This paper has presented the protocol for the randomized controlled trial of the Families for Health program. The trial has completed its set-up phase and started recruiting families from March 2012. In this section we will firstly discuss the changes made to the original protocol and then the challenges to implementation.
Two changes have been made to the original protocol. Firstly, two months after recruitment had started, the inclusion criterion for children was changed from being aged between 7 and 11 to being between 6 and 11. Our original recruitment strategy was based on the age range in junior schools; it was felt that extending the age range down to 6 years might alienate older children and prevent provision of a safe space for the 6 year olds. However, during the early stages of recruitment a number of parents of 6-year-old children contacted the trial team with an interest in taking part. These children were eligible for usual care and were keen to take part in the trial. At the same time the parents of some 11-year-old children reported that their children were focusing on the transition to secondary school and did not want to start a primary school age program. Families for Health V2 with its focus on active games, the activities around healthy eating and circle time, suits younger children very well, and including 6-year-old children compensated for the low levels of recruitment from 11 year olds.
During the time period between submission of the grant application and the award of the grant and start-up of the trial, the provision of usual care in the trial localities changed. By the time the trial started all three localities had started to provide evidence-based interventions for overweight and obese children. It was felt to be important to capture parents’ views on usual care as well as their views of Families for Health, so a further protocol amendment was requested to enable interviews with parents and children allocated to usual care, to compare the experience of families in the two arms.
Challenges to the implementation of the study included securing excess treatment costs to fund the delivery of the intervention, the recruitment of families, changes to usual care and introducing strategies to maintain blinding of researchers.
Excess treatment costs are ‘the patient care costs which would continue to be incurred (by the NHS) if the patient care service in question continued to be provided after the R&D study had stopped’, and are the responsibility of the NHS . At a time of a major NHS reorganization in which public health services are transferring from the NHS to local authorities, securing the excess treatment costs to cover the running of the Families for Health intervention (facilitators, venues and consumables) at the three localities was complex. As there was uncertainty about the budget to meet excess treatment costs, these costs were identified by the Public Health Departments from their 2011/12 budget whilst they were still located in NHS Coventry and NHS Warwickshire Primary Care Trusts, and transferred to the University of Warwick in March 2012. Wolverhampton identified funds for groups running in 2012/13 but reorganization may prove an issue for funding groups in 2013/14. As the excess treatment costs were transferred by two sites, this means the University of Warwick has to employ the facilitators of the Families for Health intervention for both Coventry and Warwickshire. This introduced additional complexity for the set-up of the research including obtaining research passports.
The NHS reorganization has had a further effect on the trial because many of the staff in the employ of the PCTs, who are either supporting the trial or who have been trained as facilitators, are experiencing stress relating to the reorganization, for example, due to actual or threatened redundancy. This has meant that the pool of trained facilitators is declining and finding facilitators to run groups is more challenging. It has also meant that those supporting the delivery of the trial within PCTs, who have offered unflagging help to date, inevitably have other major pressures on their time and cannot offer the same level of support.
However, the main challenge of the trial relates to recruitment. Families for Health V2 is a closed group intervention (that is, families start the intervention together), and as such it is necessary to have a group of 8 to 12 families randomized to that arm in order for the intervention to run. As the program involves school age children, takes 10 weeks to run and tends to lose families if there is a significant break in provision, it is only practical to start the program three times a year – at the beginning of the school terms. As approximately half the families will be randomized to usual care, it is necessary to recruit around 20 families willing to be randomized in any locality in order for the intervention to run. This has been challenging and may mean that some groups run with a sub-optimal number of families. For example, the first group in Coventry ran in May 2012 with eight families. Running a group at reduced capacity may impact on effectiveness, as reported in the WATCH IT feasibility RCT [15
], and will impact on cost effectiveness.
The National Child Measurement Programme is relatively new and has been developed in different localities in a variety of ways. The planned recruitment via the NCMP was problematic as recruitment to the trial started in March 2012 when most of the PCTs had completed their annual NCMP measurements for the year. Coventry PCT re-sent a letter to a selected group of families advising them of the research study, and this was helpful in recruiting families. Future Families for Health programs are planned to coincide with the timing of the NCMP in each locality. Additional recruitment methods have also been added to the protocol including leaflet drops, attendance at health events where offers of height and weight measurements are made to families, and paid advertisements.
Another key issue is changes to usual care. When the protocol was submitted for funding (January 2010) very few options were open to support families with childhood obesity. However, much has changed over the last 2 to 3 years as local services strive to meet new policy targets and objectives and invest significantly in the development and evaluation of programs. These changes mean that Families for Health V2 is now being assessed against a variety of possible approaches to supporting families. The usual care that children and their families in the control arm receive will be carefully documented. If significant numbers of control families take up usual care offers, our trial may be underpowered because the estimated difference in BMI z-scores between control and intervention groups is likely to be reduced.
A further challenge is to maintain the blinding of investigators who are taking the research measurements to the families’ allocations. The nature of the 3-month visit unblinds the researcher as both the responses to the CSRI [37
] and the interview are likely to reveal the group allocation. Strategies in place to maintain blinding include letters sent in advance to the parents to ask them not to reveal their allocation to the researchers until required. During the 3-month follow-up visit, researchers take all anthropometric measurements first and the other questionnaires are then completed, prior to the CSRI and then the interview (if required). To date, researchers who are involved in data collection at the 3-month follow-up have successfully remained blinded to the allocation group until the CSRI was administered in all but one family. The researcher who carries out the 3-month follow-up subsequently becomes unblinded so that a different researcher will carry out the 12-month follow-up with this family. The trial administrator will keep a record of which researcher is unblinded to which families, and follow-up visits will be set up accordingly.
This is an ongoing trial (62 of 120 families recruited).