The study methods have been reported in detail elsewhere [19] and are presented briefly below. Following study commencement, three key protocol amendments were made: (1) Follow-up was reduced from 12 to 11 months due to an unexpected delay in the ethical approval process (approved by the Trial Steering Committee); (2) Additional consent was requested at follow-up to access HENRY programme attendance data; and (3) Telephone interviews were conducted with parents in the process evaluation to replace focus groups, which were not well attended. The research was approved by the University of Leeds School of Medicine Research Ethics Committee (MREC: 16–107) and registered on the clinical trial website (clinicaltrials.gov) #NCT03333733.
Design
A UK multi-centre, open labelled, two group, prospective, cluster randomised, controlled feasibility study, with process evaluation between July 2017 and November 2019.
Objectives
Primary objectives were: (1) To determine whether it is feasible to recruit local authorities/service providers that are willing to nominate children’s centres to be involved in the research (allowing randomisation in place of the usual selection of centres in most need); (2) To assess the time required to train and certify staff to competently deliver HENRY programmes, in order to propose a clear process and timeline for a definitive trial; (3) To determine whether it is possible to recruit parents to the study, at the same rate in both control and intervention children’s centres; (4) To explore local area preferences for HENRY commissioning, provision and delivery via postal screening questionnaires and qualitative data collection in areas currently delivering HENRY; and (5) To explore potential sources and risks of contamination, including the degree to which parents use multiple centres, the level of contamination resulting from social networks (control and HENRY parents sharing knowledge) and the possibility of HENRY trained facilitators sharing knowledge with control centres.
Secondary objectives of the feasibility study were: (1) To examine the acceptability and completeness of the proposed methods of data collection to ensure they are feasible for a definitive trial; (2) To gather data to allow estimation of the sample size requirements for the definitive trial; (3) To assess the acceptability of the design for parents, commissioners and centre staff; particularly related to withholding HENRY training in control centres until the end of the study; and (4) To determine the practicalities of delivering the required number of HENRY programmes within the trial period in regards to programme implementation.
Recruitment, Participants And Setting
Local authorities planning to commission HENRY during the study period were invited to take part in the study initially via conversations with HENRY national office. In addition, ‘Expressions of interest’ forms were sent out to all local authorities across England and Wales. Children’s centres within each local authority were nominated by commissioning leads within each area. Parent recruitment took place in children’s centres post randomisation, once centres had agreed to take part via invitations from staff and through poster advertisements. Centre staff asked potential participants to provide consent for their contact details to be shared with field researchers from NatCen (http://natcen.ac.uk/) (a social research unit working across the UK), who then made contact to arrange a visit at the parents’ home or children centre. Eligibility was confirmed and participants consented to take part in the study during these visits prior to collection of data. Eligible and consenting participants were registered by authorised the NatCen researcher using the CTRU automated 24-hour registration system (Gen 24). Measurements and questionnaires were completed at baseline (within a 6 week window prior to each HENRY programme (over 5 periods of programme delivery)) and follow up at 11 months. NatCen researchers attended a briefing prior to baseline and follow up data collection where they completed a practical assessment of data collection procedures for height, weight and waist circumference. Following data collection, researchers sent completed booklets to the NatCen head office for data checking and completed booklets were then sent to CTRU. Observations of data collection were conducted by research field leads to ensure protocols were adhered to. Participants received a £10 shopping voucher per visit.
Eligibility Criteria
To be eligible, local authorities had to allow randomisation of four/six children centres and be willing to have their staff trained to deliver HENRY or use external teams outside of the centres. Local authorities had to be HENRY ‘naive’, meaning they had never delivered HENRY or trained staff to deliver HENRY prior to the study, or contain clusters of children’s centres that were naïve to HENRY. For the purposes of this study, HENRY naïve clusters were defined as a group of centres that did not include any centres that were either (a) delivering HENRY at the time of the study, or within the past two years, or (b) had been trained to deliver HENRY within the past two years. Local authorities that were not situated in an area with coverage of NatCen were not eligible. Eligible centres were children’s centre or early years setting, ready to start the HENRY programmes within four weeks of training completion, each capable of delivering three programmes throughout the study. Centre managers had to agree to support participant recruitment in their centre. Children’s centres were excluded from the study if they had delivered HENRY programmes, or their staff had attended HENRY training in the last two years. Eligible parents were parents or carers of a preschool child (aged six months to five years), willing to attend the programme (in HENRY centres), provide data for the study, and be English speaking or agree to take an interpreter to the intervention and data collection. Parents were excluded if they had severe learning difficulties, their child was tube fed or had known clinical conditions likely to affect growth, or had attended the HENRY programme for a previous child. In HENRY centres, only parents who enrolled to attend a HENRY programme were eligible to take part. No restriction on attendance at any other programmes was enforced in either treatment allocation.
Randomisation And Minimisation
Cluster randomisation of children’s centres was chosen to reduce contamination, as core training for the intervention was intended to impact across the whole centre and the programme was delivered to groups of parents. Six centres in each local authority were randomised to the intervention or control in a 1:1 allocation ratio (HENRY; control). Minimisation ensured that treatment groups were balanced for size of children’s centre (number of permanent centre members of staff not including staff using the centre such as health visitors and nursery workers) (≤ eight/ > eight members of staff); area level ethnicity (< 80% / ≥ 80% White British (census data)); and area level deprivation (≤ 10% / >10% ranking within Index of Multiple Deprivation at the Lower Super Output Area). Randomisation was carried out by the statistician (MC) at the Clinical Trials Research Unit (CTRU) at the University of Leeds. Local authorities and children’s centres were notified of treatment allocation directly from researchers at the CTRU.
Blinding
It was not possible to blind children’s centre staff and participants to treatment allocation, but data collection researchers were blinded to treatment allocation. By design, they only became unblinded during the final follow up interview question, when requesting consent to share HENRY attendance data. Prior unblinding incidents were recorded by NatCen staff, and then, attempts were made to assign a different follow up interviewer.
Intervention
Parents in the intervention arm attended the eight-week HENRY (Health, Exercise, Nutrition for the Really Young) programme delivered in children centre’s to groups of 8–10 parents. Programme details are documented elsewhere [14, 19, 20]. In brief, it aims to provide parents with the skills and knowledge to support a healthy lifestyle in preschool children and their families. Training for intervention delivery is split into two stages: (1) Centre level training: equipping staff with skills and knowledge to promote and provide healthy nutrition in early years settings and to support parents to provide heathy lifestyle and nutrition for their families; and 2) Practitioner level training to deliver HENRY programme to families: training staff to deliver the eight week programme. Both types of training are underpinned by a combination of proven models of behaviour change, including the Family Partnership Model, motivational interviewing, and solution focused support. Topics covered in the HENRY programme include eating habits, balancing healthy meals and snacks, child appropriate portion sizes, emotional wellbeing, parenting skills and activity. Services provided by children centres that were deemed similar to those of HENRY (e.g., parenting, healthy eating), as well as services attended by study participants, were recorded.
Control Group
In children’s centres randomised to the control group, standard care continued. These centres delivered all their usual programmes (including programmes such as ‘stay and play’, ‘cook and eat’, baby massage and other parenting courses) and staff did not undertake the HENRY programme centre or facilitator training. Following all data collection, centres in the control group received free HENRY training to enable delivery of the HENRY sessions. Parents in control centres were offered attendance at HENRY once the data collection for the study had ended (i.e. a waiting list). Similar to the HENRY groups, we recorded services provided by children centres deemed similar to those of HENRY, as well as services attended by study participants.
Measurements
Feasibility objectives
were assessed through analysis of process data and data routinely collected data from HENRY central office. HENRY data were transferred to CTRU via a secure transfer encrypted system using a data specification document provided by CTRU.
Outcome data
were collected at baseline and 11 month follow up using paper based Case Report Forms (CRF), which were interview administered by trained NatCen researchers within family homes, (or at the children’s centre, according to parent preference). Completed CRFs were checked for quality assurance by NatCen researchers prior to returning to the CTRU. The CRFs contained no identifiable information (unique identifiers only). Unscheduled forms, such as participant withdrawal or researcher un-blinding, were mailed directly from the NatCen interviewers to the CTRU as the event occurred.
Feasibility Outcomes
-
(a) Recruitment rate of local authorities and parents, including: number of local authorities / parents screened for eligibility, reasons for ineligibility and numbers (and reason for) declining.
-
(b) Training and quality assurance, including: length of time taken to train and certify staff to deliver HENRY programme.
-
(c) Contamination identification and risk, ascertained using a mixed methods design. Quantitative data determined the number of parents registered to control centres who attended centres running HENRY, the extent to which HENRY trained facilitators worked in or visited control centres, and the types of other programmes delivered in centres that focused on lifestyle change. Qualitative interviews sought to find evidence of sharing of HENRY messages by staff and/or parents in HENRY centres to staff and/or parents allocated to control centres.
-
(d) Acceptability of the data collection protocol, assessed from qualitative interviews (including feedback about the methods and measures) and quantitative process data, including the number, proportion and timing of parent withdrawals from HENRY programme (intervention centres only), follow-up data collection (parents registered to HENRY and control centres), reasons for withdrawal and the amount of missing data.
-
(e) Determination of sample size for a definitive trial, estimated using feasibility data of gender adjusted Body Mass Index (BMI) in both arms, difference between arms and 95% confidence intervals, estimation of clustering effect (ICC) and cluster size.
-
(f) Intervention compliance and implementation, estimated from routinely gathered data on the timing of delivery of first HENRY programme (plus reasons if delayed); number of HENRY courses delivered per centre; attendance rates at HENRY programmes and reasons for absence.
Outcome Measures
Primary outcome
Reference child BMI z-score (age and gender adjusted height(m)/weight(kg)2) measured by NatCen interviewers.
Secondary Outcomes:
-
(a) Primary caregiver BMI (measured height(m)/weight(kg)2) and waist circumference (cm);
-
(b) Family eating/ activities, via the validated Golan Family Eating and Activity Habits Questionnaire [21];
-
(c) Parenting self-efficacy, via the Dumka Parenting Self Agency Measure [22];
-
(d) Feeding, via the Baughcum pre-schooler feeding questionnaire [23];
-
(e) Dental health via a bespoke questionnaire based on the Dental Health Survey of Children and Young People by the University of Leeds, School of Dentistry to measure the potential wider impact of HENRY on a child’s dental health (Dental Questionnaire);
-
(f) Centre policy and practices via a bespoke environment questionnaire;
-
(g) Quality of life (EQ-5D) [24]; and
-
(h) health care resource use data (via a bespoke questionnaire) for the child and the parent within the NHS (health services, hospital, social services) as well as time-off work in relation with HENRY.
Progression Rules For Definitive Trial
In order to progress to the definitive Phase III trial, the following criteria were applied:
Green (fully feasible):
-
Recruitment of two local authorities within 12 months;
-
Randomisation of at least 12 children’s centres within 12 months;
-
An average of at least 4 parents registered per programme (or control group equivalent)
Amber (modifications required):
Red (not feasible):
-
Recruitment of less than two local authorities (and their service providers (if applicable)) within 12 months;
-
Randomisation of less than 8 children’s centres within 12 months;
-
Less than 3 parents registered per programme.
Sample Size
The planned sample size for the feasibility study was 120 parents across 2 local authorities. These numbers provide sufficient confidence that a phase III multi-site trial could be successfully conducted [26, 27] and meet the recommendation that at least 60 participants per group are required when estimating study summary measures [28]. A formal power calculation was not appropriate as effectiveness is not being evaluated.
Analyses
Quantitative analyses focussed on descriptive statistics and confidence interval estimation. Continuous outcomes were summarised using the mean, standard deviation, median and interquartile range. Categorical outcomes were summarised using the number and percentage of responses for each category. The number of parents or children with missing outcome data was reported for each outcome. Outcomes were presented overall, by arm and where relevant by children’s centre (cluster) for baseline and follow-up time points, using the intention-to-treat population. No formal hypothesis testing was conducted as part of this feasibility study.
Analysis of the qualitative data was conducted in NVivo data analysis software [29] and was guided by a deductive organising framework devised from the process evaluation research objectives (contamination, implementation and acceptability) and interview and focus group topic guides. Inductive thematic analysis was then applied to identify codes and sub-codes, and potential relationships between these codes [30]. Qualitative data were coded by two members of the research team, who conducted an independent check of 10% of the transcripts. Analysis was discussed in team meetings and consensus was reached for the content of themes and their influence on the subsequent design of a definitive trial.