Food insecurity among disabled adults

Abstract Background The relationship between disability and food insecurity is under-researched. Risk of food insecurity may vary by type and number of disabilities. We examine the hypotheses that (i) a higher number of disabilities increases risk of food insecurity and (ii) associations of physical disabilities, mental/cognitive disabilities or a combination of both types with food insecurity may differ in strength. Methods Data came from the fifth wave of the UK’s Food Standards Agency’s Food and You survey (2018), which contains detailed information on disability and household food insecurity. We used logistic and multinomial logistic regression to model the number and type of disabilities as predictors for food insecurity outcomes, controlling for socio-demographic factors. Results Both type and number of disabilities predicted food insecurity. Every additional disability was associated with higher odds of food insecurity [odds ratio (OR): 1.60, 95% confidence interval (CI): 1.40–1.83]. Among people with a disability, every additional disability was associated with 19% higher odds of food insecurity (OR: 1.19, 95% CI: 1.05–1.34). People with both physical and mental/cognitive disabilities had increased odds of severe food insecurity (OR: 8.97, 95% CI: 3.54–22.7). Conclusion Number and type of disabilities are associated with higher risk of food insecurity. A combination of physical and mental/cognitive disabilities, as well as having multiple disabilities are each independently associated with higher risk of food insecurity. Policy-makers may thus consider using targeted and tailored policies to reduce barriers to social and financial inclusion of disabled people to reduce food insecurity.


Introduction
2][3] According to the biopsychosocial model of disability, 4 disability is understood as an interaction between a person and social context.Thus, the relationship between disability and food insecurity may reflect the fact that disabled people are at higher risk of socio-economic disadvantage and exclusion 5,6 due to facing significant barriers to education, work, adequate income and financial security. 7,8,9Disabled people also endure higher costs of living and are more likely to experience ill-health. 10,11Lower socio-economic status and ill-health have both been shown to increase the risk of food insecurity. 12,13hile studies in high-income countries have found that food insecurity is strongly associated with mental, 2,14,15 physical and chronic illness, 16,17,18 research looking at the relationship between disability and food insecurity is limited and of mixed quality. 3This relationship is likely to be bidirectional, as food insecurity may increase the risk of physical disability, while at the same time, poor health among disabled people may lead to food insecurity. 19,20A limited number of studies in high-income countries suggest that the type and intensity of disability, as well as chronicity of food insecurity, may be important to understand the relationship between disability and food insecurity.Previous studies have focused primarily on the USA or have faced important limitations regarding the measurement and modelling of food insecurity.Studies in the USA and Canada have found that work-limiting disabilities are associated with food insecurity, as well as being disabled and of working age. 2,17,21Functional disabilities such as mobility limitations, barriers particularly faced by physically disabled people, as well as barriers as a result of cognitive impairments, have been associated with a higher risk of food insecurity. 22,23Studies also suggest that some groups of disabled people, such as people with lower or more insecure incomes, may be at higher risk of facing food insecurity. 12,24However, few studies have examined how different types of disability as well as the number of disabilities relate to the risk of food insecurity, particularly in the UK context.A better understanding of how different experiences of disability relate to food insecurity and to what extent food insecurity risk varies among disabled people is critical for developing targeted and tailored policies and programmes for reducing food insecurity among disabled people.
In this paper, we explore the hypothesis that categories and number of disabilities are independently associated with food insecurity.This hypothesis is motivated by literature suggesting that different types of disability pose different barriers and facilitators to inclusion and equality of access to adequate food.For example, mentally/cognitively disabled people may face particular knowledge, information and income and work barriers, such as difficulty building or maintaining social networks and facing discrimination and stigma. 257][28] Based on this literature, we focus on two hypotheses.Firstly, we hypothesize that having multiple disabilities puts people at higher risk of food insecurity due to more and/or a higher intensity of barriers to access and participation that increase social disadvantage, 29 e.g.public transport and supermarket access, exclusion from secure and sufficient income, higher likelihood of poorer health. 30Second, we hypothesize that mentally/cognitively disabled people have higher risk of food insecurity than people who are physically disabled due to lack of parity in terms of social support and support services which could lead to higher unmet need 31 and that people with a combination of categories will experience higher risk of food insecurity than people who have only physical or only mental disabilities.

Data source and sample
Data came from the UK's Food Standards Agency's Food & You survey (F&Y), a repeated cross-sectional, representative survey of adults aged 16 and over in England, Wales and Northern Ireland. 32The survey uses random probability sampling and face-to-face computer-assisted personal interviewing.At the time of analysis, it was the only nationally representative dataset in the UK containing an internationally agreed measure of household food insecurity: the United States Department of Agriculture (USDA)'s Adult Food Security Survey Module. 33Data from Wave 5 of F&Y, conducted in 2018, were used, as this wave collected detailed information about disability. 32Of the 6346 eligible addresses approached, the response rate was 48.2%, resulting in a sample size of 3059 adults. 32

Number and type of disability
In line with the definition of disability in the UK Equality Act, 34 respondents were asked if they had any physical or mental health conditions or illnesses lasting or expected to last for 12 months or more.Respondents who answered affirmatively were then asked whether any conditions or illnesses affected them in any of nine specific areas: mental health; social or behavioural problems; memory; learning, understanding or concentration problems; vision; hearing; mobility; dexterity; or stamina or breathing or fatigue.
Respondents could indicate if their condition affected them in other areas or 'none of the above'.Based on this information, we constructed a continuous variable that indicated the number of areas of disability, which ranged from 0 to 8. To provide visualization of the age-adjusted relationship between number of disability areas and food insecurity (figure 1), we also created a four-level categorical variable (0: no disabilities; 1-2, 3-4 and 5 or more).
We created a separate categorical variable that captured the broad type of disability individuals experienced.Subcategories were as follows: no disability; physical (vision, hearing, mobility, dexterity, stamina/breathing/fatigue) disability only; mental or cognitive (social/behavioural, memory, learning, understanding/ concentration) disability only; or both physical and cognitive/ mental disability.We combined cognitive and mental disabilities into a single category based on prior literature (i.e.disabling mental ill-health can be considered a cognitive limitation), and we expected barriers and impacts associated with these types of disability to share common mechanisms. 2,23We grouped a range of physical disabilities into one category, each of which may have different associations with food insecurity. 22Unfortunately, low sample sizes for each individual physical condition precluded a more refined analysis for specific disabilities.
People who indicated having physical or mental health conditions or illnesses but who did not provide information on domains affected by them (i.e.selected 'other') were excluded (n ¼ 124), as it was not possible to accurately establish the number of areas in which disability was experienced.

Food insecurity
Food insecurity was measured by the USDA's 10-item Adult Food Security module, a validated scale that aims to capture the prevalence of food insecurity in the general population.According to standard USDA practice, food insecurity is identified by three or more affirmative responses to questions on the module, and severe food insecurity is identified by six or more affirmative responses.At this level, respondents have indicated experiences of going without food.In addition to examining food insecurity and severe food insecurity outcomes, a measure of chronicity of food insecurity was derived from the first three module questions which ask respondents how often they worried about running out of food; how often food actually ran out; and how often they could not afford to eat balanced meals.Respondents who indicated 'Never true' for all three questions were coded as not experiencing food insecurity; respondents who indicated 'sometimes true' to at least one question, but did not indicate 'often' in any of the questions, were coded as 'sometimes experiencing food insecurity'.Finally, respondents who indicated 'often true' to at least one question were coded as 'often experiencing food insecurity'.Ten respondents were excluded as they provided no information on these questions.

Statistical analysis
Logistic regression was used to model the probability of 'any' and 'severe' food insecurity as a function of (i) the number of disabilities and (ii) the type of disability, controlling for socio-demographic variables.In a third model, we examine if, conditional on having at least one disability, the number of disabilities is associated with increased odds of food insecurity.This model captures the risk associated with each additional disability among those who have multiple disabilities.We use the same modelling strategy but a multinomial regression model to examine the chronicity of food insecurity as outcome variable.

Sensitivity analyses
Relationships between disability and food insecurity may be stronger among younger people.Disability becomes more prevalent in older ages, affecting a wider range of socio-demographic groups.In addition, older people may become eligible for pensions or other welfare programmes, which reduces their risk of food insecurity.These factors may mean that disability is less strongly associated with food insecurity at older ages.To examine this, we present results stratified by age, using two alternative cut-offs: age 55 (the age at which claims for disability benefits start to increase rapidly) and age 65 (a common age cut-off used to define older age in the ageing literature).
All analyses use survey weights provided in the F&Y data to account for sampling design and stratification.

Descriptive statistics
Table 1 summarizes key descriptive statistics and shows that 22% of respondents reported a disability.Thirteen percent of the sample had a physical (but not mental/cognitive) disability; 4% had a mentally/cognitive (but not physical) disability, while 5% had both a physical and mental/cognitive disability.Higher proportions of respondents who had a mental/cognitive disability and both physical and mental/cognitive disabilities reported any, severe and more chronic food insecurity than respondents who had physical or no disabilities.There were higher proportions of people in older age groups among physically disabled people, while mental/cognitive disabilities were more concentrated at ages 18-64.
Disability was associated with several forms of social and economic disadvantage.Disabled people were less likely to be in work and more likely to be retired (if physically disabled only), unemployed or not working for other reasons (if mentally/cognitive disabled only).For people who reported combined disabilities, larger proportions were either retired or not working.Disabled people were less likely to have achieved a degree qualification and were more likely to have an annual income below £25 999.
Table 2 shows the results of logistic regression models that adjusted for socio-economic and demographic controls.Model 1 shows that every additional disability was associated with higher odds of food insecurity [odds ratio (OR): 1.60, 95% CI: 1.40-1.83].Model 2 shows there was an increased risk of food insecurity for people who had a physical (OR: 2.58, 95% CI: 1.45-4.60),mental/cognitive (OR: 3.17, 95% CI: 1.85-7.47)or both physical and mental/cognitive disability (OR: 6.21, 95% CI: 3.22-12.0).Among disabled people (model 3), each additional disability conferred a 19% increased odds (OR: 1.19, 95% CI: 1.05-1.34) of food insecurity.In models that used severe food insecurity as the outcome (see Supplementary table SA1), number of disabilities and a combination of physical and mental/cognitive disability predicted severe food insecurity, but not physical or mental/cognitive disabilities on their own.

Chronicity of food insecurity
Results from multinomial regression analyses examining chronicity food insecurity as the outcome are presented in Supplementary table SA2.Both number and each type of disability were associated with less frequent food insecurity as well as chronic food insecurity.Among disabled people (model 3), however, an increasing number of disabilities was significantly associated with chronic food insecurity (OR: 1.27, 95% CI: 1.09-1.49)but not less frequent food insecurity.

Sensitivity analysis
Table 3 reports results from sensitivity analyses examining whether associations between disability and food insecurity differ between older and younger adults.We observed stronger associations between number of disabilities, physical disabilities and a combination of physical and mental/cognitive disabilities with food insecurity for younger age groups (defined as <55 and <65), with these relationships becoming non-significant for adults 65þ.However, among older adults (55þ or 65þ), the odds of food insecurity were particularly high for people with mental/cognitive disabilities, though confidence intervals were large (e.g.OR for mental/cognitive disability 3.36 [95% CI: 1.65-6.87]among <65; OR: 12.6 [1.22-130]  for 65þ).

Discussion
This study adds to the current literature by examining how number and type of disabilities are associated with food insecurity.Our results suggest that physical and mental/cognitive disabilities are differentially associated with food insecurity, with a combination of mental and physical disabilities conferring particularly high risk.We also observed that each additional disability conferred higher risk of food insecurity, even conditional on having any disability.These associations were generally stronger for working-age people.

Food insecurity among disabled adults 595
Our findings shed new light on the relationship between disability and food insecurity.We expected different types of disability to have different associations with food insecurity, reflecting the fact that underlying mechanisms might differ due to the heterogeneity of disability experience.Though both physical and mental disabilities were associated with food insecurity, a combination of both was more strongly associated with food insecurity, including severe food insecurity and chronic food insecurity.These results suggest different mechanisms may underlie associations between physical disability and mental-cognitive disability with food insecurity and that when combined, food access is particularly compromised.
Similarly, our findings that an increasing number of disabilities was associated with higher risk of food insecurity could reflect increasing barriers to resources important to achieve food security, 9 including economic stability. 35Research has shown that people on low incomes develop coping strategies to try to avoid food insecurity and shopping around for cheaper food and discounts is often a coping mechanism to secure an adequate diet. 36However, for some disabled people, this may be more difficult to implement or may not be an option at all. 3 Such 'coping mechanisms' may become more complex, more costly in terms of finance, impact on other areas of life, and less possible to pursue with an increasing number of disabilities.If increasing numbers of disabilities reflect increased barriers, disabled people may experience an intersection of both multiplied and new disadvantages. 29Additionally, barriers may become even harder to navigate if someone has a mental/cognitive disability as well as a physical disability, as suggested by the observed association of a combination of categories with severe food insecurity.
Multiple disabilities may also reflect an increased likelihood of experiencing chronic disadvantage, poverty and marginalization.This may also be a particular concern for people who have lifelong and work-limiting disability who may experience more discrimination and be less likely to build up long-term social or financial assets. 9,11We observed that relationships between disability and food insecurity were generally weaker at older ages.This may be  Greater than 65 adjusted for sex, ethnicity, highest level of qualification, household income and household composition.
Food insecurity among disabled adults 597 due to more effective support services designed for pensioners as well as more generous social security programmes.Disability also becomes more prevalent at older age so may be less closely tied to socio-economic disadvantage.We observed a strong relationship between mental/cognitive disability and food insecurity among older people, however.This group may face unique barriers to information and knowledge related to food access and may face more substantial barriers than physically disabled people to accessing support services. 14,379][40] This may be a particular risk at older ages when social isolation can be more of a concern.

Strengths and limitations
Our study uses an internationally standardized measure of food security assessed in a nationally representative sample and incorporated measures of the number and type of disabilities.However, several important limitations should be considered.Our study is based on a relatively small sample size, and our results are based on a cross-sectional analysis that only examined associations, rather than causal relationships.Food insecurity is correlated with other forms of social and economic disadvantage, which may confound the relationship with disability.Socio-demographic variables were limited in the dataset.In particular, age was only provided in age brackets.The crude measure of household income available in the dataset meant that it was not possible to equivalize income by household size.However, we note that including controls for the size of the household did not alter our results.In addition, the limited measures of financial hardship also meant that it was not possible to explore whether insufficiency of income explains the relationship between disability and food insecurity, especially as it does not account for the additional costs of living associated with disabilities.Though the biopsychosocial model of disability informed our conceptualization of how disability relates to food insecurity, variables reflecting the social contexts of disabled people were limited in the dataset.Future research would benefit from further exploring the role of social contexts in conceptualizing disability and food insecurity.Our measurement of food insecurity measures food insecurity as a result of economic affordability and may not account for non-financial barriers that also reduce disabled people's food access.However, sufficient financial resources can help to overcome other access barriers, e.g.transport, meal preparation, carers, help with shopping.Importantly, it is an internationally agreed, robust, standardized measure.The internal reliability of the food insecurity scale has been examined in other countries but not in this sample.
The measure of disability available in the dataset did not assess impacts of impairments on activities of daily living nor the severity of disability.Nor did our data allow us to test whether more specific disabilities beyond the broad categorizations of physical disabilities and mental/cognitive disabilities relate differently to food insecurity.Some research in the USA has found that functional and sensory disabilities may not relate to food insecurity in the same ways among older adults. 22There was also only a general range of cognitive and mental conditions captured; in particular, mental conditions did not distinguish between common mental disorders and severe psychiatric disorders.However, a strength of this measure is that it is in line with the standard ONS harmonization question for impairments.
Given that our findings point to a significant role for the number of disabilities, future research would benefit from understanding more about how disability severity and types relate to food insecurity outcomes.

Table 1
Descriptive statistics for non-disabled people and by disability category (n ¼ 3609)

Table 2
Odds of food insecurity for number, category and number if disabled

Table 3
Odds of food insecurity by number and category for adults <55 years of age and 55þ years of age and for adults <65 years of age and 65þ years of age Less than 65 adjusted for age, sex, ethnicity, highest level of qualification, work status, household income and household composition.