Health inequities in Australia

Although Australians have high average levels of health by world standards, significant health inequities remain. By most measures, health outcomes for Australia's Indigenous Aboriginal and Torres Strait Islander (hereafter ‘Aboriginal’) peoples are markedly worse than for other Australians. Other population groups with worse than average health include people with low socioeconomic status (SES), those living in regional, rural or remote areas, sole parents, and people living with disability or mental illness. In Australia, as with most countries, many indicators of health status or risk, including: tobacco use, obesity, diabetes, heart disease, premature mortality, and mental illness, follow a social gradient (Australian Institute of Health and Welfare [AIHW], 2014).

The Australian health policy environment

Under Australia's federal system, responsibilities in health policy are split between the Commonwealth (national or ‘federal’) government and eight regional State or Territory (hereafter, ‘State’) governments. The Commonwealth government has far greater revenue raising powers than the States through taxation. With the States, it jointly funds public hospital services and directly funds other medical services through Medicare, the national public health insurance system. States manage public hospitals and community health services within their respective jurisdictions. Around 57% of Australian adults also hold some private health insurance, which is subsidised by the Commonwealth government. These mixed arrangements commonly lead to blame-shifting between levels of government about perceived problems in health service delivery (Woodward et al., Reference Woodward, Parkin and Summers2010). Health policy often bridges this national/state government divide and policies on the same topic are common across jurisdictions. From 2007 to 2013 a degree of bipartisan political support at national and state levels emerged for closing the gap in health outcomes between Aboriginal and non-Aboriginal Australians.

Australian governments of the last several decades have been formed under either the centre-left Australian Labor Party (ALP), or a conservative, centre-right Coalition of the Liberal Party and the National Party. At the national level, both Labor and Coalition governments have implemented neoliberal economic reforms over the last 30 years such as labour market deregulation and privatisation of public services. However, the ALP has sought to combine these with compensatory social spending, including the establishment of Medicare (Spies-Butcher, Reference Spies-Butcher2014). Electoral support in Australia began to turn markedly in favour of conservative Coalition governments from around 2010, after a period of Labor dominance from 2002. Nationally, the ALP held government from 2007 to 2013. The majority of policies analysed in this research were developed under Labor governments, and all were current within this six-year period.

Conceptual frameworks and evidence base

Our critical analysis of Australian health policy drew on several key sources to provide conceptual frameworks for understanding the intersection between evidence on SDH/HE and public policy, including the role of the health sector. We drew on Whitehead and Dahlgren's (Reference Whitehead and Dahlgren2006) analysis of policy responses to health inequities, which claims that policy actions to improve ‘health equity’ can be formulated in various ways: to improve equity of access to health care, improve health among the most disadvantaged groups, reduce health gaps, or to level up health gradients. They argue that adoption of one or other of these formulations will determine quite different policy strategies, and that optimal responses to health inequities will include elements of all these approaches. We drew on the framework for policy action on SDH/HE from the CSDH Final Report (2008) including the need to address SDH both in people's immediate circumstances and in the higher level political factors determining the unequal distribution of social and economic resources. We drew on Baum's (Reference Baum2008) analysis of policy responses required to effectively address SDH and reduce health inequities in Australia, and of the historical dominance of biomedical and individualised behavioural approaches to health in Australian health policy. In relation to evidence on SDH/HE, we drew on Baum's (Reference Baum2008) review of evidence on SDH/HE in Australia and the research team's knowledge of research literature on SDH/HE in Australia and globally.

Document selection

From December 2012 to early 2013 the research team compiled a complete data set of all the then-current, strategic health policies from all nine Australian national and state governments. We developed and applied criteria to identify health department policies that described principles, goals, objectives and strategies for action in any area of departmental responsibility for population health or health services, and that were recognised at the time as current policy. We also selected some policies from ‘higher-level’ agencies such as the department of the government leader and Cabinet. We included these when they addressed a recognised health issue and assigned responsibilities to several departments, including the health department. Two team members audited each health department website to create a provisional list of policies meeting our criteria. A senior staff member in each respective department checked this list for currency and completeness so it could be adjusted to be more accurate. In each case, this resulted in only one or two policies being added or removed. Electronic copies of all listed policies were then compiled to create a dataset of 266 documents. We grouped the documents into four broad categories: whole-of-government or whole-of-department policies; health issue policies; population group policies; and policies on health care delivery.

Document analysis

We used document analysis because it is recognised as a potentially ‘rich source of information about. . . organizations and programs’ (Patton, Reference Patton2002: 293), which can be fruitfully combined with other methods (Bowen, Reference Bowen2009). Recent literature shows document analysis has been used to assess: uptake of research evidence in health policies (e.g. Rosella et al., Reference Rosella, Wilson, Crowcroft, Chu, Upshur and Willison2013); the extent to which health policies address equity issues (e.g. Keleher, Reference Keleher2013); and uptake of evidence on SDH/HE (e.g. Borrell et al., Reference Borrell, Morrison, Burstrom, Pons-Vigués, Hoffmann and Gandarillas2013).

We analysed each individual policy (with documents grouped by jurisdiction) using qualitative analysis software (NVivo 10). The research team developed an a priori coding framework for content analysis to code text into categories related to our research questions (Bowen, Reference Bowen2009), which we tested and revised to produce the final structure. In the initial phase, selected individual policies were coded separately by two team members and outcomes discussed by the team to ensure consistent interpretation. We constructed the coding framework drawing on Whitehead and Dahlgren (Reference Whitehead and Dahlgren2006), Baum (Reference Baum2008) and Carter et al. (Reference Carter, Hooker and Davey2009) and structured it to identify sections of text in each document conveying:

1. (a) Normative values;

2. (b) General acknowledgement or more detailed auditing of evidence on SDH/HE;

3. (c) Goals for improvements in population health;

4. (d) Objectives for improved operational performance; and

5. (e) Strategies for action.

Following Carter et al. (Reference Carter, Hooker and Davey2009), we identified text which broadly acknowledged SDH and/or health inequities, or which audited salient empirical evidence on: differences in health, risk factors or access to health services between population groups associated with differences in social, economic or cultural circumstances; or socially-produced changes in population characteristics over time relevant to burden of disease, e.g. rise in obesity rates.

Those parts of the structure concerned with goals, objectives and strategies are shown in Table 1, below. In relation to goals for health improvement, sub-codes were designed to assess the extent to which policies aimed to improve health equitably. Drawing on Whitehead and Dahlgren (Reference Whitehead and Dahlgren2006), codes distinguished between goals according to whether they aimed to improve health in the general population, or within a designated group; or to close a health gap, or level-up a health gradient. Further, in relation to groups, we distinguished between ‘equity’, ‘illness’ and other groups. Illness groups were those defined in the text according to an existing illness. We identified equity groups as any group defined in the text, where (according to our knowledge) that group is subject to some form of systemic social or locational disadvantage and evidence suggests the disadvantage causally contributes to worse health outcomes within the group than in the wider population.

TABLE 1.

Coding Structure – Health goals, objectives and strategies

In our coding, drawing on Baum (Reference Baum2008) and evidence on health inequities in Australia (e.g. AIHW, 2014) the equity groups we recognised included: people in rural/regional/remote areas; low SES groups; Aboriginal people; lesbian, gay, bisexual, transgender and intersex (LGBTI) groups; people with a disability; people in prison; victims of family violence; and refugees. We distinguished between goals to improve health in such groups, and goals to reduce a health gap or gradient because, as Whitehead and Dahlgren (Reference Whitehead and Dahlgren2006) point out, one could achieve the former without necessarily achieving the latter.

With both objectives and strategies, we sub-coded these in each case according to 13 domains of activity as shown in Table 1. For example an objective on workforce might be to increase retention of Aboriginal people in the health workforce; a strategy in the same domain might be to improve career pathways for Aboriginal employees. Application of these categories and analysis of the relative number of items coded in each allowed us to assess the overall extent to which objectives and strategies in one policy or a group of policies were concerned with health care delivery, individualised forms of disease prevention or health promotion, or addressing SDH/HE outside of health care.

We also cross-coded each identified strategy according to our interpretation of the intended health outcome, using the same categories applied to health goals. Thus, as coding progressed, we could assess strategies in one policy or a group of policies according to the kinds of strategies proposed, and the extent to which they addressed health equity.

We coded strategies as addressing disease prevention or health promotion, but not addressing SDH/HE, when they intended to change individuals’ attributes, choices or behaviours, so that they are more protected against environmental or social risk factors, or replace patterns of unhealthy behaviour with healthy ones. We also included strategies intended to intervene in highly targeted ways to break a link between individuals and a health risk. Strategies recognised as addressing SDH were any we interpreted as intended to change social, economic or cultural environments in order to improve health.

Evidence on SDH/HE

We found that SDH/HE were generally acknowledged and/or that salient evidence was audited in almost all of the policies analysed. While this varied in ‘depth’ from policy to policy, each jurisdiction's policy set contained some documents which recognised SDH/HE explicitly in general terms, and some which audited salient evidence in considerable depth. This passage provides a typical example of general acknowledgement:

Here is an example of more detailed auditing of evidence:

Coding of those individual policies that did most in this area typically identified between 50 and 100 sections of text in each policy either acknowledging SDH/HE or auditing relevant evidence. Commonly recognised social determinants included factors related to: SES (e.g. income); settings (e.g. urban environments); psychosocial conditions (e.g. social support); socio-cultural conditions (e.g. gender, racism); location (e.g. remoteness); and early child development (e.g. family environment).

Each jurisdiction's set of policies audited evidence of health inequities affecting all the equity groups listed earlier. Out of the 266 policies in the data set, 48 focused specifically on issues related to Aboriginal peoples’ health, and all audited evidence on health inequities between Aboriginal and non-Aboriginal Australians. This was the largest group of policies focused on any one population group or health issue.

However, policies only rarely referred to evidence on social gradients in health and, then, usually without detail. Policies generally did not recognise broader policy settings affecting the overall distribution of socioeconomic resources as determinants of health inequities. Thus, evidence on health inequities was predominantly represented as a problem of relatively poor health within specified disadvantaged groups. This is consistent with findings of research in the UK (Smith et al., Reference Smith, Hunter, Blackman, Elliott, Greene and Harrington2009) but, until now, has not been adequately identified as a feature of Australian health policy.

We also found that policies within and across jurisdictions varied in the views they presented on the causes of health inequities. It was common to find simple recognition of (what we recognised as) a health inequity without any accompanying explicit or implicit explanation of its causes or determinants. Where descriptions did offer explanations of a health inequity, these variously incorporated social-causal, health-care, individual-behavioural and social-selection forms of explanation. These respectively assert that health inequities are predominantly caused by differences in: socioeconomic or cultural conditions (outside health care); access to health care; behavioural choices; or health status (wherein existing illness leads to social disadvantage) (Solar and Irwin, Reference Solar and Irwin2010). Examples of each kind of explanation are shown in Table 2.

TABLE 2.

Explanations of health inequities

Within individual policies, or in some particular types of policy, we also found that use of health-care or individual-behavioural explanations of health inequities was followed with strategies preferentially focused on those kinds of interventions. For example, despite significant evidence on social determinants of inequities in cancer risk factors, such as smoking, in Australia (e.g. Thomas et al., Reference Thomas, Briggs, Anderson and Cunningham2008), the four cancer policies in our dataset acknowledged these inequities but typically explained them in terms of differences in individuals’ behavioural choices or health literacy; providing a rationale for a predominantly behaviourist approach to prevention in the strategies which followed. In one case, the South Australia Statewide Cancer Control Plan 2011–2015 recognised evidence on inequities in cancer incidence and mortality between Aboriginal and non-Aboriginal South Australians but also went on to explain cancer incidence as largely a product of individuals’ lifestyle choices. The Plan claimed that: ‘Current estimates indicate that around half of all cancers are due to lifestyle factors and therefore are largely preventable.’ The equity-related preventive strategies which followed focused on improving health literacy on cancer risks and participation in cancer screening among equity groups, especially Aboriginal people.

Similarly, Victoria's Cancer Action Plan 2008–2011 acknowledged that ‘smoking rates are highest among some of the most disadvantaged groups in Victoria’, but went on to make the following claim, as a preface to strategies focused on individualised behaviour change using methods such as social marketing:

The implication that strategies to improve lifestyle choices among disadvantaged groups are an appropriate response to health inequities is also inconsistent with research showing that individualised health promotion strategies such as social marketing campaigns tend, in fact, to exacerbate health inequities (Lorenc et al., Reference Lorenc, Petticrew, Welch and Tugwell2013). Individualised conceptions of health can also shift attention away from social conditions affecting capabilities for healthy behaviours among disadvantaged groups (Dagdeviren et al., Reference Dagdeviren, Donoghue and Promberger2015).

Social selection explanations were used in many of the 31 mental health policies analysed, but rarely used elsewhere. Although these policies often recognised SDH/HE affecting mental health across the broader population, they used social selection explanations to reframe the problem as established mental illness leading to social disadvantage. The strategies proposed then focused on health care or support services for the relatively small group of people with more severe forms of illness and did not address means of keeping the broader population mentally healthy.

Thus, our research shows that evidence on health inequities that are known to be determined in large part by social factors can be recognised in policy but nevertheless explained in ways that shift the locus of the problem and presumptive ‘need’ for policy action onto health care services, individualised health promotion programmes, or support services for people with existing mental illness.

Addressing health inequities

Health goals within each jurisdiction's set of policies were most often framed in terms of improved health for the population at large. However, in each case, around a quarter to a third of all goals coded addressed health inequities. The ways intended improvements in health equity were expressed were consistent with Whitehead and Dahlgren's (Reference Whitehead and Dahlgren2006) types of policy approaches to health inequities. Most common were goals to improve health within an equity group, and then goals to reduce or eliminate a health gap. The great majority of the latter goals were focused on the health gap between Aboriginal and non-Aboriginal Australians. For example:

In relation to Whitehead and Dahlgren's third type of approach, however, only eight policies out of 266 clearly articulated a goal to level-up a health gradient, such as the example below:

Even then, the wording of some of these goals, despite referring to the social gradient, represented health inequities as largely a problem of poor health among the most disadvantaged groups. For example:

Whitehead and Dahlgren (Reference Whitehead and Dahlgren2006) also recognised that a principle of health equity can be defined and operationalised in policy as equity of access to health care. This approach was present in the objectives and strategies of all jurisdictions’ policies. At the national level, the principle of equity of access to public health care according to need was spelt out in national policies such as the National Health Reform Agreement (2011) for Commonwealth funding to States for public hospital and emergency services:

Several of the States’ health care policies included strategies to distribute resources for health care or disease prevention services according to need across their whole population. For example:

Given that many quantitative measures of health follow a social gradient in Australia (AIHW, 2014), we interpreted these strategies as addressing health inequities across the social gradient, whether or not they were explicitly presented in those terms. These strategies to distribute healthcare according to need were the only strategies we identified as addressing health inequities as a gradient.

It is also important to note that one of the key national health care strategies found in these policies was health system reform (under the then Labor government) to shift emphasis from episodic acute care in public hospitals to improved health promotion, disease prevention, and managed care in primary health care (PHC); in keeping with World Health Organisation recommendations (2008). The WHO (2008) argued that policy emphasis on universal, accessible and appropriate PHC is important for supporting health equity (Starfield, Reference Starfield2012) and this is borne out in Australian research (Hajizadeh et al., Reference Hajizadeh, Connelly and Butler2012).

Among all the strategies we identified as addressing health inequities, within each jurisdiction and in the data set as a whole, the most common were those that proposed action to improve access to health care for specific equity groups or disadvantaged areas; whether through infrastructure development in under-served areas, targeted services, improved collaboration between services, improved cultural safety of services, or other means.

The second most common set of actions proposed to address health inequities, in all jurisdictions, were targeted, individualised interventions to prevent disease, promote healthy behaviours, or minimise harm; such as targeted programmes in immunisation, smoking cessation or child protection.

Overall, our findings suggest that Australian health policy at the time of our research contained strategies to address equity in health care combining some elements of Whitehead and Dahlgren's (Reference Whitehead and Dahlgren2006) formulation: a universal approach in primary and acute care systems, aims to strengthen PHC, and targeted strategies to improve access for equity groups. However, there is also reason for concern, given that policies frequently represented and addressed health inequities as a problem of poor health among specific disadvantaged groups and did relatively little to address (or call for cross-sectoral action to address) social gradients in health.

As found in research in other jurisdictions (Smith et al., Reference Smith, Hunter, Blackman, Elliott, Greene and Harrington2009), despite the widespread acknowledgements of SDH/HE and auditing of relevant evidence noted earlier, each jurisdiction's set of strategies to address health inequities predominantly consisted of health care or individualised promotion/prevention strategies.

Social determinants of health and health inequities

Despite the overall, dominant focus on health care and individualised promotion/prevention, each jurisdiction's policies included a lesser group of strategies to address SDH/HE outside health care, suggesting some attention to the broader stewardship role proposed for the health sector by the CSDH (2008). Some of these were directed toward the general population, or broad groups such as children or older people; others focused on equity groups, illness groups, or other minority groups.

We also identified several policy mechanisms in use to stimulate cross-sectoral actions (within or across levels of government) to address health issues; some in health department policies, others in the policies of a head-of-government agency. Many of the strategies on SDH/HE we identified were within these kinds of policies. Instances of these mechanisms at the state level included a ‘Health in All Policies’ programme, public health legislation, and policy based in the Department of Premier and Cabinet. At the national level, a number of policies formed under the Council of Australian Governments were strongly cross-sectoral in the relevant sense. Other national or state government agencies most commonly featuring as partners with the health sector in cross-sectoral policies included those with responsibility for public education, social services, social housing, justice and corrections, and urban planning. State-level cross-sectoral policies also frequently identified local (municipal) governments as partners on health issues, with a focus on strategies in community development, early child development, food security, urban planning and sustainability. These findings indicate that at least some Australian policy makers have taken up, and are seeking ways to operationalise, a whole-of-government approach to address SDH/HE. South Australia and Tasmania were the two jurisdictions doing most in this direction.

Across the dataset and in each jurisdiction we found strategies addressing social determinants in relation to: SES, settings, psychosocial or socio-cultural conditions, location, and early child development. Gender and sexuality were mainly addressed as factors influencing access to health care. Strategies addressing SDH in relation to general population health or broad population groups included strategies in urban planning, food environments, parenting support, food and water safety, and regulation of tobacco industry practices. All jurisdictions supported prescriptive regulation of tobacco but only proposed self-regulation strategies in the food and alcohol sectors, despite the evidence that this does little to change corporate behaviours, and that corporations actively promote self-regulation to avoid regulation by governments (Freudenberg, Reference Freudenberg2014).

It was relatively common for policies on mental health (in all jurisdictions) to include strategies on SDH focused on an illness group. These typically focused on providing support services in housing or employment for the relatively small group of people with more severe, on-going illness; rather than aiming to prevent illness.

Strategies specifically addressing social determinants of health inequities were most apparent in nationally-led policies addressing the health gap between Aboriginal and non-Aboriginal people, especially in early child development and education, employment, housing, and environmental health. Funding was tied to performance measures to progressively eliminate gaps in life expectancy, child mortality and educational outcomes over a 20-year period. However, although other key determinants, such as racism and high incarceration rates or, on the positive side, cultural identities and strong community governance structures, are also significant (Anderson et al., Reference Anderson, Baum and Bentley2007), we found few instances of strategies addressing these.

Beyond this, many of the strategies we identified as addressing SDH were targeted at equity groups, including strategies in: housing, employment support, family support and parenting, community development, and food security. In disability policies, determinants addressed included physical access issues, social inclusion and stigma.

Overall, we saw the range of actions on SDH/HE as a positive, but also noted that the scope of strategies was often limited relative to the nature of the determinant addressed. Examples include: improving menu choices in school canteens (food environments); behaviourist health programmes for parents (family environments for children); or housing support for people with established mental illness (housing affordability).

Furthermore, despite frequent recognition of evidence of the relationship between SES and health we did not identify any attempts within these policies to raise broader issues underpinning socioeconomic determinants of health inequities such as income distribution, taxation structures, welfare payments, education funding, housing affordability, and disparities in employment status as policy areas for possible action, or as the focus for health sector advocacy for action to other sectors. This is a policy silence when judged against research suggesting that concerted policy action to reduce socioeconomic inequalities in Australia could result in considerable savings in public expenditure, by improving health (Brown et al., Reference Brown, Thurecht and Nepal2012). Furthermore, evidence indicates that socioeconomic inequities have in fact increased in Australia over the last decade (OECD Directorate for Employment Labour and Social Affairs, 2011). Other SDH/HE not addressed or addressed in only marginal ways included pollution, new media environments and trade liberalisation. Health risks arising from climate change were only addressed in terms of health service responses to extreme weather events or changes in contagious disease vectors.

Limitations of the research

The analysis of health policy documents employed in this research necessarily took the goals, objectives and strategies identified in those policies as written, and did not assess policy implementation. Furthermore, the view offered here of Australian governments’ actions on SDH/HE does not take account of concurrent policy activity in other sectors (not involving the health sector). Other research is required to investigate how government actions across multiple sectors interact to affect health and health equity. Finally, a focus on existing ‘finished’ policy documents does not consider how policy development processes facilitate or inhibit SDH/HE getting onto the policy agenda. In another part of our research, not reported here, we have conducted case studies on individual policies, selected from our data set as examples of good practice on SDH/HE, to explore in more depth what enabled this to occur during policy development.