Early positive approaches to support for families of young children with intellectual disability: the E-PAtS feasibility RCT

Inclusion criteria

• Aged 1.5–5 years (up to the day before the child’s 6th birthday).

• An administrative label of any severity of intellectual disability (learning disability/ learning difficulties in UK terminology), referring to identification of the child within the education, health or social care systems as having intellectual disability, or as eligible for receipt of specialist intellectual disability services, or diagnoses indicating the presence of intellectual disability for younger children (e.g. ‘global developmental delay’).

• A VABS44 composite score of < 80 (i.e. allowing for measurement error but still indicating significant developmental delay) at the time of the screening interview.

Exclusion criteria

• The child is placed in a 24-hour residential placement at baseline.

• The child is placed in a foster placement that is due to end before the 12-month post-randomisation follow-up data collection point.

• The child had current child protection concerns identified.

The family unit and participants/family carers had to meet the following criteria.

Inclusion criteria

• A biological, step, adoptive or foster (if placement was planned to extend to 12 months’ follow-up) parent or adult family carer, including older siblings, grandparents or other family members who live in the family home.

• Main family carer was available to attend the E-PAtS intervention.

• Aged ≥ 18 years.

• Sufficient level of English language ability, enabling (verbal) completion of proposed outcome measures.

Exclusion criteria

• Enrolled in a group-delivered or individually delivered parenting programme outside the study at baseline (main family carer only).

• Enrolled in a programme of personal psychological therapeutic support at baseline (given that the E-PAtS programme is focused on family carer well-being).

• If any family carer in the family had already participated in an E-PAtS intervention.

• The family was recognised to be in a state of current crisis and unable to cope, indicated by a score of 9 or 10 on the 10-point Brief Family Distress Scale45 (assessed by primary family carer report only). Families in a state of crisis presented with needs that could not be addressed in a proactive programme and required urgent case management. Alternative forms of support were recommended to families in crisis.

Belfast

Mencap Northern Ireland’s family support service has been running for 3.5 years. The focus of the work is to support families with children aged 0–7 years with intellectual disability, developmental delay and/or autism. The majority of the work with families takes place at the Mencap Children’s Centre in Belfast, where families can access a variety of supports. Prior to the current study, the E-PAtS programme had been delivered in Belfast on six occasions. A family support manager employed by Mencap Northern Ireland was available to support intervention implementation and research.

Barnet

Barnet Mencap (London, UK) provides a range of services and runs campaigns alongside people with intellectual disability, and people with autism and their families to secure good services and support in the borough. This includes provision of parenting programmes to families of children with intellectual disability. Barnet Mencap took part in piloting an early version of the E-PAtS programme, but had not otherwise delivered the programme (in its current form) prior to taking part in the study. A small degree of managerial support was available to support intervention implementation and research activities.

Derry

Family support services by Mencap Northern Ireland are also provided in Derry, largely with a community-based focus (and without a specific centre, as there is in Belfast). A family support worker operates within a wide range of local community, voluntary and statutory organisations to build links with relevant families. Prior to the current study, the E-PAtS programme had been delivered in Derry on one occasion. Support for the intervention implementation and research activities was available through the same manager employed by Mencap Northern Ireland, as in Belfast, and a manager based at Derry.

Method of family recruitment

Participants were referred to the study team either directly or indirectly by service provider organisations (Mencap Northern Ireland and Barnet Mencap, referred to hereafter as Mencap) that were instructed to use a multipoint recruitment strategy, including established referral routes, local and national charitable support organisations, local authority services, special schools and nurseries, after school/weekend services for children with special educational needs and disabilities, parent/family support groups, social media, advertising in the media in local areas and self-referral (Table 8).

The majority (92.7%) of families were referred directly by Mencap, either through referrals or parent sessions arranged by Mencap. In sites 2 and 3, the majority of families were referred directly through referrals. In site 1, however, the parent sessions were a successful referral route. Only 7.3% of families were referred indirectly through advertising, word of mouth, etc. Most participants were described in qualitative interviews as being recruited through Mencap, either directly or at a parent session. Some participants stated that they were recruited by social media or word of mouth:

So I first heard about it from Mencap because they support us and so I was told about it and I was told about the information.

We were at a Mencap support group, kind of meeting and they mentioned it and they just thought that we would be good for it.

I think it was on the Mencap Facebook [Facebook, Inc., Menlo Park, CA, USA] page they had advertised it and then a few parents then had said ‘we are going to it’. So it was through word of mouth then that I sort of registered interest and came along.

Mencap was asked to complete logs for all potentially eligible families approached. However, these were not completed because of the time pressures during the recruitment period. Therefore, reporting of the number of families directly approached by Mencap or the proportion of approached families that were subsequently referred to the study team either directly or indirectly cannot be reported, and this is a limitation of this feasibility study.

Feasibility and acceptability of recruitment strategies

In qualitative interviews, both family carers and programme delivery staff highlighted two key recruitment pathways: (1) personal contact with a practitioner (most often a member of Mencap staff) or (2) through parents finding out about the programme from Mencap social media sites (e.g. Facebook). One practitioner (facilitator, site 1) suggested that recruitment needed to be done by charities embedded within the local community, but it would be important to identify and involve various groups and not just rely on one organisation, as there is a risk that recruitment might reach only those already attending support groups.

Facilitators did not identify significant challenges with recruiting family carers and the main methods described above appeared to be feasible to implement. Facilitators reported that randomisation (and therefore possible allocation to the control group) was seen as a barrier to participation when explaining the study to family carers, but not to the extent that it appeared to undermine the feasibility or acceptability of a full RCT.

Recruitment rates and participant retention

Recruitment rates and retention at 3 and 12 months post-randomisation are reported quantitatively. Reflections from the study team, in addition to barriers to and facilitators of recruitment and retention, as detailed in qualitative interviews, are also described (Figure 1).

FIGURE 1.

The E-PAtS CONSORT flow diagram. a, A total of 150 families were contacted in the study. One of the sites recruited families in two rounds and some of the families (n = 16) referred from this site were referred for both the first and second round (if they were not recruited on the first round). Reproduced with permission from Coulman et al. 41 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

Reproduced with permission from Coulman et al. 41 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original.

Recruitment of families

The study recruitment target was defined as recruitment of 64 families/clusters by two to four service provider organisations that deliver one or two intervention group sessions each. Initially, two service provider organisations (sites 1 and 2) were recruited to recruit 32 clusters/families and deliver two intervention group sessions each. The recruitment period was scheduled for the summer term of 2018, immediately prior to the intervention start dates, which, in turn, were timed to coincide with school term dates to maximise attendance at group sessions. However, because of unforeseen variations in school holiday dates in site 1, the recruitment period at site 1 was shortened to complete recruitment and deliver the intervention prior to the start of school summer holidays. To accomplish this, staff resource was reallocated between sites and recruitment periods were condensed. Despite this, recruitment targets were exceeded in site 1, with 38 clusters recruited in 6 weeks (26 March 2018–1 May 2018). However, recruitment targets were not met in site 2, with 15 clusters recruited in 5 weeks (18 April 2018–18 May 2018). Despite 41 referrals of interested families in site 2, 22.0% of these families were not recruited because they were unable to attend sessions. This was much higher than in site 1 (6.3%). Organisation of the intervention sessions differed in each site, which may have had an impact on recruitment rates. Although site 1 scheduled group sessions around the availability of recruited participants, there was limited flexibility in the timings of the intervention sessions in site 2 because of venue restrictions.

A second recruitment period in the summer of 2018 was approved by the SSC to meet the initial study recruitment target. Recruitment was paused in site 1, as it was considered that the potential study population could have been exhausted during the initial recruitment period. However, recruitment continued in site 2 to deliver to one intervention group and to assess recruitment feasibility. Amendments to the organisation of the intervention group sessions were introduced, including increased flexibility in the times/days of the intervention sessions, to improve participant availability. A third site (site 3) was recruited to deliver to one intervention group to further demonstrate recruitment feasibility.

In site 2, 16 families that were interested in taking part but who were not recruited in phase 1 were recontacted and 11 new families/clusters were referred to the study team (one indirectly and 10 directly by Mencap). Although the recruitment target was set at approximately 16 families, only seven families were recruited to the study in 11 weeks (15 June 2018–29 August 2018). In site 3, 14 families were recruited in 5 weeks (11 July 2018–13 August 2018) and recruitment was stopped because the recruitment target was met.

In total, 74 families and 95 participants were recruited and randomised in three sites, delivering between one and three intervention group sessions across two recruitment phases.

Participants’ barriers to and facilitators of taking part in the study

Overwhelmingly, participants stated that their motivation to take part in the study was to support their child with disabilities, including to educate themselves on how to deal with specific challenges that they were currently facing with their child. However, family carer facilitators also reported altruistic reasons to help similar families:

. . . because I really want to be there for my son, in terms of physically, emotionally and psychologically so I really want to know more in order to teach him so he can be better.

As I said before its things like toilet training and what have you. That will be one of our challenges and for that alone I would probably go and I would feel like I have this good opportunity and I wouldn’t want to miss it.

Do you know what, what really motivated me as well, was the fact that I felt like it was going to really make a difference, you know, not only to our life, but also like taking part in the study might help other people.

Participants described that they had experienced limited support for families of children with learning disabilities and that they were motivated to promote any additional support. Positive reviews from other family carers also encouraged participants to take part in the study:

I have heard very good things about it. I would be keen to do it.

I couldn’t have spoken about [name] without crying, my heart just broke for him because there was nothing available. He does attend . . . all the medical side of things, the therapies are fine and between the 45-minute sessions half the time he will not engage so I was just very fearful of the unknown and what am I going to do to help this child so I was very grateful for the E-PAtS course at that time.

Sometimes just for families with children with learning disabilities there’s little support, there is a bit of support, you know, you have to fight for quite a lot. I think it is good there are people out there proactively trying to make things easier. I’m very supportive of it in that regard.

Participant retention at 3 months

Eighty-four per cent (n = 31) and 78% (n = 29) of intervention and control families, respectively, provided follow-up data at 3 months, including either completion of the Five Minute Speech Sample (FMSS) by telephone interview or questionnaire completion by telephone, face to face or posting completed questionnaires (Table 9).

Barriers specific to 3-month follow-up data collection

The FMSS was used at baseline and the 3-month follow-up for all study participants. This measure asked parents to ‘talk about their child for 5 minutes without interruption, about their thoughts and feelings about the subject and how they get along together’. Participants were generally apprehensive about completing the FMSS and thought that the experience would be difficult and uncomfortable. At the 3-month follow-up, some participants reported that they declined to complete the measure because of its demanding nature. The response rate for completion of the FMSS reduced from 100% at baseline to 62% at the 3-month follow-up. It is therefore possible that the FMSS deterred participants from completing the 3-month follow-up, affecting retention at this point. This also demonstrated that the FMSS was unacceptable to some study participants. Therefore, it was decided that the FMSS would not be repeated at the 12-month follow-up. Furthermore, an administrative error in the phase 1 3-month follow-up data collection affected up to 18 participants in site 1, in that participants may have been posted the researcher case report form. The case report form contained the scoring measures used by the researcher to score the FMSS and included terms such as ‘critical comments’ and ‘negative relationship’. It was considered that this may have caused distress and put participants off completing the FMSS. All participants were contacted by letter and by telephone to apologise. No participant stated at the time or in qualitative interviews subsequently that this affected their decision to complete the FMSS. However, the impact of this error on FMSS completion rates is unknown.

Participant retention at 12 months

Seventy-six per cent (n = 28) and 68% (n = 25) of intervention and control families, respectively, completed follow-up data collection at 12 months, including either completion of the VABS by telephone interview or questionnaire completion by telephone, face to face or posting completed questionnaires (Table 10).

Barriers to all follow-up data collection

Participants described barriers to participation in the follow-up when they were contacted to arrange the relevant follow-up, including unavailability because of holidays, busy times of the year, such as Christmas, and family illness/hospitalisation.

Furthermore, the method of follow-up data collection was complicated by the type of proposed outcome measures and the requirement of blinding researchers to participant allocation. As a result, participants selecting postal data collection were still required to complete the FMSS (at the 3-month follow-up) or VABS (at the 12-month follow-up) by telephone and those participants selecting face-to-face or telephone data collection were required to complete the Group Cohesion Scale, a measure that identified the participant allocation (at the 3-month follow-up) with a different member of the team on a separate occasion. This increased participant burden, complicated the data collection procedure and may have had an impact on retention rates. However, this was not evidenced in qualitative interviews with participants.

Acceptability of participant recruitment

Overall, participants’ experience of being recruited was positive, including their interactions with the research team and the provision of information about the study. Participants appeared, almost universally, to have a clear idea of what they had ‘signed up to’ and what they would be expected to do as part of the research. Some participants were less clear on the specific tasks and timing (e.g. when and how many questionnaires would need to be completed). This was linked to the perceived complexity and extensive amount of information provided to participants:

There was a lot of information. Whatever we were told and whatever was happening at the time we were perfectly happy with, you know. If there was to be follow-ups or the questionnaires, to tell you the truth I can’t remember now how often the questionnaire lands on the mat, we fill it in but I couldn’t say to you we are due a questionnaire in November.

Main family carer, site 1

The experience of being recruited into the study contrasted with that of interactions with other professionals/services with whom participants had dealt. Participation in the study (and, for families in the intervention group, the E-PAtS programme) followed an explicit invitation to do so. Family carers contrasted this with their experience of trying to access services provided by other agencies/services, when they had often encountered numerous barriers (e.g. waiting lists, bureaucracy, needing to wait until a child had reached a certain age). Being approached and offered a service (or at the very least participation in a research study) therefore represented a different form of interaction that participants valued.

Acceptability of randomisation and the randomisation process

Participants were randomised on a 1 : 1 ratio to the intervention or control group. All control participants were offered the opportunity to attend the intervention subsequent to the 12-month follow-up data collection time point, if they had chosen to be offered the intervention in this way during recruitment. Participants generally understood the concept of randomisation and found the concept of randomisation in this setting acceptable. However, some participants spoke of their confusion and lack of clarity with regard to the process of study allocation (e.g. when control participants would be able to attend the course) and requested clearer information. Facilitators echoed this and felt that, as a result, randomisation was off-putting for participants. Although some participants described their happiness at being allocated to the intervention group and not having to wait 12 months to attend the E-PAtS group sessions, some control participants spoke of how disappointed they were not to be able to attend the group sessions immediately. Furthermore, participants spoke of practical issues if they were required to wait 12 months to attend the intervention sessions (e.g. changes in circumstances, such as the child starting school or alterations in the parents’ working situations, or changes in the child’s age may impact the participant’s ability to attend or eligibility to attend, respectively). However, follow-up rates at 3 months were comparable in intervention and control groups, and therefore the concerns raised may not have affected participants’ continued involvement in the study:

I think when I was initially putting my name down I just wanted to get the classes, but when it was explained I was happy enough. I mean, if everybody can’t do it then everybody can’t do it, and it was randomly selected.

I didn’t want to wait another year because you just don’t know what is going to happen in that year do you . . . I felt like I had won the lottery.

Yeah and I think the fact that it was the study with the control, that they might do it and they might not, I think that put people off.

. . .when I first heard of the E-PAtS study I thought ‘happy days’, this is finally something that can help us here and we need this help because we don’t have a lot of family support, we don’t have other services showing any interest in us’ and to then find out it was being put off for a year it was hard to accept.

I suppose I was like this was a bit of a misunderstanding for me because they were telling me yes and then they said I had to wait, so it wasn’t explained very well.

Because after finishing the study you have to wait 1 year which is probably the time that [name] is in full time school. Then I have to get a job and I have no idea.

Acceptability of method of data collection

Generally, participants described their experience of completing the questionnaires as ‘fine’. Participants were satisfied with the logistics of completing the questionnaires, which included telephone, face-to-face or postal options. Refer to Table 11 for completion rates of follow-up data. Over the course the study, three family carers changed their method of questionnaire completion (two from postal to telephone, and one from telephone to postal):

Sending the questionnaire was good, I had time for filling the form in, to send it back. The phone calls was not a problem at all.

Acceptability of study pathway options

Primary participants selected one of two pathways, if randomised to the control arm. Families in pathway A were offered the E-PAtS programme subsequent to the 12-month follow-up and families in pathway B were not offered the E-PAtS programme. Only one carer selected pathway B, demonstrating that being offered the option of attending the E-PAtS programme on a waiting list basis was the most acceptable process to participants. When questioned regarding the acceptability of this study process, participants expressed confusion. Some participants did not differentiate between randomisation and choice of study path. However, participants did suggest that they chose pathway A to keep their options open and to allow them the choice to attend the E-PAtS programme on a waiting list basis if allocated to the control arm. Participants also implied that research assistants may have encouraged this choice. However, participants’ change in circumstances may have had an impact on their decision. Participants questioned the suitability of the waiting list E-PAtS course if their child would no longer be ‘eligible’ because of their age at the time of the 12-month follow-up. This concern was predicted by the study team and so all participants were offered the E-PAtS programme if they were randomised to the control arm and their child was just outside age range by the time of their E-PAtS intervention:

Well, they said it’s best to option like that so I made that option because it’s nothing otherwise.

I don’t know if there’s much point in doing this because my daughter would have been 5 practically by that stage. I know from a research point of view you have to go for a random group. From a personal perspective in terms of what I was going to get out of it I would have felt that having to wait would have taken away from the value of it for me.

Acceptability of study questionnaires

Participants expressed that completion of the questionnaires was time-consuming and this was off-putting. One participant said that, because of the demands of parenting a child with intellectual disability, more time may be needed to complete the questionnaires:

Long! Long! It was fine! The interviewer was very good and everything but it was long.

. . . giving people the time like 1 month because I mean normal people can do something in 1 week, whereas we can’t, we are kind of I would say busier than other mums, so we want to do something but we can’t, sometimes I even forget to eat lunch.

The impact of study questionnaires

Participants found questions examining parents’ well-being difficult to answer. However, being asked these questions also resulted in parents feeling acknowledged. Parents described directing all of their energy onto their child to the point that they neglect their own needs. Therefore, the questionnaires required the participants to consider the impact of parenting a child with intellectual disability on their own and their families’ lives. Participants stated that completing the questionnaires made them realise that caring for a child with intellectual disability has an impact on the whole family and not just the parents:

. . . reading that questionnaire actually made me realise how this affects everybody around us, us and [elder child] now I see that [study child] has a huge impact on [elder child], it’s a huge impact, and it’s not a good one as well so it was interesting.

. . . the first thing what you think of is about him, you worry about the child and everything, you never think how it is actually affecting the whole family and everything you do and how your life is changed, so it was very interesting.

To be asked those questions, how do you cope, and at the time when I filled in those questions I was in a really dark place, it was really difficult to cope with, but actually for somebody to reach out and say ‘how are you feeling’, was kind of a good acknowledgement, you know what I mean?

I suppose it’s always easier to answer questions about your daughter, than maybe, about yourself, and your own feelings, that side kind of, it wasn’t the question, it’s just, sometimes it’s harder to think about your own feelings about something. It’s quite a sensitive subject.

Participants described how answering measures about their child instigated reflection, both on how difficult their situation may be and on how delayed their child may be. For some parents who had not received a clinical diagnosis, the questionnaires provided a means to compare the development of their child to children without diagnoses of intellectual disability. Participants spoke about how ‘emotional’, ‘upsetting’ and ‘depressing’ completing the questionnaires was and how the self-examination required to answer the questions was difficult:

. . . when you sit down and fill a questionnaire in for a few days afterwards you’re a bit like ‘God our life is terrible’, things that we didn’t realise we had because you’re kind of putting it down on paper so you go over them again.

I almost felt . . . it’s hard to describe but it was just a wee bit depressing, to fill it in about your child what they could and couldn’t do and realising that how far behind she was. It was sort of a bit more of a wake-up call of the things she couldn’t do that she should have been able to do I suppose.

I remember feeling, yeah, I remember feeling really down afterwards, because suddenly I realised how little he spoke. There was a list of probably about a hundred words there and I think I ticked about seven and they weren’t clear words. That suddenly made me think ‘my god this is actually real’.

Participants described completion of the questionnaires individually, as a couple and with other family members as a positive thing:

It was emotional! You know it’s one thing living it but when you see things written down like some of the questions, it really made you think and stuff, but you know it was a bit of a roller coaster. When we did our first questionnaire, my husband and I, we set aside a time when we could both sit down at the table and he does his, I do mine and then we both kind of share experiences and stuff and we both kind of had a little cry. The thing is you carry on everything, you do the best you can and you’re so focused on stopping meltdowns and keeping them safe and all that you don’t realise really how hard your day is sometimes, so there was a lot of reflection on that.

I think when you get a huge questionnaire like that it does bring it home to you, it’s quite profound when you look at it in black and white and you see exactly the extent of the things you’re dealing with which is different from your other children, but I think that is a good idea, a really good idea.

Answering questions regarding their child’s development provided participants with an opportunity to consider their child’s development and realise areas of concern, resulting in participants feeling more confident and knowledgeable about their child’s needs. Participants also took comfort at follow-up data collection when their child had advanced/made progress in certain areas:

Well, they actually gave me more knowledge onto what I was actually doing because before then I was kind of clueless . . . Yeah, it gave me a kind of confidence to know that I’ve made progress in some certain areas.

Impact of the Vineland Adaptive Behaviour Scales report

Feedback on the VABS report was varied. Although some participants described regularly receiving similar documentation, others expressed that the language used in the report was too clinical and should be simplified for parents. One participant expressed disappointment, as they considered the report to be ‘compensation’ for taking part in the study:

I got it and it was like double Dutch to me; I didn’t understand what the report was trying to say. I looked at it, it was too difficult. It was very scientific . . . I really struggled to get anything worthwhile out of it and as I just said I was a bit disappointed.

That report is not good and that’s supposed to be . . . well for me in my head, rightly or wrongly, that was supposed to be, not compensation, but it was supposed to be . . . we didn’t get on the real thing so here is something in return for giving up some of your time and then that didn’t. I didn’t find it useful. I couldn’t really understand it.

There were conflicting views on the content of the report. Some families felt that the report was extremely accurate, whereas other questioned the accuracy. Some participants thought that the report was generic and that they would have preferred a more personalised report, whereas others stated that the report was personalised and specific to their child’s needs:

Incredible! I couldn’t believe how it felt like they really knew him.

No. I think it goes for any child, their mum knows them best and there’s a lot of jargon, a lot of words and it didn’t feel personal to him. It went off data that I provided, it wasn’t like a handwritten report from someone who knew him so no I didn’t make any use of that.

Participants described the practical advice provided in the report and how these recommendations had been applied at home. The report, for example, was utilised by some parents when communicating with education and child-care services to improve understanding of the child’s needs in these settings. The report was also circulated to health-care professionals and was used as evidence in Disability Living Allowance (UK social security benefit scheme) applications:

I thought it was useful to receive it and it gave us a bit of focus on things we could work on with her.

Even when my son wanted to go to school I showed them this report as well and it helped the school as well.

It also is useful for any other forms, like DLA [Disability Living Allowance] . . . the information was good because it really highlighted the additional needs that he has that sometimes you as parents forget about because you just get on with it.

As for the study questionnaires, the VABS report similarly highlighted the challenges parents experienced with their child, and parents described difficulty when faced with this information:

It kind of hit home as well. It was quite . . . I was a wee bit numb after. It was . . . what’s the word, it kind of highlighted it from there, the evenings that we talked about it. It was real. It kind of reaffirms everything that you know, they’d got our kids all down to a tee, but it was kind of a wee bit hard to read sometimes because you don’t really think about that, you’re just getting on with every day but no, it was definitely good.

Data completeness of proposed outcome measures

The parent-reported proposed outcome measures were well completed (Table 12). The progression criteria target was to reconsider a measure if < 70% of collected data were useable. When at least one item of a measure had been completed, all items were useable for the majority of measures. The Sibling Relationship Questionnaire (SRQ) was the most poorly completed, with 84% of the questionnaires being useable at 12 months post randomisation. The items most frequently left unanswered across time points were ‘how much do the sibling and the child tell each other everything?’, ‘how much do the sibling and the child share secrets and private feelings?’ and ‘how much do the sibling and child insult and call each other names?’ These items require a certain level of communication ability, which may not have been acquired because of the young the age of the child or their intellectual disability.

Analysis of proposed outcome measures

Table 13 summarises information on the return of questionnaires and suitable summary statistics depending on the distribution of the data for each proposed outcome (i.e. family carer-reported outcome measures) measured at baseline, 3 months post randomisation and 12 months post randomisation by trial arm. The score range of the WEMWBS, the proposed primary outcome measure, is 14–70, with higher scores indicating higher levels of mental well-being. The mean UK population norm is 51. Ninety-five per cent of participants completed the measure at baseline, all to a useable extent (n = 90). Forty-nine per cent of control arm participants (n = 23) and 56% of intervention arm participants (n = 27) completed the measure at the 12-month time point. Again, the forms were well completed and all of those returned were useable. In the control group, the mean WEMWBS score was 43.2 (SD 8.9) at baseline and 43.4 (SD 11.0) 12 months post randomisation. In the intervention group, scores were 43.9 (SD 10.6) and 46.5 (SD 10.9) at baseline and 12 months post randomisation, respectively.

Higher scores on the anxiety and depression subscales of Hospital Anxiety and Depression Scale (HADS) indicate higher levels of anxiety and depression. The score range is 0–21 for both subscales. The mean baseline score of the anxiety component was 10.6 (SD 3.7) in the control group and 10.1 (SD 4.4) in the intervention group. This compares with the UK average of 6.14. At 12 months post randomisation, mean scores were 9.9 (SD 4.4) and 8.0 (SD 4.4) in the control and intervention groups, respectively. Similarly, depression scores in the participants were higher than the UK population norm (3.68) and some reduction in score was observed in the intervention group, which was not reflected in the control group. All completed questionnaires were of a useable standard. See Appendix 4 for further information on proposed secondary outcomes.

Multilevel regression analysis comparing intervention with control group, adjusting for baseline score and site

The results from the exploratory multilevel regression are given in Table 14 for each of the parent-reported secondary outcome measures. The mean WEMWBS score in the intervention group was 3.96 points higher than in the control group (95% CI –1.39 to 9.32) at 12 months post randomisation. As previously stated, this is a feasibility study and so it was not powered to detect an effect size of a given value. However, the authors of the measure indicate that, when comparing groups, half a standard difference is said to be meaningful (approximately 5 points in this instance) and in the case of individuals a ± 3-point change in score is said to be meaningful.

By further example, the anxiety score, as measured by the HADS, was 1.62 lower in the intervention group (95% CI 3.39 lower to 0.15 higher) and depression score was 1.30 lower (95% CI 2.89 lower to 0.28 higher). With the exception of the ‘count of formal sources’ section of the Family Support Scale, effect sizes in all secondary outcome measures are suggestive of positively favouring the intervention group. Even in this instance, however, despite the number of formal sources of help available being lower in the intervention arm than in the control arm at 12 months, the helpfulness rating [rated between 0 (not at all helpful) and 4 (extremely helpful)] of these sources when they existed was 0.40 higher in the intervention group (95% CI 0.22 lower to 1.02 higher).

Residual plots of the models were checked and the normality assumption was reasonably met; however, in a larger study where the estimates of between-group differences is of primary importance, variable transformation may be required.

Acceptability of proposed outcome measures

Participants described that questions sometimes required adoption of a best-fit approach; for example, some parents thought that questions were not applicable because of the age/developmental stage of the child:

. . . like it always is with questionnaires, the answers didn’t always fit. You felt that you were choosing the answer but really your answer needed a bit more nuance.

Participants described difficulty in recalling data required to answer the resource use/health economics questions:

. . . we did need to sit down and check both our diaries ‘did you take him to that doctor, that was a year ago’, or ‘there’s a speech and language that you took him to’, ‘oh and there’s one that you took him to an eye appointment’ so we would be relatively organised because we have to be because with so many appointments we don’t want to miss anything but I can imagine, I don’t know your demographics, but I would assume that if you were maybe a single mum or a different type of person than us it would have been quite hard to find all the information because it went quite far back.

Participants identified an error in the scale of the parenting sense of competence measure:

So the tend to agree or tend to disagree actually sounded more negative or more positive than the other one. So the scales weren’t great and some of the questions were a bit ‘what do you mean by that?’

Participants said that the language section of the Child Behaviour Checklist was difficult to complete because of the volume of words and difficulties in recall:

But in terms of the questions, I think none of the questions were particularly hard apart from maybe the number of words he had and what words he had.

One participant questioned the relevance of the finance questions in the study:

I didn’t know why it was relevant to ask about the financial things ‘could you find this X amount of money in this time?’ The purpose of the study it’s about the child and it’s about his needs and the family adapt and help manage the child so I don’t know why financial came into it, that would be one thing. Disability doesn’t really discriminate against class. Everything else was grand. Similar questions have been asked of me, by therapists, but I did find it a bit intrusive. Why do I need to go over . . . ? I don’t know what relevance it has. Maybe it’s about social background and how well the child comes on and maybe if people . . . I don’t know.

The FMSS was used at baseline with all study participants. Participants were generally apprehensive prior to completion of the FMSS. For example, some participants thought that the experience would be difficult and uncomfortable. However, there were contrasting views regarding the difficulty of completing the measure. Although some participants described the experience as ‘OK’ or ‘fine’ and less demanding than they expected, other participants described the experience as difficult and ‘emotional’. Some participants did not like the measure being unstructured and felt unnerved, as though they were being tested. At the 3-month follow-up, some participants described how they declined to complete the measure because it was so demanding:

Yeah. I think the second time I declined to do that bit just because I didn’t feel I had 2 minutes worth of talk in me.

I was thinking it was gonna be difficult and the first time when I done the recording I actually prepared little notes for myself but once I started talking and she told me ‘that’s it, finished’ and I was like ‘Oh, OK . . . it was fine. I was actually surprised at how a lot I had to say about [study child].

It is also possible that completing the measure over the telephone was not appropriate. One participant described the lack of body language/gestures from the researcher while completing the measure over the telephone intimidating:

But it is a wee bit unnerving because you are not doing it face to face so you don’t get any reaction you just have silence at the other end of the phone. So it can be a bit unnerving. You feel like you are being tested. Five minutes seems to be quite a long time when you’re talking about the child . . .

Recruitment of service provider organisations

Service providers for this study had been identified through the research team’s contacts and Mencap had provided a supporting letter for the funding application. Therefore, no process of service provider recruitment was required during the feasibility study.

Facilitators selected to deliver intervention

Although E-PAtS had been previously delivered in Northern Ireland, a new cohort of facilitators was recruited and trained to deliver the intervention at all three sites. In accordance with E-PAtS protocols (see Chapter 2), facilitators included both professionals and family carers. Facilitators were recruited by Mencap Northern Ireland and the local Mencap Barnet group, with support from the E-PAtS programme developers. Facilitator job descriptions were available that detailed desirable skills, attributes and experiences for facilitators. Professional facilitators were all pre-existing members of staff within the provider organisations. Family carer facilitators were parents of children or adults with a disability who were subsequently paid by provider organisations to attend training and deliver the intervention. As is standard for the E-PAtS programme, facilitator training was conducted in group format (with a combination of family carer facilitators and professional facilitators) over a 5-day period and was led by an E-PAtS programme trainer. One training group was co-ordinated in site 2 and another covered both sites 1 and 3. Training comprised theory and practical exercises and all facilitator pairs were supervised by an E-PAtS programme trainer during their first delivery of the programme (which included all interventions in the trial at sites 1 and 3 and the first intervention group at site 2).

Acceptability of facilitator training

Facilitators reported having enjoyed the training. They commented positively on the quality and depth of the information and the fact that information provision took place alongside seeking the views/experiences of trainee facilitators. These aspects of the training had helped those participating to ‘make sense’ of the information provided and the rationale for the different activities that were delivered to family carers during the programme:

We learnt quite quick but it was quite in depth so we went into quite a lot of psychology and the basis behind what makes the training work, not what we were going to teach the parents particularly but to give that understanding of the background that formed it and the science behind it I suppose and that was very useful to draw on during our training . . .

Guidance on practising strategies to manage emotions (of the self and others) in the groups was seen as useful. Training covered (and emphasised) the creation of a suitable environment for running the groups (e.g. not judging people), which was valued. These positive assessments of the training course led participants to feel that they had received the training and information they needed to be able to deliver programme sessions to families.

Intervention adherence

The E-PAtS intervention was attended by:

• main family carers recruited to the study

• second family carers recruited to the study

• second family carers or family members not recruited to the study (i.e. non-study attendees).

Main family carers were eligible if they could attend the E-PAtS intervention. Eligibility for second family carers was not dependent on availability for intervention sessions. However, second family carers were invited to E-PAtS group sessions. In addition, some family members who did not take part in the study (for unknown reasons) later expressed an interest in attending the E-PAtS group sessions. These family members attended as non-study participants and, for these family members, only attendance data were collected.

Adherence to the E-PAtS programme was defined as at least one family carer from a family attending five of the eight E-PAtS sessions (see Chapter 2). According to this definition, 26 out of 37 families adhered to the intervention (70.3%). In terms of session attendance, four families attended zero sessions (10.8%) and seven families attended all eight sessions (18.9%). The modal number of sessions attended by a family was seven (11/37 families, 29.7%) (Table 15).

Warwick–Edinburgh Mental Wellbeing Scale scores at baseline and 12 months by adherence and attendance

Figures 2–7 illustrate WEMWBS scores by adherence and attendance. Descriptively, there is some evidence to suggest that average WEMWBS scores are higher with adherence and increasing attendance. However, higher average baseline scores for those with adherence or attendance was also observed.

FIGURE 2.

Warwick–Edinburgh Mental Wellbeing Scale scores at baseline and at 12 months post randomisation by adherence. Plot created for those participants with both baseline and 12-month WEMWBS data available.

FIGURE 3.

Warwick–Edinburgh Mental Wellbeing Scale scores at baseline and at 12 months post randomisation by family attendance. Plot created for those participants with both baseline and 12-month WEMWBS data available.

FIGURE 4.

Warwick–Edinburgh Mental Wellbeing Scale scores at baseline and at 12 months post-randomisation by main family carer attendance. Plot created for those participants with both baseline and 12-month WEMWBS data available.

FIGURE 5.

Relationship between adherence group and availability of WEMWBS at 12 months.

FIGURE 6.

Relationship between family attendance and availability of WEMWBS at 12 months.

FIGURE 7.

Relationship between main family carer attendance and availability of WEMWBS at 12 months.

Relationship between adherence, attendance and availability of proposed outcome data

Figures 5–7 illustrate that adherence and attendance were associated with the availability of the WEMWBS at 12 months. This pattern was clearest for adherence, with availability of WEMWBS scores at 12 months being higher for participants with higher levels of adherence.

Exploratory complier-average causal effects analyses

Table 16 provides the original analysis of the proposed primary outcome for a definitive trial (i.e. WEMWBS at 12 months): an equivalent analysis using a single-level model with SEs inflated for clustering of participants within families and findings from the CACE analyses. The models imply that the effect of receiving the intervention (rather than just being randomised to it) either as adherence or as increasing attendance is associated with further gains in well-being at 12 months post randomisation.

Barriers to and facilitators of E-PAtS attendance

A number of barriers to participants attending sessions were discussed in qualitative interviews. Responsibilities at home, including child care, or if another family member had an event or illness/medical appointment, resulted in participants not attending sessions. In addition, work responsibilities, including working shift patterns or difficulties arranging time off with employers, were a barrier. These were also provided as reasons why partners did not attend the sessions, as one family carer may have been required to work or stay at home. Finally, the length of the journey to the venue and/or unreliable public transport affected attendance rates:

It was just getting somebody to look after [name] I suppose because he does have complex needs and I can’t just leave him with anybody. That was probably the biggest challenge.

. . . our older children do various after school activities and things like that, football Thursday nights or the night it was on . . .

The only thing for me would be the availability of it because of my shift pattern, I mean my shift pattern is a terrible shift pattern, but it just meant I didn’t get the full benefit out of it.

Fidelity: evaluation of session recordings

A total of seven observation recordings (two of which were video-recordings and the remainder were audio-recordings) were available for analysis. These covered sessions 4 (n = 1), 5 (n = 2), 6 (n = 2) and 8 (n = 2) and all intervention deliver sites (site 1, n = 4; site 2, n = 1; and site 3, n = 2). One of the recordings covered an entire session, with the majority capturing part of a session.

On average, 95.7% (range 88–100%) of checklist items were observed to occur in recordings, with 100% of items observed during four of these recordings (Table 17).

Five (27.8%) out of a total of 18 group sessions were not recorded because of technical issues or errors. Two (33.3%) out of six groups did not provide consent to record the group sessions. Evidence from qualitative interviews suggested that participants may not have felt comfortable with the recordings because of the sensitive nature of the group discussion:

I think maybe definitely not the videos, where you’re filming people but recording, that’s OK, but I think you need to take the video, because nobody wants their face in that, especially if they’re talking about something so important and close to them, they want people . . . their reaction and how they’re gonna speak, they want people to feel free to do it. If they’re being filmed that’s a big no, no to me. If it was me sitting there I wouldn’t want to be on camera talking about somebody like that.

Fidelity to curriculum components

Facilitators were asked to complete a session summary form at the end of each session to indicate what session components were completed. Per cent of components delivered ranged from 85% to 100% (97.1% overall) (Table 18).

Factors affecting implementation fidelity

Data from qualitative interviews with facilitators indicated that facilitators believed that they had succeeded in delivering most session activities with good fidelity. Facilitators appeared to be committed to delivering the programme as intended and felt confident in doing so. This was linked partly to receipt of the facilitator training, which had helped them to understand the purpose and ordering of activities:

I actually fully understood the aims and how to get those key messages across. I understood, along with my parent [facilitator], the function behind each session and how they all sort of linked in and carried through and we were able (from the feedback we got at the time) we were able to actually get those messages across.

Practitioner

The provision of a facilitator manual was also seen as helpful in enabling smooth progress through key activities. A number of facilitators described the importance of the time they spend preparing for each session in enabling them to deliver programme activities effectively:

I personally took a lot of time to prepare because that just is how I am, I’m not good at . . . I like to do a lot of preparation before, I’m not comfortable with ad hoc and just doing things. I much prefer to really prepare on things so it was fine. I think if I’d have had to do it without much preparation that would have worried me, but I had a chance every week to have a good think. I used to spend a day before planning what I was going to say and how that was going to go, that was always needed for me. I’ve never thought ‘we don’t need to sit down’. It always helps me.

Facilitator

The importance of creating effective partnerships between co-facilitators was also highlighted:

. . . with the first person that I facilitated with we just we worked really well together, I think. It was quite a natural partnership I think between me and [name of co-facilitator] and we were agreeable to keep to time and we did manage to get most the information that we wanted to from each session, I don’t think there was any week that we didn’t get through it all and we pretty much kept to time.

Facilitator

We had a great relationship with loads of preparation time together and everything went smoothly. We felt the same about how each session went and difficulties. We did have a couple of difficulties with one carer and we were able to come to an agreement on how to handle it and it was fine, it was all good.

Facilitator

Facilitators described being aware that some flexibility in the focus of group discussions, such as on an aspect of the material that had particular relevance for them, was permitted, as long as all intended intervention activities were covered:

I think [trainer] was happy enough if you were asked a direct question or you knew something yourself that could help a family or if you could explain something in a different way, I think he was more than happy for you to do that. As long as you cover the content and don’t leave great chunks of it out, as long as you get through all the content, the way you do it I think he led us to believe that was fine. There has to be room, you’re not just standing there like a robot reading it, you’re interacting with people and respond to how or what questions they’re asking you, I don’t think that would ever be an issue.

Facilitator

A key skill that facilitators described employing was deciding to what extent group discussions around a particular point (including those not directly related to the focus of the intervention session) should be allowed to continue before bringing participants back to the planned material:

Err well I mean we managed to cover everything that was in the course. As I said sometimes these digressions to other more pressing issues that were on people’s minds. So, someone would come along and say, ‘I visited such a such school yesterday and I’m getting this advice, and I’m getting this advice from people’ and then we’d go on to that topic for a bit, but I don’t think you can change the course of that it’s just you need to have a little bit of flexibility to allow people to offload, I guess, what’s really on their mind.

Facilitator

When children of participants were present in the session, this could make covering planned activities more difficult. Likewise, practical issues, such as problems with Wi-Fi connections (for access to programme videos), were also mentioned by facilitators.

Intervention receipt

Parents were overwhelmingly positive about the group atmosphere. Participants enjoyed the informal and friendly environment:

I did expect when I went to be in a big huge board room and there would be lots of people there and it would be very formal and everything . . . but it was very intimate, everyone was very friendly, everyone was on first name terms – it was just appreciated.

Peer support was an important aspect of the group sessions. Participants described how the interactive nature of the group allowed participants to share valuable experiences with one another in a confidential manner. Participants attending the group with other parents who had experienced similar situations to themselves was considered an important aspect of the course.

Participants described how they appreciated that they did not feel judged and were supported by the facilitators and other members of the group. Facilitators spoke about how the group members were very supportive of one another, despite the sensitive content of the course:

I think everybody was prepared to open up and speak frankly and there was an understanding that what was said in the room stayed in the room and I think people did feel able to open up and kind of wanted to support each other.

We were openly talking about our problems without thinking that we were being judged or anything.

. . . there was a lot of emotion in every group that I had, there was a lot of different emotion and upset, and a lot of support.

Facilitator

However, in contrast, it was suggested that a lack of cohesion in one group may have been due to variations among group members in the current challenges being faced by parents and the child’s clinical diagnosis:

I thought it would have been better to be honest. I think it was because the children were so different. Again, in an ideal world if you were able to separate the group into children with autism and behavioural challenges, and children with complex needs and medical issues I think there would have been more cohesion between group members, but we were all so different.

The group sessions were described as informative, but interactive and not didactic. Participants described the programme content positively and found the handouts a useful reference tool at the time and subsequent to the course end. Participants were satisfied with facilitators’ knowledge and willingness to provide extra resource when requested or helpful. Practical activities seemed popular and beneficial among group members, particularly an activity in session 4 that was used to demonstrate the frustration children may experience when trying to communicate. Furthermore, participants appeared to appreciate the role-play exercises, despite initial apprehension:

. . . it’s fantastic that you’re able to have almost like a manual to go back to, to sort of just refer to.

. . . it was easy for us to ask any questions and they answered them. Sometimes they tried to find even much more information than was in the handouts.

. . . you had to try and communicate what is on this page for the other person to draw and it was virtually impossible and that really had us in tears, I was like ‘oh I feel so sorry for him!’ That was massive. That really . . . I found very, very hard but also a brilliant way of understanding where he was at.

The course was evidence based and this allowed parents to understand the reasons behind certain behaviours that their child showed:

I liked the grounding of the topic in research and evidence.

However, some parents described how the information provided was quite basic or that they were being provided with information of which they were already aware. There was conflicting feedback regarding the relevance of some of the programme content. Although some strategies (e.g. sleeping/eating) were considered irrelevant for children with certain medical conditions, other parents described how all of the topics were relevant:

. . . there was quite a lot of information I already knew from various things I’ve read a lot myself, been to various things.

. . . he’s physically sick and has to be up on suction so he wants to sleep but because of his medical interventions he can’t. So there’s no quick fix for that.

Participants were overwhelmingly positive about the E-PAtS programme facilitators. Participants praised facilitators’ professionalism and described how well the groups were facilitated, particularly when faced with difficult situations in the group or dominant personalities. Participants appreciated the preparation involved in organising the groups, such as the professional presentation of information and additional touches, such as refreshments. Additional information, often local information and not part of the course content, was also provided by facilitators if requested:

They were very friendly, very open, you know. They don’t judge anybody, so it was very OK. They did it very well, they handled it very, very well.

So yes, there’s a big dynamic but the course leaders handled it very well.

It was clear that the people who were delivering the course had spent ages getting the presentation set up and supplied biscuits and good coffee.

Facilitators were described as empathetic and genuine. Participants valued family carer facilitators sharing past experiences and ideas, and were comforted by the fact that facilitators had ‘walked the road themselves’:

. . . the way they talked about their own experiences, you were sitting there thinking ‘oh my goodness, they’ve been through this’ and it was lovely to hear, it made you warm inside (I know that sounds corny!) but it made you think ‘oh I can do this, they’ve been through this, they’re out the other side (I suppose) of this’ and it was really reassuring.

. . . very sympathetic to everyone and the stage that everyone was going through so that was nice rather than someone who had just been there and who basically didn’t have a clue what you were going through.

Participants also appreciated the additional effort facilitators made to source local and individual information:

What was interesting was that a lot of the stuff that we could get support with was based in Mencap and that London area, so I couldn’t access any of that support because I don’t live there, I live in Kent. So what was lovely the following week the lady came back with a list of all the stuff in my local area. She spent time looking for it in kind of Folkestone and Dover which is where I live, which is really lovely, you know.

Facilitators reported that most participants were receptive to the content of the group sessions; however, a minority of participants did not engage fully. Strategies taught in the group sessions were attempted at home and discussed subsequently in group sessions. Most participants appeared comfortable to share experiences in the group setting and facilitators described speaking to participants individually if they felt it was necessary:

. . . we had some who were really appreciative of the information we were feeding to them and then we had one or two that I felt like perhaps it was not necessary for them or it might have been various reasons that they felt we were not giving them anything new or they’d heard it all before.

I’ve always worried before I delivered that somebody would say ‘oh that’s not going to work for me’ or ‘that’s stupid that won’t work’ that kind of negative opinion but it never came up, people were always very receptive to things.

. . . that was important as well, to try and help them say what they needed to say but if they wanted to do it later on a 1 : 1 basis that was fine.

Perceived aims of the E-PAtS programme

Facilitators were questioned about what they considered the intervention aims to be. Empowering family carers in a person-centred course was considered a key aim to enable parents to ‘become experts’ on their child and not just to learn through didactic teaching. Peer support was considered an important mechanism by which the E-PAtS programme works. For example, facilitators spoke of the value of understanding other parents’ struggles and learning strategies from other parents. Additionally, facilitators recognised that family carers may have previously felt isolated prior to the course and that it was valuable to meet other family carers in similar situations. A key aim of the E-PAtS programme was to provide information and also signposting to support services currently available. Family carers were taught to understand their child’s more difficult behaviours and to have realistic expectations following the course. Finally, the importance of family carers looking after themselves was instilled throughout the course:

It wasn’t experts lecturing parents, because I think parent/carers have had enough of that. It was sharing their own experience and their own struggles and valuing the parents input which I think is quite strong in the ethos of E-PAtS.

. . . at the end of every module it was ‘you can only do so much, so look after yourself first’.

. . . the big thing is breaking the isolation for people who may have been the first experience they have of having conversations with other parents.

I suppose for me it was providing the information to people around the particular areas that were covered in the course. So, you know, the sleep, challenging behaviour, the communication, the life skills, looking after yourself.

They’re not unrealistic. There’s no magic wand at all . . . We wouldn’t say to someone ‘we’ve got a magic wand; we’re gonna cure these sleep problems’.

Key mechanisms through which the E-PAtS programme works

In qualitative interviews, family carers and facilitators described four main processes through which learning and potential impacts from the E-PAtS programme occurred: (1) peer support, (2) direct learning from intervention activities/materials, (3) the sharing of learning with other family members and (4) signposting to sources of further support. First, as noted above, peer support appeared to comprise multiple dimensions, including reducing isolation, creating a safe space in which to share experiences, the normalisation of these experiences and the opportunity to learn from others in a similar situation. Group dynamics and interaction therefore helped to generate learning and support, which, to some extent, transcended or cut across specific activities and developed during the course. When participants missed one weekly session, facilitators would often provide a recap or catch-up the following week, and some group discussions carried over from one week to the next:

. . . at the end we had one lady who came to probably the last two sessions so we had to try and catch up with her, put her up to speed on what we’d been discussing up to that point. So it was like, even though the other parents were happy for us to do that because then they could sit there and reflect on what they were talking about before and we didn’t make it long-winded, we just kept it short and brief, just so that one parent . . . and we said ‘do you mind?’ and they were ‘oh no, no, we don’t mind because we can’t remember some of it so we don’t mind hearing it again’.

Facilitator

The second process (i.e. direct learning from E-PAtS programme activities) comprised the learning of new insights and strategies that parents/carers were introduced to during the programme by the facilitators, including the importance of self-care, understanding their child’s behaviour and, more generally, reflecting on the way in which they interacted with their child:

I didn’t even know that brushing teeth had different steps. I thought it was just brushing teeth, I didn’t know it was like opening the toothpaste and then putting a little bit on it and putting it in the mouth, it has different levels. I didn’t know these things so they told us and we found out how difficult it might be for my child.

1B01A

Third, and linked to individual-level learning, many shared this learning from the programme with family members, including encouraging them to adopt strategies acquired during the E-PAtS sessions:

That was a nice feeling as a facilitator that people came back and people were saying ‘this is amazing you know’, people were emotional and saying we went home and shared this with grannys and grandads and they used it; for example, the challenging behaviour section I remember one mummy said ‘my parents just think he is being naughty’, her child was autistic, ‘but when I showed them this information and explained it to them they kind of got it for the first time’. That’s incredible that that was going to change her life and potentially her child’s life because it was just looking at things in a different way.

Practitioner

There appeared to be some variation, however, in the response of family members to the sharing of learning from the programme and the ease with which it could be integrated within everyday home environments:

I did discuss with him but by the time I got home everything I said to him was ‘we had 3 hours of sessions’ and everything I said to him was less than 15 minutes because when I was telling him about it ‘OK, this is what we should not do, this is what we should do’ observation and all that, ‘OK, OK, OK but you understand everything I’ve said to you, right?’ and he said ‘yes’, ‘OK carry on’, it’s just the way they are, ‘alright . . .’. He didn’t really show any interest to learn from me, so he didn’t encourage me to say more sadly.

I shared it with the dad because I’m telling him ‘he has [?]’ so I’ve been able to share some stuff with him which he understands very well now.

Fourth, signposting of services was an important way in which families were linked to additional support and resources:

The first session as well was lots of signposting to organisations so having that information and knowing where they can go. A lot of them were already plugged in to [site 2] Mencap services so I think they would have already known who they could go to to ask for support, but hopefully E-PAtS has just kind of reiterated that. But for a parent who perhaps hadn’t done anything with Mencap before then it would have provided them with a lot of information that signposting.

Facilitator

I think what I think generally what I think is good about E-PAtS has really done lots for parents like me because it has allowed me to see that I am not alone, and having other parents share their experiences and it has really helped me calm down, knowing that I’m not alone. If I need to, if I’m worried or feel down at any point in time, I can easily approach the pre-school team and that’s really helped.

As can be seen from the above quotation, the sense of being connected to others encompassed professional support services and also a broader feeling of being connected to others.

Perceived impacts of the E-PAtS programme

Participants described improved mental health and positivity as a result of attending the E-PAtS group sessions. Participants described learning to ‘take care of themselves’ by allocating time for their own well-being. Participants described increased confidence both when managing their child’s challenging behaviour and when advocating for their child to medical professionals. Understanding why their child may demonstrate challenging behaviour resulted in participants handling their child’s ‘meltdowns’ better through improved patience and an overall ability to cope:

I was suffering from depression but now I’ve stopped taking my medication.

One of the daddy’s said ‘I haven’t been out on my motorbike in years and I just took myself out on my motorbike for a while’.

Facilitator

I’m not intimidated by doctors; if I don’t feel my son is getting the care he needs I will say something.

I don’t worry so much about people judging me if he’s having a meltdown in the middle of TK Maxx [Watford, UK] because I won’t get him a dinosaur he’s holding.

I’ve been able to know certain behaviours that come with autism, so I’m not surprised any more when he acts in certain ways.

There are times I go on the internet just to see what can I do and everything in there is stories, stories, stories. I don’t know what to do, I was just confused. So I cried, ‘what do I do? A beautiful little boy, what do I do with you?’. So when I went to the sessions it was so helpful, you cannot imagine.

Peer and professional support resulted in participants feeling acknowledged and ‘not alone’. Attending the E-PAtS programme resulted in improved parent–child relationships with siblings of the child with intellectual disability. For example, participants described improvements in communication and increased duration of quality time spent with other children.

Participants described acting on specific information provided in E-PAtS group sessions. Strategies for assisting with living skills, including sleep, brushing teeth, eating and dressing, were introduced at home. Parents also described breaking down steps of individual tasks, such as brushing teeth, to make difficult tasks more manageable for the child. Participants described establishing routines to encourage positive behaviours and discourage challenging behaviours. Participants encouraged independence by allowing the child to attempt to dress, wash or feed themselves, and encouraged other family members to do the same. Furthermore, participants described actively involving siblings in the child’s care, which in turn has had a positive impact on the child’s behaviour and sibling relationships:

. . . it was kind of the living skills, and the idea of breaking them down into parts of the task . . . I’d have to do that [brush teeth] for him, but he can go in the downstairs toilet and get his toothbrush and toothpaste and bring it to me and afterwards he can put it back and after I’ve brushed his teeth I can let him brush his teeth after.

. . . with the likes of routine ‘and put this into place’ because now on her [Grandma] wall I have this chart that we can now go down and be like ‘no, this is what we’re gonna put into practice’.

But my mum would do everything . . . and I’d say ‘no no let him do it himself’ and there are things he can do even if you need to load the spoon for him it doesn’t mean he can’t put it in his mouth.

Facilitators and participants described how attending the E-PAtS programme sessions had a positive impact on communication, including using symbols, with their child and their child’s behaviour:

. . . this one family it made such a difference to them, it really, really did and to see the change in their child from one week to the next with a few simple strategies, one of which was remaining calm when the child is doing what he’s doing and you’re afraid of him hurting themselves, not running in there and being as excitable and anxious as the child but just bring it all down a notch, that was the main thing that she took away from that and it had such a big impact, that was brilliant.

. . . some of the people there had children who had ASD [autism spectrum disorder] who had presented with things that [study child] presented with, I was given tips and stuff, so I can’t pinpoint exactly what, but certainly it had a really positive impact at home.

Sharing of knowledge from the E-PAtS programme with others – which was identified as a key programme mechanism – enabled individuals and organisations to adopt new practices or to act as a support mechanism for parents:

And that was again highlighted so it’s OK to struggle at times but to reach out and get support and talk to other parents, that was a big thing that really helped, talking to other parents.

It was great because now my child’s school has now brought in that aspect of it. They’re using that a lot now at school. . . . to help with his communication and his choices and things throughout the day.

Yes. I shared it with the dad . . . which he understands very well now.

Participant perceptions of usual practice

The E-PAtS course was described as unique to other courses available to family carers of children with intellectual disability. Facilitators and participants valued that the E-PAtS programme was designed and delivered specifically to parents of children with intellectual disability and was not a generic parenting course. Participants valued attending and sharing experiences and knowledge of local support available with other parents of children with intellectual disability. The co-delivery aspect was considered to differentiate the E-PAtS from other courses available (i.e. participants welcomed the interactive course delivery, compared with the ‘lecturing’ styles of other courses). Participants and facilitators described the E-PAtS programme as a positive course in the way the content was organised, the way the course was structured and the course’s child focus. The E-PAtS programme differed from other courses available because it was developed specifically for children aged < 5 years and dealt with addressing challenging behaviour:

. . . other programmes I’ve delivered and we’ve had a range of ages and the one that I can think of particularly, the ones with the older kids really scared the ones with the younger kids because they were saying ‘mine does this’ sort of thing but I think because they’re all of a similar age that’s a very positive thing as well.

. . . not many courses talk about challenging behaviour and I think it’s an area that is very . . . it can cause a lot of distress to parents and children and it’s very needed to be understood, to be able to deal with it really.

You get sent on courses or sent to groups where you might be the only parent there with a child with additional needs and that is really tough and generally people were saying we just don’t go, we just don’t do it, why would we because it just makes us feel worse than before we went.

. . . very much someone standing, a professional person standing and speaking in a lecture format and that’s fine. I think what E-PAtS brought that was really different was that opportunity for discussion around those sorts of things and to talk about personal experience and share that and learn from others, and, as I say, as much about getting advice and tips and that support generally, and understanding.

Service provider organisation survey results

The survey that was distributed to organisations that were representative of those likely to be invited to provide the E-PAtS programme within a future trial produced 15 responses. There was significant interest in running the E-PAtS programme, which was seen to have strong fit with the organisations’ roles and purpose. Although delivery of the intervention was broadly seen as feasible, much lower levels of perceived feasibility were reported for the provision of funds to train facilitators and, to some extent, other aspects of funding needed. There was significant variation (mainly linked to organisational structure/purpose) as to whether or not respondents thought fulfilling staff requirements would be difficult to achieve.

The majority of organisations (13/14 organisations that answered the relevant question) indicated that they were somewhat or very likely to participate in a future RCT. When respondents indicated uncertainty concerning participation in a future trial, this was linked to two main issues: (1) the feasibility of securing additional/external funding for training and programme delivery and (2) the need to consult with and gain approval from senior management. Although some respondents did raise concerns about the trial design (and randomisation of participants to a control condition), this did not appear to be a major obstacle to participation. The key barriers that would need to be addressed were the resources required to deliver the programme and the provision of sufficient time to address this prior to the start of programme implementation.

Suggested improvements to the E-PAtS programme

Parents and facilitators expressed different views on the age of the child for whom the E-PAtS programme may be suitable. Some parents/facilitators thought that parents of older children or children who had received a diagnosis would benefit from the course, whereas others suggested that the course was introductory and more suited to parents of younger children. Some participants also suggested that group members would benefit from groups that were focused on a specific disorder (e.g. different groups for autism, Down syndrome, etc).

Suggested improvements to the course content included improvements to the presentation of the handbook and the slides, which were considered to be too scientific and sometimes not appropriate, specifically the developmental table in the communication session 4. However, the developmental table in the session has since been removed from the E-PAtS programme content. Some participants suggested that the course sessions were too long and could be shortened without compromising the content. Some groups experienced a double session (structured to help fit all sessions into available time frames for families and organisations) during delivery and this was sometimes experienced as too long, and it was proposed that sessions should not be combined.

Facilitators and participants who facilitated/attended smaller groups expressed the view that larger groups would be preferable and emphasised the importance of local groups so that the course is accessible to more parents. Guest speakers who were experts in certain areas of child development, optional one-to-one sessions and follow-up group sessions were also suggested as desirable.

Cost estimation for delivery of the E-PAtS programme

The results of the cost estimation for delivery of the E-PAtS programme are provided for each study site in Table 19 (as of 17 April 2020). The cost components are aggregated into three categories: (1) staff costs, including supervision, planning, direct delivery, meetings with professionals, telephone calls and paper work; (2) travel costs, which were based on the distances reported by programme facilitators and modes of transport; and (3) other costs, such as costs of refreshments, reimbursed participant travel and child-care costs associated with delivering the programme sessions.

Table 19 shows the mean intervention costs per session delivered at each site. These varied from £91.37 (site 1) to £109.55 (site 3). The main driver of the intervention costs per session delivered was staff costs, which were valued at £61.25 in site 1 and £96.11 in site 3. Initial training costs (set-up) were £3426.67 in site 3 and £3174.17 in site 1 to prepare new E-PAtS programme facilitators.

Broader resource consequences associated with the E-PAtS programme

Data were collected about broader resource use associated with the E-PAtS programme at 3 and 12 months. These data included parents’ and children’s use of hospital care services (i.e. hospital inpatient stays, hospital day centres, hospital accident and emergency visits and hospital outpatient clinic appointments), children’s development centres and children’s day centres, community-based health care [i.e. general practitioner (GP) services, health visitors, paediatricians, ophthalmologists, audiologists, speech and language therapies, physiotherapists, psychologists, psychiatrists, dentists, child mental health services, private child care, nurseries and occupational therapists], community-based social care (i.e. counsellors, family or patient support, Women’s Aid, housing department services and Citizens Advice bureau services) and medicines. The resource use questionnaires were piloted to evaluate participants’ comprehension levels of the resource use questions.

Analyses of broader resource consequences

Resource use items were summarised by trial arm (intervention and control) and follow-up period. Means and SEs for resource use values by resource category were estimated by trial allocation group for each of the follow-up periods, namely months 1–3 and 4–12.

Tables 20 and 21 present resource use values for parents and their children, with complete data by trial allocation and study period. With respect to children, the resource values are presented for subcategories of resource item, including (1) health-care services, such as GP services, health visitors, paediatricians, ophthalmologists, audiologists, speech and language therapies, physiotherapists, psychologists, psychiatrists, dentists, child mental health services, private child care, nurseries, occupational therapists and other services, and (2) hospital services, such as hospital inpatient stays, hospital day centres, hospital accident and emergency, hospital outpatient clinics, children’s development centres and children’s day centres. With respect to parents, legal and social services, such as counsellors, family or patient support, Women’s Aid, housing department services and Citizens Advice bureau services, were incorporated in addition to health-care services.

Nurseries and private child-care services were commonly used among children with complete data. Private child-care services were used an average of 4.81 times per child in the intervention group and 2.96 times per child in the control group over the first 3 months. Over months 1–4, nurseries were used 17.68 times in the intervention group and 20.50 times in the control group. Twenty-one children (65.63%) in the intervention arm used medications over months 1–3, compared with 12 children (46.15%) in the control arm. Use of hospital services, such as inpatient, day case and outpatient services, was relatively low in both trial arms. GP surgery use was the most common form of community health-care services for parents. Among parents, the mean number of GP surgery contacts was 1.44 in the intervention group compared with 0.50 in the control group over months 9–12.

Sources of unit costs for broader resource consequences

Resource inputs were valued using various secondary sources for unit costs. Sources of unit costs for inpatient hospital admissions over the study time horizon and other hospital services, such as day care, outpatient care and accident and emergency visits, were obtained from national tariffs (NHS reference costs trusts schedules). 66 Unit costs for community-based health and social services were derived from a range of national compendia, including the Personal Social Services Research Unit unit cost compendium67 and NHS reference costs trusts schedules. 66 The costs of using private child care and nurseries were obtained from the Childcare Costs Survey 2015. 68

Sources of medication costs per dose were largely obtained from the British National Formulary69 or the British National Formulary for Children. 70 Table 22 provides a compendium of unit costs for the broader resource consequences that were valued at 2018/19 prices (GBP) and inflated, where necessary, using the NHS Hospital and Community Health Services Pay and Prices Index. 67

Analysis of health-related quality-of-life outcomes

The health-related quality of life of the parents was assessed using the EQ-5D-5L obtained at 3 months’ and 12 months’ follow-up. The EQ-5D-5L has two measurement components. The first component is a descriptive system that categorises health-related quality of life into the following five dimensions: (1) ‘mobility’, (2) ‘self-care’, (3) ‘usual activities’, (4) ‘pain/discomfort’ and (5) ‘anxiety/depression’. Participants report their level of function for each dimension within one of the following five ordinal levels: (1) no problems, (2) slight problems, (3) moderate problems, (4) severe problems and (5) extreme problems/unable to perform. The crosswalk value set for the EQ-5D-5L was used to value responses to construct an EQ-5D-5L utility score for each respondent. 50 The utility scores are indexed at 0 or 1 (0 represents death, 1 represents full health and values below 0 indicate health states that are worse than death). The utility scores can be converted into quality-adjusted life-years (QALYs) using an area under the baseline-adjusted utility curve and can be calculated using linear interpolation between utility scores at baseline, 3 months and 12 months. The second part of the EuroQol-5 Dimensions (EQ-5D) is a visual analogue scale (VAS), which ranges from 100 (best health condition) to 0 (worst health condition) and describes the participant’s own assessment of their health status.

We report descriptive statistics for EQ-5D-5L utility scores and VAS scores at each time point. Similarly, we report descriptive statistics for levels of function for each of the five dimensions of the EQ-5D-5L, as well as suboptimal (less than optimal) levels of function (defined as some, moderate, severe or extreme problems) by trial arm and follow-up point.

Table 23 shows that health-related quality-of-life outcomes were marginally better for the intervention group than for the control group. The mean EQ-5D-5L score was marginally higher for the intervention group than for the control group at 3 months (0.803 vs. 0.746) and 12 months (0.838 vs. 0.812). In addition, the mean EQ-5D VAS score was marginally higher for the intervention group than for the control group at 3 months (74.22 vs. 74.00) and at 12 months (74.56 vs. 69.77). The table shows broadly similar patterns in less than optimal levels of function between the intervention and control groups for each of the five dimensions of the EQ-5D-5L, at 3 and 12 months, other than for self-care at 3 months.

Expression of cost-effectiveness of the E-PAtS programme

The cost-effectiveness of health and social care programmes can be expressed in terms of incremental cost per unit change in a primary outcome measure, which is called the incremental cost-effectiveness ratio. Outcome measures within the feasibility study included the WEMWBS. However, this is not a preference-based measure that permits the estimation of QALYs amenable to cost-effectiveness decision-making. 80

Combining disparate outcomes for parents and children in a single preference-based outcome measure will be a challenge for a future economic evaluation of the E-PAtS programme. It is anticipated that QALY-based approaches will not be able to capture the disparate effects of the E-PAtS programme on both parents and children, partly because of an absence of a validated multiattribute utility measure for early childhood and partly because of methodological challenges surrounding aggregation of disparate benefits for both parents and children in a single metric.

Discrete choice experiments (DCEs) offer a preference-based approach for valuing the potentially disparate effects of the E-PAtS programme. DCEs are broadly used in health economics to address a wide range of health policy-related questions. For example, they can be used to improve adherence to public health programmes or to quantify the trade-offs that trial participants are willing to make between different features of health care. 81 The approach stemmed from random utility theory,82 which assumes that respondents behave with the aim of maximising the utility of given choices. Sustaining adherence to a social intervention programme, such as the E-PAtS programme, has been a main task for improving the implementation quality of the intervention. 83 The evidence revealed by a DCE can be used for policy-makers in delivering the E-PAtS programme. Given the results of a DCE revealing the preferences of trial participants, delivering a bespoke programme based on the DCE results will enhance the possibility of adherence to the programme.

One of the purposes of this feasibility study was to examine factors associated with the adherence to and reach of the intervention for a future trial. Various approaches have been recommended to identify attributes for DCEs, including literature reviews, expert opinion, qualitative research and evidence from other existing studies. On the basis of the qualitative interviews performed as part of the E-PAtS intervention and the preferences stated by the trial participants, four attributes relating to the adherence and reach of the E-PAtS programme were identified: (1) location, (2) timing of the intervention, (3) method and (4) frequency. In the same manner, two levels for each of the attributes can be delineated to explore factors associated with the adherence to the E-PAtS programme:

• location (home or drop-in clinic)

• acceptability of randomisation (i.e. timing, now or to be offered the E-PAtS programme at a later date)

• method of data collection (face to face, telephone, alternate between face to face and telephone)

• frequency (weekly or monthly).

It is anticipated that the utility function for the DCE in a future definitive trial-based economic evaluation of the E-PAtS programme will be expressed within a mixed logit model. 83 All attributes will also be coded as a dummy, with one level for each attribute coded as the reference group. If a cost attribute is additionally introduced into a DCE, a future evaluation can aim to estimate marginal rates of substitution between changes in the cost attribute and the remaining attributes. 84

It is crucial to structure an efficient design in constructing choice sets of a DCE questionnaire. D-efficiency, which is an indicator of statistical efficiency minimising the determinant of the covariance matrix, needs to be considered. 84,85 The information with regard to priors for coefficients can be obtained from the feasibility study. To list all possible choices (i.e. a full factorial design) would be impractical for future trial participants, as this would require too many choice sets. For example, four attributes with four levels would result in 256 possible scenarios (4⁴ = 256). It is also important to arrive at a balanced design that allows for each level to be chosen equally (i.e. levels of attributes need to be chosen an equal number of times, thereby minimising the variance in the parameter estimates). Therefore, a fractional factorial design rather than full fractional design will be considered to reduce the number of selected scenarios.

A DCE can be incorporated into an economic evaluation of the E-PAtS programme with a view to informing a cost–benefit analysis. 86 Cost–benefit analysis is not widely used in health economic evaluations as it is cumbersome to put monetary values onto health outcomes. However, by comparing the costs between delivery options delineated by variations in facilitator, location and timing, method and frequency, the results of the economic evaluation can be expressed as a net monetary benefit (NMB) within the frame of a cost–benefit analysis.

The NMB can be constructed by use of difference in costs and effects, which is analogous to an incremental cost-effectiveness ratio in a cost-effectiveness analysis. The NMB can be expressed as:

where ΔE is the effectiveness difference between comparators, λ is the maximum willingness to pay and ΔC is the cost difference between trial arms. When incorporating a DCE into an economic evaluation, ΔE is the regression coefficient of the outcome measures, and λ can be obtained in the survey as a price proxy of salary or income,85 or by directly asking the willingness to pay for the chosen option. 86 In general, a positive NMB means that the chosen treatment option is cost-effective as the costs are outweighed by the benefits of the intervention. 87

In conclusion, a preference elicitation approach, such as a DCE, is likely to be required to supplement the economic evaluation of the E-PAtS programme. This will quantify participants’ preferences for the heterogeneous outcome measures used in the E-PAtS programme. By using a DCE, decision-makers should be able to compare the trade-offs between different outcomes. The DCE results can be helpful in developing the effective delivery of the E-PAtS programme.

Using routine data for economic evaluations

There are several advantages of using routine data for a future economic evaluation of the E-PAtS programme. First, routine data can be used to validate and complement resource use data from the trial data collection instruments. In clinical trials, broader resource use data can be collected using health economic questionnaires completed by trial participants. The questionnaires request information about use of hospital services, community-based health and social care services, and medications. Use of routinely collected data enables researchers to track patients prospectively and retrospectively. 89 Linkage with routine data can also provide information about the medical history of patients. Therefore, researchers involved in future evaluations of the E-PAtS programme can refer to routine data when self-reported questionnaires provide insufficient information about participants. In addition, routine data can be used to inform long-term epidemiological models of disease or health status progression.

Second, routine data can supplement incomplete resource use information observed in trial-based economic evaluations. As the data offers information on past medical records, routine data can be used when designing an economic analysis that starts after patient or participant follow-up is completed. The routine data can be used for modelling parameters of costs in designing a study for which resource use data are not collected. Likewise, routinely collected data can be useful when baseline resource use data are needed but absent in a clinical trial that has already started. 90

However, existing routine data does not always provide information on social care. For an economic evaluation of the E-PAtS programme, collecting data on social care service use is essential, as children with intellectual disabilities commonly use various social care services in addition to hospital services. The State of Caring Survey and the Adult Social Care Survey offer information about social care, but they are limited in that the data are presented at an area level rather than an individual level. 91 Likewise, the NPD provides information about education and social care data in England, but the data are limited to children. As an example, in an end-of-life care trial, social care data that were obtained from local authorities were linked with routine data at an individual level to secondary health-care data. 91 Therefore, contrary to the aforementioned advantages of using routine data, it is likely that an element of prospective data collection will be required when conducting an economic evaluation that encompasses a social care element.

Based on the advantages of routine data, a future, trial-based economic evaluation of the E-PAtS programme may extract key resource use items from the Mental Health Minimum Data Set (MHMDS) and Hospital Episode Statistics. The MHMDS deals with children and young people’s access to psychological therapies, intellectual disabilities or autism services. This data set is based on mental health currencies and payments, formerly Payment by Results, and so the data can be helpful for mental health policy-makers and researchers who work on policy evaluations. The MHMDS also provides information regarding children and young people who access psychological therapies and other outpatient clinics for intellectual disabilities and services in the community.

Hospital Episode Statistics data cover a wide range of information about patients admitted to NHS hospitals in England (e.g. clinical information about diagnoses, patient age, gender, race and ethnicity, methods of admission and discharge, and geographical information of the admitted patients). In principle, Hospital Episode Statistics data cover all secondary care attendances in England.

Potential data providers and their data sets

Routine data sources can provide information regarding community care service use, such as use of GP services and hospital services (e.g. inpatient days).

Table 24 summarises potential data sources of interest for a future trial, both for resource outcomes and for longer-term follow-up. We have identified data providers in England that could be linked to on a national level. Other data providers that cover only one local area or only a percentage of the population have not been included, as a future trial would need to link all participants as efficiently as possible using the fewest number of data providers.

Timelines for accessing data

One of the main challenges of using routine data is the time it takes to apply and receive approval for access. 92 For example, the process for accessing hospital data from NHS Digital took more than 1 year for the Building Blocks Trial. 92 In addition, application approvals often rely on other parts of the research process, such as ethics approvals and other requirements (e.g. legal approvals), for data linkage. Therefore, the length of time to access routine data can often far exceed initial expectations.

Data providers do not make information on timelines generally available, as the time from submitting an application to receipt of data will depend on the complexity of the project as well as how much more detail is required in the application when submitted. Our recommendations for a faster approval process include the following:

• talking to data providers as early as possible about the project

• asking for documents (data flows, consent material) to be reviewed by data providers prior to other approvals being submitted (ethics)

• reading the guidance available from data providers on completing the application

• responding in a timely manner when contacted by the data provider.

‘Where access is of no additional cost an application will be made’

Originally, it was intended to apply for data from providers where there is no cost for data (e.g. the NPD). However, the decision was made to not consent participants to linkage of data as part of this feasibility study and, instead, to explore the acceptability of consent with current participants at 12 months. It was agreed that the time required to apply for data would outweigh anything that could be learnt by carrying out this process. Other projects in the Centre for Trials Research have applied for and accessed data from both NHS Digital and the NPD, and therefore nothing more could be learnt by doing this again for this project. 92–94 It was not intended for the data to be reported as part of the feasibility study-proposed outcomes (i.e. data on education outcomes) and so it would be an inefficient use of resource (both trial staff and public resource at the data provider end). Requests for data from data providers need to demonstrate the public benefit associated with the release of data and this could not be justified or demonstrated in this case.

Logistics of linking, transferring and storing routine data

For a future trial, the following logistical considerations will need to be incorporated into the final design.

Acceptability of routine data collection (quantitative)

Table 25 presents participants’ views on the acceptability of their routinely collected data being used in a future study. For further information, see Appendices 5 and 6. Forty-seven participants answered the questions surrounding routine data collection. Around one-third of participants were aware that researchers were able to request access to hospital data, with slightly fewer participants being aware that school and social care data can be requested. The intervention participants were less comfortable with the idea of agreeing to their/their child’s routine data being accessed. For example, 46% of intervention participants said that they would be ‘not at all comfortable’ or ‘not very comfortable’ with agreeing to their hospital data being accessed in a future study (compared with 9% of control participants). Thirty-eight per cent of the intervention group said they would have been less likely to take part in the E-PAtS programme if consent to access their routine hospital data had been requested. Only 9% of control participants answered that they would be less likely to take part in the study.

Acceptability of routine data collection (qualitative)

Questions were included in the interview schedule to explore the acceptability of a future trial linking to routine data sources. The questions on routine data collection/linkage appeared to be hard for some participants to understand, including what was intended by this activity, how data would be accessed, and the issue of whether or not some forms of data (e.g. GP visits, contact with social services) would be more or less sensitive.

There appeared to be considerable variation in the extent to which participants would be happy to take part in a study where routine data linkage would take place. Overall, although some people did have concerns about taking part in a trial of the E-PAtS programme, which included routine data collection, it was not a significant enough issue to mean that they would refuse to participate. Some participants understood and were in favour of the idea of routine data collection reducing the number of questions that they themselves had to answer. Several also expressed a desire to help contribute to research that would meet the needs of children and their families, which might therefore over-ride concerns about the use of their personal data.

There was also variation in the extent to which participants saw all routine data as similar (e.g. in terms of acceptability of researchers accessing it). In particular, some participants saw data on contact with social services as much more sensitive than number of GP visits, for instance. Similarly, there was a mix of views on the extent to which the acceptability and sensitivity of accessing data differed according to whether information related to the participant or to their child. For a number of participants, the decision to consent to data linkage would depend on knowing why the data were needed and how they would be kept secure. It was often difficult for participants to see the relevance of some of the data that might be collected during a future trial using data linkage. For example, some parents did not see why data on health service utilisation would be collected, as they are not immediately related to participation in the E-PAtS programme.

Overall, participants appeared to think that the researchers would access named files from the GP surgeries, etc, and that when they worked with the data researchers might know whose data they were using:

I didn’t really think that they would hand stuff out like that, is that not all supposed to be confidential?