Incidence of Food Insecurity Risk Factors Within a Large, Obesogenic Regional Community: A Population-based Study


 Aim: This descriptive study examined the incidence of food insecurity indicators in a large regional community, in Queensland, Australia, where high rates of obesity and corresponding obesity-driven non-communicable diseases are present. The implications of food insecurity in a large regional community are due to findings within the literature that demonstrate an association between mild-to-moderate food insecurity and a higher incidence of overweight or obesity. This concept was termed the ‘food insecurity obesity paradox’ within the literature. Subject and Methods: Six key food insecurity risk factors were examined using chi-square analysis and odds ratio, including education level, unemployment, single-parent status, rental status, Indigenous heritage, and young people. Publicly available information from the Australian Bureau of Statistics was utilized for data analysis.Results: Analysis indicated that social inequity, associated with food insecurity may result in higher obesity levels within regional Australia. These findings were even more pronounced among females, young people, and Indigenous community members, who had a higher incidence of corresponding social and cultural determinants that influenced experience of food system inequity. Conclusion: The present study has indicated that high rates of obesity in ‘at risk’ communities, government policies designed to create social mobility to address social inequity, may be effective in decreasing obesity rates and the associated corresponding chronic disease burden.


Introduction
The worldwide high incidence of obesity is contributing to rising and unprecedented burden of chronic disease on healthcare systems and increases pressures on nite resources within healthcare systems across the world (Australian Institute of Health and Wellbeing [AIHW] 2018a). Reducing this rise in obesity levels has been identi ed by the World Health Organisation (WHO) as a global priority in 2018, stating that it is one of the biggest health challenges of the 21 st century (2018a). This public health challenge is one of the key factors leading to signi cant increases in rates of non-communicable chronic diseases such as coronary artery disease, Type 2 diabetes, and stroke in many regional areas in Queensland, Australia, and throughout the world (AIHW 2018a). The prevalence of obesity and severe obesity are projected to increase to 35% by 2025 in Australia compared to 19% in 1995 (Hayes et al, 2017).
The scienti c literature has demonstrated that obesity rates may be meaningfully in uenced by food insecurity in speci c instances. Mild-to-moderate food insecurity which is de ned as a reduction in the quality, variety, or desirability of dietary intake (Franklin et al, 2012; Dhurandhar 2016) does not involve hunger; however, it can disrupt nutritional intake. However it is not clear why the food insecurity paradox is correlated with obesity, although it is proposed that high calories (withlow nutritional value food) is consumed in lower socio-economic and food insecure regions and these patterns are in uenced by social and cultural determinants (Hayes et al, 2017).
This study was conducted to better understand rates of food insecurity risk factors in the Ipswich community, a large regional community in South East Queensland, Australia (Ipswich City Council, 2017). Food insecurity risk factors were chosen due to the lack of research undertaken to assess rates of food insecurity in this region.
Ipswich community was selected as a case study because people residing in Ipswich experience a disproportionately high level of obesity, do not consume adequate amounts of fruit and vegetables, and have a higher incidence of non-communicable chronic diseases such as diabetes, heart disease, stroke and cancer

Literature Review
Australia has experienced a rapid and sustained increase in obesity levels, reaching 63% of the adult population (Australian Bureau of Statistics [ABS], 2015), with 71% of men and 56% of women being classi ed as overweight or obese (AIHW, 2018a). These increasing obesity rates are causing a signi cant resource drain and adversely impacting health care service optimization. The AIHW (2018a) states that being overweight or obese signi cantly increases the risk of mortality from all causes by 31% with each 5kg/m 2 increase in body mass index (BMI).
Obesity is implicated as a salient risk factor for the majority of the signi cant disease burdens occurring in Australia and across the Western world (AIHW, 2018a). Adults in the lowest socioeconomic geographic areas were 34% more likely to be obese (AIHW, 2018a). In fact, adults living in disadvantaged areas in Queensland were 49% more likely to be overweight than those in advantaged areas. Children in disadvantaged areas in Queensland were 25% more likely to be obese (Queensland Health 2016).
Further, it is well established that some demographic subsets within Australia are faring worse in terms of obesity rates. The Aboriginal and Torres Strait Islander Health Performance Framework report showed that Aboriginal or Torres Strait Islanders were much more likely to be overweight (43%) and obese compared to 28% of the non-Indigenous population (AIHW, 2018a; AIHW, 2018b) and overweight accounts for 70% of the Indigenous population over the age of 18, compared to 63% of the non-Indigenous population (AIHW, 2018a; AIHW, 2018b). Fifty-eight percent of all deaths in the Aboriginal and Torres Strait Islander population in Australia, between 2011 and 2015 have been classi ed by the National Healthcare Agreement 2015 Standards as 'avoidable' and consist of non-communicable, obesity-driven mortality outcomes (AIHW, 2018b).
Research undertaken by Egen and colleagues in the United States of America (USA), using median household incomes, demonstrated a link between socioeconomic disparities and signi cant differences in rates of obesity and life expectancy (Egen, Beatty, Blackely, Brown, & Wykoff, 2017). This comprehensive, longitudinal study found that social-economic stressors such as low income are associated with higher rates of obesity and lower life expectancy (Egen, et.al., 2017). This nding is supported by a large sample size, where Cook and colleagues (2017) established that Asian-American children and adolescents who were from a low socio-economic (SES) ethnic group were signi cantly more likely to be overweight than those in the high or middle SES group.
Whilst these correlations between social determinants and higher obesity rates are clear, the speci c causation remains complex, which reinforces the critical need to undertake further research that explores obesity within a social model of health lens, rather than a biomedical model of health. One theory that potentially accounts for these associations is the 'Food Insecurity Obesity Paradox', which describes how low-to-moderate food insecurity is linked with high obesity rates. There are four pillars to food security; availability of food, access to food, affordability of that food, and the use of food (Charlton, 2016). Charlton (2016, p. 73) de nes food security as "the physical, social and economic ability to access su cient, safe and nutritious food". Auckland, King, Murray, and Saunders (2015 p. vii), however, attempts to de ne food security in a much broader sense as being when "all citizens obtain a safe, personally acceptable, nutritious diet through a sustainable food system that maximizes health choices, community self-reliance and equal access for everyone". This latter de nition of food security takes into account the broader social and cultural context of food choices, accessibility, use, and availability in an equitable food system.
There are varying degrees of food insecurity; Franklin and colleagues stated that food insecurity can range from hunger to mild food insecurity (Franklin et al., 2012). Dhurandar (2016) identi ed that food insecurity is mild-tomoderate when food can be accessed most of the time; however, this may not be the most nutritional food for the household which leads to a disrupted eating pattern. The 'Food Insecurity Obesity Paradox' states that those who are experiencing food insecurity, not driven by hunger, but rather social and cultural constructs that are determining their food choices, are more likely to experience obesity (Ramsey, et al., 2011). Research clearly indicates that mild-to-moderate food insecurity is directly associated with an increased risk of being overweight or obese (Dhuranda Multiple studies from within Australia and USA strongly demonstrate the link between challenging socioeconomic drivers, mild-to-moderate food insecurity, and obesity (Ramsey, et al., 2012b;Rosier, 2012). There is further evidence suggesting that this is particularly prevalent among women, with one study demonstrating women from food-insecure households were up to two BMI units heavier than women from food-secure households (Martin & Ferris, 2007;Franklin, et al,., 2012). The risk of women being in food-insecure households was directly related to single-parent status, renting, lower-income, and lower educational attainment, resulting in women within Australia experiencing a higher percentage of not just one, but multiple food insecurity risk factors (Martin et al., 2017;McDonald, 2011). Swinburne et al., (2020) demonstrated that the levels of obesity are greatly disproportionate in populations in low-income countries, particularly women. Additionally, after analyzing socioeconomic characteristics for both adults and childhood obesity within food insecure households, Martin and Ferris (2007) found that girls are at two times the risk of being overweight or obese than boys, if their parents are obese.
Data for food insecurity in regional Australia is sparse and incomplete. It is apparent from this body of research, that the 'Food Insecurity Obesity Paradox' is prevalent in many lower, socio-demographic areas around the world. The 'Food Insecurity Obesity Paradox' was identi ed in the literature as early as 2004 by Burns as a 'hidden crisis' occurring within Australian communities. The ndings in this review of literature have determined the food insecurity risk factors that have been further analyzed in this study.

Data And Methodology
The detailed, longitudinal characterization of the Ipswich community to determine food insecurity risk factors was undertaken with publicly available data on the ABS website from 2006, 2011, and 2016 census (ABS, 2017a, 2017c, 2017d). The sample utilized in this study was from the census community pro le data collected from 2006, 2011, and 2016, which represents the three most recent census periods (ABS, 2017a, 2017c, 2017d). The community pro le for the local government area (LGA) of Ipswich was extracted from the ABS website (ABS, 2017a). The 2016 census data included 200,123 people and a geographical land area of 108,497.8 hectares (ABS, 2017a). This population rose from 140,182 people in 2006, with an increase of 42.76% over the ten years (ABS, 2017d). The boundaries of the LGA of Ipswich did not geographically change over this period.
The community population pro les containing data for LGA provides speci c data about Aboriginal and Torres Strait Islander peoples, time series, place of paid work, social community data, and working population data (ABS, 2017b). These pro les allow researchers accessing the data to compare and contrast LGA's and overall comparative Australian statistics.

Data Management and Analysis:
Initially, the relevant food insecurity risk factor population data was extracted from the ABS community pro les for Australia and Ipswich. The data was 'cleaned' to ensure that a population would not be counted twice (Abbott, 2016). For example, when ascertaining the relevant data for the Indigenous population variable for Ipswich, the number of Indigenous people in Ipswich was subtracted from the overall population for Ipswich, so that the Indigenous population was not counted twice.
The Chi-Square test including the p-value was then calculated within a Microsoft Excel 2016 spreadsheet using the Chi-Squared formula (Abbott, 2016). This was done by the following steps: a. Finding the variables of interest (e.g., Indigenous Australians and the other population of Australia) within the ABS Community Pro les; b. Extracting the relevant data (i.e. the number of persons within the populations who do and do not have certain food insecurity risk factor); c. Chi-Squared test of independence was then calculated (see below for formula), with p values of < .05 used as statistically signi cance was calculated using an excel spreadsheet. The expected count was also calculated;\ d. This was replicated for the 2006, 2011, and 2016 census data points to evaluate for trends over time.
An additional consideration with large sample sizes is that of Type I error, i.e. the inaccurate rejection of the null hypothesis or "false positive" (Nickerson, 2011). When examining relationships and differences when using large samples, statistically signi cant results are often found that do not re ect a practical signi cance (Khalilzadeh & Tasci, 2017). To provide a more accurate interpretation of signi cant results, effect sizes need to be utilized in result interpretation. However, due to the complex measurement that is often involved in social science, studies with large sample size often report smaller effect sizes, making an accurate assessment of practical signi cance di cult (Levine, Asada, & Carpenter, 2009;Slavin & Smith, 2009). As such, to compensate for both shortcomings in signi cance and effect size, the odds ratio (i.e., the odds of an outcome across conditions) was used as a measure of effect size in this analysis (Levine et al., 2010).

Results
The results are presented according to the identi ed food insecurity risk factors including education level, level of unemployment, single-parent status, renting status, and the Indigenous population and young people. Additionally, when data was available to represent gender, the gender comparison for the Ipswich region has also been presented within the corresponding section.

Overall percentage analysis results
The initial analysis determined the percentage of the population within Australia and Ipswich with at least one food insecurity risk factor, over the three previous census point data collections. Table 1.1 displays these percentages. Overall, it is clear from the percentage data represented that Ipswich had a higher percentage of the population who were experiencing these food insecurity risk factors. These percentages were higher than the national average in almost all categories over the three census dates. However, further analysis was required to understand the statistical signi cance and the relationship between these variables, hence, the ChiSquared test and p-value were calculated and are presented and discussed below.

Chi-Square test and percentage difference results
A Chi-Square test of independence was calculated on the following variables outlined below, to examine the statistical signi cance between these variables (Abbott, 2016). This analysis aims to ascertain if the Ipswich population had a higher proportion of people with food insecurity risk factors than the general population in Australia. This is presented in table form, to demonstrate relationships of expected and actual population counts and includes the Chi-Squared analysis value, the p-value, and the odds ratio value.
The following section analyses the young person (15-24 years of age) population in Ipswich in combination with education, single parent, and employment food insecurity risk factors. This analysis has been undertaken to determine if two food insecurity risk factors were prevalent in the Ipswich young person population and comparisons made to the overall Australian young person population when analyzing each risk factor.
The following section analyses the Indigenous population in Ipswich in combination with education, single parent, rental and employment status food insecurity risk factors. This analysis has been undertaken to determine if two food insecurity risk factors were prevalent in the Ipswich young person population and comparisons made to the overall Australian young person population when analyzing each risk factor.

Discussion
The detailed, longitudinal characterization of the Ipswich community (as a case study) resulted in a comprehensive understanding of the food insecurity risk factors occurring within the region. These results were then juxtaposed to the overall Australian population. This demonstrated that the Ipswich region was at risk of higher rates of multiple food insecurity risk factors than the rest of the country. The analysis of this data has demonstrated that Ipswich experiences ve out of six food insecurity risk factors at a higher rate than the rest of the Australian population with the exception of unemployment (speci cally not in labour force dmographics). Additionally, many of these risk factors were signi cantly more likely in the female population in Ipswich. Further analysis also indicated that young people and the Indigenous population groups were likely to experience more than one food insecurity risk factor. The results also demonstrated that for the Indigenous population in Ipswich, there was a further likelihood of a compounding risk factor for rental status, single-parent status, unemployment for females, and being a young person. This is signi cant for the Indigenous population in Ipswich, as this leads to not only a high risk of food insecurity but additionally, these have all been identi ed as signi cant social determinants that lead to poorer overall health outcomes including increased rates of non-communicable diseases and higher mortality rates (Talbot & Verrinder, 2010; Su et al, 2011).
Further, results indicated that Indigenous people from Ipswich were more likely to have completed year twelve than the Indigenous population in the rest of Australia. This may re ect the signi cant diverse policy initiatives and resources by the state government that has been implemented in the Ipswich region to increase year 12 completion rates for Indigenous people. This does demonstrate how social policy has been customized and prioritized for this group in a community that has been identi ed as at risk of social inequity. However, the Indigenous population in Ipswich were signi cantly less likely to have completed year 12 than the non-Indigenous population in Ipswich. Whilst the data is suggesting that this is improving, it is still a point of disadvantage and a food insecurity risk factor for the Indigenous population of Ipswich.
Overall, the Indigenous population in Ipswich is over two times more likely to be a single parent than the non-Indigenous population in Ipswich. Whilst there was no data available for gender differentiation for this risk factor, the overall single parent statistics in the Ipswich region did demonstrate a signi cantly higher risk for females.
Furthermore, this may indicate that the Indigenous female population in Ipswich has multiple, potentially compounding, risk factors for food inequity and the corresponding, socially determined poorer health outcomes (Rumbold & Dickson-Swift, 2012). Additionally, the high rates of renting for the Indigenous population in Ipswich is of concern as the literature demonstrates that renting status may represent a cascade of disadvantage and social inequity which may be compounded by nancial insecurity (McDonald, 2011).
Indigenous females in the Ipswich region were signi cantly more likely to be unemployed. This is also re ected in overall employment data in the region, where females in the overall population were signi cantly more likely to be unemployed than males. This may be due to historical and culturally contructed family care and domestice responsibilities disproportionate to males (McDonald, 2011). Hence, unemployment is a food insecurity risk factor for both Indigenous and non-Indigenous females in the Ipswich region, which leads to food system and health inequities for this segment of the population.
Additionally, the data analyzed also revealed that there were higher rates of young people (15-24 years of age) in the Ipswich Indigenous population than the overall Australian Indigenous population. The literature demonstrates that young people are less likely to be nancially secure and hence, more likely to be renters which is an additional risk factor (McDonald, 2011). The signi cance of potentially multiple social risk factors that may lead to food insecurity and poorer health outcomes is therefore evident within this population group in Ipswich.
The analysis of gender in the available data revealed a signi cant difference in the unemployment levels between Indigenous males and Indigenous females in the Ipswich region. Indigenous females in the Ipswich region were signi cantly more likely to be unemployed. This is also re ected in the overall employment data in the region, where females in the overall population were signi cantly more likely to be unemployed than males. Additionally, young people who are females are signi cantly more likely to be single parents than males in the Ipswich region. This demonstrates that young females in the Ipswich region are at higher risk of multiple and signi cant food insecurity risk factors that may lead to a higher risk of food insecurity, obesity, and associated non-communicable disease burden (Martin et al., 2007).
The impact of these social determinants in the Aboriginal and Torres Strait Islander community has resulted in and been perpetuated by transgenerational trauma, social and health inequity, and cultural degradation caused by gross class and racial divides and government policy aimed at segregation, over the history of white settlement in Australia (Walsh- Dilley, Woldford & McCarthy, 2016). The social, health and cultural systems that have been forced upon the Indigenous peoples of Australia have ostracised this group of people from their traditional food systems, food sources, and cultural and social structures resulting in gross health and nutritional inequities (Rosier, 2012).
Research (McDonald, 2011) demonstrated that young people are at risk of a cascade of disadvantage and social inequity driven by socio-economic factors perpetuated by nancial insecurity. The ndings from the analysis of food insecurity risk factors for the Ipswich region indicated that there was a higher rate of young people in the community than the overall population in Australia. The analysis revealed the relationships between young people and employment, single-parent status, and education which demonstrated that young people may be at risk of more than one food insecurity risk factor, which may result in social, health, and food inequity (Rumbold et al., 2012;Talbot et al., 2018).
However, the ndings presented in this paper also re ected that young people in Australia are much less likely to be single-parents than those in the over 25 years of age group. This may potentially re ect the high divorce rates Australia is experiencing, which occurs predominately outside of this age group (McDonald, 2011). However, in Ipswich, young people are over two times more likely to be single-parents relative to the rest of Australia. This is a signi cant risk factor for the young people in the Ipswich community as this social determinant is correlated with food insecurity risk and further health disparities (Charlton, 2016;Ramsey et al., 2012) and coupled with age, is a compounding risk. Additionally, young people who are females are signi cantly more likely to be single-parents than males in the Ipswich region. This demonstrates that young females in the Ipswich region are at higher risk of multiple and signi cant food insecurity risk factors that may lead to a higher risk of food insecurity, obesity, and associated non-communicable disease burden such as Type 2 Diabetes (PHN HNA, 2018; Martin et al., 2007).
When analyzing the risk factor of education for young people, the data re ected that young people across Australia, including in Ipswich, are more likely to have completed year 12 as opposed to people over the age of 25. This is re ective of the general trend in society, of more people completing secondary education (McDonald, 2011). However, in Ipswich, young people are less likely to complete their secondary education compared to the rest of Australia. This is a signi cant additional risk factor for young people in the Ipswich region. This risk factor, however, was more signi cant for males in Ipswich than females, which may indicate the fact that more young males in lower SES areas leave school before year 12 to enter trades (McDonald, 2011).
The results from this study have established that for young people in the Ipswich region, the unemployment trend is increasing and the single-parent status is signi cantly higher than the rest of the Australian population in this age group, particularly for young women. The young people population group in Ipswich is also more likely to have not completed year 12 than compared to young people in Australia. This leads to young people in Ipswich at risk of multiple and at times signi cant risk factors which may be resulting in food system inequity. This has signi cant social health policy implications for strategies that attempt to in uence nutritional intake within the Ipswich region for this population.
One positive trend that is occurring in the Ipswich region for young people demonstrates that the gap has closed between males and females in relation to employment, with no difference in the employment levels for either male or female young people by 2016. A factor that may in uence the rates of young people entering the workforce may be single-parent status or if they are staying at home parents. Those people, who are more likely to be women (Franklin et al., 2012), would not be re ected in unemployment data which may potentially in uence this trend. Hence, these gures may not be re ective of what is occurring in the Ipswich region. Further crosssectional analysis would be required to ascertain if this is a more signi cant problem than what is re ected in the data analyzed in this study.
The social inequities driving food insecurity include education, single-parent status, rental status, and the Indigenous and young people demographics within the Ipswich community are signi cant. Additionally, there are several signi cant food insecurity risk factors for women, Indigenous people and young people in the Ipswich region and the data does indicate that potentially concurrent risk factors may be occurring within these population groups. The literature indicates that these social inequities lead to a higher risk of food insecurity, higher risk of being overweight or obese, higher mortality rates, and higher non-communicable disease rates (AIHW, 2018; Talbot et al., 2010; WHO, 2018a).

Strengths, Limitations And Opportunities
A strength of this research is that it accessed a very large, robust, publically available data set from the ABS, which gave a clear data in relation to the prevalence of food insecurity risk factors for the Ipswich community.
Further opportunities exist to extend this research to include the next census data release to further understand trends occurring within the Ipswich community. This data set provided information on inequities occurring within Ipswich which is not only applicable to the food system and has wider health implications for the community. Social inequities have been linked to overall lower life expectancy, higher morbidity rates and poorer disease survival rates. This has broad implications for future practice, research and social health policy.
Whilst this data idneitifed the prevalence of food insecurity risk factors, there has however been no studies conducted within the Ipswich community of self reported food insecurity rates. Further research is required to understand this phenomena that is occurring in the community and the relationship to obesity rates.

Conclusion
Whilst this data highlights 5 out of 6 risk factors are more pronounced in the Ipswich community than the rest of Australia, this does not directly demonstrate the degree to which food insecurity is present or direct causation between these risk factors and high obesity rates. However, many risk factors for food insecurity have been identi ed within the Ipswich community that are occurring at a higher rate than elsewhere in Australia. This has several implications for the development of social health policy. The research suggests that government policies based on increasing social mobility in lower socio-economic regions may be required to address the obesity epidemic. Food system inequities require a long-term, sustained, and strategic leadership approach to prioritize the most vulnerable within society, which will need bi-partisan government support, multi-sectoral collaboration, and strong leadership to slowly enhance food system equity within the community. Friel and colleagues (2015) argue that an effective and sustainable response to increasing levels of obesity must include both broader structural approaches that address social inequity. This includes understanding the socio-economic demographics of the community in which interventions are embedded and in uencing food insecurity risk factors. The de nition of a young person was determined by using the 15 -24 age group in the ABS census data.