Cultural adaptation of a children’s weight management programme: Child weigHt mANaGement for Ethnically diverse communities (CHANGE) study

The study took place in Birmingham, the second largest UK city with a population of 1.1 million. Forty-two percent of all residents are from minority ethnic communities. Pakistani and Bangladeshi children comprise 26% of the Birmingham population aged 0–15 years [21]. At the time of the study a group-based child weight management programme, First Steps, was available across the city. The programme was delivered as weekly one-hour sessions over 5–7 weeks in community venues, covering nutrition education, physical activity promotion and the promotion of positive lifestyle behaviour changes. The programme was aimed at parents/carers; children attended only the first and last sessions to have their heights and weights measured. All families resident in Birmingham with a child aged 4–11 years with excess weight (BMI over the 91st centile of the UK 1990 growth reference charts [22]) and able to participate in a group setting were eligible to attend the programme. Children could be referred to the programme by a health professional, the child’s school, or families could self-refer. Children identified as having excess weight through the National Child Measurement Programme (a surveillance programme to provide data on weight indicators in primary school-aged children) were also referred to the programme.

First Steps was a children’s weight management programme developed by the service providers, based on their previous experience of delivering evidence-based programmes [23, 24], and tailored for the local population. Given Birmingham’s cultural diversity, there was a focus on parental engagement with access to interpreters, and programme material had a high pictorial content and referred to culturally appropriate foods. Despite this, Pakistani and Bangladeshi families who started the programme were less likely to complete it than families of other ethnicities (40% of Pakistani and Bangladeshi families completed it compared with 65% of families from other ethnic groups). Data collected routinely at the first and last sessions indicated that children achieved an average reduction in BMI z-score of 0.1 at programme end. This is in line with reported differences in BMI z-scores between intervention and control groups in randomised controlled trials of behavioural child weight management programmes [6]. Given the existing tailoring to the local population and evidence of effect on children’s weight, the programme provided a good foundation on which to develop a further culturally adapted programme, with the particular intention of increasing retention of families from Pakistani and Bangladeshi communities in the programme.

The theoretical and modelling stages of the UK Medical Research Council (MRC) framework for the development and evaluation of complex health interventions [25, 26] guided the cultural adaptation process. The adaptation process was informed by three main information sources: 1) data from a qualitative study exploring the experiences and viewpoints of Pakistani and Bangladeshi families who had participated in or who had initially agreed but then declined to participate in the First Steps programme; 2) local information from the First Steps programme providers; and 3) existing children’s weight management literature. Two specific theoretical frameworks were used in parallel in the adaptation process: a framework for the development of behaviour change interventions, and a programme theory and adaptation typology to guide the adaptation of health promotion programmes for minority ethnic groups [15, 27]. An advisory panel comprising Pakistani and Bangladeshi parents of primary school-aged children also provided advice during the adaptation process. Ethical approval was received from the Edgbaston Local Research Ethics Committee in July 2014 (14/WM/1036).

Community Researchers from Pakistani and Bangladeshi communities in Birmingham with qualitative research experience (AA; female and of Pakistani heritage, and MB and SK; both female and of Bangladeshi heritage) were recruited to assist the core research team (TG (research fellow in Public Health with mixed methods research experience) and LG (lecturer in healthcare anthropology with extensive qualitative research experience); both female and of white British heritage) in undertaking this qualitative data collection. The Community Researchers did not have pre-existing relationships with participants prior to the study, but were able to communicate in Urdu, Bengali or Sylheti where necessary, and understand the cultural context of participating families.

The First Steps programme provider (Birmingham Community Healthcare NHS Trust) identified all Pakistani and Bangladeshi families who had been invited to take part in the programme from September 2013 to July 2014. The families were categorised into either: (i) attended 60% or more of the First Steps programme (‘completers’); (ii) started the First Steps programme but attended less than 60% (‘non-completers’); or (iii) did not attend the programme (‘non-attenders’). Parents from completing families were invited to participate in a focus group (FG) at a community venue. FGs were the preferred method of data collection as they explicitly use group interaction as a way of stimulating discussion [28]. However, we recognised that parents from non-attending and non-completing families may find it challenging to attend a FG, and so they were invited to participate in a one-to-one interview, which gave greater flexibility for them in terms of the timing and venue of the interview. Face to face interviews were preferred but telephone interviews were offered if this was not possible. We aimed to recruit 15 ‘non-completers’ and 15 ‘non-attenders’ to participate in interviews, and to hold 3–5 FGs with ‘completers’, with a contingency of recruiting more participants if data saturation was not felt to be achieved. All participants received a £10 shopping voucher following successful completion of the interview/FG.

Parents were initially contacted by telephone and a participant information pack was posted to those who expressed an interest in study participation. A further telephone call was made and if the parent agreed to participate, an interview or attendance at a FG was arranged. Parents who did not speak English were telephoned by a Community Researcher in their preferred language.

Interviews took place in the participant’s home and FGs at a convenient community location. Participants gave written informed consent and completed a short questionnaire before the interview or FG commenced. The interviews and FGs were conducted either by a core researcher or Community researcher in the participant’s preferred language. An additional researcher was present as an observer at the FGs. Semi-structured interview and focus group schedules, informed by literature and input from the study Parent Advisory Panel, were used to guide discussions. The research questions that were explored are shown in Table 1. Interviews and FGs were audio-recorded and transcribed. Community researchers translated and transcribed interviews and FGs that were not conducted in English. A sample of translated transcripts was checked using the audio-recording by an independent researcher with the relevant language skills.

Data analysis was conducted using NVivo 10 (QSR International Pty Ltd. Version 10, 2012) and was guided by thematic analysis approaches [29]. Two researchers (TG and LG) reviewed 50% each of the transcripts independently and identified codes to apply to the data. The researchers discussed their coding and agreed on a final coding framework, which they then applied to all transcripts. Overarching themes were identified, which included commonalities and differences between the three participant groups.

Direct observation of the First Steps children’s weight management programme was undertaken by a researcher (TG) to assess structure, content, delivery and participant response. In addition, a series of consultations were undertaken with the two service managers over a period of 3 months to enable an understanding of the existing infrastructure and processes. The managers were also asked to identify any issues with the existing programme from their perspective.

A comprehensive guideline on managing overweight and obesity in children was published in 2013 by the UK National Institute for Health and Care Excellence (NICE) [5]. Two evidence reviews were undertaken to support development of this guideline, focusing on: 1) the effectiveness and cost-effectiveness of interventions to manage children’s weight [30]; and 2) the barriers and facilitators to implementing weight management programmes for children [31]. In addition a systematic review of behaviour change techniques that are effective in influencing obesity-related behaviours in children was published in 2013 [32]. These reviews, together with more recent evidence on effective children’s obesity interventions, informed the planning of the adapted programme to ensure that it was coherent with established evidence.

The adaptation process was guided in parallel by two theoretical frameworks: The Behaviour Change Wheel (BCW) by Michie et al. [27, 33] and the Typology of Cultural Adaptation and Programme Theory of health promotion interventions by Liu et al [15]. The BCW has been developed from 19 behaviour change frameworks and was used to ensure the target behaviours, pathways to change, and adaptations made to address these were clearly articulated. Three target behaviours requiring change were identified; the first was programme attendance and the other two were behaviours that directly influence weight (dietary intake and physical activity). The capability, opportunity, motivation and behaviour (COM-B) model at the centre of the BCW enabled us to gain a theoretical understanding of the factors preventing Pakistani and Bangladeshi families from adopting the desired behaviours. This was achieved through the mapping of qualitative data from parents onto the different elements of the COM-B model (physical and psychological capability, physical and social opportunity, and reflective and automatic motivation). From this understanding of the factors influencing the target behaviours identified for change we were able to select the relevant intervention functions (categories of mechanisms by which interventions may have their effects) from the nine outlined in the BCW that correspond to elements of the COM-B model. This informed the detailed intervention planning.

The second framework, the Typology of Cultural Adaptation and Programme Theory proposed by Liu et al. [15], ensured that appropriate cultural adaptations were considered for inclusion in the adapted intervention across all aspects of the programme and at all stages of the programme cycle (i.e. conception/planning, promotion, recruitment, implementation, retention, evaluation, outcome, and dissemination). The 46-item typology has been constructed from a systematic review of health promotion programmes targeting smoking, diet and physical activity, which have been adapted for minority ethnic groups. The typology was used to identify the most appropriate type of cultural adaptations to address the themes identified in the qualitative data obtained from Pakistani and Bangladeshi parents.

The identified BCW intervention functions and the types of cultural adaptation provided the outline for the detailed planning of the adapted programme. The local information from direct observation and the service providers, and the relevant literature were used to further inform the process. Consideration was also given to the flexibility of programme delivery to ensure suitability for children of different ages. The planning process was iterative to ensure that the final programme design was coherent with: a) the identified intervention functions and adaptation types; b) the qualitative data; c) local service information; and d) the children’s weight management literature. Figure 1 summarises the intervention adaptation methodology.

In total, 31 parents/carers participated in interviews and 12 participated in FGs. All participants were Muslim, 36 (84%) were Pakistani and 37 (86%) were female. Twenty-one participants were ‘non-attenders’, 9 were ‘non-completers’ and 13 were ‘completers’. Participant characteristics are shown in Table 2.

Of the 31 interviews, 27 were conducted face to face and 4 by telephone. Six interviews were in Urdu and 3 in Bengali. Length of interviews ranged from 15 to 47 min (average: 28 min). Once interviewees seemed to have no further comments, interviews were drawn to a close. Four FGs were completed. A further 3 were arranged but no participants attended. Two of the FGs were attended by 4 participants with the remainder having 2 participants. Two FGs were conducted in Urdu. The FG length ranged from 35 to 50 min.

Several themes emerged from the data. There was coherence across the three groups on several themes, but some themes were more prominent in some groups than others. Important logistical barriers to attending a family community weight management programme were raised by all participants. The majority of families reported that to attend the programme it would need to be in a close, familiar location at a convenient time. Some parents were concerned about children missing school and identified weekends as the most convenient time to attend, whilst others felt that children could take time out of school to attend. In contrast, after school sessions were commonly considered to be impractical due to many of the children attending religious classes at their local mosque at this time. This practice was also raised as a barrier to finding time for being physically active. Caring for younger siblings was cited as a barrier to attending by some parents, although it was also observed that younger siblings were often brought along to the sessions. Language barriers to participation existed for some parents from Pakistani and Bangladeshi communities who did not speak English. These were highlighted as a problem by some non-attenders at the initial recruitment stage. Once participants attended the programme, language barriers were less of an issue, particularly if interpreters were present (all participants were asked whether they required an interpreter prior to commencing the programme). Several English speaking participants discussed supporting other parents in the group who were struggling to understand.

The focus of the programme being on weight and obesity, rather than a positive focus on health was also a barrier. Some parents, particularly those who had not attended or completed the programme, considered that their child did not have a weight problem, or felt that they could not do anything to address their child’s weight. These families engaged less with the programme as its focus was on weight loss. However, data from these parents indicated that they recognised the value of healthy lifestyles and wanted to encourage their children to adopt healthy behaviours. Some parents who did not attend or complete the programme also highlighted that children were sensitive about attending for “weigh ins”.

Another group of important themes related to the target audience, content and delivery of the programme. Most parents felt that a programme involving children in all the sessions would be of more value, as they felt that children need to learn how to change their behaviour first hand, and would respond more positively to messages relating to behaviour change if they were given by someone other than their parents. Interactivity within the programme was highlighted as important. Completing participants spoke of the value of the interactive elements of the programme. However, non-completing participants perceived that there was little interactive content and disliked the ‘classroom’ format of sessions. They also reported that they disliked receiving a high volume of written information. Many of the participants expressed that there needed to be much more physical activity content in the programme, particularly getting the children to participate in physical activities during the sessions. They identified a range of barriers to physical activities in their daily lives which they thought should be addressed through the sessions. The group setting and the ability to share ideas and experiences amongst attending families was highly valued by many participants who had attended the programme. Some participants who had not attended or completed the programme felt that they were not going to gain anything new from it, and that they already had a good idea what was ‘good’ and ‘bad’ for their children, particularly in terms of their diet. This viewpoint differed in several of the completing participants, who felt they had gained new nutritional knowledge, and also advice on how to apply this in their daily lives. Although the First Steps programme included references to South Asian foods, some parents felt that the nutritional content could be made more relevant to their traditional diets, whilst other participants acknowledged the importance of also talking about Western foods, as their children’s diets encompassed both traditional and Western foods. There were mixed views regarding the cooking methods of traditional foods; some participants felt there was opportunity to learn about healthier cooking methods (e.g. using less oil), but others felt that they would not change their cooking methods. There was also concern about children’s intake of ‘junk’ foods, which they felt needed to be addressed. Finally, several parents who had attended part or all of the programme expressed difficulty in ensuring their children adhered to the changes that they instigated at home, especially in relation to food, and therefore, they felt they needed help with overcoming this issue.

Apart from dietary and language factors and the time spent attending religious classes, no other emerging themes related explicitly to Pakistani and Bangladeshi culture. The more prominent issues identified in the data were the difficulties and competing priorities which families had to face in their daily lives (e.g. juggling demands of siblings, busy family lives, and perceived safety issues in local communities), and the impact of these on the ability to undertake healthy behaviours. The emergent themes and examples of data to illustrate these themes are shown in Table 3.

The UK NICE guideline on managing overweight in children and young people (PH47) [5], published in 2013, presented several evidence-based recommendations regarding children’s weight management service provision. The recommendations were taken into consideration during the detailed planning phase of programme adaptation to ensure that the finalised programme was consistent with the guideline (see Table 4). The guideline, along with other relevant literature [5, 34], emphasised the importance of parental involvement in child weight management programmes, and the need for elements that address both diet and physical activity [35, 36]. Therefore, these important aspects were included in the adapted programme. The behaviour change techniques that have been identified as effective in obesity interventions for children [32] (provision of information on the consequences of behaviour to the individual; environmental restructuring; prompting practice; prompting identification of role models or advocates; stress management/emotional control training; and general communication skills training) were also considered for inclusion in the adapted programme.

A researcher (TG) observed two programmes, delivered by different facilitators (all sessions of one programme and two sessions of the other programme). Observations broadly concurred with the qualitative data. Particularly evident were: the lack of interactive activities for participants; the large volume of written information handed out; and the heavy focus on nutritional knowledge, with less emphasis on skills around food preparation and feeding practices, and little physical activity content. Goal setting was incorporated in the programme sessions but was not always well implemented. The programme managers also identified that the didactic delivery and volume of written information were problematic.

Through the mapping of the COM-B elements to the qualitative data, the intervention functions of enablement and education were identified as appropriate to address all target behaviours. Environmental restructuring, persuasion and incentivisation were identified as functions to address programme attendance, and training was identified as a function to address physical activity and healthy eating. Modelling was also identified as a way to address the physical activity behavioural target.

From the parallel process of mapping the 46-item cultural adaptation typology [15] to the qualitative themes, several types of cultural adaptation and the stages at which they could be applied in the programme cycle were identified. This process ensured that there was explicit consideration of how adaptations to the programme were culturally appropriate to the target population. The qualitative themes, mapped COM-B components, intervention functions, cultural adaptations and programme cycle stage, and corresponding NICE guideline recommendations are presented in Table 4.

Following application of the two guiding frameworks to the qualitative data, specific adaptations were planned by two members of the research team (TG and MP). This planning was also informed by the research evidence and local programme information. Further consultation with the programme managers took place at this point so that they could comment on the feasibility of delivery of the planned programme. The specific adaptations are outlined in the right-hand column of Table 4. To further illustrate how the adaptation process was undertaken, an example of the process is given in Additional file 1. When the adaptation process was completed, the planned intervention programme was presented to the Parent Advisory Panel for feedback.

A summary of modifications made as a result of the adaptation process is provided below. The adapted intervention programme is reported in more detail using the Template for Intervention Description and Replication (TIDieR) checklist [37] (see Additional file 2).