What is the perceived policy problem?: Health inequalities and HTR populations

The imperative to tackle systematic inequalities in health outcomes has increasingly come to form a central stated aim of international, national and local health policies (Graham, Reference Graham2007; Judge et al., Reference Judge, Platt, Costongs and Jurczak2006; Marmot et al., Reference Marmot, Atkinson, Bell, Black, Broadfoot, Cumberlege, Diamond, Gilmore, Ham, Meacher and Mulgan2010; Benach et al., Reference Benach, Malmusi, Yasui, Martinez and Muntaner2011). Thus, in discussing how health inequalities have been conceptualised and targeted in the UK and Europe, Graham notes that health policy has largely moved away from a focus on average measures of population health to a more relative, distribution-sensitive approach. As a result, targets for reducing inequalities have become widespread as policy tools.

However, within the UK, across Europe and within the US a growing body of evidence points to confusion around the nature of the goal of tackling inequitable health outcomes and perforce the strategies required to achieve them (Graham, Reference Graham2007; Judge et al., Reference Judge, Platt, Costongs and Jurczak2006; Bauld et al., Reference Bauld, Day and Judge2008; Krieger, Reference Krieger2008). In particular, Graham illustrated the central confusion between the goals of improving the health of the most disadvantaged, reducing the gap between the most and least affluent and tackling the overall health gradient of the population. She argued that, whilst elements of all three goals were to be found in the policy documents of the UK's New Labour government (1997–2010), action focused on addressing the health disadvantages of socio-economically deprived individuals or communities and on reducing the gap between these and either the most affluent or the population average. Systematic actions to address differentials across the entire health gradient have been far less apparent. Graham posits that a focus on addressing the health of the disadvantaged, whilst important in its own right, reinforces a belief that the health problems faced by those in disadvantaged neighbourhoods are divorced from systematic social and economic inequities across the gradient and result from deficits in poor people themselves. Focusing on reducing the gap between the least well-off and the average is deemed more appropriate but also problematic because it does not address the fact that disproportionate wealth (as well as disproportionate lack of wealth) leads to an inequitable distribution of health (Lynch et al., Reference Lynch, Davey Smith, Kaplan and House2000).

The proof of the pudding, perhaps, is in the eating. Despite the increased focus on health inequalities as a policy problem and significant levels of investment in health improvement programmes that aimed to reduce the gap, the systematic variation in health outcomes between rich and poor are found to be intransigent (Thomas et al., Reference Thomas, Dorling and Davey Smith2010). Within the broad public health academic community there is consensus that, for the most part, efforts to tackle inequalities in health need to focus on material and social determinants that lie beyond the scope of health care systems (Baum et al., Reference Baum, Begin, Houweling and Taylor2009; Dorling, Reference Dorling2010; Benach et al., Reference Benach, Malmusi, Yasui, Martinez and Muntaner2011).

However, it is also argued that the NHS itself has a part to play in reducing health inequalities through a systematic focus on reducing differential access to, and usage of, health care services (Marmot et al., Reference Marmot, Atkinson, Bell, Black, Broadfoot, Cumberlege, Diamond, Gilmore, Ham, Meacher and Mulgan2010). Within this context, there has been a recent focus on augmenting the specific role of health services in preventing excess morbidity and mortality within deprived populations (Marmot et al., Reference Marmot, Atkinson, Bell, Black, Broadfoot, Cumberlege, Diamond, Gilmore, Ham, Meacher and Mulgan2010). In England, for example, this is signalled in the Health Inequalities Progress Report (Department of Health, 2007a) and in the establishment of the National Support Team for Health Inequalities (Department of Health, 2007b). In Scotland, a cross-governmental ministerial inquiry into health inequalities resulted in the ‘Equally Well’ report and implementation plan (Scottish Government, 2008a and b) which endorsed the role of preventive or anticipatory care approaches to improving the health of the most vulnerable populations. Furthermore, the notion of equity ‘proofing’ health care provision has become more mainstream (Marmot et al., Reference Marmot, Atkinson, Bell, Black, Broadfoot, Cumberlege, Diamond, Gilmore, Ham, Meacher and Mulgan2010; Department of Health, 2008; Scottish Government, 2008a).

Across the UK, health policies identify explicit roles for the national health service (and primary care in particular) in identifying and engaging those most at risk of preventable illness and least likely to access preventive services. To date, this targeted approach has been placed within the context of a more universal focus on prevention and exemplifies what Marmot et al. (Reference Marmot, Atkinson, Bell, Black, Broadfoot, Cumberlege, Diamond, Gilmore, Ham, Meacher and Mulgan2010) have described as proportionate universalism, whereby those who are deemed to be ‘hard to reach’ are the recipients of particular interventions that aim to differentially target them and increase their engagement in universally provided services. In other words, in the absence of equity in the myriad social determinants of health, they aim to provide a level playing field in terms of service receipt in relation to health need.

Simply stated, then, the policy problem is twofold: at the macro level there is the problem that material disadvantage leads to poorer health outcomes; at the micro level there is the problem that those in deprived circumstances do not, on average, engage with preventive health services to the same level as those who are more affluent, and, thus, pre-existing disadvantage is compounded. It is the latter that places an emphasis on the need to engage the HTR.

Who are the HTR and why don't they engage?

The term ‘hard-to-reach’ is used in policy discourse as a means of referring to those individuals, and groups of individuals, who are deemed to be ‘in need’ but do not make use of either universally provided services or of more targeted interventions. It is a term that is utilised across public policy domains and across countries. Both implicitly and explicitly it is the target group for policies and programmes that aim to increase engagement in education (Carpentier and Lall, Reference Carpentier and Lall2005), health (Kovandzic et al., Reference Kovandzic, Chew-Graham, Reeve, Edwards, Peters, Edge, Aseem, Gask and Dowrick2011) involvement in civic life (Brackertz, Reference Brackertz2007) and with the workplace (Yeandle and Pearson, Reference Yeandle and Pearson2001). What these diverse policy interventions have in common is the more general view that access to, and utilisation of, services needs to be augmented for certain populations. However, the types of mechanism employed by such policies and programmes range from the delivery of modified and targeted services (such as those developed through Sure Start – NESS Research Team, 2004) through the use of financial ‘sticks’ (for example, the current UK government's use of benefit withdrawal to encourage ‘employability’ (Lindsay, Reference Lindsay, Yeates, Haux, Jawad and Kilkey2011) to ‘carrots’ (exemplified by the growing use of conditional cash transfer programmes in Latin America (Bastagli, Reference Bastagli2009), where transfers are utilised to incentivise school attendance and utilisation of health services).

However, definitional problems abound. In her analysis of whether conditional cash transfer programmes promote resilience, Bastagli (Reference Bastagli2009) identifies problems in developing appropriate parameters to help identify who programmes should be targeting. In describing the use of the term HTR within the context of Australian local government planning, Brackertz (Reference Brackertz2007) argues that policy documents utilise the term to indicate a homogeneous group of constituents. She argues instead that, to be useful, hard-to-reachness should be more clearly categorised as demographic (such as place of residence), cultural (for example, social invisibility, or language spoken), behavioural and attitudinal (for example, those with previously bad experiences of participation) or structural (by which she refers to the way in which organisational practices shape access). These more nuanced definitions are largely absent from, or hidden in, the key health policy documents that aim to tackle the policy problem of hard-to-reachness.

To expand our understanding of what is meant by, and who is conceptualised as, HTR, it is helpful to consider another term that has gained currency as a means of identifying individuals and communities who are viewed as difficult to serve through either universal or targeted public service provision – social exclusion. In the last decade there has been much discussion about this concept (Burchardt et al., Reference Burchardt, Le Grand, Piachaud, Hills, Le Grand and Piachaud2002; Levitas, Reference Levitas1998; Mathieson et al., Reference Mathieson, Popay, Enoch, Escorel, Hernandez, Johnston and Rispel2008). Here we identify two ways in which thinking about social exclusion might illuminate what is meant by hard-to-reachness.

First, we argue that the two concepts relate implicitly to the same group of individuals and communities. For example, social exclusion was defined by the UK Cabinet Office's Social Exclusion Task Force (SETF) as:

Although the above definition does not include poor health, it shares two key features with the concept of hard-to-reachness. First, it is one that can be applied at either individual or area level. Second, both terms imply homogeneity within the disadvantaged group (Brackertz, Reference Brackertz2007; Mathieson et al., Reference Mathieson, Popay, Enoch, Escorel, Hernandez, Johnston and Rispel2008).

Critics argue, however, that social exclusion has suffered from poor conceptualisation and a lack of definitional clarity. For example, it is unclear whether it is an outcome or a process; whether its defining characteristics are necessary and sufficient conditions for one to be excluded; and whether it results from structural determinants or is primarily an issue of agency (Burchardt et al., Reference Burchardt, Le Grand, Piachaud, Hills, Le Grand and Piachaud2002; Mathieson et al., Reference Mathieson, Popay, Enoch, Escorel, Hernandez, Johnston and Rispel2008).

A second way in which writings about social exclusion offer useful lessons to those interested in the concept of hard-to-reachness is in the degree to which the term can be viewed as morally charged. Labonte (Reference Labonte2006) writes convincingly that terms such as social exclusion do not explicitly acknowledge the structural determinants of inequality and can, implicitly at least, carry the suggestion that those who are not included are partly to blame for their condition. The potential for social exclusion as a concept to give off different moral messages is most aptly described by the ground-breaking work of Levitas (Reference Levitas1998). She argues that it is an ambiguous concept that acts as a means of garnering support from a wide array of political bedfellows, including those whose analysis is socialist and redistributive and who view ‘the excluded’ as having been systematically marginalised by social and economic structures, those from a neo-liberal persuasion who employ ‘the excluded’ as part of a moral underclass discourse and those from a New Labour perspective who perceive paid work as the route from exclusion to inclusion (Levitas, Reference Levitas1998). Linguistically, hard-to-reachness is arguably more clear-cut. Regardless of policy intent (as we discuss later), the term carries an explicit message that it is people (individuals or communities) who are ‘hard’ to reach in the face of services that have been ‘reaching out’ to them (Brackertz, Reference Brackertz2007).

So, what does it mean to be HTR in relation to health services? A very high proportion of the population is known to be registered with a general practice – estimated to be around 98 per cent (RCGP, 2004) and 90 per cent of NHS consultations take place in primary care (BMA Scotland, 2010). In 2007/2008, there were approximately 21.8 million patient contacts with GPs and practice nurses in Scotland, just over four contacts per person per year for the entire Scottish population (BMA Scotland, 2010). Nonetheless, the sub-population not registered inevitably has a higher representation of those with complex health and social needs such as the homeless, women fleeing abuse, transient populations and those who have been actively de-selected from practice lists. This group is likely to have an over-representation of young men, homeless people and those with chronic drug, alcohol and mental health problems (Shiner, Reference Shiner1995; Crane and Warnes, Reference Crane and Warnes2001); their needs will be different.

It is also known that the levels of consultations in poor areas are higher than in more affluent areas but that, crucially, this is not true once standardised for levels of need (Mercer and Watt, Reference Mercer and Watt2007). There is also evidence that general levels of consultation mask another important difference between the most and least deprived: attendance at screening, immunisation and preventive services is lower among those living in the poorest communities (Dixon-Woods et al., Reference Dixon-Woods, Kirk, Agarwal, Annandale, Arthur and Harvey2005). Furthermore, use of preventive services is also gendered, with fewer men taking up care (Goddard and Smith, Reference Goddard and Smith2001). From the supply side, there is evidence that primary care is not adequately resourced in the most deprived areas (Watt, Reference Watt2002; Mackay et al., Reference Mackay, Sutton and Watt2005). As a consequence, consultations are shorter, and shorter consultations are known to be associated with lower patient satisfaction (Freeman et al., Reference Freeman, Horder, Howie, Hungin, Hill and Shah2002). There is, therefore, a strong case for bolstering primary care capacity to engage those individuals and groups of individuals in preventive services. The work of researchers such as Barlow et al. (Reference Barlow, Kirkpatrick, Stewart-Brown and Davis2005), however, provides an example of how the decision of individuals to make use of services is driven by a range of factors. In a study of why women did not make use of an antenatal care intervention, they found that non-usage was caused by a lack of interest in, or misgivings about, the particular service, by a sense of being over-burdened by other aspects of daily living, by a lack of trust in services or by holding discrepant views about their vulnerability/need for the intervention.

A more theoretically interesting picture of why some groups do not make use of health services in the same way as others is found in the work of Dixon-Woods et al. (Reference Dixon-Woods, Kirk, Agarwal, Annandale, Arthur and Harvey2005). In their review of the access to health literature, they discuss the notion of candidacy which offers a helpful conceptual framework for understanding the dynamic relationship between health services and their particular patient groups. In brief, candidacy is argued to be a negotiated social construct whereby, on the basis of cultural and personal knowledge about symptoms, normative views of risk behaviours and past experience of health service encounters, people make decisions about whether their current state makes them a ‘candidate’ to use a service. Evidence supports the view that perceptions of risk are shaped, for example, by gender (Galdas et al., Reference Galdas, Cheater and Marshall2005). If individuals are in doubt of their candidacy, they may not attend services or may be susceptible to signs from their general practitioner that consultation times are pressured and therefore refrain from disclosing pertinent health worries. Once candidacy is acknowledged and presented to a general practitioner or other healthcare professional, it is then there to be reinforced or challenged as patients move through the healthcare system. In relation to hard-to-reachness the concept is potentially potent as a means of understanding the importance of structural factors, organisational systems and relationships in influencing the reach and engagement of a preventative approach such as Keep Well (KW). Thus, it helps to emphasise the bidirectional nature of hard-to-reachness, with services struggling to legitimise preventive services to a vulnerable population that questions its own candidacy. Finally, the concept is also useful in helping to view hard-to-reachness as a dynamic rather than a fixed category over time and across individual health decisions.

In summary, the concept of hard-to-reachness is problematic. It is poorly defined and insufficiently problematised in terms of how it is created or can be reduced. It has obvious connections to other dominant policy ‘problems’ such as social exclusion and health inequalities, yet the links and overlaps are poorly theorised. The aim of ensuring that those in disadvantaged circumstances are engaged in preventive health care is, nonetheless, fundamentally and morally important as a partial contribution to a wider endeavour of reducing health inequalities. Having set the scene in terms of hard-to-reachness as a conceptual puzzle we now go on to explore the policy and practice challenges it raises for those charged with its development and delivery at national and local levels of the preventive health programme, Keep Well. In the final section of the paper, we connect these challenges back to the literatures considered thus far.

The case-study: Keep Well

Keep Well (KW) is a health improvement programme implemented through primary care and aimed at tackling inequalities in cardiovascular morbidity and mortality (Scottish Executive, 2005). Notionally, the programme was modelled on the work of Julian Tudor-Hart who, through a combination of indentifying health need in routine consultations, continuity of care and empowering therapeutic patient consultations, demonstrated significant reductions in cardiovascular morbidity in comparison to other neighbouring areas in a close-knit but economically deprived community in Wales (Tudor-Hart, Reference Tudor Hart1971). The first wave of KW pilots was launched in 2006; another four waves ensued before a decision was taken in 2010 to roll out key components across Scotland. Box 1 summarises key features of the programme.

The logic of the programme was consistent with the wider policy thinking described above. In its simplest terms, this policy logic can be represented as follows: cardiovascular health outcomes are unequally distributed across the population, and areas experiencing multiple deprivation also demonstrate inequitable levels of poor health; part of the problem of ill-health in these areas is explained by lower than average uptake of preventive health services (both in relation to supply and demand); the NHS can help to alleviate this set of problems through increasing such engagement; but, to do so, it needs to capture that part of the population not currently maximising service use – the hard to reach.