Addressing the social determinants of inequities in physical activity and sedentary behaviours

Abstract

Participation in both physical activity and sedentary behaviours follow a social gradient, such that those who are more advantaged are more likely to be regularly physically active, less likely to be sedentary, and less likely to experience the adverse health outcomes associated with inactive lifestyles than their less advantaged peers. The aim of this paper is to provide, in a format that will support policymakers and practitioners, an overview of the current evidence base and highlight promising approaches for promoting physical activity and reducing sedentary behaviours equitably at each level of ‘Fair Foundations: The VicHealth framework for health equity’. A rapid review was undertaken in February–April 2014. Electronic databases (Medline, PsychINFO, SportsDISCUS, CINAHL, Scopus, Web of Science, Cochrane Library, Global Health and Embase) were searched using a pre-defined search strategy and grey literature searches of websites of key relevant organizations were undertaken. The majority of included studies focussed on approaches targeting behaviour change at the individual level, with fewer focussing on daily living conditions or broader socioeconomic, political and cultural contexts. While many gaps in the evidence base remain, particularly in relation to reducing sedentary behaviour, promising approaches for promoting physical activity equitably across the three levels of the Fair Foundations framework include: community-wide approaches; support for local and state governments to develop policies and practices; neighbourhood designs (including parks) that are conducive to physical activity; investment in early childhood interventions; school programmes; peer- or group-based programmes; and targeted motivational, cognitive-behavioural, and/or mediated individual-level approaches.

active living, inequalities in health, physical activities, sedentary behaviour

INTRODUCTION

Habitual physical activity (PA) confers important health benefits, including reduced risk and morbidity associated with cardiovascular disease, overweight and obesity, hypertension, Type 2 diabetes, some cancers, osteoporosis and musculoskeletal impairments, and depression (Warburton et al., 2006). Despite this, most Australian adults (57%) do not meet the recommendations of 150–300 min per week of moderate-intensity PA (Australian Bureau of Statistics, 2013). Over half (60%) of Australian children aged 5–17 years average the recommended 60 min of moderate to vigorous PA per day; however just 19% of those children meet the recommended 60 min every day (Australian Bureau of Statistics, 2013). Furthermore, in Australia and other developed countries, sedentary behaviours (SB)—those typified by low energy expenditure (e.g. sitting, driving, watching TV)—are pervasive, and pose similar health risks, independent of PA (Healy et al., 2008). Australian adults spend on average 39 h a week in sedentary behaviours, with television viewing (almost 13 h/week) the most prevalent behaviour; children spent on average just over 2 h/day watching television (Australian Bureau of Statistics, 2013).

THE SOCIAL DISTRIBUTION OF PA, SB AND ASSOCIATED HEALTH OUTCOMES

This review focuses primarily on discretionary PA, i.e. leisure-time PA (active recreation, active play and sport) and transport-related PA (active transport such as walking/cycling). In Australia and other developed nations, discretionary PA, SB and associated health outcomes are socially distributed, such that those who are more advantaged (typically, those with higher education levels; higher incomes; higher status occupations; living in more affluent neighbourhoods) are more likely to be regularly physically active and less likely to experience adverse health outcomes than their disadvantaged peers (Australian Bureau of Statistics, 2013). Further inequities exist in PA participation according to other social stratifying characteristics such as gender (with women generally less active than men), disability, ethnic background and Aboriginality (e.g., Australian Bureau of Statistics, 2012). The causes of these social inequities are not completely understood, but are likely complex, rooted in broad structural and contextual factors, as well as in individual and interpersonal factors.

This review aims to summarize, using as a basis ‘Fair Foundations: The VicHealth framework for health equity’, existing evidence on social determinants of inequities in PA and SB, and identify promising approaches to promoting equity in PA and related health outcomes at each layer. The Fair Foundations framework identifies three layers of influence and entry points for action in the social determinants of health inequities: Socio-economic, political and cultural contexts; daily living conditions; and individual health-related factors. The short time-frames, broad scope and rapid nature of this expedited review preclude a comprehensive systematic literature search. Rather, the focus was on the identification of key evidence-based recommendations and entry points for action, with illustrative examples.

METHODS

The literature search was undertaken in February–April 2014, and comprised four phases: The initial search was conducted by one author (K.D.), followed by a review of all identified titles (4265) conducted independently by two researchers (K.D., M.J.). Abstracts of all papers whose titles were deemed potentially eligible by either author (875) were included in the abstract screening phase, conducted by four reviewers (K.B., A.C., K.D., M.J.) independently. Full-text articles (390) were reviewed and 90 deemed to be eligible for inclusion, with data extraction from these 90 papers conducted by the same four reviewers.

• Analysis of key existing evidence syntheses and recommendations for the promotion of PA and reduction of SB in the general population. An equity lens was applied to these recommendations, with consideration of evidence on the potential for differential impacts of the recommended actions across social groups. In applying such a lens, the authors considered evidence, where it existed, of the accessibility, acceptability, effectiveness and use of interventions or programmes by disadvantaged groups in the population;

• A search of published peer-reviewed literature using a pre-defined search strategy across nine bibliographic databases and citation indices; Medline, PsychINFO, SportsDISCUS, CINAHL, Scopus, Web of Science, Cochrane Library, Global Health and Embase (see Appendix 1 for search strategy);

• Manual searches of reference lists of articles identified in the above search; and

• ‘Grey literature’ searches using Google Scholar and of websites of key relevant organizations, e.g. World Health Organisation; Australian Federal and State Government departments; Robert Wood Johnson Foundation; Public Health England; US Government Departments; and key Australian NGOs, including VicHealth and the Heart Foundation. The search focused on identifying relevant reviews, interventions, policies and programme evaluations concentrating on, or potentially applicable to, the Australian context.

RESULTS

The results described below incorporate an analysis of existing syntheses and recommendations for increasing PA or reducing SB in the population generally (see Methods section, Phase i), noting that most do not consider inequity specifically. These existing syntheses include reports of the US Centers for Disease Control and Prevention (Centers for Disease Control and Prevention, 2001, 2011; Kahn et al., 2002; Zaza et al., 2005; Heath et al., 2006); and the UK National Institute for Health and Clinical Excellence (NICE Public Health Collaborating Centre, 2006). We apply an equity lens to these existing recommendations, reporting on the potential for differential effects across social groups. In most cases, there is no empirical evidence on differential effectiveness, hence consideration is given to the potential differential impact based on known determinants and barriers to PA across social groups, and how these recommendations might be implemented in an equity-sensitive manner. In addition, the results described below incorporate a review of primary studies identified from the search described in the Methods section (Phase ii-iv), which covers studies and interventions where there is evidence of effectiveness in varied social groups. Research into public health implications of SB is in its infancy. The limited available evidence relating to approaches for reducing SB equitably is incorporated below.

WHAT CAN BE DONE TO ADDRESS THE SOCIAL DETERMINANTS OF PA AND SB?

Actions addressing socioeconomic, political and cultural contexts

Participation in PA and SB is determined not only by proximal factors such as individual motivations and local environments, but also by broader socioeconomic, political and cultural contexts. Such factors include policies and legislation that contribute to inequities in the distribution of determinants of PA and SB, such as ‘walkable’ neighbourhoods, crime rates, or recreational facilities, as well as social norms regarding the role and value of PA. Addressing this layer of the Fair Foundations framework is politically, logistically and economically challenging, because such changes typically require a combination of political will; public support; substantial financial and human resource; and time. Nonetheless, several examples of initiatives at this layer of the framework were identified and are described below.

Governance/partnership approaches

One potential model of collaboration in governance for PA promotion was the early whole-of-government approach implemented in 1996–2002 in NSW, Australia. In 1996, the Premier established the NSW PA Taskforce to develop a comprehensive strategic plan for PA promotion in NSW. Every government department was instructed to promote PA, with a particular focus on policy and environmental change. While the Taskforce achieved key outcomes including the collaborative development of a strategic plan, and jointly auspiced initiatives such as a mass media campaign, no evaluation of resulting changes in PA was undertaken. Nonetheless, the Taskforce led to the establishment of the Premier's Council for Active Living (Premier's Council for Active Living NSW, 2012), which achieved several key outcomes that might be considered evaluative markers of success (e.g. development of the NSW bike plan; and of a framework for local councils to create supportive environments for active living). Similar state-wide government approaches have been implemented in Western Australia (Government of Western Australia, 2012), South Australia (Government of South Australia, 2014) and Tasmania (Tasmanian Government, 2014). Again comprehensive direct evaluations of these are lacking; such evaluations are challenging due to their scale, multiple components, and complexities of assessing against appropriate control sites. However, potential effects can be indirectly examined via regular PA monitoring or surveillance surveys over the implementation period. For example, monitoring in WA indicated that adults' PA participation in 2009 was higher than in 1999 or 2002, but whether this was attributable in part to Be Active WA initiatives cannot be established conclusively. Similarly, the impact of these governance approaches across varied social groups cannot currently be determined.

Another example of governance and policy actions involving partnerships that could lead to enhanced PA opportunities is policy regarding sharing of government, school and community facilities for PA. Evidence suggests that facility partnerships can successfully meet the sporting/recreational needs of school students and the wider community, including in socioeconomically disadvantaged communities (Sport and Recreation New Zealand, 2011).

Community-wide campaigns

Community-wide campaigns are consistently recommended as an effective means of promoting PA in the general population in evidence-based reviews (Centers for Disease Control and Prevention, 2001, 2011; Kahn et al., 2002; Zaza et al., 2005; Heath et al., 2006). These are typically large-scale, highly visible, multicomponent campaigns involving multiple sectors and partnerships that deliver messages via television, radio and newspapers. They include components such as community events; support groups; PA counselling; risk factor screening and education at worksites, schools, and community health centres. They also include policy and environmental changes, such as the installation of walking paths, or opening school facilities for public use. A review of 10 community-wide campaigns internationally demonstrated their effectiveness in increasing rates of PA engagement (Zaza et al., 2005). However, the effectiveness of community-based campaigns in increasing PA amongst disadvantaged groups is less well-studied and evidence less consistent (Bock et al., 2014).

Economic instruments

Relatively few studies have examined economic or fiscal approaches for promoting PA or SB change (Mitchell et al., 2013; Shemilt et al., 2013) and there has been little consideration of an equity perspective on such approaches. Taxes or tax exemptions are examples of such economic instruments. For instance, taxes can penalize individuals financially for behaviours that reduce PA or increase SB, such as driving. Two studies assessed congestion taxes in Sweden (Bergman et al., 2010). While evidence was somewhat inconclusive, and social inequities in impact were not reported, it was suggested that a road tax might increase PA and decrease sedentary time (Bergman et al., 2010).

One example of an incentivization health promotion programme is the Vitality program, run by South Africa's largest private health insurer, Discovery Health, with over 2.5 million beneficiaries. The programme includes health risk assessments, subsidized gym memberships and weight loss programmes with incentives, including rebates on healthy food purchases. Evaluation data (Lambert and Kolbe-Alexander, 2013) suggest that increased participation in fitness-related activities was associated with lower healthcare expenditure. However, less than one-third of members utilized the gym benefit regularly, suggesting that the incentives may not be sufficient to promote activity. In addition, any initiative involving physical activity incentivization via private health insurance in Australia could potentially exacerbate socioeconomic inequities, since private health insurance is strongly socioeconomically patterned (Australian Bureau of Statistics, 2010).

Transportation policies and practices

Transportation policies and practices can promote PA by improving infrastructure, safety or providing incentives to encourage walking or cycling as active transport modes that may be combined with public transport. Potentially, these approaches can shift perceived norms about the environment and travel behaviour, with flow-on effects to PA norms (e.g. the normative nature of active commuting). Whilst intervention evidence is limited by methodological issues including lack of appropriate control sites (NICE Public Health Collaborating Centre, 2006), existing interventions that promote active transport using traffic calming; installation of shared trails; road closures/restrictions; road tolls (see also economic instruments); cycling infrastructure; and the creation of safe routes to school show promise for effectiveness and sustainability (Kahn et al., 2002; NICE Public Health Collaborating Centre, 2006; Pucher et al., 2010).

A longitudinal study of walking/cycling infrastructure (including traffic-free bridges and feeder routes) at three sites in the UK reported sustained use over 2 years following construction (Goodman et al., 2013). However the infrastructure was used mainly for recreational walking rather than for active transport and may have attracted those who already walked/cycled rather than promoting behaviour change. Furthermore, despite being accessible to a socioeconomically diverse population, the infrastructure was used more by more socioeconomically advantaged adults.

Several major cities, worldwide, have introduced bike-share schemes to promote sustainable transport. In London, social inequalities were identified in the uptake and usage of the scheme during the 9 months post-implementation (Ogilvie and Goodman, 2012). Users of the scheme tended to be males residing in more socioeconomically advantaged areas where cycling was already pervasive (Ogilvie and Goodman, 2012). Whilst the potential of bike-share schemes to ‘normalize’ cycling as a transport mode rather than being an activity exclusively for lycra-clad athletes has been identified (Goodman et al., 2013), their uptake in Australian cities has been poor, possibly due to mandatory helmet legislation in Australia posing a barrier to ad hoc cycling trips (Fishman et al., 2012).

An example of successful transformation of the urban environment and transportation policy to promote active living comes from Columbia. Bogotá is home to the world's largest Ciclovía Recreativa, a free community-based programme in which streets are closed temporarily to motorized transport, allowing exclusive access to walkers, runners, rollerbladers and cyclists. The model has demonstrated cost-benefits and has inspired similar programmes in 1500 cities worldwide (Montes et al., 2012). Whilst these programmes attract many participants, empirical evidence of effectiveness is limited. However, one study reported that among women, participants in Ciclovia events were seven times more likely to be physically active than non-participants (Gomez et al., 2004). The Bogotá program was considered promising for socioeconomically disadvantaged groups, but attracted more men than women. Overall, across the transport literature, there is insufficient evidence to draw conclusions about the differential impact of transport policy interventions on PA across social groups.

Daily living conditions

The conditions into which people are born, live, work and play impact on their PA opportunities. Initiatives aimed at improving these conditions that may be relevant to PA promotion may focus on access to and quality of early childhood settings such as childcare/preschool; education; workplaces; and local neighbourhoods. Our review identified no published PA-related evaluations that have directly intervened on these levels in order to redress inequalities in living conditions; however, several studies have aimed to promote PA or reduce SB in these settings, as described below.

Most settings-based initiatives described in this review use multi-component approaches, typically combining environmental level supports (e.g. provision of sports equipment) with social support (e.g. teacher/employer endorsement/support) and also individually-targeted education/health promotion. As such, it is impossible to determine exactly which component(s) lead to behaviour change. Nonetheless, the success reported by numerous reviewed studies, including those targeting disadvantaged individuals/communities, suggests that multicomponent settings-based approaches show promise for reducing inequities in PA and SB.

Create neighbourhood environmental designs to promote active living

Physical environments that support active lifestyles may be provided via policies to increase or enhance public open space/green space; establish or improve footpaths or lighting; create or refurbishing playgrounds; and zoning that encourages mixed land use (Heath et al., 2006; Hunter et al., 2015). Such environmental approaches are appealing as they can reach those at risk of inactivity due to their social position. However, little empirical evidence exists of the differential effects of these approaches in promoting PA or reducing SB across social groups. Natural experiments represent a valuable methodology for investigating the impact of such environmental initiatives on behaviour change. Refurbishment of public open space in one socioeconomically disadvantaged neighbourhood in Victoria represents one of few controlled natural experiments that have demonstrated a positive impact on PA amongst children and adults (Veitch et al., 2012).

Initiatives that create or enhance access to PA venues, combined with informational outreach activities, are particularly promising. By combining environmental (e.g. construction of walking trails) and individual components (e.g. education) these approaches fit into both upper layers of the VicHealth Fair Foundations framework. Their effectiveness has been demonstrated among diverse population groups, including disadvantaged ethnic or socioeconomic groups (King et al., 1988; Lewis et al., 1993; Kahn et al., 2002).

Group-based and social support programmes

PA intervention strategies may target changes to the social rather than the physical environment. Social support for PA can be provided via simple interventions that require few resources and are implementable in diverse settings. Typical components include establishing ‘buddy’ systems, making ‘contracts’ with others to be active, peer/professional coaching, walking groups that provide companionship and support or discussion groups to address barriers to participation. Three systematic reviews, including two that focused specifically on socioeconomically disadvantaged participants, concluded that peer- or group-based interventions are effective for increasing PA (Webel et al., 2010; Cleland et al., 2012).

Early childhood settings

Playgroups, childcare centres, kindergarten and early parents' groups are settings that potentially influence young children's active play and sedentary time. Centre-based early childhood environments represent a key setting for PA promotion, by providing access to a large proportion of preschool-aged children (3–5 years). However, Australian childcare services often lack a PA policy or staff trained in PA, while around a third of services allow children to participate in screen-based behaviours daily (Wolfenden et al., 2010).

A systematic review of obesity prevention interventions in early childhood identified that most were conducted in the preschool/childcare setting, or at home (Hesketh and Campbell, 2010). Half (12 studies) targeted socioeconomically disadvantaged children, though some of those targeted diet rather than PA or SB. Despite strong designs, many studies set in preschool/childcare centres reported no evidence of intervention effects on obesity-related behaviours. Most of these centre-based studies lacked a parental component, which is suggested as potentially vital (Hesketh and Campbell, 2010).

Two obesity prevention trials targeting infants (Wen et al., 2012; Campbell et al., 2013) with relevance to social determinants were identified (Victoria, Australia and NSW, Australia). One of these (Cameron et al., 2013; Campbell et al., 2013) included a large proportion of socioeconomically disadvantaged families, indicated by low maternal education. The dietician-delivered intervention showed promise in reducing children's television viewing equally in families of low and high socioeconomic status (SES). However, there was no impact on PA. The second study indicated that a nurse-delivered intervention delivered to parents in socioeconomically disadvantaged communities was effective in reducing BMI among 2 year-olds, and showed positive effects on TV viewing, as well as mothers' PA (Wen et al., 2012).

School settings

School settings show potential for addressing social inequities in PA. Importantly, the secondary analyses of three diverse school-based PA interventions undertaken by De Bourdeaudhuij et al. (De Bourdeaudhuij et al., 2011) suggest that European school-based programmes to promote PA are unlikely to widen or exacerbate existing inequalities. School-based physical education (PE) interventions may involve policy or curriculum changes aimed at increasing the amount of time children are active at school (e.g. additional PE classes, increasing the time that children are active). PE-based approaches are effective for increasing PA and fitness amongst children in diverse populations and settings (Kahn et al., 2002) and may help reduce PA inequities (Salmon et al., 2011; Morgan et al., 2012; Wright et al., 2013; Vander Ploeg et al., 2014).

A literature review (van Sluijs et al., 2007) identified only three (of a total of 57) interventions that targeted PA amongst socioeconomically disadvantaged children, and all three showed significant changes in PA. Elements of these effective interventions included lessons on PA and decreasing screen-time, a focus on school sports activity, in-class exercises, an activity week, and parental information services. Nonetheless, given the few studies identified, the need for further evidence on effective PA interventions for socioeconomically disadvantaged children was highlighted.

Environmental strategies may also increase PA after school and to/from school (active transport). Most active transport interventions that target children have focused on the school journey. A review (Chillon et al., 2011) of such interventions reported that strategies to encourage active transport to/from school have included the use of travel coordinators to develop travel plans, safe routes to school, walking school buses and one-off events such as ‘Walk Safely to School’ days. In total, 14 interventions were identified but because they varied in size, scope, focus and design quality, the most effective intervention strategies could not be determined clearly. Among the interventions, no changes in overall PA levels were observed, but self-reported active transport to and from school increased. The review did not report on socioeconomic variations in effectiveness.

One particular intervention to promote children's active transport on the school journey is the Walking School Bus (WSB), comprising a group of children who walk together to school along a set route, led by an adult ‘driver’ with an adult ‘conductor’ at the rear. Children are picked up or dropped off at ‘bus stops’ near home. The programme was pilot-tested in Victoria between 2001 and 2002. VicHealth (VicHealth, 2007) reported that as well as increasing participants' PA levels, further benefits included opportunities for social interaction and reduced traffic congestion around schools. However, WSBs were less prevalent in socioeconomically disadvantaged areas, due to a lack of volunteer parents. A WSB program was found to increase the numbers of children walking to school among urban, low-income elementary school students in Seattle, USA (Mendoza et al., 2009).

Workplace settings

Evidence on the effectiveness of workplace settings for promoting PA or reducing SB, including amongst socially disadvantaged populations, is mixed (Hutchinson and Wilson, 2012; Magnée et al., 2013). Two reviews have examined strategies to reduce workplace sitting (Chau et al., 2010; Neuhaus et al., 2014). The first did not find any workplace intervention study in which the primary goal was to reduce sitting, reflecting the recency of this field of research (Chau et al., 2010). Of the six studies reviewed, none showed significant intervention effects on decreasing sitting time. The second review found that 13 of the 38 included studies reported on sedentary time as an outcome of activity-permissive workstations in office workers, with a pooled effect size of −77 min of sedentary time per 8 h day (Neuhaus et al., 2014). More evidence on best practice approaches to reducing social inequities in sitting behaviours in offices and in other occupations and contexts is required.

Healthcare settings

These include primary health care settings such as General practice. Initiatives include the Green Prescription in New Zealand in which a primary care professional discusses and ‘prescribes’ a PA goal to the patient, with follow-up support from local sports foundations and exercise specialists. Evidence from systematic reviews (World Health Organization, 2009; Orrow et al., 2012) shows that minimal contact interventions, such as health checks or single visit counselling are typically ineffective, but more intensive approaches supported by targeted information, or coordinated with other stakeholders have shown positive effects on PA participation. However, effectiveness across social groups is not well-established. For example, one review (Orrow et al., 2012) demonstrated that most participants in primary healthcare delivered PA interventions were white and socioeconomic characteristics were poorly reported.

Community settings

As distinct from community-wide campaigns, which involve large-scale, multi-component, whole-of-community initiatives, programmes delivered in community settings have had mixed effects on increasing PA participation. In Australia, one example of a community-based approach to promoting physical activity was that undertaken by the Rumbalara Football Netball Club, which has developed and delivered a number of culturally appropriate PA programmes for Aboriginal people of different ages, and has been successful in attracting more players, both Aboriginal and non-Aboriginal, and both men and women (Doyle et al., 2013). In a recent review of the differential effectiveness by socioeconomic position of 26 Dutch obesity-related lifestyle interventions (Magnée et al., 2013), the few interventions that were successful among socioeconomically disadvantaged groups tended to be conducted in the community setting.

Other intervention approaches that may be categorized under this layer of influence are church-based programmes. Church or faith-based organizations provide a supportive and accessible setting for health promotion, particularly to African Americans in the USA. For example, the Health-e-AME project (Wilcox et al., 2007) trained church health directors and PA co-ordinators to organize and deliver church-based programmes, often incorporating spiritual and religious components. This study, and similar church-based interventions such as 8 Steps to Fitness (Bopp et al., 2009), reported positive changes in health and PA using faith-based methodology; however, their relevance to the Australian context is unknown.

Individual health-related factors

The most commonly reported approaches aimed at promoting PA in Australia and internationally involve interventions that are individually targeted, focusing on individual attitudinal and behaviour change. These approaches can be broadly categorized into strategies targeting intentions (e.g. point of decision prompts); motivation level; broader cognitive behavioural strategies; and mediated approaches, including those delivered via novel technologies. One meta-analysis of physical activity intervention strategies demonstrated that behavioural (rather than cognitive) approaches, and face-to-face interventions targeting individuals (rather than mediated approaches or interventions targeting communities) promoted larger increases in physical activity, and that the effects were similar across sex, racial/ethnic and education groups (Conn et al., 2011).

Some individually focused PA programmes and interventions have specifically targeted socially disadvantaged individuals; the most promising have tailored strategies according to participants' needs. However, focusing solely on the most disadvantaged will not have a major impact on reducing health inequalities (Marmot, 2010). However, there is some evidence of their effectiveness for promoting PA behaviours among socially disadvantaged individuals in the short-term.

Point-of-decision prompts

These include signs posted alongside escalators or lifts to encourage stair use by including information about related health benefits. Aesthetic enhancements (e.g. music, artworks) may be introduced but are not deemed necessary for effectiveness (Kahn et al., 2002). Point-of-decision prompts appear equally effective for promoting PA among men and women, across ethnic minority groups (as long as they are appropriately tailored); and in diverse settings (Kahn et al., 2002; Soler et al., 2010).

Motivation-based interventions/counselling

Motivational interviewing (e.g. by health care workers) may be an effective approach for increasing PA amongst socioeconomically disadvantaged participants (Hardcastle et al., 2012). In Denmark, lifestyle counselling has also been shown to be effective across socioeconomic groups in achieving increases in PA, or attenuating age-related decreases in PA (Aadahl et al., 2011).

Pedometers offer an affordable and accessible technology that may be appropriate for increasing motivation for PA amongst socially disadvantaged groups, as the devices are simplistic in output and low-literacy friendly, and immediately understandable to end-users. They are effective in promoting PA in general populations; however, little is known about effectiveness in socially disadvantaged groups (Tudor-Locke and Lutes, 2009).

Cognitive behavioural approaches

These approaches not only target motivation but typically involve broader behaviour change strategies such as intention formation, goal setting, self-monitoring of behaviour, creating or strengthening social support networks, reinforcing behaviour through self-reward, addressing barriers, problem-solving and relapse prevention. Behaviour change strategies can be tailored to an individual's specific PA preferences, barriers, social contexts and readiness for change. This intervention approach has been found to be effective with diverse populations including different socioeconomic groups and in different settings (e.g. communities, workplaces, schools, health and fitness settings) (Kahn et al., 2002). While some cognitive behavioural approaches are delivered face-to-face, others are delivered via various media (see mediated approaches below). These approaches may be more likely to show long-term effects when complemented by environments and policies that support individual behaviour change.

Mediated approaches

There is some evidence supporting the use of mediated approaches (i.e. PA promotion support/resources delivered via media such as print, telephone, mass media or web-based/novel technology) in promoting PA amongst socially disadvantaged participants. These approaches are more likely to be successful if they are individually tailored. For example, a US-based study showed that internet/video-delivered intervention resulted in increased moderate-to-vigorous PA participation amongst children in low-middle income schools (Frenn et al., 2003). Among older adults, a telephone-assisted PA promotion programme (Hooker et al., 2005) achieved reasonable buy-in from a diverse range of local agencies leading the initiative, and was relatively successful in promoting PA, although there were signs of reduced effectiveness amongst socially disadvantaged groups, who were highly sedentary.

Overall, mass media campaigns alone are not effective in promoting increased PA (Kahn et al., 2002), but may play a role in promoting increased self-efficacy and/or knowledge/attitudes related to PA, and such increases do not seem to overly favour more advantaged groups (Thomas, 2012).

Digital/novel delivery approaches

Modern technology provides further mediated approaches to delivering interventions. Evidence suggests that PA interventions delivered via the internet or mobile devices show promise for increasing PA [e.g. (Fanning et al., 2012; Blackman et al., 2013)], although a recent Cochrane review of mobile-telephone supported interventions suggested that the availability of studies using randomized controlled trials was insufficient to determine if these approaches could influence PA or other health behaviours (Vodopivec-Jamsek et al., 2012). That review highlighted that this body of literature does not typically describe the target population or indicate how representative the study samples are of a larger population. Hence inferences cannot be made regarding who may be likely to benefit from such interventions based on different demographic or socioeconomic factors. Similarly, it is unclear which population subgroups may be more or less likely to engage in internet or mobile-telephone supported (mHealth) PA interventions. This is especially important given that those studies that did examine sample characteristics found that non-participants were less educated and, if they did engage in the study, had greater difficulty in operating technology. This is an encouraging area for future research, considering the saturation of mobile phone technology across all social groups within Australia, with almost every adult Australian up to age 60 owning a mobile phone (Wajcman et al., 2007). These technologies have potential to create tailored, flexible, interactive PA programmes that meet the specific requirements of individual users, including those of different socioeconomic or cultural backgrounds, or persons with a disability.

DISCUSSION

This review identifies a lack of evidence of the differential impact of PA or SB interventions across social groups. It highlights many research needs, including importantly the need for intervention studies and policy and programme evaluations to build in an examination of differential impacts across social groups. Further studies identifying effective approaches for equitably reducing SB are particularly required. In addition, the bulk of existing evidence is derived from individual, family and some settings-based interventions, with limited evidence about whether broader government or whole of community approaches reach or impact different social groups equally. This is unfortunate, since government and whole of community campaigns may be particularly salient amongst socioeconomically disadvantaged groups as they may provide access to free or low-cost programmes and services that can benefit individuals with limited resources to engage with other type of interventions. Nonetheless, based on the available evidence, promising potential intervention strategies for addressing the social inequities in PA or SB that are applicable in the Australian context are summarized below.

Recommendations focused within the socioeconomic, political and cultural context

• Multicomponent, appropriately tailored whole-of-community campaigns for promoting increased PA and reduced SB (see also layers below; these approaches can potentially address and impact at several layers given their multilevel nature)

• Support/incentives for local and state governments and community organizations in socioeconomically disadvantaged communities to develop policies and practices to promote PA programmes for individuals across the life course, using existing infrastructure (e.g. active parks programs; walking groups)

• Neighbourhood designs that are walkable and support PA. For example, zoning that promotes walkability

• Transport policy (e.g. investments in pedestrian and cycling facilities/infrastructure; road closures or restrictions on use) shows promise, although more evidence is needed on the impact across social groups

Recommendations focused within daily living conditions

• Multicomponent, tailored, community-wide approaches (also see layer above; these can potentially address and impact at several layers)

• Creation or enhancement of access to PA venues (e.g. walking paths, recreational facilities, shared use facilities between communities and schools), combined with informational outreach activities (e.g. risk factor screening and support groups)

• Investment in early childhood interventions, such as expert-delivered parent support for obesity preventive behaviours

• Establishment, refurbishment and maintenance of suitable parks in disadvantaged communities

• Multicomponent school-based programmes to promote PA and reduce SB during and outside of school hours

• Peer- and group-based social support programmes that encourage peer support and shared experiences

• Investment in physical education to ensure all schools, including those in disadvantaged areas, engage students in at least mandated levels of quality physical education

Recommendations focused within individual health-related factors

Despite the paucity of evidence on the differential impact that PA or SB interventions may have across social groups, findings from this review suggest that there are several promising approaches. Future research should focus on these identified approaches, and should be high-quality, rigorous, and cost-effective. Interventions aimed at promoting PA or reducing SB should also incorporate evaluations of the differential impact across social groups, in order to further consolidate the evidence base.

• Motivational, cognitive-behavioural, and/or mediated individual-level approaches for promoting PA are promising if appropriately tailored. Ideally these would be incorporated as part of a multilevel approach, complemented with broader social and environmental supports

• Tailored, culturally-sensitive point-of-decision prompts to promote stair use in workplaces/public settings which attract people from different social groups

FUNDING

This work was commissioned by VicHealth. K.B. is supported by a National Health & Medical Research Council (NHMRC) Principal Research Fellowship, ID 1042442. The contents of this manuscript are the responsibility of the authors and do not reflect the views of NHMRC. K.D. is supported by a Deakin University Postgraduate Scholarship.

APPENDIX 1: SEARCH STRATEGY

The search strategy outlined below was undertaken in the following databases and citation indices: Medline, PsychINFO, SportsDISCUS, CINAHL, Scopus, Web of Science, Cochrane Library, Global Health, and Embase. Search terms were modified slightly according to database parameters.

Searches were combined as follows to identify papers in which the terms appear in article titles, abstracts and/or as keywords:

(a) AND (b) AND [(c) or (d) or (e)] AND (f)

Search terms

(a) Physical activity: Physical* activ*, Exercis*, Sedentar*, Physical* Inactiv*, Sport*, Active travel, Active transport*, Active commut*, Walk*, Cycl*, Bicycl*, Active free play, Active play, Active recreation, Sport* Club, Sport* organization*, Incidental activ*, Active Leisure, Screen time.

(b) Intervention: Intervention, Policy, Program*, Health promotion, Trial, Evaluat*, Review.

(c) Socioeconomic, political and cultural context: Socio*, Economic*, Education*, School*, Politic*, Policy, Policies, Cultur*, Social*, Environment*, Income, Occupation, Occupational, Social class, Transport*, Infrastructure*, Resource*, Inequalit*, Welfare, Norm, Norms, Value*, Labour, Land use, Housing, Gender, Race, Ethnicity, Aboriginal*, Disability.

(d) Daily living conditions: Income, Housing, Poverty, Neighbourhood, Neighborhood, Work*, School*, Social protection, Social network, Social connection, Social capital, Social cohesion, Social support, Social trust, Participat*, Social determinant*, Safety, Crime, Incivilities, Pollution, Litter, Vandalism, Traffic, Sport* equipment, Employment, Employed, Childcare, Walkability, Street connectivity, Park*, Public open space, Green space, Aesthetics, Footpath*, Destination*, Amenit*, Walking path*.

(e) Individual determinants of physical activity: Knowledge, Attitude, Awareness, Behaviour*, Behavior*, Beliefs, Barrier*, Self-efficacy, Intentions, Perception*, Perceived, Self-regulation, Self-monitoring, Skills, Habit.

(f) Equity: Affordab*, Access*, Advantage*, Disadvantage*, Availab*, Equit*, Inequit*, Inequalit*, Equalit*, Disparit*.

Author notes

‘Fair Foundations: The VicHealth framework for health equity’ was developed by VicHealth under the leadership of author O'Rourke. It was published in 2013. It is a conceptual and planning framework adapted from work done by the WHO Commission on the Social Determinants of Health (Solar and Irwin, 2010). Social determinants of health inequities are depicted as three layers of influence – socioeconomic, political and cultural context; daily living conditions; and individual health-related factors. These determinants and their unequal distribution according to social position, result in differences in health status between population groups that are avoidable and unfair. The layers of influence also provide practical entry points for action (VicHealth, 2013). Fair Foundations can be accessed at www.vichealth.vic.gov.au.