Abstract

Introduction. This article is a meta-ethnographic analysis of qualitative studies to explore complex social ecological aspects (individual, microsystem, mesosystem, exosystem, and macrosystem) of physical activity among socioeconomically disadvantaged groups living in industrialized countries. Method. Using MeSH keywords, we searched major electronic databases including Medline, EMBASE, CINAHL, and PsycINFO in order to identify relevant publications published between January 2000 and October 2018. We included 19 qualitative studies which met inclusion criteria and were focused on physical activity determinants among adults. Results. Determinants emerging from these studies were grouped into six themes: (i) urban environment, (ii) financial constraints, (iii) work-life integration, (iv) community engagement, (v) social support, and (vi) psychosocial factors. After conceptualising these six themes into a social ecological model, we identified potential research gaps for physical activity among adults with low socioeconomic status living in industrialized countries. Conclusion. Our major insight was that, in industrialized countries, physical activity overlooks potential strengths to maintain health and well-being of those people with low socioeconomic status. A more complex understanding of contradictions between positive and deficit frames would lead to more critical insights of research gaps of physical activity in adult population with low socioeconomic status.

1. Introduction

Physical inactivity is increasingly recognized as one of the leading causes of mortality worldwide [13]. Evidence shows that, annually over 5 million people worldwide die due to low or insufficient physical activity (PA), which accounts for 6% of global deaths [3, 4]. Insufficient PA can impact considerably on health and productivity and is a cause of many chronic diseases such as coronary heart disease, type 2 diabetes, breast cancer, and colon cancer [1, 3, 5, 6]. Regular PA is a protective factor not only for leading chronic diseases [4] but for a range of important disease risk factors such as hypertension, overweight, and obesity [7]. Recently published evidence shows that there is still over a quarter of the global population who do not undertake recommended levels of PA, and this proportion is higher among women than men [8]. In 2018, the World Health Organization launched its’ Global Action Plan on Physical Activity (GAPPA) 2018–2030 emphasizing “More people active for a healthier world,” which aims to achieve reduction of 15% in physical inactivity levels by 2030 [1]. The targets and policy recommendations of GAPPA are envisioned to contribute to the achievement of a number of Sustainable Development Goals (SDGs) [9].

PA varies across sociodemographic groups; not all adults have equal access to PA opportunities as part of their daily lifestyle [1012]. People of low socioeconomic status are more likely to have poorer health and shorter life expectancy than people of higher socioeconomic status [10, 12], attributed in part to a lower prevalence of PA [13]. Models for PA promotion identify individual factors such as age, sex, health status, self-efficacy, and motivation [14, 15]. Health promotion models also recognise the role of contextual influences and the effect of interactions between individual, social, and physical environmental factors [1618]. The relevance of these models extends to mass media campaigns for PA, where message design processes need to be supported by environmental and policy actions that enable and reinforce the adoption of behaviour change [19].

Whilst the literature commonly focuses on socioeconomic status related disparities in PA [13, 20, 21], it should be acknowledged that culture, gender, disability, and a host of other sociodemographic factors intersect to impact movement opportunities in different contexts [22]. Therefore, the focus on the term “socioeconomic” in the current literature, particularly when it is defined by rigid economic measures, may have excluded relevant studies. The intent of this investigation was to explore broadly, across contexts, how qualitative literature has elucidated the relationships between socioeconomic disadvantage and PA.

The Social Ecological Model (SEM) has been well recognized worldwide and been used broadly in health sectors [2325] including for the improvement of PA among a range of populations and in different settings [23, 2629]. The SEM emphasizes understanding the multifaceted and interactive effects of multiple factors, including those that are personal and environmental, upon behaviour [18, 23, 30]. Studies have emphasized the need for understanding and social and contextual correlates in order to ensure appropriate use of SEM for achieving PA outcomes [23, 2629]. However, the SEM has been critiqued for a lack of “sufficient specificity to guide conceptualisation of specific problems or to identify appropriate interventions” [31]. Bauman et al. argued that SEM can evolve to become more context specific when evidence generation is planned for the purpose of creating more focused structural models [14, 15]. When the SEM becomes context specific, it can better “expose and account for complexity of sociocultural and environmental effects” to guide coordinated interventions [32]. The disparities in PA across the sociodemographic spectrum have been explored using both quantitative and qualitative research methods. Meta-analyses of quantitative studies concerning PA within low socioeconomic groups have provided important insights to further explore this phenomenon [13, 33].

Evidence suggests that disparities exist across personal, social, and environmental determinants of PA, which contribute to social disparities in the achievement of PA targets [20, 34]. The available evidence provides a compelling case for paying close attention to socioeconomic and cultural disparities in formulating policies and developing intervention approaches for PA promotion [2, 35, 36]. To explain PA inequalities between socioeconomic groups, SEM can take an account of the socioeconomic and cultural backgrounds of population subgroups with regards to their decision-making [37].

A number of philosophical approaches to synthesizing the qualitative data have been in practice [38, 39]. Some are based on analysis methods used in primary research, and most use either integrative or interpretative approach to synthesizing qualitative evidence base [38, 39]. Meta-ethnography is an interpretive approach originally developed by Noblit and Hare [40], and data from primary studies are synthesised to achieve new conceptual understandings or produce new models or theories [41]. The synthesis approach has the potential to advance level analyses to generate new evidence base, find a new research question, and reduce duplication in research [42, 43]. This approach has been used in a range of sectors including public health [44, 45]. Meta-ethnography is an opportunity for qualitative research to expand on theoretical approaches to the promotion of PA involving individual, social, and environmental influences [17, 46]. To this end, our philosophical approach was interpretative in nature and sought to be objective and systematic by applying key criteria to ensure reliability and validity of our interpretive inferences [47, 48].

The aim of this study was to undertake a meta-ethnographic analysis of qualitative studies to explore complex social ecological aspects of PA among socioeconomically disadvantaged groups living in industrialized countries. We explored and summarised the breadth of qualitative findings across contexts, to capture an overarching social ecological account of what qualitative research have determined in relation to socioecological determinants and to use this to highlight strengths and gaps in PA research. The intent was for future research to benefit from a broad summative understanding of what is known qualitatively about individual, social, and environmental influences on PA across complex societal systems (cultures, countries, ages, and settings).

2. Methods

2.1. Study Design

In this study, we adopted Noblit and Hare’s seven-step approach to synthesise findings from qualitative studies into a social ecological frame [40]. The seven steps included: getting started; deciding what is relevant to the initial interest; reading the studies; determining how the studies are related; translating the studies into one another; synthesizing translations, and expressing the synthesis. We categorized the seven-step process into three major stages including (i) selecting studies, (ii) synthesizing translations, and (iii) presenting the synthesis. Firstly, we identified the research interest and selected articles that were closely related to our initial interest. Secondly, we assembled all the studies together and determined how they are related to each other. We then translated the studies with each other based on a comparative approach. In addition, we determined and synthesized the commonalities and differences between each account and derived a new framework that not only maintains the central concept of individual interpretation but also reveals a more comprehensive explanation in comparison to what each part alone implies. Thirdly, we presented our new metaphor (i.e., proposed social ecological model for PA) in the form of a diagram to facilitate understanding of PA among the disadvantaged population groups.

2.2. Search of Relevant Literature and Study Selection

We conducted searches of four major online databases including Medline, EMBASE, CINAHL, and PsycINFO using MeSH keywords for Medline and CINAHL, subject headings for EMBASE, and a standard keyword search for all four databases. The key search words and their combination were based on the study’s aim. MeSH keywords and subject headings used search terms based on “physical activity,” “determinants,” and “socioeconomic status.” The final search was as follows: [[“Physical activity” [MeSH/Subject heading] and “Determinants” [MeSH/Subject heading] and “Socioeconomic status” [MeSH/Subject heading] and “qualitative”] or [“Physical activity” [Standard keyword] and “Determinants” [Standard keyword] and “Socioeconomic status” [Standard keyword] and “qualitative”]] for each individual database. As MeSH and subject heading words change over time, replication of the study should commit to the three concepts “Physical Activity,” “Determinants,” and “Socioeconomic status” when selecting various MeSH and subject headings.

2.3. Inclusion/Exclusion Criteria and Study Selection

Studies were included in the review based on the following criteria: (i) they focused on PA among socioeconomic disadvantaged communities, (ii) were qualitative in design, (iii) focused on adults (>18 years but <65 years), (iv) were published in English, and (v) were published between January 2000 through October 2018. The decision to include or exclude a study was made using an EndNote database in a three-phase process after removing duplicates: first reading the title, then the abstract, and finally the full text. Additionally, references for each of the studies were screened manually to identify any additional articles that were not detected by our database searches.

2.4. Screening Process and Data Extraction

The initial screening of articles was undertaken by HQ, and final check was done by LR. The list of references of the selected articles was manually checked for any missing articles. The data were independently extracted by HQ and were final checked by LR using a piloted form that included research aim, authors’ details, year of publication, country, sampling method, sample size, population ethnicity, data collection methodology, health topics discussed in addition to PA, views expressed by study respondents (first-order constructs), the interpretations of these views by study authors (second-order constructs), and limitations.

2.5. Thematic Analysis

The process for thematic analysis was derived from the study by Malpass et al. [49]. Both reviewers (HQ and LR) read twice over the articles chronologically to identify emerging themes. The identified themes represented commonalities between study authors’ interpretations of relevant data across individual articles. A data extraction form was developed to record the views expressed by study respondents (first-order constructs) and the interpretations of these views by study authors (second-order constructs) for each individual article. The reviewers then independently consolidated the first- and second-order constructs of the individual articles into a summary definition (translation) to produce a third-order construct. Following this, reviewers compared their summary definitions and their respective second-order constructs and worked collaboratively to synthesise them into third-order constructs.

The constructs were further reviewed by members of the research team (LR, BS, and AR) to provide new perspectives on third-order constructs. This process facilitated a consensus about a “line of argument,” a technique to synthesise translations [44]. The reviewers used third-order labels—labels are names of a group of constructs—as suggested frames for crafting a “line of argument.” The “line of argument” explains the rationale behind the suggested third-order label; different reviewers will craft different variations of a “line of argument.” Suggested third-order labels were located within one of the individual, microsystem, mesosystem, exosystem, or macrosystem layers. The key articles most influential in the construct of the summary definition of the first- and second-order constructs are bolded in Table 1.

2.6. Critical Appraisal

The quality of the studies included in the review was assessed using a checklist for assessing reporting standards of qualitative studies, derived from the “Consolidated Criteria for Reporting Qualitative Research (COREQ)” [50]. Nineteen checklist items relevant to the quality assessment, particularly for this study purpose, were selected as dimensions to form the criteria. Examples included (1) discussion of reflexivity, (2) statement of methodological theory that underpinned the study, and (3) reporting how participants were selected. The dimensions were scored as 0 if not present and 1 if present. The reviewers cross checked their independent assessments before reaching consensus on reporting standard scores. The maximum quality scores sum up to 19 points, and based on the discussion within authors and agreement up on, the quality levels were categorised into three groups: high (14–19); medium (7–13); and low (0–6) (Table 2). Effort was made to assess the quality of the studies in terms of how researchers have presented their thoughts and understanding to the topic based on their critical reflection of the subjective observation.

The qualitative data processing and analyses presented in this study adhered to the Standards for Reporting Qualitative Research (SRQR) [51], and strategies were employed to enhance the trustworthiness (credibility, transferability, dependability, confirmability, and transferability) of the study findings [52, 53]. This included checking the data for accuracy, organising meetings for completeness of information (HQ and LR), using team meeting for inclusion/exclusion consensus, and providing adequate information about the studies included (see Appendix-A, SRQR checklist as a supplementary document).

3. Results and Discussion

3.1. Description of Studies

Nineteen articles were included for synthesis. Their characteristics are shown in Table 3. Eleven studies were conducted in the United States, three in the United Kingdom, two in Australia, two in the Netherlands, and one in Canada. Fourteen studies used focus groups, and five used in-depth interviews. The age of participants ranged from 14–89 years. The age range of participants in some studies was not clearly defined and included adolescents and elderly people, as well as adults.

3.2. Description of Themes

Our metasynthesis of first- and second-order constructs generated twelve third-order constructs and six primary themes/third-order labels (Table 1). The primary themes have been organised within a social ecological model that is shown in Figure 1. We ordered the primary themes from broadest (macrosystem) to narrowest (individual) in the description to help visualise authors’ analysis.

3.3. Urban Environment: Macrosystem to Mesosystem

Participation in PA is influenced by accessibility to resources that support PA. People desire to travel to parks, public sports centres, or local private health clubs to pursue PA but lack access to reliable transportation such as private cars, public transport, and bicycle infrastructure [54, 56, 57, 60, 69]. In addition, the availability of peers in close proximity and the location influences walking and other forms of PA for transport [63]. If parents have the means to reach PA facilities, they would need more convenient opening times that take into account working hours and make childcare services accessible [69]. The recurring theme of lack of reliable transportation and transport facilities within close proximity may arise from urban planning.

The effect of poor urban planning is further compounded by the aesthetic feel of the neighbourhood environment. People consistently reported being exposed to feelings of neglect and depression [54, 56, 59, 61, 70, 71]. In one study, mothers reported being unable to overlook the poor state of the environment or escape the sadness of living in a place that has seemingly been left to decline [54]. Other studies highlight the importance of well-maintained paths [55, 59, 66, 71, 72], such as sidewalks with sufficient streetlights and properly paved and resurfaced roads [59]. This, along with parks, make the neighbourhood look green and inviting [61] rather than vandalised and decaying [66].

The suitability of a place for PA also takes into account perceptions of neighbourhood disorder, particularly in relation to social crime and violence. Analysis reveals that the perceived crime rate of a community can significantly affect how people use community resources for recreation. Issues related to crime and violence as reported, included the prevalence of drug use [56, 58, 59, 65, 69], police presence [58, 59, 66], muggings and homicide [59], gangs, bullying and abduction by strangers [65, 70, 71] and gunshots, and vandalism [66], especially within dark unlit areas [61, 66, 69]. Noncriminal environmental stressors in the neighbourhood included accounts of traffic-related issues such as dangerous roads and careless drivers [54, 59, 65, 66], stray dogs [55, 59, 66], poor street lighting [5557, 59], homes being situated on highways and lack of sidewalks [57], broken glass on the ground [65], and the presence of dirty needles and cigarettes [69].

3.4. Financial Constraints: Macrosystem

PA participation is often affected by whether people can afford the fees and resources required. Sports equipment and sports club contributions are considered high expenditures, especially by single mothers and people who rely on social payments for day-to-day living [61]. Refreshments [64] and childcare services [64, 65, 69] are further costs incurred by parents. Parents report having to prioritise spending on their children for sports [64, 69], which means forgoing expenditure on themselves [65]. As such, demand for low-cost sports facilities and gymnasiums is evident across multiple studies [61, 64, 65, 71] with one study highlighting that cost is the number one cause of physical inactivity for a number of people [61]. There is recognition of the potential value of low-cost PA facilities such as subsidised gyms [56, 64, 6971] and opportunity to walk and run in the local area [64]. However, the complexities of discounted fees were revealed with the finding that even reduced cost membership contracts may be too constraining to access the PA services [64].

3.5. Work-Life Integration: Exosystem

PA participation can be affected by the nature of the time constraints upon socially disadvantaged adults. Inflexible work hours [56, 63, 66, 67, 72] and constant family responsibilities [64, 66, 70] are major factors that also contribute to feelings of exhaustion [56, 62, 69]. Put simply, time and energy expenditures are prioritised for work, school, and family, and PA is perceived as an additional burden. It is evident that there are social differences in the scope that individuals have to balance work-life demands and that coping mechanisms are exhausted by personal responsibilities.

3.6. Community Engagement: Mesosystem

PA interventions have had limited impact due to poor design. It has been reported that conventional marketing of PA programs has been hindered by delivery in a narrow range of languages and with limited resources devoted to communicating about local activities [64, 69, 71]. Although lack of resources for multilingual services may prevent tailoring PA activities effectively [66], friendly nonprejudicial and socialisation environments can build trust within the communities [69]. This highlights that community engagement strategies should be tailored to meet the needs of culturally diverse groups.

3.7. Social Support: Microsystem

PA is heavily influenced by encouragement from a person’s main sources of social support such as friends, parents, family members, coaches, and health professionals [63]. However, lacking an exercise companion is not always a barrier to PA because some people prefer to exercise alone [56, 70]. People have mixed opinions about exercising with co-workers during workplace funded fitness programs [64] further complicating interpersonal relationships in group fitness programs and how they affect the social experience. Evidently, social support from community networks is a motivator to engage in PA because of the encouragement and sense of security that it provides [59].

It is reported that PA can be influenced by negative social experiences. Some people can be made to feel guilty by family members who consider PA to be a selfish and low-priority activity [56, 69]. There can even be weight-related teasing [69], social misconceptions, and cultural stigma within and between families [65, 66, 71]. This suggests the possibility that people are isolated from physical activities because of family culture.

3.8. Psychosocial Factors: Individual

PA can be affected by poor self-image. Some people feel flabby in physical appearance [55] and have preconceived notions that people in fitness clubs are mostly slim and physically fit [64]. Body image issues also mean difficulty finding appropriate clothing to wear in public [6971]. Self-esteem can determine how people feel around physically fit peers. For some, transport-related walking feels like a chore and is described as a negative experience that is burdensome and unnecessary [54]. This is compounded by perceived low levels of ability to participate in certain types of exercises because of advancing age, chronic health conditions [55], physical disabilities, poor mental health [66], fatigue, physical discomfort, or low fitness levels [62, 63]. Interestingly, people refer to boredom and stress and being “stuck in a rut” or simply feeling embarrassed rather than refer to psychological barriers as issues of mental health [65]. Some studies highlight that some people perceive themselves as having low physical competence because of previous experiences with PA. One study discusses the concept of “mastery experiences” meaning that successes are associated with ongoing participation [55]. Conversely, negative experiences discourage participation [55, 56, 70, 71]. This suggests that people need self-belief through recognised improvement in performance.

4. Implications for Future Research

This study employed meta-ethnography to highlight how understandings of PA participation in low socioeconomic status populations have been reported in the literature. Findings reveal how broader themes (macrosystem) have more sources than narrower themes (individual), and this emphasizes the needs for more research focused on individual-level influences on PA in socioeconomically disadvantaged communities.

Across all nineteen studies included in this review, we were unable to discover any evidence of the perceived value of PA as a positive social construct in socioeconomically disadvantaged communities. We discovered that the literature has focused on the barriers for PA rather than what supports PA among the socioeconomically disadvantaged populations. However, the literature has shifted to positively framed asset-based approaches to promoting health [73, 74].

It appears important for the future promotion of PA to consider how resources are made available and promoted to target communities. For example, qualitative research may benefit from exploring informal and less-structured forms of PA. Furthermore, qualitative research can explore seemingly unconventional activities to examine the extent to which these can be recognised as offering viable sources of PA. Future meta-ethnographies may benefit from research that can adopt a salutogenic lens [73, 74] to deliver more positive insights about what people in socioeconomically disadvantaged communities do well to sustain PA. This would require a sense of coherence research orientation that probes the necessary life experiences needed to confront stressors on healthy living [75], believing that the challenge to cope with stressors are understood (comprehensibility), believing that the resources to cope with stressors are available (manageability), and wishing to be motivated to cope with stressors (meaningfulness), removing the focus on only PA barriers as influencing PA behaviour.

5. Limitations

In line with the aim of this study, we summarise what is known qualitatively about individual, social, and environmental influences upon PA within and across complex societal systems. This meta-ethnography therefore used a broad-brush approach to compare commonalties across selected socioeconomically disadvantaged population groups. Because of this, the findings of this study may not be grounded or contextually relevant to all population groups. The search criteria limiting the results to qualitative studies may have excluded relevant studies without the word qualitative as a keyword. Additionally, socioeconomic disadvantage is not readily defined, nor is distinct from other forms of disadvantage or exclusion [22]. We recommend future meta-ethnographies explore more broadly concepts that intersect to form class structures. Nonetheless, our findings are useful for future qualitative researchers to hypothesise new research directions based on where they see hidden complexities and intersectional insights.

6. Conclusion

In this study, we propose a social ecological framework by conducting a meta-ethnography of PA determinants in low socioeconomic status communities residing in industrialized countries. Future qualitative researchers can use findings from this meta-ethnography to theorise the potential complexities and intersections that would illuminate the interconnectedness of influences on PA in people with low socioeconomic status. We believe that a more complex understanding of contradictions between positive and deficit frames would lead to more critical insights concerning research gaps of PA in low socioeconomic status.

Conflicts of Interest

The authors declare that they have no conflicts of interest.