Preschoolers in the Playground: a pilot cluster randomised controlled trial of a physical activity intervention for children aged 18 months to 4 years

Theoretical underpinning

The utility of basing health promotion interventions on sound theoretical frameworks is well expounded. 32 A systematic review33 conducted by members of our team (CDS, HJM) reported that the predominant behaviour change theory used in successful childhood obesity prevention interventions is Bandura’s social cognitive theory. 34,35 This theory describes behaviour change as an interaction between personal, environmental and behavioural factors. Personal and environmental factors provide the framework for understanding behaviour. The personal concepts include skills, self-efficacy, self-control and outcome expectancies whereas the environmental concepts include availability (e.g. provision of space for physical activity) and opportunity for social support (e.g. a group setting). The review also reported that providing information on behaviour–health links appeared to be an important component in childhood obesity prevention and physical activity interventions.

An outdoor setting

Enhancing environmental and cultural practices that support children to be more active throughout the day are thought to be promising strategies to prevent childhood obesity, particularly if the children perceive the activities as being fun. 36 Time spent outdoors has been shown to correlate with physical activity levels in preschool children30 and outdoor play is associated with a lower risk of overweight. 37 School and preschool playgrounds provided a potential opportunity to intervene to increase physical activity levels. Indeed, in school-aged children, outdoor playground interventions have been shown to increase daily physical activity. 38,39 Furthermore, some studies have reported that factors related to increased MVPA in playground interventions include greater provision of equipment38,39 and greater play space per child. 40 However, a recent systematic review concluded that further research is needed to elucidate which intervention strategies and playground characteristics are most effective. 41 Interventions in preschool playgrounds supervised by preschool teachers have had mixed success. Adding portable play equipment in a US preschool playground increased physical activity levels in 3- to 5-year-olds. 42 In contrast, in Belgium, no change in physical activity levels of 4- and 5-year-olds was reported after providing playground markings or play equipment or both in the preschool playground. 43 The authors concluded that creating an activity-friendly environment may not be sufficient to promote physical activity in preschoolers and regular infusions of different equipment with more guidance and encouragement from adults to play in an active way are required. Currently, there is no evidence regarding preschool playground interventions in the UK.

Seasonality and weather conditions can present barriers to participation in outdoor play for children. 44,45 Emerging literature is beginning to focus on the importance of season in physical activity promotion and weight management. Two recent studies in the UK21 and Denmark46 have examined seasonal variations in preschool children’s physical activity levels. In both studies, overall, children were significantly less active during the winter months than in the spring, summer and autumn. The Danish study also examined differences in activity patterns across seasons according to body mass index (BMI) status. Children in the highest tertile of BMI had a flatter activity profile throughout the year than children in the lowest tertile. Those in the lowest tertile showed a curved pattern, with activity peaking between June and August at around 200 counts per minute (measured via accelerometry) greater than their higher BMI counterparts. Although not statistically significant, these different activity patterns between normal and overweight people have also been observed in adult populations. 47 Thus, the establishment of annual patterns of physical activity, with levels peaking in the summer months, may begin during the preschool years and possibly track into childhood, contributing towards the prevention of childhood obesity. This emerging research highlights the need to consider seasonality when designing and evaluating the effectiveness of interventions. The PiP intervention was conducted in three waves that were staggered across the year to examine the impact of season on the feasibility and acceptability of the intervention.

Parental involvement

Parental engagement should be a key part of any intervention in the preschool years48 and the suggestion that adult encouragement is required to increase active play is supported by findings from a systematic review of preschool obesity prevention interventions. 49 The 12 interventions reported in this systematic review were conducted in a variety of settings (preschool/childcare, home, group, primary care and mixed settings) and included a physical activity component. Home-based interventions appeared to be the most successful at increasing physical activity despite the small sample sizes of studies and poor adherence to the interventions. This is perhaps because of parental involvement in the interventions, which has been suggested to be vital for facilitating behaviour change during the early years. 33 Mothers in particular are considered to play a major role in establishing a child’s healthy lifestyle behaviours50 and this, in part, may be indirectly through modelling of behaviours. 51

Since the PiP pilot trial started in September 2012 more physical activity interventions for preschoolers have reported a parent or family component. A recent German preschool physical activity intervention with strong and active parental involvement found statistically significant increases in physical activity, reductions in sedentary time and improved perceived general health and quality of life 1 year post intervention. 52 Using a participatory approach, parents were supported to develop and implement their own physical activity-promoting activities for their children in a preschool centre. This study is a good example of capitalising on a setting where preschool children and their parents congregate and involving parents in activities to promote and model physical activity. Another study from the UK demonstrated a statistically significant increase in preschoolers’ physical activity levels immediately after receiving a 10-week active play programme that targeted the whole family. 53 Families were invited to attend five sessions of active play and parent education. Although these recent studies are promising, the changes in physical activity have not been clinically meaningful and thus further research and intervention development is required for this age group.

A further intervention in Belgium (ToyBox) used a preschool-based curriculum plus passive family involvement (giving newsletters, tip cards and posters) approach. 54 This intervention demonstrated effectiveness for boys or children from high-socioeconomic status (SES) preschools only and not girls or children from low-SES preschools. Given that boys and children of a higher SES level have higher levels of physical activity,10,30 this intervention actually generated further health inequalities. These studies with a parental component indicate that active parental involvement rather than just passive parent education may be required to elicit improvements in preschoolers’ physical activity levels and prevent further health inequalities. The development of preschool physical activity interventions that focus on increasing physical activity levels in groups with particularly low levels, including girls, low-SES children and ethnic minority groups, requires particular attention.

Site of delivery of the intervention and timing of the sessions

Local primary school playgrounds were chosen as the site for delivering the PiP intervention. It was anticipated that using the school setting for the intervention would overcome some of the barriers to participation in physical activity for preschool children reported by parents in the focus groups. For example, the intervention would be available locally and would fit into family routines and the playground is an enclosed space that can be supervised easily. Furthermore, the intervention would be accessible to hard-to-reach families, who often do not attend other community-based children’s activities such as those held at children’s centres. Primary school playgrounds (or children’s centre playgrounds when located on the primary school site) were made available to parents and preschool children at times when parents were likely to be attending school (e.g. dropping off or picking up older children from school or young children from nursery). These times were chosen to capitalise on times of the day that parents are already out of the house and minimised the number of journeys to activities that parents had to make (identified as a barrier in focus groups). Six PiP sessions per week for 30 weeks (three school terms) were available to families. This number of sessions was chosen to meet the need for flexibility, which was highlighted at the focus groups. Families were encouraged to attend at least three of these sessions each week. Each PiP session lasted for 30 minutes. There were two phases of the intervention, the initiation phase (10 weeks: one school term), which was supervised and facilitated by a member of school staff, and the maintenance phase, which was unsupervised (20 weeks: two school terms).

Intervention facilitators

Following consultations with two staff members from two schools, parental involvement workers or nursery teachers were suggested as appropriate facilitators to deliver the intervention. It was thought that by training existing staff the intervention would be more sustainable. The initiation phase (10 weeks) was facilitated by parental involvement workers and/or nursery teachers employed by schools/children’s centres. These facilitators undertook a 2-hour training session using the PiP manual and were accompanied during the first session by the lead researcher (SEB) who developed the PiP intervention. Subsequent telephone support from the lead researcher was offered to facilitators during the intervention period as and when they requested it.

The Preschoolers in the Playground manual

The PiP manual was developed as a resource for the facilitators. It included a background section explaining how the intervention had been developed, an explanation of the facilitators’ role and what was expected of them, an equipment list detailing the play equipment that would be required to deliver the intervention (all play equipment was standard school physical education equipment), information about the importance of physical activity in the early years, guidelines for physical activity in preschoolers and information about giving effective praise, encouragement and positive reinforcement. The manual also described the session timings and structure of the sessions. A guide on how to deliver the first session was also provided for support. The manual included a description of the ‘statue game’ used in sessions to gather the group together, a description of 20 structured play games called PiP-PoPs, five handouts to be photocopied for parents and instructions on facilitating 10 discussions (one per week) with parents about physical activity-related topics (more detail is given in Content of the Preschoolers in the Playground intervention and Appendix 2). At the back of the manual was a paperwork section that included a register of attendance and adverse event reporting forms. The manual was reviewed by five early years workers and minor revisions were made before the start of the intervention (e.g. changing the order of the contents in the manual). The manual was also modified during the delivery of the intervention following feedback from the facilitators delivering the intervention (e.g. language was clarified).

Structure of the Preschoolers in the Playground sessions

Each PiP session lasted for 30 minutes and included two 5-minute structured play activities (PiP-PoPs), 15 minutes of free play (during this time handouts were given out and guided discussions conducted) and 5 minutes for an active tidy-up. Play activities were designed to be of light physical activity (LPA) and MVPA. There is some evidence to suggest that as little as an extra 10 minutes per day of MVPA significantly reduces fat mass. 2 Furthermore, 56 minutes per day of MVPA in boys and 42 minutes in girls aged 5–8 years has been reported to improve metabolic status assessed from a composite score of various CVD risk factors. 61 Given the sporadic nature of physical activity in very young children, and that between 25% and 70% of total physical activity (TPA) in preschoolers has been classified as MVPA,23,62,63 it was estimated that during a 30-minute session the children would spend at least 10 minutes in MVPA and that this could contribute towards the amount of MVPA required for health benefits. Furthermore, it was thought that a 30-minute session would be more acceptable to both schools and families.

During the maintenance phase (20 weeks) playgrounds remained available to parents and their preschool children six times a week for 30 minutes at specific times to coincide with the ‘school run’, as allocated by the schools/children’s centres. For the purposes of the pilot trial, the PiP facilitator took a register of attendees at the beginning of the sessions but did not give formal supervision.

Content of the Preschoolers in the Playground intervention

The following sections provide a detailed description of the content of the PiP intervention. Table 1 summarises the content, evidence for the content and associated behaviour change techniques for both the initiation and the maintenance phases of the intervention.

Changes to the intervention

Before the PiP intervention commenced, a change was made to the duration of the intervention, from 36 weeks (12 weeks in the initiation phase and 24 weeks in the maintenance phase) to 30 weeks (10 weeks in the initiation phase and 20 weeks in the maintenance phase); blocks of 10 weeks were better suited to school term times. During the PiP intervention, changes were made to the content of the intervention. These included the addition of weekly, take-home play equipment during the initiation phase, with £15 spent per child. This action implemented the behaviour change techniques of providing rewards for participation and skill rehearsal. The take-home play equipment included items that would allow parents and children to recreate at home the PiP-PoPs that they had learned in the sessions. The kit included:

• PiP kit bag

• PiP-PoP game cards

• hoop

• skipping rope

• beanbag

• ball

• bat

• collector bag

• chalk

• water pot and brush.

A further change was the addition of telephone support provided by facilitators to parents who, for unknown reasons, had not attended three consecutive sessions (one week of the intervention). A change to the delivery of the intervention was also made: schools were given the option to reduce the number of sessions offered to four per week after the first 3 weeks of the intervention if two sessions were poorly attended.

School recruitment

The study was initially publicised to primary schools through advertising and presentations at education conferences and events hosted for Bradford primary school staff. Letters were also sent to eligible schools (those located in the two poorest quintiles of the IMD for Bradford) (see Appendix 1). Schools were asked to contact the researchers if they were interested in taking part in the study. Interested schools were then contacted and visited to discuss the study further. Schools that were located in different areas of the city of Bradford (Bradford South, Bradford West, Bradford East and Shipley) were selected to take part. Each study site completed a study site agreement, which was signed by the head teacher and chair of governors. To express thanks for participating in the pilot trial, all schools received a donation of £200 towards play equipment. Because the children were of preschool age and were not attending the primary schools, those affiliated to the schools in the control arm did not have access to the play equipment during the trial period.

Participant recruitment

Families from participating schools, and families who used children’s services affiliated with the schools (e.g. nearby children’s centres or nurseries), were approached to take part in the trial. Recruitment was carried out through letters home with school-going children and through face-to-face conversations with community research assistants in school playgrounds and children’s centres or nurseries. A participant information sheet (see Appendix 3) was given to participants before they attended a baseline measurement appointment. Written informed consent was obtained before any data were collected at the baseline measurement session. To account for the linguistic diversity among the study population, research staff recruiting families and conducting measurements and questionnaires were bilingual and undertook these tasks in either English or Urdu. 68 Urdu is the language spoken and understood by Pakistanis; it is also spoken and understood in many parts of Bangladesh. For Bangladeshi participants who did not fully understand Urdu, the consent form was audio recorded in Sylheti and participants were asked to sign a paper copy of the consent form (translations and transliterations were prepared in collaboration with Born in Bradford staff, according to standard Born in Bradford procedures68). All families received a £10 voucher for a children’s shop (Mothercare or the Early Learning Centre) as a thank-you for attending each study measurement session (baseline and 10, 30 and 52 weeks’ follow-up) and for participating in a qualitative interview.

Quantitative data

Health economics data

Qualitative data

Interviews were carried out with parents in both the intervention arm and the control arm and with PiP facilitators and head teachers in the intervention arm. The methods used to carry out the interviews are described here. How the data collected were used to meet each study objective is described in Methodology for each study objective.

All of the interviews used a semistructured topic guide (see Appendix 6) to ensure consistency. The questions were as open-ended as possible to enable participants to raise issues that were important to them. The wording and order of the topics covered also varied between interviews to ensure that each interview was informal and open-ended to suit individual participants. 76 Interviews were audio recorded digitally with the exception of the wave 1 interviews with parents (see the following section).

Methodology for each study objective

Recruitment of schools

In total, 27 schools were approached to take part in the study, 13 with predominantly South Asian pupils and 14 with predominantly white pupils. Of these, 15 (56%) expressed an interest in taking part (11 with predominantly South Asian pupils and four with predominantly white pupils). Nine did not respond to the invitation to take part and three declined. The reasons for declining were no space in the playground during school time (n = 1) and changes in school management (n = 2). Ten schools were recruited to the study (six with predominantly South Asian pupils and four with predominantly white pupils); this equates to 37% of those who were approached to take part. Table 7 shows the characteristics of the recruited schools. The smallest school had one class per year group and the largest had three classes per year group. All schools had a high proportion of children receiving FSMs. The schools with a high proportion of South Asian pupils also had very high numbers of children who spoke English as an additional language.

Recruitment of participants

In total, 341 children were screened for inclusion in the study and the parents of 305 (89%) of these children agreed to be contacted further about the study. The reasons for parents declining to be contacted further about the study are summarised in Table 8. The main reason for declining was that parents were too busy.

The CONSORT diagram in Figure 2 illustrates the flow of participants through the study. Of the 305 parents who agreed to be contacted about the study, 164 (54%) consented to take part in the study. The ethnicity of the recruited parents by consent status is summarised in Table 9.

FIGURE 2.

Cluster CONSORT flow diagram.

Table 9 shows that, of those who agreed to be contacted further about the study, a higher proportion of the white parents (66%) than the South Asian (52%) or other (18%) parents then consented to take part. In total, 47 (15%) parents did not state their ethnicity and 22 (47%) of these consented to take part in the study.

Randomisation and participant characteristics

Five schools and 83 children were randomised to receive the PiP intervention and five schools and 81 children were randomised to the control arm. The characteristics of the participating children and parents are shown in Table 10. The mean age of the children in the study was 2.8 (SD 0.7) years and the mean age of parents attending the baseline measurement session was 30.6 (SD 5.6) years. There was an equal proportion of female and male children (53% and 48%, respectively) in the study and there were more South Asian (57%) than white (37%) or other (6%) children. A similar pattern of ethnicity was seen for the parents at baseline (South Asian 54%, white 44% and other 2%).

Randomisation achieved balance across age, gender and parental ethnicity and attrition did not impact on the balance achieved at baseline, as seen in the ‘as-analysed’ sample (those included in the primary analysis). There was a slight imbalance in children’s ethnicity, with a higher percentage of white children in the intervention group. Table 11 summarises gender by ethnicity of the children in the study. In total, 45% of the white children were female, 57% of the South Asian children were female and 60% of the other children were female.

Recruitment and characteristics of those who participated in the qualitative component of the study

Head teachers’ reasons for taking part in the study (Box 1)

In discussing the reasons for agreeing to their school taking part in the PiP study, both head teachers identified perceived benefits for the children and the parents: linking health with education, engaging the parents in ‘meaningful activities’ with their children and promoting learning in children. Although both of the schools were either currently, or had previously been, involved with other Born in Bradford initiatives and were supportive of the Born in Bradford programme (see www.borninbradford.nhs.uk; accessed 6 March 2015), this was not the reason for participating in the PiP study. Similarly, although the £200 that the schools had received to spend on play equipment had been used to buy equipment for use after the end of the study, this was also not the reason why they had chosen to take part.

Parents’ reasons for taking part in the study (Box 2)

Parents offered a variety of reasons why they had agreed to participate in the PiP study. These were most commonly related to the perceived benefits of the study rather than what study participation entailed. The majority of parents explained that they viewed the PiP study as being ‘good’ for their child, which was why they had agreed to take part. When asked to elaborate on this, most talked about wanting to be involved in something that might directly benefit their child, other children or their child’s school. For some this was simply about the study’s focus on children, whereas for others it was related to the focus on physical activity. Two parents discussed how they saw the PiP study as an opportunity to introduce their child to the school environment before they started nursery in the next academic year. For many the PiP study appeared to be trusted to deliver these perceived benefits because it was part of the wider Born in Bradford programme. Several parents explained that they were already taking part in the Born in Bradford programme and so they were familiar with this type of research and/or they wanted to ‘give something back’.

Only one parent commented that the requirements of the PiP study had influenced her decision to participate. She was already involved with the Born in Bradford programme and perceived the PiP study to be relatively straightforward. She also explained that she had felt sorry for the community research assistant standing in the snow with a clipboard. Parents were not directly asked if being in the intervention or the control arm would affect their decision to take part and only one parent stated that they would not have taken part if put in the control arm because she had particularly wanted her daughter to attend the playground sessions.

Parents’ views on recruitment and study information (Box 3)

All but one parent described being approached by the community research assistant in the school playground or at the nursery or children’s centre to take part in the PiP study. The one exception was a mother who had learnt about the study from her sister-in-law and sought out the community research assistant to volunteer to take part. Parents were generally positive about being approached in this way and this appeared to be associated with the friendly manner of the community research assistants and the assumption that the school/nursery/children’s centre had endorsed the PiP study. Several parents explained that they had received a subsequent follow-up telephone call from a PiP researcher to make an appointment for the first measurement session. One parent had telephoned the Born in Bradford office herself as she was so keen to take part.

Over half of the parents interviewed had no particular recollection of the written study information (participant information sheet; see Appendix 3) that they had received and were unable to comment on it other than saying that they thought it had been ‘OK’. Two parents for whom English was not their first language said that they had only seen the participant information sheet in English and had been unable to read it but that the community research assistant had explained the study to them verbally, which had helped with their understanding. Other parents commented that the study information was clear both from the conversation with the community research assistant and in the written information that they were given to take home. Two parents commented that they had felt very informed about the study, each for different reasons. One had read the leaflet five times and the other explained that the PiP team had telephoned/met with her on many occasions to discuss the study (this parent’s son did not attend any of the PiP sessions and so she had received telephone calls to encourage attendance). Despite parents expressing confidence that they understood the PiP study, with further probing it was clear that many had limited understanding about the processes of the study (see Parents’ understanding of the study procedures).

Finally, parents in waves 2 and 3 were shown the participant information and consent forms during the interview. They still had little to say about this written information although commonly described liking that the study was explained step-by-step, using a question and answer format and bullet points. Only one parent made a suggestion for improving these resources, which was to include more colour and pictures to make them look more appealing.

Parents’ understanding of the study procedures (Box 4)

During the interviews it was clear that, although most parents claimed to understand the design and processes of the PiP study, in reality many were quite confused, particularly about the concept of randomisation and its implication for their participation. One parent explained that she had not really understood what the study entailed when she turned up for the first measurement session. Several parents did not understand the interview question, ‘What did you think about being randomised to the two groups in the study?’ Others understood randomisation to mean ‘picked by chance’ (P13); however, they did not appear to recognise its relevance to the PiP study. Examples of confusion included thinking that randomisation was about the PiP sessions being offered in the morning or afternoon, uncertainty about which trial arm the child’s school was allocated to despite having discussed attendance at the PiP sessions and thinking that the schools with the most children were chosen for the PiP intervention. Only one parent (in the control arm) commented that randomisation seemed unfair.

The PiP facilitators had mixed views about how well they thought the parents at their school had understood the study. Indeed, one facilitator believed that the low attendance at the PiP sessions was because of a lack of understanding.

Using the accelerometer (Box 5)

Most parents said that using the accelerometer belt was easy and this judgement seemed to be based primarily on how they perceived it was for the child to wear the belt. Parents generally described their child as ‘being fine’ with wearing the belt. Some explained how their child had really liked wearing it. Examples included a boy who thought that he was a ‘Power Ranger’ when wearing it and did not want to take it off at night or return it to the community research assistant at the end of the week. Another boy thought that it was his ‘Sportacus’ belt and liked the flashing light on it and a girl was described as feeling ‘important’ when wearing the belt. There were also examples of children being less keen to wear the belt, particularly when tired or unwell. The comfort or discomfort of the belt also made a difference, with one parent describing how she distracted her daughter when putting the belt on as it kept swivelling around, and tightening it to stop this was uncomfortable for the little girl. Indeed, several parents mentioned that the belt either moved or pinched the child when tight. Another practical issue mentioned by one father was that a small part of the accelerometer kept coming off (and could be swallowed by the child). One parent suggested that the accelerometer could look more child friendly, perhaps by being a different colour.

Parents were asked to ensure that their child wore the belt for 7 days over a week and to complete a diary of when the belt was worn. A few mentioned that they had sometimes forgotten to put the belt on their child and in some cases the child had reminded them. As with handling the belt, completing the diary appeared to be ‘fine’ for parents. Once again, the community research assistants were commended on their instructions and support for parents.

Measurement sessions (Box 6)

There were mixed views about the location of the measurement sessions, with a few parents not minding where they took place, some liking the sessions being at school and others preferring the sessions to be at home. When a preference was expressed this seemed to be related to convenience, for example tied in with the school drop-off or coming to the house when a parent was unable to attend the sessions in the school. The happiness of the child and the comfort of the surroundings were also mentioned. All of the parents from one school in wave 2 commented that the room was very small or that they had been moved from room to room during the second (10-week) measurement session.

Parents were generally accepting about the length of time that the measurement sessions took, despite describing some of the sessions as taking up to an hour. Reasons offered for these longer sessions included the volume of paperwork to complete, poor organisation (being moved from room to room during the session) and the tablet not working. They also tended to report that their children were fine or even enjoyed being measured and several commended the approach of the community research assistants who were observed as engaging well with the children, making the session fun and having a friendly approach. Two parents described slightly different experiences: one explained that his son had been grumpy at the second session but attributed that to him being ill; another discussed that her daughter had been a bit wary. Nevertheless, both of these children completed all of the measurements.

The one component of the measurement sessions that many parents were less positive about was the set of questions about themselves (rather than about their child). There were several issues mentioned: confusion as to the purpose of these questions (rather than focusing on the child’s physical activity); the perception that they were very long and repetitive; and their personal nature, particularly the mental health questions but also the questions on income (these questions are from the ComQol-A5). One mother for whom English was her second language said that she would have preferred the questions to be translated for her. A couple of parents spoke about a preference to complete the questions on paper or using a tablet rather than being asked them verbally; this again was related to the personal nature of the questions.

Feasibility of accelerometry

The number and proportion of data meeting the wear-time criterion (any 3 days with at least 6 hours of wear) overall, by trial arm and by ethnicity are presented in Table 18.

Overall, 68.9% of baseline data, 53.0% of 10-week data and 53.0% of 52-week data met the wear-time criterion. There was little difference across trial arms but a higher percentage of South Asian children met the wear-time criterion than white children. Only 64 children (39%), 29 from the intervention arm and 35 from the control arm, provided accelerometer data that met the wear-time criterion at both baseline and 52 weeks.

Feasibility of other health outcomes

The feasibility of the anthropometric measurements, the ComQol-A5 and the General Self-Efficacy Scale in terms of completion rates at each time point is presented in Table 19.

From Table 19 we can see that the quality of life and self-efficacy questionnaires had the highest response rates at baseline, week 52 and both baseline and week 52. The ComQol-A5 had 69% of data available at both baseline and week 52 and the General Self-Efficacy Scale had 79% of data available at both baseline and week 52. The anthropometry data response rates for those with baseline and week 52 data ranged from 56% for upper arm circumference to 65% for body mass, with height (62%), BMI z-value (57%) and waist circumference (59%) falling between these two figures. Higher response rates were seen among South Asian children than among white children for all measures apart for upper arm circumference.

Feasibility of conducting an economic analysis

The overall completion rates for both the EQ-5D and the PedsQL were at least 70% at each time point, whereas those for resource items did not fall below 80% (Table 20).

Attendance to the intervention

Attendance data for the PiP intervention sessions were available for 83 children from the five intervention schools and are summarised in Tables 21 and 22 by initiation and maintenance phase.

Across the intervention schools 54 of the 83 children (65%) attended one or more sessions during the initiation phase and 13 of the 83 children (16%) attended one or more sessions during the maintenance phase. Among those children who attended at least one session during the initiation phase, the average number of sessions attended was 9.1 (SD 9.6 sessions), with a median of three sessions attended (minimum–maximum 1–29 sessions), and among those children who attended at least one session during the maintenance phase, the average number of sessions attended was 5.2 (SD 3.4 sessions), with a median of 4.5 sessions attended (minimum–maximum 1–11 sessions). Attendance was much higher during the summer term initiation phase than during the autumn or winter/spring initiation phase. Attendance was also much higher in the initiation phase than in the maintenance phase. It was recommended that children attend 30 sessions in the initiation phase and 60 sessions in the maintenance phase (n = 90 overall). None of the schools provided the second term of the maintenance phase and, thus, the maximum number of sessions that families could have attended was 60.

The recommended number of sessions per week was three; in weeks 1–9, between six and 10 children (11.1–18.5%) attended three sessions. In week 10, no children attended the recommended number of sessions. Just under half of the participants (n = 23, 42.6%) attended one session in the first week; between three and 15 pupils (5.6–27.8%) attended one session per week in the remaining 9 weeks. The proportion of participants attending no sessions a week rose from 35.2% in the first week to 72.2% in week 10 and, with the exception of week 3, was between 51.9% and 68.5% in the interim weeks. Attendance data for the maintenance phase were available for three of the five intervention schools. In total, 13 children attended at least one maintenance session. No participants attended more than the recommended three sessions per week and only in weeks 1 and 7 did any children attend the recommended number of sessions (in both cases one child attended the three sessions). Mostly, participants did not attend any sessions per week (between 46.2% and 76.9%); the exceptions were weeks 2, 7 and 8 when around half attended one session (n = 6, 46.2%). Table 23 presents the attendance data by ethnicity for those children who attended at least one session.

Attendance at the PiP intervention sessions was higher among South Asian children than among white children. Among those children who attended at least one session, the average number of sessions attended for white children was 10.7 (SD 12.1 sessions), with a median of three sessions attended (minimum–maximum 1–33 sessions), and for South Asian children was 11.3 (SD 11.8 sessions), with a median of 4.5 sessions attended (minimum–maximum 1–35 sessions).

Three children withdrew from the intervention (one white and two South Asian). The reasons given for withdrawal were (1) that it was too cold, (2) that the child was injured and (3) that the mother was heavily pregnant. Throughout the duration of the intervention there were no adverse events reported.

Acceptability of the intervention (Box 8)

Only five parents who were interviewed in the intervention arm had actually attended any PiP sessions; two of these had attended most sessions (and provide the most detail about the sessions), two had attended a few/four sessions and one father had attended only once, explaining that he was not impressed with the delivery of the session. The views presented here are, therefore, based on a very small number of parents.

Four of the five parents were very positive about the PiP sessions. They saw them as fun for the children and commented on the variety of games and equipment that the children had enjoyed playing with during free play and in the PiP-PoPs. The two parents who had frequently attended the sessions identified how their children had developed over the 10 weeks, not only in initiating playing active games but also in developing confidence, learning colours, sharing toys with other children and making friends with children who would also be starting nursery in the following September. The wave 3 parents reported that they had received a piece of play equipment at each PiP session that they attended, which they then used at home.

The PiP facilitators also spoke about how they thought the children had enjoyed the sessions, offering examples of how children started playing with equipment whilst the facilitator was setting it up for the session and how they looked sad when the session ended.

Parents also spoke about the benefits to themselves of attending the sessions, namely making new friends, enjoying doing something with their child, being more active themselves and learning how to undertake active play with their child at home. One mother had enjoyed translating for the other parents. Parents did not talk particularly about the leaflets that they had received at the PiP sessions, with only one mother mentioning them (she could only remember the leaflet about what to wear in wet weather). Finally, the parent from wave 1 who had attended a few PiP sessions in the winter months observed that the sessions would have been better if more parents and children had attended.

The PiP facilitators also offered some suggestions of how parents had benefited from coming to the PiP sessions. They talked mainly about how they had seen parents develop in confidence in playing with their children, illustrated by them using the PiP-PoPs during the free-play sessions. One facilitator described how she had undertaken lots of modelling with parents at the start of the 10 weeks but gradually over time she had been able to do less as the parents knew what to do themselves. Two PiP facilitators also mentioned that these were parents and children who did not know each other before attending and so it was nice that they had met new people. Facilitating the transition to nursery was also mentioned in this regard. Finally, the wave 2 and wave 3 PiP facilitators spoke about the play equipment that the parents received for attending the sessions that they could then use at home, a further perceived benefit of participation.

Reasons for not attending the intervention (Boxes 9 and 10)

Of the parents from the intervention arm who were interviewed, four had not attended any PiP sessions, two had attended ‘a few’ sessions, two had attended regularly and two (parents of the same child) attended one session. Parents were asked directly why they had not attended the PiP sessions and why they thought other parents may not have attended. The PiP facilitators were also asked for their views on the reasons for low attendance to the sessions.

Parents offered a variety of reasons for not attending the PiP sessions. Some of these were related to family issues rather than to the PiP sessions, for example competing priorities or the child and/or other family members being ill. One mother also described stopping attending as she no longer worked at the school and did not want to go there. This parent’s child had asthma, which also deterred her from attending as the child was unable to be out in the cold. The PiP facilitators also acknowledged that lots of children were unwell in the winter months.

The timing and location of the PiP sessions were also discussed by several parents, particularly in the context of their children’s routines. Some parents talked about the afternoon sessions clashing with the time that their child would be having a nap whereas for other parents the sessions clashed with child-minder days or nursery; one mother said that she would ask if her daughter could leave nursery 20 minutes early to attend PiP sessions. Another parent (who had been recruited to the study through a children’s centre) had children attending a different school and so could not get to the sessions in time. The PiP facilitators also alluded to difficulties with the timing of the sessions for some of the parents, particularly mentioning children’s nap times and competing commitments of full-time employment and English lessons (for South Asian parents). They were able to identify the sessions that were less popular and in one school they had changed the timing of the afternoon sessions so that they did not clash with a particular activity running at the nursery: however, this did not improve attendance.

For the parents who attended regularly the timings and location worked as anticipated, linking with school drop-off and pick-up times. One of the facilitators described how one mum had ‘made it work’ by attending two sessions on one day each week to fit all three sessions in around her employment. One of the PiP facilitators in a wave 3 school, where attendance at the PiP sessions was better, commented that the timing of the sessions worked well. Another described how the nursery had been flexible in letting a boy come out of nursery to attend the PiP session.

The bad weather was discussed by all seven intervention arm parents in waves 1 and 2 and was offered as a reason why ‘other parents and children’ may not have attended. Interestingly, although none of these parents had attended regularly and four had not attended any sessions, they did not see the weather as the reason why ‘they’ themselves had not attended. The PiP facilitators also talked about the negative influence of the weather, suggesting that it was to be expected that parents did not return when they had been sat on the cold, wet ground outside in ‘freezing’ weather. A facilitator in wave 3 described the opposite problem of parents not bringing their children because it was too hot. A head teacher from a wave 1 school also alluded to the cold weather as a reason why parents in his school had not attended.

Finally, two parents said that they did not know about the PiP sessions and would have gone if they had known (the PiP facilitator at this school said that she had telephoned these parents on several occasions to encourage them to attend). In fact, it was clear from the interviews with all of the PiP facilitators that they had regularly texted or telephoned parents to try to encourage them to attend and they did not accept that parents were unaware of the PiP sessions.

The PiP facilitators offered some additional reasons why they thought that the parents had not attended the PiP sessions (which parents themselves did not mention). Some were related to the study, namely a lack of promotion of the study at the schools, the ‘burden’ of the measurement sessions and that these sessions might involve home visits, the embarrassment of being watched doing the PiP sessions in the playground and the challenge of bringing two children to the sessions. In one school there had been a staff change and the facilitator who took over the running of the sessions suggested that, because of wider timetabling issues, the original facilitator did not know the parents very well, which did not help with attendance. Another facilitator also commented on the importance of knowing the parents and that the ones who attended were the ones that she knew and the ones who had not come she did not know.

Finally, the other reasons related more to a general nervousness of parents and reluctance to take part in school activities. One facilitator described how the school relies on incentives (e.g. refreshments) to encourage parents to come to events. Related to this was the suggestion by two PiP facilitators that parents had attended only to receive the gift vouchers and that once they realised that these were only for the measurement sessions they had stopped coming. Both of the head teachers who were interviewed spoke about the children who would most benefit from attending being the ones whose families are the most difficult to engage in school activities.

Acceptability of the maintenance sessions (Box 11)

Following 10 weeks of PiP sessions led by a PiP facilitator, the next term’s sessions were to be led by parents, without supervision. PiP facilitators were asked in their interviews whether or not they thought that parents would continue to attend once they were no longer leading the sessions. Two PiP facilitators were not confident that the sessions would work without them being there because they believed that someone needed to take the lead. Another two facilitators were more positive because they thought that there were some parents who were regular attenders who were very capable of running the sessions, but they were less sure that parents with less confidence would continue to attend.

The likelihood of the PiP sessions running in the maintenance phase was also explored with parents interviewed in wave 3. Both parents, who had attended regularly, thought that the maintenance sessions could work because as a group of parents they now knew what to do. These were the same parents who the facilitator had identified as being capable of leading the sessions.

Playground and local environment

Table 25 displays the summary statistics of the playgrounds and local environments of the schools by trial arm and overall.

In total, 70% of schools had a concrete playground surface, 60% had fixed play equipment and 60% had a grassed area available. Playground markings were present in 90% of schools, 50% had a covered area in the playground and 60% were located in the nursery. The average proximity to a busy road was 246 (SD 289) metres, with a median proximity of 188 (minimum–maximum 0–975) metres. The average proximity to a local park with fixed play equipment was 374 (SD 321) metres, with a median proximity of 288 (minimum–maximum 50–950) metres. There were large differences between trial arms in the average proximity to a busy road and the average proximity to a local park with fixed play equipment, with the intervention schools being further away from busy roads and in closer proximity to a local park with fixed play equipment. However, the number of schools in the comparison is low.

Head teacher views on the capability and capacity of schools to deliver the intervention (Box 12)

When asked about the capability and capacity of the school to deliver the PiP intervention the wave 3 school head teacher explained that it had taken a little time to organise the timing of the sessions and that a building project that was happening at the time had reduced the space available. Both head teachers described how they had incorporated the work into their staff’s existing workloads. One head teacher was clear that he viewed this as a cost to the school because it took that person away from their normal work; furthermore, this member of staff was more senior than a nursery nurse. The other head teacher described how the school was well staffed with a good budget so this was less of a concern. For different reasons neither head teacher saw delivering the intervention to be sustainable in the long term. The first head teacher stated that the Born in Bradford programme would need to provide staff to continue to deliver the PiP sessions. The second head teacher explained that parents at her school do not ‘stick with’ any initiatives for very long and so the PiP sessions would need to recruit new families continually to run long term.

Preschoolers in the Playground facilitators’ experiences of delivering the PiP sessions (Boxes 15 and 16)

All of the PiP facilitators said that they had enjoyed running the PiP sessions, with those in the wave 1 and 2 schools suggesting that it would have been better had more parents attended as this would have created a different atmosphere.

The PiP facilitators were asked to describe how they had delivered the PiP sessions, reflecting on how they had used the suggested format (in the training manual) of warm-up, PiP-PoP, free play, PiP-PoP. What emerged clearly was that they had been flexible in their delivery, using their experience of working with children as well as getting to know the children to adapt the activities. Half of the PiP facilitators explained that they would decide when to initiate the free play and PiP-PoPs according to how engaged (or not) the children and their parents were. Specifically, they did not break up free play ‘for the sake of it’ if it was going well, and conversely the PiP-PoPs were used as a way of re-engaging the families when they appeared bored during free play.

A number of factors informed which PiP-PoPs the PiP facilitators chose to use. Most important was what would work for the children in terms of the number of children in the session, their ages (associated physical development and concentration span) and their mood, as well as what would promote learning. All of the PiP facilitators could name the PiP-PoPs that the children in their sessions most enjoyed and they commented that they used these the most often and adapted them when they did not work. The popular PiP-PoPs were the statue game, the spaceships and planets game, Mr Crocodile, chalks and water painting, balancing along the skipping rope, the hula hoop and throwing balls. Finally, two of the PiP facilitators explained that they had started off using the PiP-PoPs that they felt most comfortable with and which others talked about, taking into account how the parents were responding to the activities.

The PiP facilitators talked about using a wide range of equipment during the PiP sessions. Some of this was required equipment highlighted in the PiP manual, for example hoops, balls, beanbags and skipping ropes. Other pieces, including a ‘windy box’, a climbing frame, a trampoline and a sand pit, the PiP facilitators introduced themselves. As with the structure of the PiP sessions, it was clear that the PiP facilitators thought carefully about which equipment they used, introducing additional equipment when the children were no longer engaging with the equipment already set out for the session, and using the same equipment that the children were being given to take home and use so that the parents and children had some ideas of how to play at home.

As part of the PiP sessions the PiP facilitators were expected to guide discussions each week and to give out a handout. These handouts focused on matters such as giving praise and encouragement and why active play is important (see Chapter 1, Guided discussions and handouts). For many of the PiP facilitators it was evident that the handouts did not feature heavily in the sessions. Indeed, one facilitator said that she did not remember seeing the leaflets and another stated that it was hard to get a discussion about the handouts going within the sessions. Others described how they had thought about which handouts to distribute, deliberately staggering the distribution of the handouts to not overload parents with too much information and ensuring that parents received the handout about what to wear in different weathers when the weather was particularly hot.

The PiP facilitators described varied parental responses to the handouts, with some parents asking questions and others ‘pulling a face’ and putting the handout in their bag. Two PiP facilitators said that some of their parents were not very confident about physical activity and that the handouts did contain new information for those parents. Another facilitator thought that it was her ‘more capable’ parents who engaged more with the handouts. The ‘Active parents have active kids’ and the ‘What to wear for outdoor play’ handouts were those most frequently mentioned as being liked by parents.

Preschoolers in the Playground facilitators views’ on previous experience, training, the Preschoolers in the Playground manual and follow-up support (Box 17)

The PiP facilitators were asked what types of skills and experience someone would need to run the PiP sessions. All of them talked about the importance of having previous experience of working with children and their parents. It meant that they were confident in leading the sessions, could ‘think on their feet’ and could adapt the sessions when needed, as well as having a good rapport with children and parents. One facilitator said that it was important to enjoy working with children. Several also explained that it had helped that they already knew many of the families coming to the PiP sessions through working at the nursery and running baby yoga. The PiP facilitators did not think that they had needed to have previous experience of physical activity because that was provided in the training and manual. Indeed, without exception the PiP facilitators described the training session and the manual as helpful in providing them with useful information, ideas and confidence to deliver the PiP sessions in a flexible way. A few acknowledged that the brief training, which really ‘just involved working through the manual’, was sufficient because they were already experienced in working with children. They also valued the offer of follow-on support from the PiP research team, although no PiP facilitators mentioned accessing this.

Physical activity

The baseline physical activity measures are summarised in Table 26.

The average time spent in MVPA per day was 63.1 (SD 24.5) minutes, in light activity was 212.7 (SD 41.9) minutes and sedentary was 302.2 (SD 78.7) minutes. The average total time spent in physical activity per day was 275.8 (SD 57.8) minutes and 96% of children achieved a total of 180 minutes or more of physical activity per day. Randomisation achieved balance across most measures. There were slight imbalances in the minutes spent in MVPA, in light activity and sedentary per day in the as-analysed comparisons, with the intervention group having slightly higher values for MVPA and light activity and being more sedentary than the control group.

Table 27 shows the physical activity data for participants at 10 and 52 weeks.

The mean number of minutes spent in MVPA per day was lower in the intervention group than in the control group at both 10 weeks [61.8 (SD 25.2) minutes vs. 62.1 (SD 22.9) minutes] and 52 weeks [72.6 (SD 30.4) minutes vs. 74.3 (SD 25.5) minutes]. The mean number of minutes spent in light activity per day was higher in the intervention group than in the control group at both 10 weeks [212.9 (SD 39.4) minutes vs. 200.4 (SD 45.1) minutes] and 52 weeks [220.9 (SD 37.4) minutes vs. 219.1 (SD 43.1) minutes]. The mean number of minutes spent sedentary per day was higher in the intervention group than in the control group at both 10 weeks [318.0 (SD 102.6) minutes vs. 303.7 (SD 69.0) minutes] and 52 weeks [335.7 (SD 102.4) minutes vs. 302.5 (SD 65.6) minutes]. Hence the intervention group appeared to be undertaking less MVPA and more light activity and to be more sedentary than the control group. The total number of minutes spent in physical activity per day was higher in the intervention group than in the control group at 10 weeks [274.6 (SD 49.3) minutes vs. 262.6 (SD 61.1) minutes] but the two groups were similar at 52 weeks [293.5 (SD 52.0) minutes vs. 293.4 (SD 59.3) minutes]. A similar pattern was seen for the percentage of children undertaking 180 minutes of physical activity a day at 10 weeks (n = 44, 100% vs. n = 41, 95.3%) and at 52 weeks (n = 41, 97.6% vs. n = 44, 97.8%).

Trial arm and ethnic group differences

Figure 3 and Table 28 summarise the accelerometer data by ethnicity. Different results are seen by ethnic group. For white children, the number of minutes spent in MVPA per day was lower in the intervention group than in the control group at both 10 and 52 weeks; however, for South Asian children it was slightly higher in the intervention group than in the control group at 10 weeks and very similar between the two groups at 52 weeks. Although these differences by ethnicity are demonstrated in the raw means, these results must be interpreted with caution, as there appears to be an imbalance in MVPA scores at baseline between the intervention group and the control group for white children for those included in the analysis (i.e. baseline and 52-week data) and these summaries are based on data for 18 white children (nine intervention and nine control children; see Figure 3). For white children, the number of minutes spent in light activity per day was lower in the intervention group than in the control group at both 10 and 52 weeks; however, for South Asian children the number of minutes spent in light activity per day was higher in the intervention group than in the control group at both 10 and 52 weeks. No differences in the number of minutes spent sedentary per day were apparent between ethnic groups. For white children, the number of minutes of TPA per day was lower in the intervention group than in the control group at 10 and 52 weeks; however, the reverse was seen for South Asian children, with more minutes of TPA per day in the intervention group than in the control group at both 10 and 52 weeks. All white and South Asian children were undertaking at least 180 minutes of physical activity a day at 10 weeks but at 52 weeks different patterns were seen by ethnic group. For white children the control group had a higher number of children undertaking at least 180 minutes of TPA per day; however, for South Asian children the intervention group had a higher number meeting the criterion.

FIGURE 3.

Mean minutes of MVPA per day at baseline and week 52 by group and ethnicity (white and South Asian). The top graphs include all children at each time point who provided enough accelerometry data. The bottom graphs include just the children who provided enough accelerometry data at both baseline and 52 weeks’ follow-up.

TABLE 28 - Summary of accelerometer data across all time points by trial arm and ethnicity

Characteristic	10 weeks, mean (SD)			52 weeks, mean (SD)
	Intervention	Control	Total	Intervention	Control	Total
White, n	14	14	28	16	12	28
South Asian, n	28	27	55	24	32	56
Minutes spent in MVPA per day
White	57.2 (24.6)	65.0 (18.0)	61.1 (21.5)	74.6 (21.3)	81.1 (21.3)	77.4 (21.1)
South Asian	63.2 (26.2)	61.3 (25.3)	62.3 (25.5)	71.5 (35.3)	71.6 (27.1)	71.6 (30.6)
Minutes spent in light activity per day
White	211.5 (43.8)	212.5 (43.4)	212.0 (42.8)	219.7 (47.8)	229.6 (37.8)	224.0 (43.3)
South Asian	213.2 (39.3)	193.1 (45.1)	203.3 (43.0)	223.3 (28.9)	214.0 (44.9)	218.0 (38.8)
Minutes spent sedentary per day
White	337.7 (111.7)	313.4 (85.3)	325.5 (98.3)	334.2 (130.4)	294.5 (78.5)	317.2 (111.2)
South Asian	308.9 (102.1)	302.1 (50.5)	305.5 (80.3)	333.4 (86.5)	304.2 (61.9)	316.7 (74.2)
Minutes of TPA per day
White	268.8 (54.2)	277.5 (57.8)	273.1 (55.1)	294.3 (58.0)	310.7 (51.2)	301.3 (54.8)
South Asian	276.4 (49.1)	254.4 (62.0)	265.6 (56.4)	294.8 (46.2)	285.6 (61.8)	289.5 (55.4)
Percentage of children spending ≥ 180 minutes per day in physical activity, n/N (%)
White	14/14 (100.0)	14/14 (100.0)	28/28 (100.0)	15/16 (93.8)	12/12 (100.0)	27/28 (96.4)
South Asian	28/28 (100.0)	25/27 (92.6)	53/55 (96.4)	24/24 (100.0)	31/32 (96.9)	55/56 (98.2)

Anthropometry

Table 29 summarises the baseline anthropometry data and Table 30 shows the 10- and 52-week follow-up data.

The average height of children in the study was 92.8 (SD 7.1) cm, with an average body mass of 14.2 (SD 2.4) kg, abdominal circumference of 50.0 (SD 3.8) cm and arm circumference of 17.1 (SD 3.1) cm. The average BMI z-value was 0.7 (SD 1.1) and the number of children with a BMI z-value of ≥ 1.04 was 53 (32.3%). Randomisation achieved balance across most anthropometry measures and attrition did not impact on the balance achieved at baseline in the as-analysed sample across most measures. There was a slight imbalance in BMI z-value in the as-analysed comparison, with the intervention group having a slightly lower BMI z-value (and subsequently a lower percentage of overweight children) than the control group.

Table 30 shows that there was little difference in average body mass and height between the intervention group and the control group at 10 and 52 weeks. The average BMI z-value at 10 and 52 weeks was slightly lower in the intervention group than in the control group. At 10 weeks the average BMI z-value in the intervention group was 0.5 (SD 1.2) and in the control group was 0.7 (SD 1.2) and at 52 weeks the average BMI z-value in the intervention group was 0.2 (SD 1.2) and in the control group was 0.3 (SD 1.2). The average waist and upper arm circumference at 10 and 52 weeks were slightly lower in the intervention group than in the control group. At 10 weeks the average waist circumference in the intervention group was 49.6 (SD 3.7) cm and in the control group was 50.4 (SD 4.8) cm and at 52 weeks the average waist circumference in the intervention group was 51.1 (SD 4.1) cm and in the control group was 52.2 (SD 4.8) cm. Similarly, at 10 weeks upper arm circumference in the intervention group was 17.0 (SD 1.4) cm and in the control group was 17.8 (SD 5.5) cm and at 52 weeks upper arm circumference in the intervention group was 17.5 (SD 2.1) cm and in the control group was 17.6 (SD 1.5) cm.

Quality of life and general self-efficacy

Table 31 shows parents’ well-being and general self-efficacy at baseline.

The average total objective quality-of-life score was 74.2 (SD 7.6) and the average general self-efficacy score was 29.7 (SD 5.6). Randomisation achieved balance across both measures and attrition did not impact on the balance achieved at baseline in the as-analysed sample across both measures.

Table 32 shows the individual domains of well-being at 10 and 52 weeks’ follow-up.

At 10 weeks the mean overall total score for objective well-being was lower in the intervention group than in the control group [intervention 72.7 (SD 9.2); control 74.0 (SD 6.7)] and a similar pattern was present at 52 weeks [intervention 72.4 (SD 7.5); control 74.5 (SD 7.1)]. However, there was little change in scores between the two time points.

Analysis of the primary outcome (mean daily moderate to vigorous physical activity at 52 weeks)

The analysis for the primary outcome measure included 64 children (39%) who had both baseline and 52-week outcome data. At baseline the average time spent in MVPA per day was 63.1 (SD 24.5) minutes. The number of minutes spent in MVPA per day was lower in the intervention arm than in the control group at 10 weeks [61.8 (SD 25.2) vs. 62.1 (SD 22.9) minutes].

At 52 weeks the number of minutes spent in MVPA per day was very similar across trial arms (mean difference 0.39, 95% CI –16.26 to 17.04 minutes), although the results are not statistically significant (p = 0.96). The ICC for the primary analysis was estimated to be zero. On inspection of the data it appeared that there was a plausible but outlying child who had a low baseline MVPA level but the highest MVPA level at 52 weeks and so we undertook a sensitivity analysis to explore the impact on the results of excluding this individual from the analysis. In the sensitivity analysis the average number of minutes spent in MVPA was higher in the control group than in the intervention group (mean difference –3.19, 95% CI –18.06 to 11.68), although the results are not statistically significant (p = 0.63). The ICC for the primary analysis excluding the outlying child was estimated to be 0.02.

Sample size calculation

A full RCT would require a sample size of 600 children at 38 schools (16 children per school). This is based on a previous study by Specker et al. 101 and data from the pilot trial. The study by Specker et al. 101 found higher average daily accelerometer counts in gross than in fine motor groups at follow-up [mean (SD) 27.7 (8.8) vs. 31.0 (8.4), p = 0.03]; this translates into an approximate effect size of 0.38. To detect an effect size of 0.3, assuming 80% power, we require 352 participants (176 per group). Using an ICC of 0.02 and an average cluster size of 15 from our pilot data we estimate that there will be a design effect of approximately 1.28 [using 1 + (n – 1) * ICC]. On this basis, to detect 0.3 of a standardised difference between the groups would require enrolling 31 schools and 452 children. However, from our pilot data the pre-/post-test correlation between MVPA at baseline and MVPA at 52 weeks was 0.6. Taking this into account suggests that 20 schools and 300 children would be sufficient. Allowing for a potential loss to follow-up of 50% for the primary outcome (based on experience from the pilot trial), the final sample size needed is 600 children and thus 38 schools (16 children per school).

Costs of health care

At baseline and at each follow-up, parents were asked to recall their child’s health and social care service use during the past 3 months. Table 35 demonstrates the mean cost of each type of health-care service at each time point. The accrued costs for each resource used have been calculated based on the number of observations that have available data for any cost item and hence n varies by group and time point.

Figure 4 summarises the mean cost per participant by randomised arm for participants for whom complete cost data were available at both week 10 and week 52 (n = 58 for the intervention group and n = 61 for the control group). Costs were extrapolated from week 11 to week 39 based on mean resource use in the other two follow-up periods as no data were collected in the trial for this period. Participants in the intervention arm tended to generate higher costs over the whole trial period but there was not a statistically significant difference in total cost per participant between trial arms (p = 0.23).

FIGURE 4.

Total resource use costs over weeks 0–10, weeks 11–39 and weeks 40–52. Intervention costs have been excluded.

Health-related quality of life

European Quality of Life-5 Dimensions questionnaire

Parents’ HRQoL at each data collection point was measured using the EQ-5D questionnaire. Table 36 shows the proportion of participants in each health state of the EQ-5D for whom HRQoL data were available (126 parents at baseline, 124 parents at 10 weeks and 125 parents at 52 weeks).

TABLE 36 - Percentage of participants in each EQ-5D dimension by trial arm at baseline and follow-up

1, no problems with; 2, some problems with; 3, a lot of problems with.

Item	Level of severity	% of participants in health state
		Intervention			Control
		Baseline (n = 70)	10 weeks (n = 64)	52 weeks (n = 61)	Baseline (n = 56)	10 weeks (n = 60)	52 weeks (n = 64)
Mobility	1	98.57	93.75	95.08	96.43	88.24	92.19
	2	1.43	6.25	4.92	3.57	11.76	7.81
	3	0.0	0.0	0.0	0.0	0.0	0.0
Self-care	1	100.00	100.00	95.08	98.21	96.0	96.88
	2	0.0	0.0	4.92	1.79	4.0	1.56
	3	0.0	0.0	0.0	0.0	0.0	1.56
Usual activities	1	92.86	90.63	93.44	96.43	90.20	92.19
	2	7.14	9.38	6.56	3.57	9.80	7.81
	3	0.0	0.0	0.0	0.0	0.0	0.0
Pain/discomfort	1	82.86	78.13	90.16	78.57	76.47	70.31
	2	11.43	14.06	3.28	17.86	11.76	23.44
	3	5.71	7.81	6.56	3.57	11.76	6.25
Anxiety/depression	1	87.14	85.94	93.44	85.71	74.51	76.56
	2	10.00	14.06	6.56	14.29	25.49	20.31
	3	2.86	0.0	0.0	0.0	0.0	3.13

Figure 5 shows the mean EQ-5D index at baseline, 10 weeks and 52 weeks by randomised arm (without MI). The estimates of the means were derived from the observations with available EQ-5D data at each time point (Table 37). HRQoL tended to decrease in both groups at 10 weeks followed by a slight increase at 52 weeks. Despite the similar pattern, overall the PiP intervention appeared to improve parents’ HRQoL marginally, leading to an overall increase in score of 0.023, whereas in the control group the EQ-5D index of parents decreased by 0.063. A different pattern was observed in the complete-case analysis, that is, when only participants with complete cost, EQ-5D and PedsQL data were taken into account in the calculation of the mean EQ-5D scores (Table 38). In this case parents’ HRQoL showed a steady improvement from baseline to week 52 in the intervention group compared with a steady decline in the control group.

FIGURE 5.

Mean EQ-5D index of HRQoL at baseline and follow-up by trial arm. The scores at each time point have been derived from all participants with available data at that time point.

TABLE 37 - Mean EQ-5D index of HRQoL at baseline and follow-up by trial arm: available cases

Difference in differences for intervention arm vs. control arm = 0.09.

Time point	Intervention		Control
	n	Mean (SD)	n	Mean (SD)
Baseline	70	0.90 (0.22)	56	0.92 (0.18)
10 weeks	64	0.89 (0.23)	50	0.85 (0.28)
52 weeks	61	0.93 (0.23)	64	0.86 (0.25)
Difference in means from baseline to 52 weeks		0.02		–0.06

TABLE 38 - Mean EQ-5D score at baseline and follow-up by trial arm: complete cases

Difference in differences for intervention arm vs. control arm = 0.17.

Time point	Intervention		Control
	n	Mean (SD)	n	Mean (SD)
Baseline	38	0.92 (0.18)	34	0.94 (0.12)
10 weeks	38	0.92 (0.19)	34	0.84 (0.28)
52 weeks	38	0.97 (0.16)	34	0.83 (0.27)
Difference in means from baseline to 52 weeks		0.05		–0.2

Pediatric Quality of Life Inventory

Parents were asked to complete the PedsQL questionnaire at baseline, 10 weeks and 52 weeks to measure the HRQoL of the infants and toddlers. In both groups children’s quality of life showed a deterioration from baseline to 10 weeks, which was followed by an improvement at 52 weeks (Figure 6). Infants and toddlers in the control arm had a higher mean score at 52 weeks than those in the intervention arm. However, as Table 39 shows, that score was lower than the baseline score and reflected a decrease in HRQoL of 0.2 units from baseline to 52 weeks, whereas children in the intervention arm demonstrated an overall improvement in HRQoL of 1.5 units over the same period.

FIGURE 6.

Mean PedsQL index of HRQoL at baseline and follow-up by trial arm. The scores at each time point have been derived from all participants with available data at that point.

TABLE 39 - Total PedsQL score at baseline and follow-up by trial arm: available cases

Difference in differences for intervention arm vs. control arm = 1.70.

Time point	Intervention		Control
	n	Mean (SD)	n	Mean (SD)
Baseline	83	89.22 (7.37)	81	91.76 (6.31)
10 weeks	62	87.96 (9.02)	67	87.28 (10.68)
52 weeks	63	90.72 (7.62)	60	91.56 (7.06)
Difference in means from baseline to 52 weeks		1.50		–0.20

Imputation of missing data

Of the 164 participants in the trial, 89 provided EQ-5D and PedsQL data at both 10 weeks and 52 weeks; 28 had missing data at 10 weeks only, 22 had missing data at 52 weeks only and 25 had missing data at both time points. MI was used to impute missing data (costs, EQ-5D scores and PedsQL scores) following the predictive mean matching approach. The imputation produced five imputed data sets, based on which the means of the outcomes were estimated.

The mean estimates of the EQ-5D index and the PedsQL index derived from the entire trial population were similar to those of the available case analysis, showing a decrease in HRQoL in both groups at 10 weeks and subsequent improvement at 52 weeks (Tables 40 and 41, respectively). Overall, however, the health gains over the entire trial period were higher for the intervention group than for the control group (change in mean score from baseline to 52 weeks for the intervention group vs. the control group: 0.0343 vs. –0.0568 for EQ-5D and 1.97 vs. 0.18 for PedsQL).

Cost-effectiveness analysis

Total population

Table 42 shows the mean incremental QALYs and mean incremental costs compared with usual practice. MI was performed to deal with missing data and a SUR model was used to estimate the mean total costs and QALYs from the imputed data sets. After adjusting for baseline differences in EQ-5D score, the PiP intervention appeared to improve parents’ HRQoL by 0.058 QALYs at an increased cost of £1137 per participant compared with the control arm. Estimates of both incremental costs and effects are associated with considerable uncertainty with all CIs crossing zero.

TABLE 42 - Results of the cost-effectiveness analysis after adjusting for differences in baseline EQ-5D score

With imputation of missing data for 17 participants who had only baseline data and 75 participants who were followed up at only one time point.

No imputation. Complete cases consist of participants with complete data for costs, QALYs and the PedsQL.

	Mean incremental QALYs (compared with control)	95% CI	Mean incremental cost (compared with control) (£)	95% CI (£)	ICER (£ per QALY)
MI, n = 164a
PiP intervention	0.06	–0.01 to 0.13	1136.62	–940.08 to 3213.33	19,588
Complete cases, n = 72b
PiP intervention	0.11	0.05 to 0.18	2686.24	–1748.32 to 7120.79	23,895

Overall, in the MI analysis the PiP intervention appeared to be borderline cost-effective with an ICER of £19,588 per QALY gained. However, this estimate is surrounded by a considerable amount of uncertainty. The probability that the PiP intervention is cost-effective is 0.507 at a cost-per-QALY threshold of £20,000, 0.656 at a threshold of £30,000 and 0.807 at a threshold of £50,000. Figure 7 illustrates the uncertainty associated with the estimate of the ICER, that is, the probability of the PiP intervention being cost-effective at various thresholds.

FIGURE 7.

Cost-effectiveness acceptability curve controlling for baseline utility: MI analysis.

Sensitivity analysis

A complete-case analysis was conducted as a sensitivity analysis. Complete cases (n = 72) consisted of participants who had no missing data for any cost, EQ-5D or PedsQL item at baseline, 10 weeks or 52 weeks. This analysis resulted in slightly higher incremental costs and QALYs for the PiP intervention (see Table 42) than with MI, which produced an ICER of £23,895 per QALY gained. The results of the complete-case analysis were not statistically significant in terms of incremental costs, with wide CIs that crossed zero.

Figure 8 captures the uncertainty associated with the estimates.

FIGURE 8.

Cost-effectiveness acceptability curve controlling for baseline utility: complete-case analysis.

Short-term links

Long-term links

Childhood obesity has been reported to pertain to a number of health conditions in adult life. Extrapolating the findings of the systematic review by Timmons et al. 14 it could be hypothesised that increased physical activity in children may positively affect their long-term HRQoL by reducing obesity levels and, in turn, the risk of future health disorders. The exploration of such links between health outcomes in children and their potential consequences in adult life was the main focus of the literature review.

Of the body of evidence included in the review, in all nine studies102–110 that explored the links between short-term outcomes and long-term outcomes obesity was the exposure measure and eight102,103,105–108,110,118 out of the nine studies reported evidence of an association with cardiometabolic morbidity in adulthood. Additional health conditions explored included asthma105,108 and various types of cancer. 105,108,109

Attendance to the initiation phase

Overall, only 65% of children attended one or more sessions during the initiation phase. The number of children attending at least one session was lowest during the winter initiation phase (wave 2) at 40% but was double this during the summer initiation phase (wave 3) at > 80%. In the autumn initiation phase (wave 1) those who ever attended came to very few sessions, whereas in the winter and summer initiation phases those who ever attended had a mean attendance of 10 sessions, equating to one session per week, with some children attending between two and three sessions per week.

Participants in the PiP intervention group were encouraged to attend three 30-minute sessions per week (equating to 90 minutes of exposure to the intervention per week). This is greater than in other similar preschool physical activity interventions: the Belgium ToyBox intervention54 and two physical activity interventions (one based in preschools156 and the other a family-focused group-based intervention57) all requested that the intervention be delivered in one 1-hour session per week. These studies reported small to moderate effects of the intervention on MVPA (up to 7.2 additional minutes of physical activity per day). Given this emerging evidence, we propose encouraging families to attend at least two and ideally three PiP sessions per week, which equates to an exposure of between 60 and 90 minutes per week. Unfortunately, these and other studies of group-based parent and child lifestyle interventions have not reported their attendance rates and therefore our data cannot be compared.

The interviews with parents showed that one clear reason for poor attendance during the initiation phase in waves 1 and 2 was the poor weather, with all seven parents interviewed mentioning this. Indeed, during the autumn of 2012 there were several extended periods of heavy rainfall and between January and April 2013 there were several weeks of significant snowfall and very low temperatures. Additionally, the PiP facilitators acknowledged that many children were unwell during the winter months and commented that running the intervention during the summer months or in an indoor environment would be better. The possibility of accommodating PiP sessions in an indoor setting within the school as an alternative during the winter months was explored in conversations with the schools (outside of the qualitative interviews); however, it was not feasible in the majority of schools because of restricted indoor space.

Another reason given for poor attendance was the sessions not fitting in with the children’s daily routines. However, the parents who regularly attended stated that the timings and location worked well for them. The timings of the sessions were agreed on by the schools following randomisation and therefore one way to overcome this would be to work with the schools to set the session times before recruiting families. Parents could then be made aware of the session timings before committing to the intervention. This mirrors what already occurs in practice for preschool-organised, group activity sessions and this would also be the case if the intervention was rolled out at scale.

The PiP facilitators suggested that attendance could have been improved by the school promoting the intervention more and by reducing the burden of the measurement sessions (which they thought discouraged parents from participating in the entire study including the PiP sessions). They reflected that it was important that the facilitators were familiar to the parents. This stresses the importance of using existing school staff to deliver the PiP sessions rather than bringing in new ‘specialist’ staff. One method often used by schools to increase familiarity is for outreach workers to visit families at home (this is sometimes part of the role of a parental involvement worker); however, this is resource intensive. Instead, a welcome ‘fun day’ session, with refreshments for parents, could be used as a way for the PiP facilitators to meet new parents before the start of the intervention. In addition, for those parents whose children use the school nursery, the facilitators could introduce themselves during drop-off/pick-up times.

Acceptability of the initiation phase

Parents and facilitators commented that the PiP sessions were varied and fun and that the children had very much enjoyed taking part. One facilitator reflected that the sessions became more enjoyable over time as everyone’s confidence improved and the children settled into the routine of the sessions. Only one parent had a negative experience of the one session he attended, reporting that it was the first session and that the facilitator did not know what she was doing.

Parents, PiP facilitators and head teachers all commented that the intervention was beneficial to children, parents and the school. Children and parents had learned new active games to play together and children had developed their confidence, learning colours, sharing and making friends. Parents reported that they had benefited from taking part; they had enjoyed doing something with their child, being more active themselves and making new friends. Both the structured PiP-PoP games and the free-play sections of the intervention were well accepted and enjoyed by children, parents and facilitators. Different children and different groups had their favourite PiP-PoP activities, which they enjoyed re-enacting during free play. Facilitators also commented on how they bought out different equipment during free play to sustain the active play.

The PiP facilitators and head teachers reflected on how parents had grown in confidence in playing with their children over the duration of the initiation phase, learning new activities and engaging with their children in a way that they had not done before, modelling good language and meeting other parents whose children would be attending the same nursery. Another benefit of the intervention reported by a facilitator at one of the wave 3 schools was that it was a good way to support transition into nursery, with parents and children becoming familiar with the staff in the nursery where many would be starting in the next academic year. The head teacher of the other wave 3 school commented that there had been better attendance at nursery following the initiation phase of the PiP study and that parents had been more involved in their child’s learning. The other head teacher interviewed commented that taking part in the PiP intervention had made the nursery staff more aware of the importance of parents and children playing outside together and planned to build this into the school’s nursery programme using the PiP manual the following year.

The PiP facilitators commented that the PiP manual and the training for the PiP intervention were easy to understand and helpful in providing them with useful information, ideas and confidence to deliver the intervention in a flexible way. They did not feel that they required any previous experience in delivering physical activity sessions but did emphasise the importance of previous experience of working with both children and parents.

There was a mixed response to the guided discussions and handouts in the intervention. One facilitator could not remember seeing the leaflets and another found it hard to guide the discussion within the short time of the sessions. This facilitator commented that she gave out the leaflets but received little feedback. Two other facilitators described the leaflets and discussions as useful but did not want to overload the parents with information. One way to improve on this could be to provide the information in a different format, such as a short film of a group of parents discussing each topic, which parents could view online and then discuss with facilitators at the PiP sessions.

Changes made to the initiation phase of the intervention and their acceptability

During waves 2 and 3 of the intervention a new behaviour change technique was introduced: families received a new piece of play equipment each week when they attended at least one session (see Chapter 1, Changes to the intervention, for a list of equipment). The wave 3 parents reported using the equipment at home and the PiP facilitator confirmed that parents had told her that they had been using the equipment at home.

A further change in wave 3 was the addition of telephone support from facilitators to parents who had, for unknown reasons, not attended three consecutive sessions (one week of the intervention). PiP facilitators reported telephoning and texting those who had not attended sessions but often they received no reply and these families did not attend. This was also the difficulty we experienced when contacting some of the parents for measurement sessions and may reflect the frequency with which some parents change their telephone numbers.

A change to the delivery of the intervention was also made: schools were given the option to reduce the number of sessions offered to four per week after the first 3 weeks of the intervention if there were two sessions that were poorly attended. This change was made before wave 3 of the intervention, in which attendance was much better, and neither of the wave 3 schools implemented this change.

Maintenance phase attendance and acceptability

During the maintenance phase only 16% of children attended any sessions. Attendance was low across all waves, ranging from 0% attendance in wave 1 to 29% of children attending any session in wave 3 (school 5). Of those children who attended at least one session, the average number of sessions attended was 5.2 (SD 3.4 sessions), with a median of 4.5 sessions (minimum–maximum 1–11 sessions). Before the maintenance phase facilitators were asked whether or not they thought that parents would attend. Two thought that the maintenance phase would not work because parents would need someone to lead them and two were more confident because they felt that parents who were regular attendees were capable of running the sessions. Subsequent conversations with facilitators during the maintenance phase about poor attendance revealed that changes in family routines and circumstances, particularly children starting nursery or preschool or changing nursery days or times but also mothers returning to work, advancing pregnancies, poor weather, lack of visibility of the intervention without a facilitator and school building work, hampered attendance during the maintenance phase. The head teacher at one school indicated that, with the maintenance phase, the intervention was too long and that parents attending the school do not engage in long-term programmes. A recent meta-analysis of the effectiveness of physical activity interventions for preschoolers highlighted the importance of having professionals lead intervention sessions. 155 This study found that interventions led by teachers had a larger effect on MVPA levels than interventions led by parents. One previous study maintained a preschool physical activity intervention over 3 years using newsletters and a radio programme to give parents information about physical activity. 163 This strategy could be used in the PiP intervention to sustain behaviour change. Families would receive a newsletter/video link during the summer holidays (after the 10-week initiation phase) including reminders about physical activity guidelines, PiP games, physical activities that they can do as a family and what is available to them locally. This newsletter/video would be repeated each school term for the subsequent year (autumn, spring and summer terms) and would include details of local physical activities available to families. These behaviour change techniques of ‘prompts and cues’ and ‘information regarding the behaviour’ fit within social cognitive theory, targeting behavioural and personal factors.

Ethnic group differences

Attendance at the PiP intervention sessions was higher among South Asian children than among white children. Among those children who attended at least one session, the average number of sessions attended for white children was 10.7 (SD 12.1 sessions), with a median of three sessions attended (minimum–maximum 1–33 sessions), and for South Asian children was 11.3 (SD 11.8 sessions), with a median of 4.5 sessions attended (minimum–maximum 1–35 sessions). The qualitative analysis found no clear differences between ethnic groups regarding the acceptability of the intervention. This pattern of disengagement has been seen throughout the trial, with fewer white participants attending measurement sessions and providing outcome data. These findings may be the result of social differences, specifically social capital. In the PiP pilot trial we found that the ethnic minority South Asian families were more willing to engage in health research and health behaviours and this may be because they have better social networks, are more involved in the community, are less isolated and are more familiar with group activities than poor white parents. Being part of a minority ethnic group and sharing culture and social support networks may be protective for health, even in economically deprived contexts. This is illustrated by research in the USA on the ‘Latina paradox’. 164 Latina women had comparable birth outcomes with white women of better SES and McGlade et al. 164 argue that this was because systems of family, friends, community members and lay health workers provide support.

Proposed changes to the intervention

To improve the intervention the following changes are proposed:

List of recommendations for a full trial and for other researchers

1. Few trials of preschool physical activity interventions conduct qualitative interviews, assess fidelity, report attendance or conduct a health economic evaluation. Including these in research studies is a recommendation for other researchers when designing evaluations of interventions in the future.

2. Build in sufficient time and resources to research plans to ensure successful follow-up of, and extra support for, hard-to-reach participants.

3. With very young children consider the impact of multiple measures over time on research fatigue and minimise measurements when it is possible to do so.

4. The ComQol-A5 questionnaire was not acceptable to parents because of the sensitive questions about mental health and income. It would not be used in a full trial of the PiP intervention.

5. Ensure that study materials are appealing to participants by using colours and pictures and that they can be easily understood. Consider supplementing the participant information sheet with an online video to describe the study procedures visually, including randomisation.

6. Ask participants to provide as many different contact details as possible.

7. Incentives for research appear to be important in deprived communities; the type of incentive that is appropriate and useful to the participant population must be considered.

8. To recruit schools into health research, consider the priorities of education professionals and describe the potential benefits that fit with these priorities, for example parents and children becoming familiar with the school.

9. Consider the motivation for families to take part in the study (in the PiP trial this was mainly social and learning) and describe these potential benefits in the study information. Be aware that these may not match up with the intended primary outcome of the study (in the PiP trial increasing physical activity).

10. Ensure that participants can identify with the research staff who are conducting recruitment, for example the researchers can translate study materials into relevant languages for participants and are not ‘too academic’.