Across the statutory and voluntary sectors outreach work has a long history as a means of engaging individuals and communities, typically those marginalised by processes of social exclusion and socioeconomic deprivation . Within the health domain, since the outbreak of the HIV and AIDS epidemic of the 1980s, outreach activities have focused on harm reduction in relation to sexual health and substance abuse, as well as in the field of mental health [1, 2]. Whilst outreach, as embodied by home visits, has historically had a place in primary care through the work of health visitors, district nurses and general practitioners, outreach more generally has not traditionally had a significant role in the delivery of primary care prevention programmes.
Nonetheless, it is now widely accepted that preventive approaches to health are disproportionately accessed by the more affluent in society, thus increasing (rather than reducing) inequalities in health outcomes between different social strata [3–6]. As a result, recent international health improvement policies have advocated the use of targeted, preventive primary care as a means of reducing risk factors in poorer individuals and communities [7–9]. In the UK, outreach work has become one part of primary care's response to this challenge within a broader set of reach and engagement strategies [10–12].
However, outreach has been described as eclectic in its purpose, client group and specific mode of practice and, as a direct result of this heterogeneity, little is known about its effectiveness. Ottoson and Green have suggested that the concept of outreach needs to be more carefully considered, with typologies of different approaches and their underlying mechanisms delineated prior to assessment of their effectiveness . In other words, outreach approaches need to be better theorised. Such a recommendation fits with a growing consensus within the evaluation community that impact assessment without adequate theory does not constitute good science [14–16].
Keep Well was launched in 2006 as a national approach to tackling cardiovascular mortality and morbidity through primary care, targeting those in areas of severe socio-economic deprivation [17, 18]. It was based on an anticipatory care approach [17, 19] and, amongst other strategies, made use of outreach to engage its target population in health checks and subsequent interventions. Using Keep Well as a case, this paper addresses the question of what types of engagement 'problem' outreach aims to solve and identifies implicit and explicit 'theories' of how these problems might be addressed using outreach approaches.
Background to outreach as an approach
According to several commentators 'outreach work' acts as an umbrella term covering a wide range of activities designed to bridge both physical and ideological gaps between users and services (for example ). The concept of users' 'non-engagement' with services is central to the discussion of outreach work and is used to signify an inability or unwillingness to engage with services that have been deemed appropriate for particular groups. This overarching 'problem' is, however, rarely unpacked to make transparent the more specific issues that constitute, or lead to, a lack of fit between users and services. Furthermore, the particular groups that outreach aims to engage have historically been identified within 'classic' areas of harm-reduction (sexual health, HIV risk-reduction, substance abuse and mental health). This raises questions about the types of problems that outreach located within primary care, and more specifically in relation to anticipatory care within the broader field of tackling health inequalities, aims to address.
More emphasis has been placed on describing models of outreach activity than on understanding the theorised links between problems and the mechanisms by which outreach aims to address these problems . Across the proliferation of outreach projects, models operate on a continuum of engagement and vary in the extent to which they aim to address problems at an individual or structural level. Authors voice concerns, therefore, that services continue to implement outreach activity without clarifying what types of outreach (with which specific mechanisms) are suited to generating positive outcomes in particular circumstances (for example ).
Keep Well as a test bed for theorising the practice of outreach for primary prevention
Keep Well was a multi-million pound investment in primary care that emerged from policy documents advocating a key role for the Scottish NHS in reducing health inequalities . Its aim was to reduce inequalities in cardiovascular morbidity and mortality through improved reach of primary prevention programmes. Initially targeted at five pilots in the most deprived areas of Scotland, it provided incentives for general practices within those pilot sites to develop reach and engagement strategies to encourage registered patients between the ages of 45 and 64 to attend for a health check and to sustain their involvement in subsequent treatment. Treatment included pharmacological interventions, lifestyle advice, motivational interviewing and referrals to wider social and material supports. The programme (as with prevention approaches more generally) was predicated on the view that, for those whose modifiable risk factors are identified early, there are evidence-based interventions that can be used to reduce the risk of clinical events. The evidence base for their role in tackling inequalities is not, however, strong.
Keep Well was based, therefore, on a model of anticipating health needs before they become manifest (thereby challenging a reactive, episodic model of care provision) and delivering continuous, integrated, preventive care with the patient as partner [19, 22]. Such a model of care has been termed anticipatory care and has its roots in the work of Van den Dool in the Netherlands and Tudor Hart in Wales [23, 24].
The challenge that the programme aimed to meet was that of engaging with patient and population groups who are traditionally less likely to be in contact with mainstream services . Across a range of health conditions those living in more deprived circumstances have been found to make less use of preventive health services than their more affluent counterparts .
The kinds of explanations that are given for this socially-patterned disparity relate to both demand and supply. In terms of factors that may help to shape demand for preventive health care, there is evidence that those in more deprived groups are less likely to act on early signs of ill-health because of the normalisation of poor health within poorer communities  and because of a concern that they will be held responsible for their illness .
On the supply side there is a long-standing literature, for example, that demonstrates the under-resourcing of primary care in relation to need  and that posits that under-resourcing stems from a lack of political and managerial understanding of the disproportionate resources required to deliver a level-playing field in relation to care never mind an augmented service .
A construct developed by Dixon-Woods and colleagues [4, 5] provides a helpful explanatory bridge between explanations that focus on expressed patient needs and health service responses. They argue that 'candidacy' [The term candidacy as utilised by Dixon Woods and colleagues is broaden than that used in other medical sociology/anthropology writings. For example, as used by Davison, Davey-Smith & Frankel  it refers more narrowly to lay views of what constitutes a typical individual with a heart problem.] for having a particular medical condition or for receiving health services is socially constructed and influenced by one's social circumstances, cultural norms and interaction with health care professionals and systems. It is argued that people living in more disadvantaged circumstances tend to normalise symptoms of ill-health because of a higher prevalence of such symptoms, and tend to manage their health as a series of minor and major crises rather than by prevention and positive maintenance . The construct of candidacy, therefore, attempts to explain the complex vulnerabilities, contexts and processes that influence how people, and healthcare systems, come to define eligibility for healthcare . Such eligibility is negotiated prior to initial contact with services and through subsequent journeys through the health service. They describe this dynamic set of factors influencing access to care as follows :
"We suggest that social patterning of perceptions of health and health services, difficulties in marshalling the practical and social resources needed to use services, a lack of alignment between the priorities of disadvantaged people and the organisational values of health services, conspire to create vulnerabilities for socially disadvantaged people in their negotiation of health services" (p. 98).
In reactive help-seeking behaviour, people present themselves to the healthcare system once they have recognised and acted on their candidacy; however, in the case of preventive health, this help-seeking is reversed in that it is the healthcare system that has identified certain features of candidacy, which are then negotiated with those individuals who fall within particular target groups . In this case recipients of invitations are asked to recognise and accept the candidacy that is presented to them.
For anticipatory care initiatives such as Keep Well, important questions therefore arise as to whether those targeted by the intervention recognise their candidacy, firstly in relation to preventive health in general, and, secondly, in relation to the specific health issue(s) that the initiative hopes to address. In respect of Keep Well, the issue is whether people see themselves as candidates for cardiovascular disease and/or for preventive treatments to reduce the risk of such disease. The key question is: to what extent is their perception of their own candidacy for disease and its prevention in alignment with those of the pilots and practices that have already defined their eligibility? Candidacy also looks, however, at how the health service can influence notions of eligibility, including a focus on access, capacity, flexibility of services as well as on the many one-to-one interactions between patients and professionals. Barriers erected by systems and individual professionals are therefore seen to be important factors that overlap with, and influence, perceptions of candidacy.
Keep Well's approach to stimulating candidacy was flexible but, at the same time, remarkably conservative and, in the main, not targeted at social and material determinants of non-engagement. Individual pilots/practices were given freedom in relation to how they contacted their target population and were able to use a range of methods introduced at different stages .
The 'problems' that outreach work in Keep Well hoped to address, however, were less explicit than the overarching notion of the need for a novel approach to reaching and engaging hard-to-reach groups. This study aims to delineate the theories (both implicit and explicit) of professionals responsible for planning and implementing outreach approaches within Keep Well. It asks: what is the 'problem' that outreach work seeks to address and how is it tackled'?