Disability and food insecurity in Aotearoa New Zealand: a population-based analysis

Food insecurity, defined as limited or uncertain access to adequate and safe foods, is a major public health challenge with significant equity implications. In 2024, an estimated 673 million people worldwide (8.2% of the global population) experienced chronic hunger (undernourishment, defined by the Food and Agriculture Organization as habitual insufficient dietary energy intake to maintain a normal, active and healthy life), with 2.3 billion people facing moderate or severe food insecurity—335 million more than before the COVID-19 pandemic.¹ The pandemic illuminated persistent hunger disparities, particularly affecting vulnerable groups, including rural populations and Indigenous peoples.²

Globally, 1.3 billion people live with disabilities.³ Disabled people are disproportionately food insecure due to economic disadvantage, limited mobility and access to food sources, discrimination, costly dietary needs and inadequate accessibility in food assistance programmes.⁴ In the United States of America, 28% of households with disabled adults have food insecurity compared to 7% where there is no disability.⁵

In Aotearoa New Zealand, food security is a public health priority.⁶ Inequities are pronounced among Māori and Pacific children⁷ and among disabled people, who make up 24% of the population and 26% of Māori.^8,9 Employment rates illustrate these disparities: only 22.5% of disabled people were employed compared with 69.8% of non-disabled people.¹⁰ Lower incomes, mobility barriers and limited transport further restrict food access.¹¹ Additionally, the 2013 Disability Survey found that 11.8% of disabled people in New Zealand reported dietary restrictions due to their disability,⁹ indicating the importance of tailored food security measures that consider their unique dietary needs.

Despite growing recognition of food insecurity as a public health challenge, research is scarce concerning the intersection of disability and food security in New Zealand. Macaulay et al.¹² examined data from the New Zealand Health Survey (NZHS) and Youth2000 surveys, highlighting persistent food insecurity among youth and children, with notable disparities linked to ethnicity and disability. Similarly, Fanslow et al.¹³ investigated intimate partner violence and found a significant association between multiple disabilities and food insecurity among women. These studies represent the limited research available on the relationship between disability and food security in New Zealand, underscoring the evident gap that our study seeks to fill within the public health evidence base. As such, this study aims to explore the association between disability and food insecurity and determine how the level, type and complexity of disabilities relate to food insecurity through an analysis of secondary data from the 2019/2020 NZHS.

Method

Data source

This study analysed data from the 2019/2020 NZHS. The survey collects comprehensive data on various aspects, such as health status, health behaviours, health service utilisation and socio-demographic factors, including children’s and adults’ information. It includes specific questions on disability status and household food security, which were essential for addressing our research questions. A full description of the survey methodology and sample has been published by the Ministry of Health.¹⁴

The 2019/2020 NZHS was utilised as the most recent NZHS with a dataset that included both adult disability and household food security data at the time of our analysis. The 2020/2021, 2021/2022 and 2022/2023 NZHS were excluded since household food security data were only available for households with children in these surveys (i.e., the household food security questionnaire was only included in the child module, not the adult module).¹⁴

Disability measurement

The NZHS disability-related questions use the Washington Group Short Set (WG-SS), which consists of six items enquiring about visual, hearing and walking difficulties, as well as difficulties in remembering or concentrating, washing or dressing and communication. Respondents may report i) no difficulty, ii) some difficulty, iii) a lot of difficulty, or iv) cannot do at all. In the NZHS, respondents who indicated “a lot of difficulty” or “cannot do at all” in the assessed domains were classified as disabled people.¹⁴ By default, those who reported “no” or “some difficulty” are considered non-disabled people. In the dataset used for this study, the “cannot do at all” category contained very few (or no) respondents across all disability domains; therefore, it was combined with “a lot of difficulty” to form a single category, “severe difficulty”, for the analyses. The number of children identified as disabled by the WG-SS in one survey year of the NZHS is too small to allow for robust conclusions. Therefore, our analysis focussed on adults only.

Food security measurement

The NZHS uses a New Zealand–specific eight-item questionnaire based on the New Zealand Food Security Model¹⁵ to assess various dimensions of food insecurity, such as food availability, accessibility, utilisation and stability, taking into account cultural preferences and dietary needs specific to the New Zealand context. To classify households as food insecure or not food insecure, a food security status variable was created as a function of the eight indicators, following the scoring protocol of Smith et al.¹⁶ Households were classified as “food insecure” if respondents answered affirmatively to at least two of the eight questions; otherwise, they were classified as “not food insecure”. An affirmative response was defined as “always” or “sometimes” for question CFS1.01, and “often” or “sometimes” for questions CFS1.02–CFS1.08. The full set of questions is provided in Appendix 2.

Other variables

Demographic and socio-economic variables are likely to play a role in food security status. Therefore, age group, sex and ethnicity, employment status, education level and New Zealand Index of Deprivation (NZDep) quintile were included in the analysis.

Prioritised ethnicity, as recorded on the NZHS database, was used for analysis. Under this classification, each person is allocated to a single group (Māori, Pacific peoples, Asian, or European/Other) even if they identify with multiple ethnicities.

Statistical methods

Exploratory data analysis

Initially, an exploratory data analysis was performed, which included contingency tables and bar plots to explore potential relationships between the six disability indicators and each of the eight household food security status indicators. Cross tables were also built to examine disabilities domains and food security status across demographics and socio-economic variables categories. After initial results suggested an association between disability and household food security, and the association of these with demographics and socio-economic variables, another round of data analysis was performed, focussing on the relationship between the household food security status variable and the six disability factors.

Regression analyses

Relationships between household food security status and disability factors were examined using modified Poisson regression.¹⁷ Models were fitted sequentially. Univariable models were used to estimate the total association between each disability domain and food security status. A multivariable model including all six disability domains was then fitted, followed by sequential adjustment for demographic covariates and potential confounders (age, sex and ethnicity) and subsequently for socio-economic covariates considered potential mediators (employment status, education level, and NZDep quintile).

All analyses accounted for the complex survey design of the NZHS, as described in the methodology report.¹⁴ Analyses were performed in R (version 4.4.1)¹⁸ using the packages tidyverse, haven, ggpubr and survey.

The Wald test (p<0.05) was used to assess overall associations. In addition, tests for linear trend were performed using ordinal coding of disability severity (0 = no difficulty, 1 = some difficulty, 2 = severe difficulty) to evaluate graded associations across levels. Model fit was assessed using the normalised sum of squared residual test and visual inspection of residual plot to identify patterns of mis-specification. Additionally, the occurrence of predicted probability exceeding 1 was examined to ensure the stability of the estimate given the binary nature of the outcome.

Results

The relationship between disability and food insecurity was explored, aiming to analyse how various dimensions of disability—specifically the level, type and complexity—are associated with food security status.

Figure 1 displays a comprehensive overview of the eight food security status indicators among people with disabilities. While only a proportion report stable access to food, 46% cannot always afford to eat, and many experience serious challenges: some regularly go without food, often run out or rely on others and food banks. Limited dietary variety is common, and food-related stress, both about cost and in social situations, is prevalent.

View Figure 1–2, Table 1–4.

Figure 2 shows that the proportion of respondents experiencing food insecurity varies by levels of disability. Food insecurity is more common among households experiencing some or severe difficulty, in contrast to those who have no difficulty in any of the disability domains. Furthermore, increasing levels of difficulty across all disability types show increases in the proportion of households experiencing food insecurity. This provides evidence that non-disabled people and their households are less likely to experience food insecurity compared with those with some degree of disability.

Disability and food security status: socio-demographic characteristics

Table 1 summarises socio-demographic characteristics according to household food security status, and disability status and type. The survey included 9,699 adult participants, representing 77% of those invited to participate. Based on survey-weighted estimates, the median age of adults in the population is 45 years, with a range from 15 to 90 years. The population has slightly more females than males. Around 20% of the adult population is estimated to be Māori or Pacific, while the majority (over 65%) are estimated to identify as European or Other, according to the New Zealand prioritised ethnicity classification. An estimated substantial proportion of adults (12.83%, or approximately 52,368 individuals) live in food-insecure households, while 8% of adults have a disability. Difficulty with walking was the most frequently reported disability, followed by challenges with memory or concentration, vision problems and hearing impairments. In terms of complex disabilities, 2.2% of individuals have two or more disabilities, complementing the information presented in Table 1.

Estimates indicate that adults not in paid employment have a higher risk of living in food-insecure households compared with those in paid employment. For example, the risk of being food insecure among those in paid employment is 10.3%, compared with 28.2% among those not working but looking for a job. Differences are also observed across education levels, with an estimated risk of 18.3% among adults with no qualifications, compared with 8.1% among those with post-secondary non-degree qualifications. A clear gradient is also observed across NZDep quintiles, with the risk increasing from 6.1% in the least deprived areas to 24.8% in the most deprived areas. Differences are more pronounced across ethnic groups. The estimated risk of food insecurity is 9.8% among those who identify as European/Other, compared with 23% among Māori and 30% among Pacific peoples.

Disability is not evenly distributed across demographic groups. Disability is more common among those not working and not looking for work (16.8%) compared with those in paid employment (3.7%). Similar patterns are observed across education levels, with higher proportions among adults with no qualifications (14.9%) compared with those with post-secondary non-degree qualifications (4.8%) or a bachelor’s degree or higher (10.7%). A clear gradient is also seen across NZDep quintiles, where the proportion with disabilities increases from 3.7% in the least deprived areas to 12.7% in the most deprived areas.

Relationship between food security status and disability

Table 2 shows that 27.4% of disabled people experience food insecurity, which is substantially higher compared with 11.6% of non-disabled people who experience food insecurity. In other words, disabled people are 2.4 times more likely than non-disabled people to experience food insecurity.

Table 3 presents the percentage of food insecurity by ethnicity and disability status.

Overall, the percentage of food-insecure people increases significantly among the disabled group compared with the non-disabled group, regardless of their ethnicity. By far, the most food insecure ethnic groups are Pacific peoples and Māori, accounting for 27% and 21% respectively within the non-disabled groups. These percentages increase dramatically within the disabled groups, reaching 57% and 35%, respectively.

Modified Poisson regression analysis

Table 4 displays relative risks (RR) and 95% confidence intervals from sequential modified Poisson regression models examining the association between disability domains and food insecurity, with progressive adjustment for demographic and socio-economic covariates.

The unadjusted models (Table 4, Model 1 column) show a strong association between each disability domain and food insecurity, with risks increasing with the severity of difficulty, using no difficulty as the reference category. After mutual adjustment for all disability domains and further adjustment for age, sex and ethnicity (Model 3), the associations were attenuated, although the disability domains remained significantly associated with food insecurity. This suggests that the observed relationships are partly explained by shared variance across disability domains and demographic factors, but that independent associations persist. Further adjustment for employment status, area deprivation and education led to additional attenuation of the associations (Model 4). The percentage change in RR (from Model 3 to Model 4) was up to 44%, suggesting that socio-economic factors partially mediate the relationship between disability and food insecurity but do not fully explain it. For example, among individuals with severe walking difficulty, the RR (compared with those with no walking difficulty) decreased from 1.82 to 1.57 after adjustment for socio-economic variables, indicating that disability may influence socio-economic position, which in turn affects food insecurity. Nevertheless, the RR remains statistically significant for all disability domains, indicating that even after accounting for socio-economic factors individuals with severe walking difficulty and their households, for example, have a 57% higher risk of experiencing food insecurity compared with those without walking difficulty. Overall design-based Wald tests confirmed that all disability domains were significantly associated with food insecurity (all p<0.001). In addition, significant linear trends (p<0.001) were observed across severity levels for all domains, indicating a consistent dose–response relationship, with increasing disability severity associated with higher risk of food insecurity.

Impact of multiple disabilities on household food insecurity

The RR for individuals with more than one disability can be estimated as the product of the corresponding RR presented in Table 4, since the modified Poisson model with a log link assumes multiplicative effects on the risk scale. An important finding from the modified Poisson model (Model 4) is that households of respondents with difficulties in two or more disability domains are more likely to experience food insecurity compared with those with just one of these difficulties. For example, Model 4 in Table 4 shows that a person with severe visual difficulty is 1.16 times more likely to experience food insecurity compared with a person with no difficulty, while someone with severe hearing difficulty is 1.20 times more likely. A person with both severe visual and hearing difficulties is therefore 1.39 times (1.16×1.20) more likely to experience food insecurity. Thus, households of people with complex disabilities (two or more disabilities) experience greater food insecurity.

Discussion

Our analysis of the 2019/2020 NZHS provides the first comprehensive examination of the relationship between disability and food insecurity among adults in New Zealand. Disabled people face more than twice the risk of food insecurity compared with non-disabled people, and this association persists after adjusting for socio-economic factors.

The 8% estimate of disabled adults in our study is lower than national estimates (24%)⁹ and global estimates (15–16%),^3,19 largely due to differing definitions: Stats NZ includes 10 disability domains, whereas our study covers six using a higher severity threshold. Despite this, our findings align with international evidence showing people with disabilities consistently experience higher rates of food insecurity.^20,21

Consistent with other New Zealand studies,^6,7,22,23 we found higher rates of food insecurity among disabled people across all indicators. “Limited variety of food due to lack of money” was the most commonly reported challenge, while “relying on others for food” was the least. Food insecurity increased with disability severity across all domains (Figure 2). However, this relationship was substantially attenuated after adjusting for socio-economic factors—RR decreased by up to 44%—suggesting that disability affects food security partly through reduced employment and income. Interventions improving economic participation for disabled people may therefore help reduce food insecurity. That said, significant associations remained after adjustment, pointing to disability-specific barriers beyond socio-economic disadvantage that warrant further investigation.

All disability types were associated with food insecurity, with compounding risk for those with multiple disabilities. Ethnic disparities were stark: 57% of disabled Pacific peoples and 35% of disabled Māori experienced food insecurity, compared with 23% of disabled European/Other. These figures reflect how disability intersects with historical and ongoing socio-economic inequities.^24,25 Further research should explore how disability, ethnicity and gender combine to shape food insecurity, particularly for Pacific and Māori women; our study was limited in examining these intersections because disability and gender were measured at the individual level while food insecurity was measured at the household level.

Public health implications

Our findings have direct implications for public health practice and policy. First, the strong association between disability and food insecurity, persisting even after adjustment for socio-economic factors, indicates that disability-specific barriers exist beyond income and employment disadvantage alone. Second, the substantial attenuation of effects when adjusting for employment and deprivation (up to 44% reduction in RR) suggests that interventions improving employment opportunities and income security for disabled people may reduce food insecurity. Third, the multiplicative risk for individuals with multiple disabilities supports a tiered approach to food security support that accounts for disability complexity. Finally, the pronounced disparities among Pacific (57%) and Māori (35%) disabled people highlight the need for approaches developed in partnership with these communities.

The study’s strengths include a large, nationally representative sample with survey weighting and standardised assessment tools enabling comparison with other New Zealand research. Additional limitations include the household-based nature of food security questions, which prevents determining whether disabled respondents are household heads or whether multiple members have disabilities. The WG-SS methodology also does not capture mental health conditions or eating/swallowing difficulties, both particularly relevant to food security.

Conclusion

Disability is significantly associated with food insecurity in New Zealand, with socio-economic factors—particularly employment—partially mediating this relationship. Multiple disabilities and greater severity amplify risk, while Pacific and Māori disabled people face the highest rates. These findings support policy approaches that address both economic participation and disability-specific barriers, recognise varying impacts across disability types and are developed in partnership with the communities most affected.

View Appendices.