following these explorations.
Collection of Usual Care data
A minimum of two scoping focus groups will be held prior to recruitment. Each will utilise a case analysis approach with clinicians providing treatment for sensory processing difficulties (one in each region). A small number of one-to-one telephone interviews may supplement focus group data as required. Focus groups will explore what is currently delivered/received as UC in the Health Boards/Trusts involved, and what if any difference exists in local provision and between regions (i.e. currently South Wales and England).
To develop a schedule for the focus groups, a brief survey will have been distributed to Occupational Therapist Practice leads (OTs) in trial regions working with the trial population, via OT service leads.
Interviews with carers of children with ASD and sensory processing difficulties (parents from both South Wales and Cornwall) will be conducted. These will utilise a time-line facilitated process(31). We will ask participants to focus on key points along a timeline, including ‘the beginning’, ‘diagnosis’ and ‘now’.
Family carer and therapist interviews
Following the six-month/post-intervention time-point, diary and artefact facilitated interviews will be conducted with all SIT therapists (5-10 interviews) and a sample of family carers in both arms (anticipated to be 10-15 in each arm before data saturation).
Primary carers may choose to be interviewed alone or with other members of their family who are involved in day-to-day care. Participants will be asked to reflect on their experience of the intervention and the usual care activities that occurred alongside it, or their experience of usual care alone.
Therapists will be sampled to achieve variation in Health Board/Trust and regional centre and will be given the choice of telephone or face-to-face interviews. Family carers will be sampled to ensure maximum variation in terms of range of ASD and sensory symptoms, Health Board/Trust, and regional centre. Interviews will take place at a location of the interviewee’s choice, often their home, or over the phone.
The interview topic guides will be developed from a review of previous research, guides used by the research team in similar studies, and with input from the multi-disciplinary research team and family carer advisors to avoid bias in topic selection and wording of questions. The topic guide will be piloted and refined as necessary. Interviews will be recorded and transcribed.
No adverse events are expected. However, adverse events will be collected, recorded and reported in accordance with Good Clinical Practice and the requirements of the research ethics committee. The Chief Investigator may carry out Urgent Safety Measures if appropriate to protect participants from immediate harm.
Sample size determination
We will recruit 216 participants in total (108 allocated to usual care, 108 allocated to the SIT intervention). This will provide 90% power at the 5% significance level (two-sided) to detect a standardised effect size of 0.5, allowing for 20% loss to follow-up at primary outcome time point (six-months post-randomisation).
Our effect size is based on means and standard deviations of the ABC-I in relevant populations found in the literature(18)(32)(33). This literature also suggests that a 25% relative difference represents a clinically meaningful difference on the ABC-I. Findings from the internal pilot will aid in confirming the accuracy of the assumptions behind the sample size calculation and could potentially lead to an adjustment of this.
The primary analysis will be based on the MITT analysis population, and will estimate the between-group mean difference in the ABC-I at six months using linear regression, adjusting for baseline ABC-I, recruitment site, severity of SP difficulty, and sex of the child. If appropriate, therapist clustering will be accounted for using mixed models. Secondary outcomes will be analysed similarly.
Subgroup analyses will be conducted, exploring any differential intervention effects by site, region, age, sex of the child, severity of SP difficulties, adaptive behaviour, socialisation, motor skills and comorbid conditions. This will be carried out by repeating the primary analysis but including each subgroup as an explanatory variable along with a subgroup x treatment arm interaction. Subgroup analyses will also be performed carer stress.
Impact of missing data and non-adherence
The impact that non-adherence to the intervention has on the Intention To Treat (ITT) findings will be investigated by estimating the Complier-Average Causal Effect (CACE) for the primary and secondary outcomes(34).
While the main trial analysis will be based on a MITT analysis population, sensitivity analyses will be carried out exploring the impact that missing data may have had on trial findings. Where outcome data are missing due to drop-out/loss to follow-up, these will be assumed to be missing at random given observed data (MAR), and multiple imputation will be used to achieve a full ITT analysis population. Additional sensitivity analyses will be conducted using joint modelling approaches (e.g. selection and/or pattern mixture models) to explore departures from a MAR assumption(35).
Exploratory mediation analyses will be conducted to examine whether any effect of the intervention on behavioural problems at one-year (all ABC subscales) is mediated through an effect on sensory sensitivities immediately post-intervention. The analyses will control for baseline measures of behavioural problems and SP difficulties, to minimise any residual confounding between mediator and outcome(36). Additional analyses will be conducted to explore the association between measures collected as part of the process evaluation and primary/key secondary outcomes. As the majority of process evaluation measures will only be collected for participants allocated to the SIT arm, the analysis will be purely associational and therefore hypothesis generating in nature.
Given the variability in the ‘usual care’ that we are likely to see, we will conduct analyses using participants in the UC arm only that explore the association between different types of usual care and clinical outcomes. Parameters we will use to characterise different types of usual care will include number of treatment contacts, therapist experience/level of training, and type of difficulty for which the therapy is intended. Regression models will be fitted using our primary and secondary outcomes, and the therapy characteristics/parameters as explanatory variables. Variables that confound the relationship between therapy characteristics and outcome (e.g. age, severity of SP difficulty) will be investigated and controlled for in the models. The interpretation of the findings from these analyses will reflect the exploratory nature of this work and will be purely associational (that is, without ascribing cause).
A Statistical Analysis Plan will provide further detail on analytical methods using for the analysis of trial outcomes, and will be finalised prior to the end of recruitment.
Qualitative data will be analysed by the qualitative team using thematic analysis(37). We will search across the data set to find repeated patterns of meaning, and identify key themes and sub-themes. We will identify contradictory data, as points of contrast as well as similarities to understand uptake and engagement with the intervention. Vital measures will be put into place to ensure validity and reliability. Double coding will be carried out until consensus is reached. Data will be managed using qualitative coding software (such as NVivo10). This qualitative component has been designed using the principles of the Critical Appraisal Skills Programme qualitative checklist, to ensure the quality of qualitative research(38).
Health economic analysis
A within trial health economic analysis will be conducted from a health and personal social services (NHS and PSS) perspective. The health economic analysis will be carried out on an ITT basis. The main analyses will compare cost and cost-effectiveness at six-months follow-up of SIT compared to UC. Total and mean costs for the SIT and UC group will be reported in a disaggregated format. Total and mean severity outcomes (ABC-I) will be reported for the intervention and control groups. The difference in mean scores between the two groups will be assessed with appropriate statistical tests. The difference in mean costs for the treatment groups will be analysed using regression analysis and bootstrapping. NHS and PSS costs (or societal costs in the secondary analyses) over the six-months will be regressed on treatment allocation, baseline ABC-I, site, severity of SP difficulty and baseline costs. We will account for clustering in the analysis.
Incremental cost-effectiveness ratio (ICER) will be calculated, defined as the difference in mean costs divided by difference in mean ABC-I as: a) the incremental cost per participant achieving a significant improvement in mean ABC-I score from an NHS/PSS perspective. Results will be plotted on a cost-effectiveness plane. Bootstrapping will be used to estimate a distribution around costs and behavioural outcomes and to estimate the confidence intervals around the incremental cost-effectiveness ratio Cost-effectiveness acceptability curves (CEAC), a recommended decision-making approach to dealing with uncertainty, will be generated by plotting these probabilities for a range of values of the ceiling ratio. Sensitivity analysis will be used to explore the sensitivity of the results from using a broader societal perspective (including NHS/PSS costs, education service costs, carer expenses and lost productivity) than a narrower NHS/PSS perspective preferred by the National Institute for Health and Care Excellence (NICE) reference case(39). Additional sensitivity analyses will build on results of the subgroup analyses.