Social determinants and lifestyle risk factors only partially explain the higher prevalence of food insecurity among Aboriginal and Torres Strait Islanders in the Australian state of Victoria: a cross-sectional study

Food insecurity has been defined as the “limited or uncertain availability of nutritionally adequate and safe foods or limited or uncertain ability to acquire acceptable foods in socially acceptable ways” [1].

The prevalence of food insecurity in developed countries ranges from 4 to 14% [2]. In Australia, a single item question; “In the last 12 months, were there any times that you ran out of food and couldn’t afford to buy more?” was first included in the 1995 National Nutrition Survey to estimate the prevalence of food insecurity [2]. Subsequently the question was incorporated into the National Health Survey program and the prevalence of food insecurity in Australia was reported to be approximately 5%; the unemployed, low income households, lone parents and young people being over-represented [3].

The National Aboriginal and Torres Strait Islander Health Survey in 2004–05 found that the prevalence of food insecurity was higher (24%) among Australians of Aboriginal and Torres Strait Islander descent [4]. In the second most populous Australian state of Victoria, the substantially higher prevalence of food insecurity among Aboriginal and Torres Strait Islanders, adjusted for age, was found to be 17.7% (crude prevalence = 20.3%) [5].

It is well established that Aboriginal and Torres Strait Islander Australians are notably more socially and economically disadvantaged than their non-Aboriginal and Torres Strait Islander counterparts [6]. There is some evidence, however, that the main cause of food insecurity among Aboriginal and Torres Strait Islanders who live in remote rural communities is food scarcity rather than economic constraint [7]. However, the state of Victoria does not have any remote rural communities. Therefore, we hypothesised that the substantial difference in prevalence of food insecurity between Aboriginal and Torres Strait Islanders and their non-Aboriginal and Torres Strait Islander counterparts in Victoria would primarily be explained by economic constraint [8].

The aim of this paper was to investigate the relationship between food insecurity and Aboriginal and Torres Strait Islander status in the state of Victoria. More specifically, this paper aims to: (1) identify factors associated with food insecurity in Victoria; (2) measure the strength of the association between Aboriginal and Torres Strait Islander status and food insecurity; and (3) describe the influence of other factors associated with food insecurity on the association between Aboriginal and Torres Strait Islander status and food insecurity.

Table 1 shows the demographic characteristics of the Victorian population by Aboriginal and Torres Strait Islander status. Overall, the Aboriginal and Torres Strait Islander population was younger with a greater proportion residing in rural Victoria, compared with the non-Aboriginal and Torres Strait Islander population.

The crude prevalence of food insecurity in Victoria (Table 2) was substantially higher among Aboriginal and Torres Strait Islanders (20.3%), compared with their non-Aboriginal and Torres Strait Islander counterparts (5.4%).

Table 2 shows the results of the univariable logistic regression analysis that sought to identify factors that were associated with food insecurity. Excessive alcohol consumption and geographic location (rurality) were not associated with food insecurity. The remaining variables all showed associations with food insecurity.

Adult Victorians of Aboriginal and Torres Strait Islander descent were more than four times more likely to have experienced at least one episode of food insecurity compared with their non-Aboriginal and Torres Strait Islander counterparts, odds ratio (OR) = 4.5(95% confidence interval (CI); 2.7-7.4).

Food insecurity was more likely to be experienced by females than males (OR = 1.4; 1.2-1.7). Age was associated with food insecurity with persons aged 18–24 years (OR = 3.4; 2.5-4.6), 25–34 years (OR = 3.7; 2.8-4.9), 35–44 years (OR = 3.0; 2.4-3.9), and 45–54 years (OR = 2.4; 1.9-3.1) being more likely to have experienced food insecurity compared with those aged 65 years and older, while those aged 55–64 years were no different to those aged 65 years and older (OR = 1.2; 0.9-1.6).

SES, measured by total annual household income, was strongly associated with food insecurity. For every decrease in income bracket, food insecurity increased by 70% (OR_trend = 1.7; 1.6-1.8).

Of the lifestyle risk factors evaluated, only smoking (OR = 3.5; 3.0-4.1) and obesity (OR = 1.6; 1.4-2.0) were associated with food insecurity.

All three indicators of social support (inability to get help from family or neighbours or friends) were associated with food insecurity; friends (OR = 3.6; 3.0-4.4), family (OR = 3.5; 2.8-4.3) and neighbours (OR = 2.5; 2.1-3.0).

All three indicators of household composition were associated with food insecurity; being a lone parent (OR = 4.0; 3.3-4.9), households with dependent children (OR = 1.5; 1.3-1.8), and household size (OR = 1.1; 1.1-1.2).

Table 3 summarises the relationship of food insecurity to Aboriginal and Torres Strait Islander status after adjustment for other factors associated with food insecurity.

Controlling for sex had no impact on the crude OR. By contrast, the adjusted OR (OR_adj) when age was taken into account was 4.0 (2.5-6.6), representing an 11.1% reduction of the crude OR.

Controlling for SES (total annual household income) reduced the crude OR by 13.3% (OR_adj = 3.9; 2.3-6.6).

Of the two lifestyle variables, controlling for smoking status reduced the crude OR by 15.6% (OR_adj = 3.8; 2.3-6.3). By contrast, controlling for obesity increased the crude OR by 8.9% (OR_adj = 4.9; 2.9-8.3).

Of the three social support variables, controlling for the inability to get help from neighbours had no impact on the crude OR. By contrast, controlling for being unable to get help from family reduced the crude OR by 8.9% (OR_adj = 4.1; 2.5-6.9) and controlling for being unable to get help from friends increased the crude OR by 6.7% (OR_adj = 4.8, 2.9-7.9).

Of the three household composition variables, controlling for being a lone parent household, households with a child or household size had no impact on the crude OR.

Subsequently, we fitted various multivariable models to examine the relative impacts of each of the domains of the factors that were associated with food insecurity. Age and sex, SES and lifestyle risk factors were the only domains that appeared to make a substantial impact as measured by a change in the crude OR. However, when the individual factors were examined, the impact of sex and obesity was small if present at all, while the impact of the inability to get help from friends appeared to cancel the impact on the inability to get help from family because their effects were in opposite directions.

We ran a multivariable model that included only the variables that changed the crude OR by 5% or more: age, smoking status, obesity, inability to get help from family and/or friends. This reduced the association of food insecurity and Aboriginal and Torres Strait Islander status by approximately 38% from OR = 4.5 (2.7-7.4) to OR_adj = 2.8 (1.6-5.0). The model provided a good fit to the data, by the F-adjusted mean residual goodness-of-fit test: F(9,26972) = 1.28, prob > F = 0.2413. When we included all the other factors that were associated with food insecurity, irrespective of whether they impacted on the crude OR, we obtained similar results.

All first order interaction terms with the primary explanatory variable were not statistically significant and therefore not included in the model.

We investigated the relationship between food insecurity and Aboriginal and Torres Strait Islander status and show that Aboriginal and Torres Strait Islanders in Victoria were markedly more likely to have experienced food insecurity than their non-Aboriginal and Torres Strait Islander counterparts. Age, household income, smoking, obesity, and the inability to get help from family and/or friends only partially explained the higher prevalence of food insecurity among Aboriginal and Torres Strait Islanders.

Remote and rural communities may have problems accessing food due to geographical barriers; resulting in food scarcity, which has been identified as a major cause of food insecurity among Aboriginal and Torres Strait Islander populations in Australia [7]. Our finding that rurality (being resident in rural Victoria) was not associated with food insecurity supports our hypothesis that food scarcity is not a significant determinant of food insecurity in the state of Victoria, and highlights the importance of not assuming that the determinants of health and well-being among Aboriginal and Torres Strait Islanders are the same across the country.

The finding that younger age is associated with food insecurity, while older age appears to be protective is consistent with the literature [8]. This may reflect the fact that Australia provides a modest safety net for the elderly that is often lacking for the younger age groups. Since the Aboriginal and Torres Strait Islander population is a much younger population, it is therefore to be expected that this would render Aboriginal and Torres Strait Islanders more vulnerable to food insecurity.

Total annual household income and smoking made the largest contributions to the relationship between food insecurity and Aboriginal and Torres Strait Islander status; reducing the OR by approximately 13% and 16%, respectively. Anglicare Australia, in their recent study of the clientele at emergency relief centres, concluded that insufficient income is the key determinant of food insecurity in Australia, as food is often the only discretionary item in a low-income household budget [8]. Our findings confirm the importance of income; as the lower the household income the higher the OR of food insecurity, with those in the lowest household income bracket being 8.4 times more likely to have experienced food insecurity compared with those in the highest household income bracket. However, while household income is often used as a proxy of SES, it does not capture all aspects of SES [13]. SES is a multi-dimensional concept and other indicators of SES include: education, occupation, wealth, and area-based composite measures. These indicators are not necessarily interchangeable as they do not always correlate highly with each other; for example, the correlation between income and education has been reported to vary by ethnic group from 0.34 to 0.58, reflecting that income can vary at similar levels of education across different ethnic groups [14]. We selected household income as a measure of SES, based on our consideration of the most plausible causal pathway between the outcome variable of food insecurity and SES; since, income provides individuals and families with the necessary material resources and determines their purchasing power for accessing goods and services. However, it is possible that the difference in prevalence of food insecurity between Aboriginal and Torres Strait Islanders and their non- Aboriginal and Torres Strait Islander counterparts, after adjusting for household income, may still reflect unmeasured socioeconomic differences not captured by household income. Moreover, income as a measure of SES has been shown to be a more unstable measure than education or occupation with a higher non-response rate than other measures [15].

The prevalence of smoking has consistently been reported to be higher in the Aboriginal and Torres Strait Islander population [5]. Hypothetically, the high cost of tobacco products might be expected to put strain on low income households and to be in direct competition with other discretionary household items. However, adjusting for smoking status only reduced the OR by 15.6%; suggesting that smoking did not make a particularly large contribution to the higher prevalence of food insecurity among Aboriginal and Torres Strait Islanders. Given that this is a cross-sectional study design, one cannot rule out the possibility of reverse causation where the stress of being subject to food insecurity causes people to smoke.

Excessive alcohol consumption was not associated and made no difference to our findings whether included or not. The wider societal belief that Aboriginal and Torres Strait Islanders are more likely to engage in excessive alcohol consumption has been the basis of negative stereotyping that has fuelled racist perceptions. This belief, however, is not supported by the evidence. Our data found that the proportion of adult Victorians who engaged in excessive alcohol consumption was no different between Aboriginal and Torres Strait Islanders compared with non-Aboriginal and Torres Strait Islanders (43.9% vs. 45.5%). By contrast, Aboriginal and Torres Strait Islander men were more likely to abstain from alcohol consumption than their non-Aboriginal and Torres Strait Islander counterparts (22.4% vs. 12.4%, p < 0.05) [5]. These findings make an important contribution to dispelling ongoing negative stereotypes and combating racism, given that racism is a key determinant of Aboriginal and Torres Strait Islander health and wellbeing [16].

Aboriginal and Torres Strait Islanders in Victoria bore a higher burden of child removal from their families than any other state in Australia due to previous government assimilation policies [17]. It is possible that our finding that Aboriginal and Torres Strait Islander Victorians were less able to get help from family than their non-Aboriginal and Torres Strait Islander counterparts may be a reflection of the success of such past policies. Since family is usually the first point of contact in times of personal crisis, lacking family support means lacking an important resource that can effectively reduce an individual’s vulnerability to negative life outcomes such as food insecurity. The assimilation policies of previous governments are now recognised as a form of institutional racism and exemplify how racism continues to be a key determinant of Aboriginal and Torres Strait Islander health and well-being, as their impacts continue to be experienced by the current generation [18].

Indigenous people in other comparable countries, such as Canada and New Zealand, have also been reported to experience a higher prevalence of food insecurity than their non-indigenous counterparts [19, 20]. A study in Canada showed similar findings to our study; where, adjusting for SES reduced but did not eliminate the difference in prevalence [19]. This study used several indicators of SES including education, income, and home ownership; controlling for more aspects of SES than our study, yet still not able to fully account for the difference in prevalence of food insecurity between the indigenous and non-indigenous populations. The study employed a more comprehensive tool to measure food insecurity than we used in our study; the US Household Food Security Survey Module (HFSSM). They concluded however, that a shortcoming of the HFSSM was that it did not collect any information about the coping strategies of the food insecure and that it was therefore possible that differences in coping strategies may also have contributed to differences in the prevalence of food insecurity. Similarly, the VPHS only contained a single question pertaining to food insecurity and did not collect any information on coping strategies; such as, the use of food banks and the rationing of food. It is therefore possible that unmeasured differences in coping strategies may have contributed to the difference in the prevalence of food insecurity between Aboriginal and Torres Strait Islanders and their non-Aboriginal and Torres Strait Islander counterparts; which would warrant investigation.

Food insecurity is a serious problem among Aboriginal and Torres Strait Islanders who reside in the Australian state of Victoria. Differences between Aboriginal and Torres Strait Islanders and their non-Aboriginal and Torres Strait Islander counterparts in age structure, SES (household income), prevalence of lifestyle risk factors and social support only partially explained the higher prevalence of food insecurity among Aboriginal and Torres Strait Islanders. Further research is needed to identify the reason(s) for the substantially higher prevalence of food insecurity among Aboriginal and Torres Strait Islanders in Victoria.

To our knowledge this is the first study of its kind to investigate the determinants of food insecurity among Aboriginal and Torres Strait Islander in the state of Victoria using a population representative approach.

Information on the health and well-being of Aboriginal and Torres Strait Islanders in Australia most often comes from national datasets and is often dominated by studies with a focus on remote rural communities. This is mainly due to pragmatic considerations relating to sample size; Aboriginal and Torres Strait Islanders only constitute 2.5% of the Australian population with 73% residing in three states: New South Wales, Queensland and Western Australia [23]. Our work therefore highlights the importance of being reminded that while a small population in size, the Aboriginal and Torres Strait Islander population is a culturally and linguistically diverse population; geographic location is an important determinant of that diversity.

Our work contributes to the overall understanding of food insecurity in Australia, which, to date, has largely been based on treating ethnicity as a risk factor for food insecurity. We confirm that income insufficiency is indeed a key determinant of food insecurity. However, we also show that income insufficiency is not in itself sufficient to explain food insecurity among Aboriginal and Torres Strait Islanders.

Future research on food insecurity among Aboriginal and Torres Strait Islanders should be directed at investigating all facets of food insecurity; including factors that influence the food supply, access to and utilisation of the food supply. Coping strategies for households at risk of food insecurity is another important area of focus; as is, understanding the role that institutional and other forms of racism (past and present) continues to play.