In recent years, the effects of psychosocial risk factors on population health have received considerable attention in both research and policy circles [1
]. Psychosocial epidemiology explores the way peoples' interactions with their social environments may influence health either directly (e.g. through biological responses to what is commonly called 'stress') or indirectly through health behaviours [5
]. This research is controversial (for example, some researchers contest the evidence that psychosocial factors are important health determinants, particularly compared to material/economic determinants of population health [6
]). Nonetheless, psychosocial theories are influential and have encouraged policy-makers to develop public health strategies that consider people's support networks, sense of control and empowerment, their sense of security, and the extent to which people participate in the local community and civic society [2
A major obstacle to reviewing this area is the lack of consensus regarding the definitions and usage of psychosocial concepts in the research literature. Martikainen et al suggest that psychosocial factors most usefully describe a bridging or 'meso-level' between individual and social structures and hence include such factors as support from social networks, control at work or in the home, effort/reward imbalance, security and autonomy, and work-family conflict [14
]. They state that a psychosocial explanation of health should describe how macro- and meso-level social processes lead to perceptions and psychological processes at the individual level. We have made this bridging role a central component of our own definition of psychosocial: seeking to exclude macro-level risk factors that are unlikely to effect health via psychological processes, and exclude psychological characteristics (e.g. depression, anxiety and type 'A' characteristics) that described individuals rather than some form of interaction between people and their social environment involving psychological processes.
In practice, identifying psychosocial characteristics can be a difficult task, the results of which may be contestable. Martikainen et al highlight one difficulty when they argue that psychosocial exposures do not necessarily invoke psychosocial processes, and may arrive at health outcomes through alternative (non-psychosocial) pathways. For example, social networks provide instrumental and material benefits as well as emotional support from friends and family; yet they consider only the latter path to qualify as a psychosocial process [14
A recent meta-review of non-health sector psychosocial interventions (conducted by the present authors) corroborated the view that psychosocial terminology is frequently employed without consensus or definition [Egan M, Thomson H, Petticrew M, Tannahill C, Kearns A, Hanlon P: What are 'psychosocial interventions' and how might they improve health?, submitted]. It also found that at the level of systematic review, most of the available evidence on the health impacts of community-based psychosocial interventions comes from studies of workplace psychosocial interventions. The review concluded by calling for better theory to guide research in population health and social epidemiology (others have made similar calls [15
]). In terms of underlying theory, research into psychosocial factors in workplace settings tends be framed around two well-known theoretical models of the workplace psychosocial environment (i.e. the 'demand control support model' and the 'effort-reward imbalance model') [3
]. In contrast, we have found research into health and wellbeing in the wider community to be open to a broader, but less consistently described, array of theoretical concepts that do not always clearly distinguish psychosocial and non-psychosocial components [16
One of the most influential theoretical frameworks derives from Putnam's work on social capital, which he defines as 'features of social organisations, such as networks, norms, and trust, that facilitate action and co-operation for mutual benefit' [16
]. Key to this approach is the hypothesis that strong social interactions between residents of a neighbourhood can benefit not only those who interact (e.g. in terms of emotional, practical and financial forms of social support), but also to neighbours who do not take part in these interactions. These latter benefits are referred to as 'externalities' and may include increased feelings of safety, reductions in anti-social behaviour, and better services and amenities in neighbourhoods where communities are cohesive enough to give individuals and groups the confidence to engage in informal social control (e.g. intervening to prevent anti-social behaviour) and civic participation (e.g. establishing youth groups, participating in local decision-making, intervening to improve the local area, etc).
Referring to Martikainen et al [14
], we would argue that some components of social capital can be considered psychosocial factors because they are likely to involve social processes that lead to perceptions and psychological processes at the individual level (e.g. community characteristics that encourage mutual trust, emotional support and participation/control). However, social capital also includes components that may be associated with 'non-psychosocial' pathways to health, such as practical support and improved local services, and with macro-level contextual factors which impact on health above and beyond psychosocial effects.
In their work on social cohesion, Stafford et al have developed a measurement tool that focuses on 8 components of neighbourhood social cohesion: (1) family ties (frequency of contact with local family); (2) friendship ties (frequency of contact with local friends); (3) participation (regular participation in local organised groups, such as social, religious, neighbourhood interest, evening classes, etc); (4) integration into wider society (contact with people in the same area and outside the local area); (5) trust (e.g. the extent to which people in the area can be trusted, being afraid to walk alone after dark); (6) attachment to neighbourhood (e.g. belief that neighbours are friendly, feeling part of the area); (7) tolerance (e.g. belief that everybody in the area should have equal rights, people in the area are tolerant of others not like them, respect for privacy); and (8) being able to rely on others for practical support (e.g. feeling comfortable asking neighbours to run errands for each other during illnesses) [17
]. Many of the components of this framework would fit our definition of 'psychosocial', but again the list does not refer exclusively to risk factors that may affect health through psychosocial processes (for example, it includes practical rather than emotional support).
MacIntyre et al have advocated a broader framework to describe the pathways in which neighbourhoods may effect health [19
]. Besides considering the physical environment, services and amenities, this framework also considers the 'socio-cultural' characteristics of a neighbourhood and its reputation, some of which may be regarded as relevant to psychosocial theories of health. Neighbourhood characteristics cover a range of risk factors including those associated with the political, economic, ethnic and religious history of a community, current norms and values, the degree of community integration, levels of crime, incivilities and other threats to personal safety, and networks of community support. Neighbourhood reputation includes how a local area is perceived by its residents, how it affects their self-esteem, who moves in and out of the area, and how the neighbourhood is perceived by service or amenity planners, providers and investors.
Siegrist and Marmot have defined the 'psychosocial environment' as the sociostructural range of opportunities that is available to an individual person to meet his or her needs of well being, productivity and positive self-experience [5
]. They emphasise the importance of self-efficacy and self-esteem. A psychosocial environment conducive to self-efficacy enables the person to experience control in terms of successful agency. A psychosocial environment conducive to self-esteem enables the person to connect him- or herself with others in a way that strengthens feelings of belonging, approval and success (in contrast to feelings of being excluded or of not getting anywhere despite one's efforts). Although his work has largely focused on workplace health, Marmot has co-authored a paper from the Whitehall study of UK civil servants (Chandola et al) that presents some evidence suggesting that low control at home associated with excessive household and family demands may have a greater adverse effect on the health of women compared with men [20
In the design of this review, we combined elements from this broader literature to develop a search strategy to identify studies of psychosocial risk factors in the community. We included those risk factors that appeared to us to fulfil the bridging role between socio-structural and psychological characteristics described above, but we reiterate the point made earlier that risk factors considered to be 'psychosocial' may potentially affect health through non-psychosocial pathways. The main sets of themes we have focused on are (a) autonomy and control, (b) involvement, participation and empowerment, (c) social capital, social cohesion, trust and belonging, (d) social support (including specific types of support: e.g. emotional), social networks and receiving positive feedback (e) social diversity and tolerance, (f) vulnerability, security or safety, and (g) demands, role conflicts or role imbalance.
We have identified and synthesised findings from systematic reviews that report data on any kind of health measure in association with any of the above psychosocial factors within a home or neighbourhood residential setting. The review includes a broad range of risk factors and health outcomes because the psychosocial epidemiological literature we have scoped is similarly varied. We do not suggest that empirical evidence is available to demonstrate how psychosocial processes explain associations between each of the specific risk factors and health outcomes identified in this review (in our discussion section we call for more evidence of this kind). We also note that whilst our initial search strategy was intended to include risk factors pertaining to all the themes referred to above, the output from the process has yielded evidence on a more limited range of factors, and so our findings do not address the full spectrum of those themes. This report summarises the evidence we identified, prioritising findings from the more robust reviews.