1.1 Obesity: Defining the problem

Obesity is a global public health issue. Carrying excess body fat increases an individual's susceptibility to diseases including cardiovascular disease, type 2 diabetes, respiratory conditions, muscular disorders and a range of psychological issues.¹ More recently, obesity has been identified as a key risk factor for hospitalization and death due to COVID-19 outcomes.^2-4 British South Asian and other economically deprived populations have exceptionally high rates of obesity compared to the wider population.⁵ For instance, the prevalance of obesity is the highest in the most deprived 10% of the population—approximately double that of the least deprived 10%.⁶ The prevalence of obesity-related disease, for example, type 2 diabetes, is reported to be up to six times higher in South Asian populations than in the wider population.^7-9 Despite this, South Asian communities are less likely to follow a healthy diet or engage in exercise compared to the general population.¹⁰ Among the British South Asian population, the rates of obesity are higher in some subgroups, such as British Pakistani women (PW).^{11, 12} Bradford has a strong British Pakistani presence¹³ and is an economically deprived city in the north of England, which has higher obesity and Type 2 diabetes rates than the UK average.^{14, 15} These rates are pronounced in more deprived areas of the district where there is a large British Pakistani community.¹⁶

It is unclear whether efforts to reduce obesity in South Asian populations have been successful. For instance, one intervention positively changed the physical activity behaviours and attitudes of individuals from low socioeconomic groups living in deprived inner-city areas of the United Kingdom¹⁷; yet, the ethnic profile of these individuals was not made explicit by the authors. Obesity reduction interventions specifically targeting British South Asian communities are scarce and the evidence from these interventions is mixed.^18-20 It is more generally argued that for weight loss interventions for South Asian populations to be effective, they need to be culturally sensitive.²¹ There is an urgent need to address obesity in British PW. Development of a research agenda that has incorporated the perspectives of this population based on their lived experience could be one way to reduce the exceptionally high rates of obesity by developing obesity interventions and projects that more accurately reflect their obesity health needs.

1.2 Coproducing research agendas with ethnic minority populations

Research agendas have historically been determined by health service providers and funders.²² Debates continue on how to include a diverse range of voices in research design and delivery as part of the patient and public involvement.²³ Marginalized groups are often uninvolved in research prioritization exercises and when they are involved, sample sizes are small and exercises are poorly conducted.²⁴ Consequently, research questions that matter to people may be unheard,²⁵ leading to failed interventions and alienation of already marginalized communities. This is especially concerning, given that these populations have disproportionately poorer health outcomes and can encounter negative experiences of services.²⁶ Partnerships between professionals and the public to collectively identify and prioritize research by facilitated debate and formal decision-making methods have been developed, such as the James Lind Alliance.²⁷ However, holding consensus groups with underrepresented Black and minority ethnic individuals and professionals together may advantage professionals, leading to the disengagement of disadvantaged groups that are underrepresented in research priority setting.^{28, 29}

Although research prioritization exercises in obesity do exist,³⁰ no research could be found identifying the obesity research priorities of PW or South Asian women more generally. A research priority exercise involving South Asian children, young people and their families sought to prioritize health topics requiring more research according to this population.³¹ Nutrition, obesity and physical activity ranked highly as priority topics requiring further research. The objective of this study, therefore, was to produce an obesity research agenda that is reflective of PW's unmet needs, with added input from a range of professional stakeholders.

2.1 Study design

Feminist participatory action research (FPAR) guided this study. FPAR provides an opportunity for women to voice their experiences of health and define health issues pertinent to them.³² This study is part of a wider research project influenced by earlier stages of the project (see Figure 1). In Phase 1, a bottom-up approach was used so that PW themselves could direct the research agenda by identifying the topic area in which they had unmet needs. PW identified obesity as the priority topic area in which to develop a research agenda by identifying obesity-related concerns and unmet needs (under review). Phase 2 gathered additional information on the issues that PW had concerning obesity to provide a more comprehensive picture of their obesity health concerns and unmet needs (linked article). The objective of Phase 3, reported in this article, had two stages. Part 1 was to generate additional input from multisectoral stakeholders (MS) as it is good practice in research priority setting to involve a range of stakeholders in the process.³³ Part 2 was a ranking exercise, which only involved PW, to determine their top 10 obesity health priorities to develop an obesity research agenda. Part 2 was organized deliberately this way to address issues of power and to ensure that women's voices were not silenced to align with the principles of FPAR, which aims to transform spaces in a manner that promotes authentic power sharing and opposes gendered, racialized and class power relations.³⁴

2.2 Survey of MS (Phase 3, Stage 1)

To align with quality research priority setting exercise practices,³³ a survey was developed to obtain the perspectives of a wide range of MS with an interest in obesity in this population. Survey development was based on a previous survey that identified the research priorities around what makes children happy and healthy in Bradford with a range of stakeholders (unpublished survey). H. I. had assisted in the development of that survey, and it had been screened by a core group of researchers to ensure suitability for determining research priorities from a range of stakeholders. Respondents were asked to list their three most significant research priorities (an unmet needs or topic area) that should be addressed in the area of overweight and obesity in PW (File S1).

2.2.3 Statements

A set of statements had been previously developed by PW from the earlier stages of the project. Also, the findings from the survey in the present study were translated into statements, to be ranked by PW in the final phase of the study. A total of 31 statements were categorized under five domains: diet; physical activity; accessing information; mental health; and other. A ranking form was developed containing the statements (see Figure 2).

2.3 Ranking exercise with PW (Phase 3, Part 2)

Only PW were invited to participate in the ranking exercise. This was to avoid difficulties that may arise conducting a joint prioritization exercise between professional stakeholders and PW, such as the stifling of women's voices due to power dynamics. Participants were also recruited as participant researchers as FPAR advocates women's involvement in data collection^{36, 37} to share control of the research processes with participants.³⁸

3.1 Phase 3: Survey to MS

A total of 159 multisectoral professionals working in Bradford completed the survey. Respondent demographics are shown in Table 2. Respondents were largely female, from Black and minority ethnic backgrounds and worked in the healthcare sector. Under profession, the ‘other’ option was selected by academics, researchers, imams and local authority workers.

Content analysis of open-ended responses identified three main themes: health behaviour barriers, education needs and mental well-being. Many of the responses did not indicate research priorities, yet were observations made by respondents, requiring more attention. The topics covered under each theme are shown in Table 3.

3.2 Phase 4: Ranking exercise results

Thirty-two women completed the ranking forms. Participants lived in the most deprived wards of Bradford, which fall within the 10% most deprived wards in England according to the UK Indices of Deprivation data.¹³ They were aged between 18 and 69 years and were mostly English-speaking. Three were proficient in spoken Urdu and could also read Urdu. See Table 4 for participants' demographics.

Table 5 below shows a summary of the 31 obesity statements ranked in order of priority. The ranking score was calculated in the first right column next to each statement. The second column shows whether the statement was identified by PW in Phase 1/2 or both, by MS in the present study or both women and stakeholders. The third column shows the number of participants that selected the statement. Some statements were selected by more participants than those that were ranked higher but were assigned a lower score, thus ranking lower overall. There were large discrepancies between statements that scored high and those that scored low. Four of the top 10 statements belonged in the domain of mental health, two statements in the domain of physical activity, two in the domain of diet, one was within the domain of accessing information and one statement in the list did not belong to any domain. The statements are ordered in terms of priority based on their ranking score.

4.1 Summary of the findings

A range of topic areas and unmet needs were identified by stakeholders. Many of the findings, however, were not research priorities, yet were observations as to why PW struggled to manage their weight and what they felt was needed by this population to overcome obesity. This may have been because some survey respondents may not have known what was meant by research priorities as was reported in other research priority setting exercises involving stakeholders from a nonclinical background.^{31, 43, 44} Findings concerned healthy behaviour barriers such as dietary, physical activity, language and interpersonal barriers, the need for obesity-related dietary education, physical activity and obesity health risks and issues concerning mental well-being. The main findings from the survey were translated into statements, to be ranked by PW. Greatest priority was given to statements concerning the need for education around mental health, physical activity and diet.

4.2 Similarities and differences between PW and MS

Similarities and differences could be found between topics identified by stakeholders and those by PW. Notably, the four statements identified by stakeholders around the reasons why women did not engage in healthy behaviours were not ranked highly by women. This could be because women may already feel that these barriers are well known. Surprisingly, the statements relating to the high cost of food and facilities being too expensive to use were not in the prioritized list, despite the women living in deprived areas. This could either be because PW did not perceive finances to be a barrier in engaging in health-promoting activities, or because they felt that other concerns were more pressing. These differences highlight the importance of involving underrepresented communities in research priority setting whose views are traditionally not considered.

PW assigned the highest rankings to four of five available statements surrounding mental health. This was surprising, given that two of the mental health-related statements were not identified by PW in the earlier studies, yet were instead identified by stakeholders in the present study. This demonstrates the importance of involving a variety of different stakeholders and sources in health research prioritization to ensure that a range of perspectives are incorporated into the process, advocated by the research priority setting literature.^{33, 45} PW in earlier studies of this project may not have considered mental health to be significant in relation to obesity, but recognized its importance when they saw it on the ranking form. Many of the stakeholders were female and from Black and minority ethnic backgrounds, reflecting the diverse ethnic profile of Bradford. This may have influenced the findings, as these respondents may have had more insight into issues encountered by PW living in deprived areas of Bradford.

4.3 Social determinants of health

The findings from this study demonstrate that obesity and overweight extend beyond the responsibility of the individual and the multifaceted layers of influence that impact an individual such as socioeconomic status, culture, religion, family influence and community networks need to be considered. Health inequalities are not always considered in obesity research, and to ensure successful intervention design, understanding the socioeconomic situations of those with obesity is essential,⁴⁶ as is the need to address structural barriers to bring about behaviour change. Dahlgren and Whitehead's⁴⁶ rainbow model provides a useful framework for exploring the interaction between the determinants and their influence on obesity in PW living in deprived areas of Bradford. Our findings can be contextualized within the layers of influence reported by the model, especially individual lifestyle factors, social and community networks and living and working conditions.

In the model, individual lifestyle factors relate to behavioural risk factors, which are potentially modifiable, such as low uptake of physical activity in PW and poor dietary habits, which are known contributors to obesity.⁴⁷ It is argued that there is an element of choice in becoming obese.⁴⁸ Our findings, however, suggest that these women do not have much choice in managing their weight due to the structural barriers that they encounter such as healthy food and exercise facilities being expensive, unsafe neighbourhoods for physical activity and a lack of culturally appropriate dietary advice.

The second layer is the collective social context, that is, mutual support from family, friends, neighbours and the local community, which are external to the individual, yet influence the level of health experienced in that population.⁴⁶ International evidence suggests that family social capital, that is, positive close relationships with household members, can have health-promoting aspects.⁴⁹ Our findings suggest that PW interact with their family, peers and community, which, in turn, influences their health behaviours. Prioritized interpersonal barriers included the need to ascertain whether PW's role in the family affects their mental health, and support needed for women who do not leave the house. However, it has been previously reported that in deprived areas of Bradford, PW are more likely to report low family social capital compared to White British women.⁵⁰

The third layer of the rainbow is an individual's ability to maintain their health and is influenced by living and working conditions, food supply and access to essential goods and services.⁴⁶ A range of structural issues were identified by survey respondents that made it difficult for women to manage their weight, and were ranked as important by PW. These included accessing weight management services for non-English-speaking and socially isolated women as well as the prioritized statements on education.

4.4 Strengths and limitations

A strength of this priority setting exercise is the incorporation of views from professional stakeholders across different sectors, including religious leaders and fitness professionals, who, to our knowledge, have previously been uninvolved in research prioritization. Another strength is that PW were solely involved in the ranking exercise rather than a joint prioritization exercise with other stakeholders, which could have stifled their voices. Instead, they were able to rank the statements without any pressure to conform. Also, PW translated the obesity health concerns and obesity unmet needs into research priorities rather than members of the research team. This reduced the risk of statements being misinterpreted. Lastly, by recruiting non-English-speaking women, we aimed to ensure that this study would be applicable to the wider British Pakistani population.

Limitations include the top 10 list, which could have been different had the priority setting exercise been set up differently such as disagreements bring resolved through facilitated debate and lists prioritized based on consensus.²⁷ However, the COVID-19 pandemic lockdown made it impossible to hold a face-to-face consensus exercise. Another limitation is that scores between some of the ranked statements were marginal, which meant that a statement could have switched positions in the list had one participant ranked differently. This needs to be considered in future research priority setting as there are implications involved in relying on prioritized lists to develop projects or interventions. Also, the researcher was not present during all 32 ranking sessions. As a result, there was uncertainty around the manner in which the participant researchers may have influenced the ranking, despite being trained on how to facilitate the exercise. Lastly, depth was lacking in some of the survey responses, making it difficult to draw conclusions from what was stated. Therefore, future research should be conducted with PW to further understand how these priorities are interpreted.