The impact of active stakeholder involvement on recruitment, retention and engagement of schools, children and their families in the cluster randomised controlled trial of the Healthy Lifestyles Programme (HeLP): a school-based intervention to prevent obesity

The cluster-randomised controlled trial of the HeLP assessed the effectiveness and cost-effectiveness of HeLP in preventing children being overweight or obese [19]. The primary outcome was change in body mass index standard deviation scores (BMI SDS) at 24 months post randomisation. Based on the trial power calculations [19] and assuming a 20% loss to follow up, we needed to recruit 28 schools of varying sizes with a total of at least 952 children to ensure that we had 24-month outcome data from 762 children for the analysis. We decided to recruit 32 schools to ensure we had a minimum of 28 schools completing the trial, each with an estimated average of 35 year-5 children. To ensure that we had a representative sample of schools in terms of socio-economic status, we aimed to have half of our recruited schools with ≥19% of pupils eligible for free school meals, which represented the national average at the start of the trial. In July 2012, following the recruitment period, half of the schools were randomised to receive HeLP and half acted as control schools following stratification by the proportion of children eligible for free school meals (<19%, ≥ 19%) and class size (one year-5 class, >1 year-5 class). For practical reasons half of the schools (eight intervention and eight control) entered the study in 2012 (cohort 1), and the other half in 2013 (cohort 2). Children were recruited in September 2012 for cohort 1 and in September 2013 for cohort 2, using an opt-out process. Children were measured at baseline (before group allocation was revealed to schools and research staff) and then at 12, 18 and 24 months post baseline (see Fig. 2).

Participation by children and their families (for parental engagement events), was assessed using registers of attendance with percentages of children or parents attending each component and/or phase calculated.

All children had a one-to-one discussion with the HC in phase 3 about their goals. It was during this interaction that the HC gave each child an engagement score between 0 and 3 (0 = disinterested/unaware goals needed to be set; 1 = reluctant/needs a lot of prompting; 2 = enthusiastic and happy to chat about goals and how they will achieve them; 3 = very enthusiastic, has discussed them at home and has clear strategies for achieving them). The HC had worked closely with the class for 10 months at this time, thus had the insight to carry out an accurate assessment. These scores were then dichotomised to create two groups (≤1 = less engaged children and >1 = engaged children).

Parental engagement was measured using two sources of data; attendance at one or more parent events and/or signature on the goal setting sheet. A score between 0 and 2 was given to each parent (0 = did not attend/did not sign; 1 = attended or signed the goal setting sheet, but not both; 2 = attended one or more events and signed the sheet). As with child engagement these scores were dichotomised to create two groups (≥1 = engaged, <1 = not engaged).

School engagement was assessed using three scores based on the HC’s interaction with the head teacher, the year 5 teacher(s) and the support staff. A score between 0 and 3 was given to each group (0 = unengaged/uncooperative; 1 = supportive; 2 = enthusiastic and supportive; 3 = very enthusiastic and used HeLP in other aspects of teaching/school activities). These scores were aggregated and then dichotomised into two groups (0–3 = less engaged school and 4–9 = engaged school).

From the outset we worked with a group of teachers, head teachers, parents and children from the early piloting of HeLP [25] who became our Project Advisory Group (PAG). Membership of this group has increased as we progressed though the piloting phases. Meetings were held when required at times that were convenient to the group (usually 4.00–6.00 p.m.) and all expenses were paid including cover for teaching staff if this was necessary. The composition of the group depended upon the focus of the meeting. Our PAG members advised us on what was feasible and acceptable when taking behavioural and anthropometric measures from 9–10-year-old children and how to communicate with parents about the research process so that they: (a) would receive the information, (b) understand it and (c) feel they were able to engage with the researchers if they had any concerns or queries. In addition, it was important for us to understand how to recruit schools and engage teachers. The head teacher in our PAG suggested we recruit schools via a regional network of primary school heads, the Devon Association of Primary Heads (DAPH) during one of their quarterly briefing sessions, and a teacher involved in the exploratory trial [26] offered to talk to heads about her experiences of being involved in the programme during this session. We were also advised to recruit schools via the Academic Learning Partnerships (ALP), which consists of clusters of approximately eight schools in a particular location. Head teachers of these partnerships meet regularly, thus providing an opportunity to target schools to increase the chances of achieving a representative sample. We presented the study at the two locality DAPH meetings and invited schools to sign up there and then to participate. We also attended two ALPs to discuss the study at one of their meetings to recruit further schools in order to meet our deprivation criteria.

Teachers and parents from our PAG were invited to be partners on our research bids and both our funded exploratory [26] and definitive trial [19] had a year-5 teacher and a parent as a co-applicant.

Table 1 details the main strategies used to engage schools and parents with the study overall to ensure that we were able to recruit sufficient schools and children across the socio-economic spectrum and, crucially, follow up as many children as possible at both 18 and 24 months. These strategies were based on the advice of our PAG and lessons learnt during the piloting phases [25, 27]. Strategies have been grouped under three headings.

Our PAG not only advised us on how best to recruit and engage participants so that they remained in the trial, but they also co-created the intervention with us as our research partners. They provided invaluable feedback on possible intervention activities and delivery methods ensuring they were acceptable and feasible for schools, children and their families. It was important that any intervention we developed did not widen existing health inequalities and had the potential to engage children and their families from across the socio-economic spectrum. We wanted to encourage children to identify with, and take ownership of, the healthy lifestyle messages and take them home to their parents and discuss them with their peers [18].

Our PAG also highlighted the importance of quality delivery by personnel who were able to engage school staff, children and their families. Teachers in our group commented that having one key contact person would help build relationships and that this person should have the necessary skills and competencies to understand the busy lives of teachers and parents. They believed that a background in working with children would be really important in understanding how children behave and respond. From these discussions the role of the HC was developed and the parents on our PAG helped recruit the HCs for the cluster RCT in addition to providing critical feedback during practise delivery of the parent assemblies. Table 2 summarises engagement strategies used within the intervention and throughout delivery.

Within a 4-month period, 36 schools signed up to participate following the two DAPH briefing events and 8 schools signed up following the two ALP meetings, giving a total of 44 primary schools, of which 36 were eligible (having at least one year-5 class of 20 or more children). There were 32 schools that we purposely sampled to ensure that trial schools represented a range of sizes (one to three year-5 classes), locations (urban and rural) and deprivation (5–53% eligible for free school meals). The remaining four schools were asked if they were prepared to go on a “wait list” in case a cohort-2 school dropped out during the waiting year. All four schools agreed. There were a total of 52 classes across the 32 schools and 1371 children were eligible to participate. There were 34 children who opted out prior to baseline measures (17 in cohort 1 and 17 in cohort 2) and 13 children had left school before baseline measures were taken, giving a total of 1324 children (see Fig. 2).

At baseline, we achieved 99% or more completeness of data collection for all anthropometric and behavioural measures at baseline (see Table 3). At 12-month follow up (immediately after delivery of the final phase of the intervention) MLQ data were collected from 95.7% (630/658) of children in cohort 1 and 97.3% of children (648/666) in cohort 2 (96.5% in total).

The 18-month follow up occurred in May/June when the children were in year 6, following their Standardised Assessment Tests. The measures taken at this time point were anthropometric, food intake and physical activity. Anthropometric data were collected by assessors blinded to group allocation. To obtain 24-month measures, children were tracked to their respective secondary schools (n = 60 schools) where only anthropometric measures were taken, by assessors blinded to group allocation (see Fig. 2). Table 3 shows the completeness of data at each time point for anthropometric and behavioural measures, respectively.

Across both cohorts 676 children were randomised to receive the intervention. Child attendance registers were kept for each component in each phase: 52% (353/676) of children had family attending at least one parent invitation. Table 4 shows the percentage of children participating in each phase of the HeLP intervention and the total percentage of children receiving key components i.e. providing children with the information, motivation and behavioural skills in the drama workshops and leading them to take action through the use of personalised goal setting and support. Intervention uptake was exceptionally high for all phases and across both cohorts.

There were 96% of children (652/676) who set goals with the HC in phase 3 and of these, 63% (411/652) had parental support (indicated by a parent signature on the goal setting sheet and/or written comments regarding how the parent would support the child in achieving their goals). In total 92% (602/652) of children were deemed to be engaged with the HeLP programme with little difference between genders (91% of boys and 94% of girls). There was no difference in IMD rank spread between the engaged and less engaged children, suggesting that HeLP was able to engage across the social spectrum.

There were 77% of parents (520/676) deemed to be engaged with HeLP. There was no difference in IMD rank spread between the two groups, suggesting that HeLP was also able to engage parents across the social spectrum. Out of the 16 intervention schools, only three were categorised as less engaged. Reasons for this included administrative and teaching challenges due to external school assessment pressures and staffing issues due to the absence of the head and/or teachers, whereby the drama workshops were viewed as an opportunity to free up teachers to do “other things”.