Study design and methods

This chapter summarises the study protocol for this pilot RCT as originally funded. Some of the material has already appeared in a published format72 and is reproduced here. Reproduced from Clemes et al.72 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Study design

This was a school-based, pilot, two-armed, cluster RCT with economic and process evaluations. Individuals (Year 5 children, aged 9–10 years) were the unit of analysis and schools (clusters) were randomly assigned to one of two conditions: (1) six manually adjustable sit–stand desks incorporated into the classroom environment (intervention condition) or (2) current practice (control condition). Baseline measurements preceded randomisation and the sit–stand desks were installed into the intervention classrooms following this. An identical set of outcome measurements were taken from all participants approximately 7 months after the baseline measures, at the end of Year 5. Figure 1 shows the schedule of enrolment, intervention and outcome measurements.

FIGURE 1 Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) diagram illustrating the design and time scales of the pilot Stand Out in Class RCT. BMI, body mass index. Arrows indicate start and end points of intervention and control period. Reproduced from Clemes et al.72 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. Includes minor additions and formatting changes to the original. (Figure not reproduced here; see the original article.)

Ethics approval

Ethics approval for this study was obtained from Loughborough University’s Ethical Advisory Committee (reference R16-P027). All members of the measurement team were required to have a current enhanced disclosure and barring service check and all employees of the Bradford Teaching Hospitals NHS Foundation Trust who undertook data collection completed online training in safeguarding children and information governance.

Study setting

The study was conducted in primary schools in Bradford, UK. Bradford was chosen as the study location given its ethnic composition (predominantly South Asian and white British) and high levels of deprivation, health inequalities and childhood morbidity.73 Half of all babies born in Bradford are of South Asian origin and 60% of babies are born into the poorest 20% of the population,73 with elevated risk factors for chronic diseases seen in South Asian children in comparison with white children.20,22 Furthermore, in 2016–17, 38% of Year 6 children in Bradford were overweight or obese according to the National Child Measurement Programme.74 The setting of this study, therefore, is fundamental in addressing the important issue of health inequalities, in that the intervention will be accessible to all children. In addition, the location enables us to pilot the intervention under challenging circumstances, meaning that, if it proves acceptable, it is likely to be transferable to most schools. At the planning stages of the trial, PPI with local teachers, parents and children demonstrated widespread support for the study, with all of those informally interviewed stating that they would be keen to try sit–stand desks in their classrooms (teachers and children).

Sample size

A recruitment target of eight primary schools, each with at least 15 child participants (approximately 50% of a typical class), was set, giving a minimum total sample of 120 participants. This matches the minimum of four clusters per arm recommended for a definitive cluster RCT75 and exceeds the target minimum sample size recommended for pilot trials.76 Consequently, the sample should be sufficient to provide clear estimates of recruitment and follow-up for a definitive trial.

School and participant recruitment and eligibility criteria

Government-funded primary schools located in the city of Bradford were invited to take part in the trial. Private and designated special schools and schools with < 25 pupils in Year 5 were not eligible to participate. Schools running a sitting time reduction programme, or with a unique characteristic that precluded comparisons with another school, were also not eligible for participation. The study was publicised to primary schools through existing local networks (e.g. Bradford Primary Improvement Partnership and Bradford’s Public Forum for Education). We aimed to recruit four schools with predominantly South Asian pupils (> 50%) and four with predominantly white British pupils (> 50%). The ethnic composition of the pupils was determined using local school census data.

The following three-stage process was adopted, informed by our earlier feasibility work and PPI with teachers and head teachers, for contacting schools about the study: (1) head teachers/senior teachers were sent an e-mail detailing the study, with the e-mail including a copy of the information sheet for schools; (2) 2 days after sending the e-mail, the school was telephoned and the reception team was asked whether or not the e-mail had been received by the head teacher/senior teachers; and (3) a follow-up telephone call was made to the school to arrange a time for the research team to meet either with the head teacher or with a senior teacher, either over the telephone or in person, to discuss the study in more detail, or to confirm that the school was not interested in participating in the study. Schools interested in participating were sent/given further information about the study, including an information sheet for teachers, and during the follow-up conversations the head teacher, or a nominated senior teacher, was requested to provide written informed consent to participate on behalf of the school. During recruitment, schools were requested to identify a designated lead teacher and were informed that they may be randomised to a current practice control condition in which they would be asked to maintain their usual classroom practice.

Following school recruitment, a member of the research team contacted the designated lead teacher directly via e-mail. The school was then provided with an invitation pack for the parents or guardians of eligible children (9- to 10-year-olds within the nominated Year 5 class of participating schools). The invitation pack contained a detailed information sheet for parents/guardians about the study, an opt-in consent form for the parent or guardian to complete and return if they were happy for their child to participate in the evaluation and an information sheet for children. The packs were given to eligible Year 5 children, who were asked to give them to their parents or guardians. In consultation with the participating Year 5 teachers within the schools, it was agreed that members of the research team, and the teacher, would be present at the school gates during school drop-off and collection times over a 4-day period (after children were given their packs to take home) to speak to parents and guardians about the study, should they have any questions. Completed consent forms were returned by pupils to their teacher, who informed the research team of the children who were to be involved in the evaluation measures.

At the beginning of the baseline measurement session, all methods were fully explained to the participants by a team member who was suitably qualified and experienced, and who was authorised to do so by the principal investigator (PI). Children were then asked to provide verbal and written assent that they were happy to participate, and this was requested again at the start of the follow-up measurement sessions. Children without parental consent or those who did not give their assent to participate in the evaluation were excluded from the evaluation measures, described in Outcome measurements, although they were still able to use the sit–stand desks in their classrooms. Any children in the intervention schools with known contraindications (e.g. those with a musculoskeletal injury, wheelchair users) that would preclude periods of standing were invited to participate in the evaluation measures and encouraged to use the sit–stand desks in a seated posture for inclusivity. These individuals were, however, excluded from the analyses.

Changes from the original protocol/grant application

The original grant application proposed that the intervention would run throughout the entire academic year (≈10 months) while the participants were in Year 5. The intended start date of the grant was originally 1 April 2016 and the desk installation was planned for September 2016. However, owing to delays at the project approval stage, the study did not officially commence until 1 November 2016 and the sit–stand desks were installed in the intervention schools between 27 February 2017 and 3 March 2017 (due to further delays caused by desk availability from the manufacturer). The intervention was therefore shorter (4.5 months) than originally proposed. As one of the project goals was to build on the previous evidence and assess the longer-term acceptability of using sit–stand desks in primary school classrooms, the study team discussed with the Trial Steering Committee (TSC) the possibility of extending the duration of the intervention until December 2017, with the desks remaining with the children as they entered Year 6. The TSC, however, recommended that the intervention be completed as originally planned in July 2017, and this recommendation was supported by expert advisors within the National Institute for Health Research (NIHR). It was felt that extending the trial into Year 6, particularly in terms of the scheduling of a further set of follow-up measures, would place added pressure on the new Year 6 teachers and schools at a critical time when the children would be preparing for and taking their Standard Assessment Tests (SATs). It was also felt that, as a pilot study, the team would still have an adequate number of data to assess the acceptability of the trial following 4.5 months.

The Stand Out in Class intervention

Six sit–stand desks were placed in one Year 5 classroom (replacing three standard desks sitting six children) in each intervention school for two school terms. The research team supported teachers in the development of a classroom rotation plan to ensure that all children in their class were exposed to the sit–stand desks for at least 1 hour per day, on average, across the week. Stools or chairs remained in the classroom and children were free to choose whether they sat or stood when using the sit–stand desks.

Teachers and pupils in the intervention classrooms received training on sit–stand desk use by the research team and teachers also received a professional development manual (see Report Supplementary Material 1) containing information on the health benefits of reducing prolonged sitting and on correct posture when standing at the desks. The manual and training focused on encouraging correct adoption of the intervention, targeting key barriers to and facilitators of sit–stand desk use. These were identified from our previous work,29,77 from the COM-B, within the BCW,78 and from the theoretical domains framework79 (e.g. self-efficacy, motivation and knowledge). Standardised behaviour change techniques (e.g. goal-setting, instruction)80 were also used during the training with teachers and pupils. A summary of the intervention components is shown in Table 1 and a logic model for the Stand Out in Class intervention, applicable for a definitive trial, is presented in Figure 2. Appendix 1, Table 23, details potential intervention barriers, solutions and hypothesised mediating processes informed by the above theoretical frameworks.

FIGURE 2 A logic model for the Stand Out in Class intervention, which is applicable for a definitive trial. Adapted with permission from Clemes et al.81 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. (Figure not reproduced here; see the original article.)

The usual practice control arm

To compare the effects of the intervention against usual practice (i.e. the provision of standard classroom desks), schools assigned to the control arm were requested to continue with their usual practice and lesson delivery, and no environmental changes were made to their classrooms. The Year 5 participants in the control schools were asked to complete the same study measurements as those in the intervention schools, at the same time points. On completion of the study, control schools were offered a report summarising the collected data of their pupils.

Allocation to treatment groups

To assess the acceptability of the intervention and proposed outcome measures for use in a definitive trial across an ethnically diverse sample, recruited schools were stratified based on the ethnic composition of their pupils (see School and participant recruitment and eligibility criteria). Following the completion of baseline measurements, schools within each stratum were randomised by an independent statistician at the Leicester Clinical Trials Unit into the two trial arms [intervention and control, with an allocation ratio of 1 : 1 using a randomisation list in SAS^® software (SAS Institute Inc., Cary, NC, USA). [SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ^® indicates USA registration.] Two schools with predominantly South Asian pupils and two schools with predominantly white British pupils were randomised into the intervention and control arms (four schools in each trial arm). This stratification enabled the examination of whether or not there were differences between South Asian and white British pupils in terms of recruitment and retention, adherence to the outcome measures and preliminary effects of the intervention on outcome measures (objective 4). The statistician performing the analyses was blinded to the schools’ allocation to the study arms, as were the community researchers undertaking the outcome measurements.

Outcome measurements

Trial feasibility-related outcomes

The primary aim of this pilot study was to establish school and participant recruitment and retention rates, the acceptability of the intervention and proposed outcome measures, intervention and measurement fidelity, and the availability and completeness of economic data for an estimation of potential cost-effectiveness, to inform the development of a definitive trial. Study uptake was monitored by recording the number of schools and pupils approached, and the number agreeing to participate (objective 1). Withdrawal rates of schools and children (objective 2) and completion rates for outcome measures were recorded (objective 3), and these feasibility outcomes were compared between ethnic groups (objective 4).

Process evaluation

A series of interviews and focus groups with teachers, children and parents were conducted following the completion of the baseline measures and randomisation to explore the acceptability of the trial procedures, including randomisation (objective 5), and the acceptability of the measurement instruments (objective 6). The acceptability of the intervention (objective 7) and children’s, teachers’ and parents’ perceptions and experiences of the intervention and outcome measures were determined during the intervention through a further set of interviews (n = 4 with teachers from the intervention schools) and focus groups (n = 4 with children from the intervention schools) (objective 8). Potential differences in the trial and intervention acceptability between ethnic groups were explored as part of the analyses from the focus groups (objective 4). Towards the end of the intervention, four one-to-one interviews were also conducted (one per intervention school) with senior staff (head teachers/deputy head teachers) to further examine the acceptability of the intervention (objective 7).

To assess intervention fidelity, intervention classrooms were observed by a member of the research team for the duration of at least half a school day during the spring and summer terms. Field notes were taken to document the occurrence of any intervention components (i.e. use of prompt cards, engagement with a standing champion; see Table 1) during the observation period. In addition, children’s postures during sit–stand desk use were recorded using a postural analysis recording system based on the portable ergonomic observation63 to assess any potential future risk of musculoskeletal injury (objectives 8 and 9). Full details of the process evaluation methods, analyses and results are described in Chapter 5.

Health- and education-related outcome measures

All health-related outcome measurements were taken twice: at baseline prior to randomisation and approximately 7 months after baseline when pupils were at the end of Year 5 (the sit–stand desks remained in the intervention classes until the end of Year 5, while the follow-up measures were taking place).

Proposed primary outcome in a full trial: activPAL-determined sitting time

The likely primary outcome in a definitive trial would be change in average daily school day sitting time. Sitting was measured objectively for 7 consecutive days during each measurement period using the activPAL3 micro accelerometer (PAL Technologies Ltd, Glasgow, UK). All activPALs were initialised and downloaded using manufacturer proprietary software (activPAL Professional v.7.2.32, PAL Technologies Ltd, Glasgow, UK) and data were processed using the freely available ProcessingPAL software (version 1.1, University of Leicester, Leicester, UK; https://github.com/UOL-COLS/ProcessingPAL). The activPAL3 was waterproofed [using a nitrile sleeve and hypoallergenic Hypafix^® (BSN medical, Hull, UK) dressing] and participants were requested to wear the device continuously (24 hours/day) on the anterior aspect of their right thigh during each measurement period. Participants were provided with a brief diary during each monitoring period, in which they were requested to document time in bed and any periods of non-wear.

The use of the activPAL to objectively measure sitting has increased in recent years82 and the device is recommended for use in interventions when the primary outcome measure is sitting.83 It is regarded as the most accurate method of assessing sitting behaviour in free-living settings,84 and has been shown to be almost 100% accurate in measuring sitting, standing, walking and postural transitions in children.85,86 The device has also been successfully used as a primary outcome measure in previous school-based sedentary behaviour interventions.29,59,67

Although total daily school day sitting time is likely to be the primary outcome for a future trial, we also extracted classroom and leisure time sitting, standing and stepping times from the activPAL data. This allowed us to specifically examine the impact of the intervention on these behaviours during class time (the environment in which the intervention was administered). Furthermore, any positive (i.e. reductions in sitting) or compensatory effects (increases in sitting) of the intervention on children’s sitting, standing or stepping time out of school hours were examined by extracting activPAL data collected during leisure time.

Periods of non-wear and sleep time were excluded from the analyses using an automated algorithm,87 (applied within the ProcessingPAL software) and supplemented with cross-checking against participants’ diary entries. Owing to the exploratory nature of this study, children were included in the analyses if they had worn the activPAL for at least 8 hours on at least 1 weekday at baseline and follow-up, as applied elsewhere.29 To assess whether or not the number of valid days of activPAL wear had an effect on the primary outcome (to inform wear time criteria in a definitive trial), sensitivity analyses were undertaken to examine the impact of the number of included days (see Statistical analysis). The variability (SD) of the data from the proposed primary outcome was used to inform a sample size calculation for a definitive trial (objective 11).

Proposed secondary outcomes in a full trial

ActiGraph-determined physical activity

Proposed secondary outcomes for a definitive trial include objectively measured physical activity. Although the activPAL provides a valid measure of posture, it has not been well validated for assessment of various physical activity intensities among children. Therefore, participants were requested to wear the ActiGraph GT3X+ accelerometer (ActiGraph, Pensacola, FL, USA) on the waist continuously (24 hours/day) for 7 consecutive days, concurrently with the activPAL, during each measurement period. The feasibility of collecting ActiGraph data, in addition to activPAL data, was examined in the present study to inform a definitive trial, in which this device could be used as a secondary outcome to examine any positive or negative (i.e. compensatory) effects of the intervention on physical activity, either during or after school hours. ActiGraphs were initialised to record data at 15-second epochs, the devices were initialised and downloaded using ActiLife version 6.13.3 (ActiGraph) and the data were processed using specifically developed and commercially available software (KineSoft version 3.3.20, Loughborough, UK). Periods of non-wear were documented in a brief diary provided with the activPAL and ActiGraph.

Waist-worn accelerometers have traditionally been considered the criterion measure of children’s physical activity.88 The ActiGraph is the most commonly used accelerometer in field-based research and has been shown to have acceptable reliability and validity in paediatric populations.89 Times spent in light (26–573 counts/15-second epoch) and moderate- to vigorous-intensity (≥ 574 counts/15-second epoch) activity throughout the day, and during and out of school hours, were extracted from the ActiGraph data using the Evenson cut-off points.90 ActiGraph-determined sedentary time (≤ 25 counts/15-second epoch) was also extracted for descriptive purposes. In a comparative study examining the classification accuracy of five different ActiGraph cut-off points for determining children’s and adolescents’ physical activity intensity, against indirect calorimetry, the Evenson cut-off points were found to be the most accurate across all intensity levels and are therefore recommended for use in this age group.91 Periods of non-wear and sleep time were excluded from the analyses during the processing of the ActiGraph data, and supplemented with cross-checking against participants’ diary entries. Owing to the 24-hour wear protocol of the ActiGraphs, a blanket removal of sleep time between 23.00 and 05.59 was undertaken when processing these data. However, to identify periods of sleep occurring outside this time period (i.e. after 06.00 and before 23.00), the three-axis acceleration data from the ActiGraph were used to detect periods of no movement. If these periods exceeded 20 minutes of zero counts, then this additional period was excluded as non-wear/sleep time. The same wear time criteria as applied to the activPAL data (a minimum of 8 hours of wear on at least 1 weekday) was also applied to the ActiGraph data.

Anthropometry

At each measurement point, children’s stature and body mass (both assessed without shoes) were measured directly using standard procedures by trained research staff. Body mass index (BMI) (kg/m²) was calculated and converted to a BMI percentile based on UK reference data.92 Body composition (percentage body fat and fat mass) was assessed via bioelectrical impedance analysis using Tanita DC-360S body composition scales (Tanita, Tokyo, Japan) which contain specific algorithms for children.

Blood pressure

Blood pressure was measured from the left arm after at least a 5-minute period of quiet sitting using a semi-automated recorder (Omron HEM-907; Omron Corporation, Kyoto, Japan) with a paediatric cuff, in accordance with current recommendations.93 Three measurements of blood pressure were taken with each measurement, separated by a 2-minute rest period. The mean systolic and diastolic blood pressures recorded from the second and third assessments were calculated and used in the analyses.

Cognitive function

A set of objective cognitive function tests were administered via a validated software package. The software was installed on school computers, enabling study participants to undertake these assessments collectively in the classroom, under the supervision of two researchers. Participants undertook a practice run-through of the cognitive function test battery a day before the test day. The test battery took children approximately 15 minutes to complete and included the Corsi block-tapping test,94 the Stroop test95 and the rapid visual information processing (RVIP) task.96

The Corsi block-tapping test measures visuospatial working memory capacity.94 Performance on the Corsi block-tapping test is linearly associated with age in typically developing children97 and the original version has exhibited a test–retest reliability coefficient of 0.7 (Pearson’s correlation coefficient) in 11-year-olds.98 The test involves participants repeating a sequence of blocks, displayed on a 3 × 3 grid, which individually change colour. The test begins with a sequence of three blocks and proceeds to a maximum of 12. The sequence length increases or decreases by one with every correctly or incorrectly reproduced sequence, respectively. The outcome measure was the mean correctly recalled sequence length, assessed across 12 attempts.

The Stroop test measures sensitivity to interference and the ability to suppress an automated response (by reading colour names in favour of naming the font colour), and is a commonly used measure of selective attention and executive function.99 The current test comprised two levels: a simple level and a complex level. The simple level involved reading a name of a colour in black font and identifying the colour using the right or left arrow keys on the keyboard. In the complex level, participants were presented with a colour name (i.e. red, green or blue) in a different font colour (e.g. the word ‘red’ in blue font) and were required to identify the colour name (i.e. red in the example given above). The average reaction time of all correct responses for the simple and complex levels were calculated. The complex reaction time measure within this computerised test has been shown to exhibit a test–retest reliability of 0.55 (Pearson’s correlation coefficient) in a combined sample of children and adults.100

The RVIP task measures sustained attention and has been shown to exhibit an internal reliability coefficient (assessed using Cronbach’s alpha) of 0.49 in school-aged children.101 The test involves detecting sequences of three consecutive even or odd numbers presented on the screen. Each number is presented individually and changes every 600 milliseconds. Outcomes from this test include the true-positive rate (i.e. the number of correctly identified sequences out of all identifiable sequences) and the miss rate (i.e. the number of missed sequences out of all identifiable sequences), with scores for both measures ranging between 0 and 1. The true-positive rate is considered a measure of accuracy, whereas the miss rate is considered a measure of alertness, with higher and lower values indicating better performance, respectively. Mean reaction times for each correctly identified sequence were also recorded.

Academic progress and attainment

Measures of participants’ academic progress across maths, reading and writing were collected using routine assessment data recorded by the schools. As collecting data relating to pupils’ progress across these key subject areas is not a requirement of the national curriculum, the participating schools did not have a standardised method of collecting these data. Through consultation with participating teachers, attainment data were dichotomised for the purposes of this study into binary classifications of whether or not each participating pupil was meeting expectations in these three key subject areas at a particular time of development and education. With teachers’ guidance, we transformed the attainment data collected by each school, for each participating pupil, into the following variables: ‘yes – pupil is meeting expectations (for maths/writing/reading) of children of that age’ or ‘no – pupil is not meeting expectations (for maths/writing/reading) of children of that age’. Each school’s attainment data collected at the start of the spring term (2016/17 academic year), coinciding with before the intervention commenced, and at the end of the school year (July 2017), were collected and transformed in this way, to ensure standardised data across the eight participating schools. The binary scores calculated for baseline were subtracted from the follow-up scores for maths, reading and writing separately, leading to a nominal variable of ‘change of subject expectation’, which led to three categorical levels: (1) decreased level of expectation (pupil was meeting expected level for the subject, but is no longer meeting expectations); (2) no change in expected level; and (3) increased level of expectation (pupil was not meeting level of expectation, but is now).

Questionnaire measures

The impact of the intervention on participants’ behaviour was assessed using the Strengths and Difficulties Questionnaire, completed by teachers at baseline and follow-up. This questionnaire, when completed by teachers, has been shown to provide a valid indicator of children’s behaviour (convergent validity: Pearson’s correlation coefficient with the Rutter Children Behaviour Questionnaire = 0.92102). In addition, children self-reported their engagement and disaffection with their own learning via the Engagement Versus Disaffection with Learning Questionnaire103 (correlation coefficients between pupil and teacher reports of the components of engagement using this measure range from 0.26 to 0.44104). Postural comfort was reported by participants using a further brief questionnaire.63

Children completed the Paediatric Quality of Life Inventory (PEDS-QL)105 and EuroQol-5 Dimensions Youth (EQ-5D-Y)106 at each measurement point, to provide a measure of self-reported quality of life. The construct validity of the PEDS-QL has been previously demonstrated, with healthy children displaying significantly higher scores on this measure than acutely or chronically ill children.105 Responses on the EQ-5D-Y have also been shown to correlate with other measures of children’s health-related quality of life (convergent validity correlation coefficients up to 0.56107). To inform the economic analysis described in more detail in Economic Analysis (objective 12), teachers and parents completed a questionnaire72 created for the purpose of this study, which assessed participants’ health- and education-related resource use at baseline and follow-up.

Basic demographic information (sex, age, ethnicity and postcode, to determine Index of Multiple Deprivation as an indicator of socioeconomic status) was collected at baseline.

School and participant appreciation

As a thank you for participating in this pilot trial, all schools (intervention and control) received a donation of £200 at the end of the study period. A £5 gift voucher was also given to each child following the completion of both the baseline and follow-up measures to encourage a timely return of the accelerometers.

Economic analysis

The availability and completeness of economic data were established as part of this pilot study (objective 12). Resource use information was collected, which included the cost of the sit–stand desks, along with participants’ health-related [e.g. general practitioner (GP) visits] and education-related resource use (e.g. requirements for additional tutoring). Proposed outcomes within a definitive trial were based on two sectors: health and education. For the former, as described in Questionnaire measures, the PEDS-QL105 and EQ-5D-Y106 were collected at baseline and follow-up to assess children’s health-related quality of life. For the latter, the Strengths and Difficulties Questionnaire completed by school teachers for each pupil was used. A preliminary cost-effectiveness analysis was conducted (see Chapter 4) to inform the value of, and to make recommendations for, the design of a full trial. All outcomes collected were presented as a within-trial analysis at baseline and follow-up; however, no formal comparison of the cost and benefit of providing sit–stand desks was conducted for within-trial outcomes. A brief scoping review was conducted to identify existing model(s) that link short-term health outcomes to long-term health effects with this model, informing a preliminary assessment of the long-term health costs and benefits associated with the intervention.

Statistical analysis

The analysis and reporting of this pilot RCT was in line with the Consolidated Standards of Reporting Trials (CONSORT) statement for cluster RCTs.108 Data were analysed on a complete-case basis. The purpose of the primary analysis was to assess the feasibility of recruitment and adherence/retention of primary schools and pupils to the sit–stand desk intervention. As this was a pilot trial, the primary analyses mainly utilised descriptive statistics. The number of schools approached, the number agreeing to participate, the proportion of children within each school with parental consent to participate in the study evaluation, the number of children completing the study protocol, retention rates and the number providing valid outcome measurement data at baseline and follow-up are described in Chapter 3. Study acceptability data are presented for the sample as a whole and stratified according to study arm (intervention and control) and ethnicity (South Asian and white British).

Although the main aim of this study was to establish acceptability, feasibility, recruitment rates and sample size to inform a definitive trial, and although effectiveness was unlikely to be established with the small sample size, we examined the proposed primary and secondary outcomes for a definitive trial to mimic practice in such a trial. The results from this analysis should, however, be treated as preliminary and interpreted with caution.109,110 As the number of clusters is low, cluster summary statistics were used rather than multilevel modelling.111,112 The analyses were conducted using children as the unit of analysis, with change in average total school day sitting time as the primary outcome. A weighted linear regression model was used to compare the intervention arms weighted by the number of participants followed up in each cluster, and adjusted for baseline total daily sitting time on school days and average weekday wear time when awake across the two time points (baseline and 7 months). To examine preliminary effects of the intervention on the secondary outcomes, the same analytical approach was adopted as for the proposed primary outcome (objective 10). These analyses were intended to provide preliminary evidence of the effectiveness of the intervention to, in part, inform the decision of whether or not a definitive trial should be undertaken.

Sensitivity analyses were performed on the proposed primary outcome for a full trial of average weekday sitting time. This was achieved by including pupils who wore the activPAL (with a minimum valid wear time of 8 hours/day) for at least 2, 3 and 4 weekdays, at baseline and follow-up.

An objective of this pilot study was to estimate the SD of the primary outcome to inform a full trial. The intraclass correlation coefficient (ICC) to inform the sample size calculation of a definitive trial was estimated from published literature. The ICC was not estimated from this pilot study because multilevel modelling was considered to be an unsuitable analysis method for the present study, due to the small number of clusters, and an ICC estimate from such a model would not be sufficiently robust to inform a sample size calculation for a definitive trial.

Data management and research governance

Anonymised data were entered into a secure and validated clinical data management system provided by the Leicester Clinical Trials Unit; this database (InferMed Macro v4; Elsevier Ltd, Oxford, UK) included a series of quality control mechanisms to ensure that the data collected were complete and accurate. The study was sponsored by Loughborough University and two groups were created to oversee the research: an independent TSC and a Project Committee. As the study was regarded as low risk, the TSC took on the role of a Data Monitoring Committee, with the intention of reviewing any serious adverse events, should they have arisen, and monitored progress with data collection. The TSC met every 6 months and included the PIs (Stacy A Clemes and Sally E Barber), an independent chairperson, two independent external members (including a statistician) and two school representatives (public members who were teachers from schools separate to those participating in the trial). The Project Committee comprised the PIs, all co-investigators and those concerned with the day-to-day running of the study, and provided update reports for the TSC.