Children's perspectives and experiences of health, diet, physical activity and weight in an urban, multi-ethnic UK population: A qualitative study

Background: Children from Black and South Asian ethnic groups are at risk for childhood obesity in the United Kingdom. To inform local action for childhood obesity prevention, it is crucial to explore the basis of ethnic disparities and consider the perspectives of children. This study aimed to understand cultural and contextual factors influencing childhood obesity in an ethnically diverse population using child-centred methodology. Methods: ‘ Draw, write and tell ’ interviews were held with children aged 9 – 10 years in Coventry, an urban, multi-ethnic city in the United Kingdom. Data were analysed thematically using framework analysis. Results: Twenty-six children participated (85% from Black or minority ethnic groups). Children's perspectives revealed universal themes around health, diet, physical activity and weight and highlighted issues specific to ethnic groups and those living in deprived areas. An underlying feature was weight-based stigmatization and group stereotyping, and an emphasis on internal factors as the cause of obesity. Children described some experiences of social disadvantage but did not regard these as a barrier to being physically active. Children identified cultural or religious practices or experiences of migration that influenced diet and physical activity. Conclusions: These findings allow a broad range of children's perspectives to inform future intervention design. In addition, the study was able to identify the many similarities and small amount of diversity in children's perspectives across ethnic groups.

genetic and physiological mechanisms, to socio-economic and structural barriers and facilitators, and factors associated with migration such as stress and acculturation (Murphy et al., 2017).
Few studies have explored the influence of parental migration status upon child adiposity in the United Kingdom, but there is some evidence that having a foreign-born mother explains some of the higher adiposity observed in children from Black ethnic groups (Martinson et al., 2015). In order to develop appropriate preventative services, local community-specific characteristics and contexts need to be understood. Qualitative approaches provide a means for exploring cultural and contextual influences upon food and physical activity behaviours, and beliefs related to body weight.
A recent systematic review of the qualitative literature found studies exploring the views of those from ethnic minority communities on childhood weight to be lacking the 'child's voice' (Chatham & Mixer, 2019). An earlier review on the perspectives of children towards body weight found methodological weaknesses in the included studies, for example, lacking features that privilege children's own framing of issues around obesity; giving little consideration to the role of ethnicity in sampling and analysis; and having a bias in sampling towards children from high socio-economic groups (Rees et al., 2009). To address ethnic and socio-economic inequalities in childhood obesity, it is imperative to understand the experiences of children from these high-risk groups.
Researchers have aimed to fill this gap in the evidence (Eyre et al., 2015;Rawlins et al., 2013). However, these studies were limited by use of focus groups and traditional interviews that do not fully account for the unique challenges of research with children (i.e., risk of group conformity, sensitive topics, wide-varying linguistic abilities and power imbalance). Child-centred approaches attempt to address these concerns.
The aim of the current study was to identify the cultural and contextual factors that influence childhood weight status in Coventry, England, through an exploration of child perspectives and experiences around health, diet, physical activity and weight.

| Study design
This study was undertaken within a multicomponent mixed methods study (Murphy, 2018). The first, quantitative phase sought to understand ethnic inequalities in obesity in primary school-aged children in Coventry, a large ethnically diverse city in the West Midlands of England, where deprivation is higher than the national average, and the proportion of school children from minority ethnic groups is 48% (Coventry City Council, 2018). The current qualitative study was conducted alongside another qualitative study to understand parents' perspectives and experiences of health, diet, physical activity and weight, with data analysed separately due to variation in study design (e.g., sampling approach and data collection methods).

| Sampling, recruitment and consent
Children aged 9-10 years were recruited through primary schools in Coventry, aiming for a sample size of 20 children. The sampling frame was designed to achieve an ethnically diverse sample, with high proportions of the largest minority ethnic groups in Coventry. Schools with a high total proportion of children from Black and minority ethnic (BME) groups (>Coventry average) and/or a high proportion of children having free school meals (FSM) (>Coventry average), based on school census data from the Local Authority, were targeted for recruitment. Of the 85 state primary schools in the academic year 2015/2016, 16 schools were invited to take part. In participating schools, all children in Year 5 (aged 9-10 years) were provided with child-friendly participant information packs informing them that a researcher (M. M.) was interested in their views on health but that they did not have to take part if they did not wish to. In schools with a high response rate, additional sampling was applied at the child level by prioritizing children from BME groups. Once children from these groups had been interviewed, children were interviewed sequentially in order of response forms received.
The researcher (M. M.) visited the school classes involved prior to the distribution of packs to explain the research in child-friendly terms and build familiarity with the potential participants. Written parental consent and verbal child assent were required. Prior to commencing the interview, children were provided with some options and phrases for ceasing the interview if they chose, and the digital recorder was demonstrated to the child, with the participant being invited to start and stop the interview by pressing the appropriate buttons. The interviewer (M. M.) checked participant understanding using two questions, adapted from Hensel et al. (2002): 'What will I be talking to you about today?' and 'what should you do if you don't want to talk to me anymore?'. This provided the opportunity to assess the participant's comprehension, with clarifications being given where children were unclear.
Parental information packs and topic guides were reviewed by a Patient and Public Involvement representative for the National Institute for Health Research (NIHR), and the pupil information packs and

Key messages
• Emphasizing the benefits of healthy behaviours for mental well-being is likely to be a motivating message for this age group, but messaging should be sensitive to avoiding the moralization of such behaviours.
• Incorporating peer-support components into 'healthy lifestyle' interventions may help to reduce the stigmatization of the overweight/obese body and provide moral support for children.
• Qualitative research with parents should be considered if seeking to culturally tailor healthy lifestyle interventions in ethnically diverse populations. data collection methods were piloted with four children from a school in Birmingham. These were subsequently adapted according to feedback and the researcher's reflections.

| Data collection
Child postcode and ethnicity (based on census categories) were collected through parental questionnaire. Child age and migration background were collected through an interview-administered questionnaire with children, using the Ethnic Background Indicators tool from the Health Behaviours in Schools-aged Children study (HBSC-EBI), validated for use in 11-year-old children (Nordahl et al., 2011).
Child perspectives were gathered through a 'draw, write and tell' technique (Angell et al., 2014) within a semi-structured oneto-one interview. This was selected as an engaging, participatory, familiar and non-threatening method of data collection providing a structured way for children to gently recall experiences and construct cognitively complex ideas, enhancing communication between researcher and child and supporting meaning-making. Children were first asked to draw a picture of a healthy child and then an unhealthy child. These images were then used as a launch pad for discussing children's views on health through the opening question 'can you tell me about your picture?'. The topic guide prompted the researcher to ask additional questions on diet, physical activity and weight when they were not raised spontaneously in the discussion. Children were advised that they could write on the paper and that they did not have to draw a picture or write if they did not wish to.
Interviews were conducted in a quiet and relatively private space (e.g., meeting room or community room) within the school during the school day or during breakfast club (before school). All interviews were conducted in English, and one child with limited English was supported by a class friend who provided translation when needed (with the agreement of both children). Drawn images, written text and discourse were captured as data. Field notes were taken to note emerging themes and challenges arising. Interviews were recorded on an audio digital recorder and fully transcribed verbatim through an external transcription service and reviewed by the researcher for accuracy and to develop familiarity.

| Data analysis
Data were analysed in NVivo v11 using a framework approach, based on the process detailed by Gale et al. (2013). Exploratory free-coding of a sample of transcripts was undertaken by one researcher (M. M.). The authors met to review codes and agree a coding framework. Once the framework was agreed, transcripts were systematically coded using this a priori framework (M. M.).
Drawn images and written text were coded concurrently with the transcripts. There remained scope to add new codes that arose from the remaining transcripts. Next, data were abstracted into a case-code matrix to support the identification and development of emerging themes. These emerging themes were studied and refined into final 'interpretative themes' that went beyond descriptive analysis by seeking possible explanations for what was happening within the data. The existing literature has been critiqued for a lack of attention to the roles of ethnicity and SES in the analysis of qualitative data. As such, we specifically set out to explore these factors, initially through interrogating the data for emerging themes relating to these a priori concepts and then through further organizing the case-code matrix into aggregated ethnic groupings and parental migration status (parents born abroad vs. parents born in the United Kingdom) to explore any diversity in perspectives. ID numbers have been assigned to verbatim quotes in the text below, enabling cross-referencing with the sociodemographic characteristics of each participant detailed in Table 2. Quotes that exemplify each sub-theme are provided in tables, whereas further quotes that reflect nuances in children's views and language are provided in the body of the text. Children's drawn images are also provided to illustrate the content of themes.

| Sample description
Six schools out of 16 invited agreed to participate, with children successfully recruited from three of these schools. Demographic characteristics relating to the proportion of children from BME groups and taking FSM, and recruitment numbers for these schools are summarized in Table 1.
Participant characteristics are provided in Table 2. Ethnicity is provided based on census categories. Where the parents of participants selected 'other', they were asked to specify their child's ethnic group in a free text box, and this self-description is detailed where provided.  Abbreviations: BME, Black and minority ethnicity; FSM, free school meals.

| Interpretative themes
Five sub-themes were identified and grouped into three broader 'universal themes': (1) conceptualizations of the 'healthy ideal'; (2) otherization of obesity; and (3) spheres of influence upon health behaviours. A summary of universal themes and sub-themes with example quotes is provided in Table 3.
3.2.1 | Universal Theme 1: Conceptualizations of the 'healthy ideal' Sub-theme 1: Health and happiness are intrinsically linked For children, health and happiness were viewed as intrinsically linked and operated bidirectionally; that is, being healthy makes you happy; and being happy makes you healthy. The converse was also that unhealthy or overweight children were viewed as unhappy; for example, 'they're really sad because they're fat' (ID3).
Happiness was seen to drive health behaviours by providing a sense of positivity and enthusiasm, which made healthy behaviours easier.  Figure 2). It was generally accepted that the unhealthy and overweight child would be stigmatized by others. Few children empathized with the unhealthy child, viewing children's unhealthy behaviours as akin to character flaws. However, one girl (ID3), who had previously described herself as 'a bit overweight', associated feelings of shame with the overweight body and empathized that these feelings would cause the overweight child to be too sad to be capable of losing weight. This child's beliefs suggest that children have the capacity to challenge stigma and that personal experience was key to this.
It's hard for a fat child to get healthy because they are scared that everyone will be like, ha, ha, ha, look he's so fat and your belly jiggles and your legs jiggle, and you might feel really ashamed of your body and then it's really hard. (ID3)

| Influence of ethnicity and SES upon beliefs
When exploring themes relating to the a priori concepts of SES and ethnicity, three additional sub-themes were identified. The subthemes and example quotes are provided in Table 4, with counts provided to indicate the number of children referring to each concept.
3.3.1 | Sub-theme 6: Socio-economic circumstances and neighbourhood deprivation References to the influence of socio-economic circumstances generally came from occasional mentions of the home environment (e.g., living in a block of flats) and material disadvantage (e.g., not owning a car), which were both seen as enabling physical activity.
Some implicit references to neighbourhood deprivation were made (e.g., descriptions of crime and safety concerns; quality of the local environment; and the number of takeaways on the high street), which were viewed as negatively influencing health behaviours. Overall, despite exploring SES as an a priori concept, it did not emerge as a substantial issue in the experiences described by children.

| Sub-theme 8: The influence of religious practices upon diet and physical activity behaviours
Although three children described specific cultural or religious celebrations or commitments related to food and use of time, these children did not express that these factors influenced their health behaviours. Two children talked about specific foods eaten during Eid celebrations, and these foods were a combination of items typically viewed as healthy (e.g., fruit) and unhealthy (e.g., fizzy drinks). of their definitions of health to operate bidirectionally as both a motivation for driving healthy behaviours and a resource for enabling healthy behaviours.
The current study found weight-based stigmatization and group stereotyping, and an emphasis on internal factors as the fundamental cause of obesity (i.e., fundamental attribution error), to be common place in children, and this appeared to form the basis for how children framed health conceptually. Similar findings have been identified by other qualitative researchers (Rees et al., 2009;Trigwell, 2011); however, the current study suggests that these negative value judgements are a result of the ways in which children moralized health as 'right' or 'good' and poor health as 'wrong'.
Other researchers have found fundamental attribution error over the causes of obesity to be common in adults (Sikorski et al., 2011), and the current study goes further to show that these views also predominate in children's beliefs. This may be a result of children simply reflecting/reinforcing the beliefs of adults (and society) around them. On the other hand, it may be that well-intended messages promoting the benefits and attainability of healthy lifestyle behaviours for optimal health, weight and happiness may inadvertently reinforce a contrasting view that those who do not fit this ideal have in some way 'failed'. This failure is seen to justify stigma and ridicule relating to weight status.
The findings also suggest that the negative value judgements assigned to overweight status were permitted by a distancing of childhood obesity from children's own lives and the 'otherization' of obesity. This was driven by children's unrealistic perceptions of the 'obese body' and the deflection of the reality of how overweight and obesity appear in the general population (or even in the children themselves). This creates a difficult scenario for those working with children. The distancing of the 'obese' body is, in some ways, potentially protective against the development of preoccupation with one's weight/body dissatisfaction and the negative effects this brings (Neumark-Sztainer et al., 2006). On the other hand, this distancing may enable greater stigmatization of others, the result of which is often social marginalization or diminished social capital (Strauss & Pollack, 2003), which may further restrict the child's ability to undertake healthy behaviours (e.g., playing with friends). The finding that stigma is both a product of and a driver of the issue has also been found in relation to depression (Boardman et al., 2011), in which the authors propose consideration of how stigma interacts with coping strategies and resilience during treatment.
A surprising finding was how children perceived aspects of their socio-economic circumstances typically viewed as barriers to physical activity (i.e., living in a flat and not having access to a motor vehicle) as enablers. This is a novel finding in children, with other research reporting only barriers related to children's socio-economic circumstances. The novelty of this finding may be because the current study sought the perspectives of children directly and recruited exclusively from deprived neighbourhoods, whereas other studies seeking to understand the role of SES upon weight-related behaviours have tended to come from researcher observations or comparative approaches rather than verbalized by children themselves (Backett-Milburn et al., 2003;Eyre et al., 2015;Irwin et al., 2007;Lofink, 2012;Pearce et al., 2009). This surprising finding also contradicts research with parents, who view living in flats and not owning a motor vehicle as barriers to physical activity (Trigwell et al., 2015)-suggesting that children's lived experiences of these familial circumstances may be quite different to those of parents.
In the current study, children identified very few barriers or facilitators to achieving a healthy weight relating to their ethnicity, religion, country of origin/parental country of origin or experiences of migration. This is surprising because it contrasts with research with parents, who highlight many ethnicity-specific influences upon diet and physical activity behaviours and perspectives towards weight than children (Cross-Bardell et al., 2015;Eyre et al., 2015;Pallan et al., 2012;Rawlins et al., 2013;Syrad et al., 2014;Trigwell et al., 2014;Trigwell et al., 2015). It may be the case that in early childhood, culturally linked experiences related to ethnicity do not generally manifest-perhaps children do not recognize these as cultural behaviours or do not register them as important enough to mention in discussion.

| Strengths and limitations
The use of a draw, write and tell technique allowed the child participants greater freedom and control to highlight issues with salience for them, achieving a potentially more valid understanding of their meanings described in their own ways and words (Noonan et al., 2016).
The study makes a significant contribution to the gap in the existing literature highlighted by Rees et al. (2009) by privileging children's views via child-led methods.
Studies of children's perspectives have been criticized in the past for their inattention to sampling and analysis, especially relating to ethnicity and SES (Rees et al., 2009;Thomas et al., 2003). This study aimed to focus on ethnicity and socio-economic circumstances a priori, thus incorporating these factors from the outset. However, although an ethnically diverse sample was achieved (85% from non-  (Adamson & Donovan, 2002), and this could also have resulted in a failure of the researcher to pick up on some important elements related to ethnicity because of a position as an 'outsider' (Ochieng, 2010).

| Implications
This research provides an insight into the salient issues for achieving a healthy weight for primary school-aged children living in deprived areas in Coventry and so provides the basis for developing community-centred preventative services that take children's perspectives into account, for example, providing strategies for being more physically active within the home. Emphasizing the benefits of a healthy weight and healthy behaviours for mental well-being is likely to be a motivating message for this age group, and programmes that improve mental well-being may help children build the personal capacity for achieving behavioural changes. Messages that include a 'peer-support' component may help to counter the social marginalization of those with unhealthy behaviours. This should be supported by messages that discourage weight-based bullying and educational components that counter the prevailing idea of the overweight child as a morbidly obese, ridiculed figure. All such messages should sensitively consider the way in which a healthy weight is portrayed and personified to avoid the unintended consequence of reinforcing unhelpful weight-based stereotyping and stigmatization.
There appeared to be little diversity in children's perspectives and experiences across ethnic groups, so support for the cultural tailoring of intervention messages and components (Liu et al., 2012) is lacking in the current findings. However, the potential influence of culture upon children's obesity-related behaviours should not be dismissed on this basis, because parental perspectives are also important to consider. The literature suggests that ethnicity is an important factor in parental perceptions of a healthy weight and parental beliefs and experiences relating to a healthy diet and physical activity; hence, local intervention development should also be informed by qualitative research with parents.
Recommendations for future research include incorporating faith and weight status into qualitative sampling frames, in order to explore how these important factors interact with ethnicity to influence childhood adiposity. Additional public involvement, considering the ethnic diversity of public representatives and incorporating school leaders, may support recruitment and interpretation of data.

| CONCLUSIONS
This study has contributed to the evidence base by exploring children's perspectives on health, diet, physical activity and weight in an ethnically diverse, deprived population, using methods that privilege children's views. The findings are valuable in allowing children's views to inform the design of future childhood health promotion initiatives, suggesting that obesity prevention and weight management interventions would benefit from focusing on the existing assets that children describe, such as friendships, and emphasizing health benefits framed by child-based motivations, for example, happiness and mental well-being. In addition, the study was able to identify the many similarities and the small amount of diversity in children's perspectives and experiences across ethnic groups, filling a gap in the literature and contrasting with some of the existing literature.