Discrimination and avoidance due to disability in Australia: evidence from a National Cross Sectional Survey

Data

Data for this study are from the 2015 Survey of Disability, Ageing and Carers (SDAC) conducted between July and December 2015. Three populations were sampled using multi-stage sampling techniques. These consisted of people living in private dwellings, in self-care retirement villages and in care accommodation. Within these populations, the ABS sought to collect information from people with a disability, individuals aged 65 years and older, and carers of people with a disability or long-term health condition.

This study utilises information collected from the household component of the survey, which includes individuals living in private residences and self-care retirement villages. Of 31,957 private dwellings contacted, 25,555 fully responded, yielding a response rate of 80%. Of the 288 people in self-care retirement villages contacted, 251 responded yielding a response rate of 87.2%. The corresponding response rate for care-accommodation was 89.4%. The module on discrimination and avoidance was included in a personal interview for all people aged 15 years and over with a disability who live in households. This reduced the sample size for analysis to 9763.

Measures

For the first time in 2015, the ABS included a module on disability discrimination and disability avoidance. Firstly, respondents were asked “In the last 12 months do you feel that you have experienced discrimination or have been treated unfairly by others because of your condition/s?” For those who responded ‘yes’, a follow up question was asked: “Who treated you unfairly or discriminated against you because of your condition/s?” A list of multiple responses was provided, that consisted of employer, work colleagues, family or friends, teacher or lecturer, health staff (e.g., GP, nurse, hospital staff), bus drivers/rail staff/taxi drivers, restaraunt/hospitality staff, sales assistants, strangers in the street, or ‘others’.

Following these questions, respondents were asked “In the last 12 months have you avoided situations because of your condition(s)?” Again, those who responded ‘yes’ were asked the following question: “What situation(s) did you avoid because of your condition(s)?” A list of multiple responses was provided including work, visiting family or friends, school, university or educational facility, medical facilities (e.g., GP, dentist, hospital), shops, banks etc., restaurants, cafes or bars, public transport, public park or recreation venue, other social situations, other public places, other.

In addition to these new measures on discrimination, the SDAC instrument collected detailed measures of the recipient’s demographic characteristics and disabling conditions. Measured disabling conditions included sensory and speech, intellectual, physical, psychosocial, head injury and other. For the first time in 2015, the SDAC measured psychosocial disabilities. This concept is broader than a psychological disability and encompasses nervous or emotional conditions, memory problems or confusion, social or behavioral difficulties, a mental illness for which help or supervision is required and brain injuries including stroke, that lead to mental illness or other cognitive problems [14]. This concept was designed by the ABS to specifically measure conditions that “have an impact on a person’s ability to participate fully in daily living and opportunities like education, employment, and social and cultural activities” [14]. The ABS also included a measure of the number of long-term health conditions. Long-term health conditions defined by the ABS included diseases or disorders lasting or likely to last more than 6 months at the time of the survey.

Statistical model

To model the association between demographic factors, health characteristics and the experiences of discrimination and avoidance, we fit logistic regression models. Using the raw logit coefficients, we calculate odds ratios (OR) which measure the change in the odds of experiencing discrimination or avoidance given a change in a covariate, once all other factors in the model are controlled.

Variable selection for all models was informed by the range of literature on discrimination outlined above. Model fit between alternative specifications was assessed using the Bayesian Information Criteria following Raftery’s procedure [16]. With all models specified, we checked the conditioning of the matrix of independent variables to investigate any collinearity influence [17]. The condition numbers and variance inflation factors (Mean VIF < 2) were very small providing support for the model specification. Final goodness-of-fit for the discrimination and avoidance models was confirmed using the Hosmer and Lemeshow test [18].

Due to the complex survey design, additional adjustments were necessary to generate correct variance estimates. Unfortunately, the ABS does not provide information on the primary selection unit due to privacy concerns. However, the ABS provides 60 replicate weights, in addition to a person weight, to adjust for sample design and non-response. Utilising an algorithm developed by Winter (2008), we employed the delete-one jackknife method to provide correct standard errors for the estimated logit coefficients [19, 20]. This provided an alternative to the standard Taylor series linearization methods when only replicate weights were available.

Avoidance and discrimination

Before turning to the key research questions of this study, it is useful to consider potential exlpanations for the higher prevalence of avoidance relative to discrimination among Australians living with a disability. Conceptually, avoidance is a broader concept than discrimination as it may capture functional limitations in engaging in certain contexts or situations. For example, facing access difficulties to public places due to mobility issues. However, the Australian Government through the Australian Human Rights Commission (AHRC) considers such avoidance behaviours to be a measure of “indirect discrimination”. Specifically, they argue:

“Discrimination can be direct, meaning a person with a disability is treated less favourable than a person without that disability in the same or similar circumstances… Indirect discrimination can happen when conditions or requirements are put in place that appear to treat everyone the same, but actually disadvantage some people because of their disability. For example, it may be indirect discrimination if the only way to enter a shop is by a set of stairs, because people with disability who use wheelchairs would be unable to enter the building.” [21].

This form of indirect disability discrimination through avoidance behaviours has rarely been examined using nationally representative data. It highlights an important measurement issue in that measuring disability discrimination alone without measures of avoidance may provide considerably lower point estimates of exclusion more generally. Indeed, the extant literature provide important evidence of the link between experiences of discrimination and avoidance behaviours. For example, empirical studies show people who experience disability discrimination are less likely to utilise health services [22–24], less likely to access preventative health services, delay or fail to fill prescriptions and delay or avoid treatment [24, 25]. This avoidance behaviour can contribute to the development of undiagnosed comorbidities, further increasing the burden of poor physical and mental health upon the individual, community and economy [24].

Variations in discrimination and avoidance by demographic characteristics

Given the relatively high levels of avoidance and sizeable minority of people with a disability reporting discrimination, this paper sought to firstly examine whether particular demographic groups were at a heightened or reduced risk of exposure, once controls for health factors were included. Although descriptive statistics have been released elsewhere, this study examined the unique contribution of a number of demographic variables to discrimination, once all factors included in the multivariate analyses were controlled for. With extensive controls for disability and demographic factors and adjusting for complex survey design, we confirm that disability discrimination and avoidance is spread unevenly throughout the population. Notably, we find that disability discrimination and avoidance decreases with age. A speculative explanation is that as the prevalence of disability increases with age, it becomes more normative and therefore socially acceptable. In contrast, disabilities are less common amongst younger people and therefore more likely to be a point of differentiation and discrimination. This hypothesis of discriminatory attitudes stems from the social psychology Social Identity Theory of the ‘in’ versus ‘out’ group mentality, developed by Tajfel and Turner in 1979 [26]. Social Identity Theory leads to intergroup social comparison where individuals in one group compare themselves against their out-group [27]. In-group versus out-group comparison frequently creates tension, conflict or discrimination. Based on an individual’s perceived commonality with other group members, one might have several group-memberships [28]. As the likelihood of gaining one or more disabilities increases as people age, more older people will self-identify within this group, thus mitigating any prejudice or discrimination.

Notably, avoidance issues were also found to decline with increasing age, reflecting several possible reasons. Firstly, Australians who are 65 years and older or 50 years and older for Aboriginal and Torres Strait Islander people, have access to the Home and Community Care program (HACC). Within this program individuals are able to access services such as nursing and allied health services, household domestic assistance and personal care thereby offsetting some of the issues that may make access to healthcare difficult and costly for younger cohorts. In addition, lower levels of perceived discrimination are also likely to be a driver of lower levels of avoidance.

There are also methodological issues which may effect the age grading of discrimination and avoidance that we observe here. Specifically, the cross-sectional design of SDAC enforces an important limitation on correctly identifying age differences in the reporting of discrimination and avoidance, particularly in the later lifecourse. As is well established in the economics of ageing literature, selective mortality removes a disproportionate number of older people with low levels of resources and poorer health from the population [29, 30]. Moreover, this selectivity effect may also operate through migrating out of the survey population through entry to aged care [29], a point that is particularly pertinent given that the SDAC module on discrimination did not include people in non-private dwellings. Longitudinal data is required to measure complex movements out of private households and further data collections would be necessary to examine the generalisability of these findings presented herein to individuals living in care accommodation and other institutions (non-private dwellings).

Once controls for age are included in the multivariable regression analyses, the likelihood of exposure to avoidance and discrimination was higher for divorced people (versus married), the unemployed (versus employed) and was lower for people with lower levels of education (versus a degree) and those born overseas. Interestingly, the broader literature on discrimination shows that demographic characteristics are associated with exposure to discrimination. For example, findings of fewer reports of discrimination among lower educated people has been cited in racial discrimination research. Cunningham and Paradies (2013) provide several reasons, specifically for education: “more educated Indigenous people (1) may have higher expectations about how others should treat them, a difference in interpretation rather than exposure; (2) are more likely to work and socialise with non-Indigenous people and hence be exposed to inter-racial discrimination … and (3) are more likely to be the targets of discrimination because they defy stereotypes” [31]. These explanations are likely to be relevant to an association between education and disability discrimination that we observe herein. Moreover, there is a considerable body of literature on the role of disability based discrimination and employment, underscoring the higher prevalence of discrimination among the unemployed. Levels of discrimination cited by the unemployed are likely partially reflective of discriminatory practices in recruitment processes that are oftentimes faced by people with disabilities [32].

Variations in discrimination and avoidance by disability characteristics

With controls for demographic characteristics included, results from the multivariate regression analyses showed having a psychosocial or physical disability significantly increased the odds of experiencing discrimination or avoidance, as did an increasing number of long-term health conditions.

The findings that people with a psychosocial disability are more likely to report discrimination or avoidance is consistent with a large body of literature on mental health stigma more generally. For example, using Danish data, Dammeyer and Chapman (2018) find that those with a mental health disability were more likely to report both discrimination and violence [33]. More broadly, a recent systematic review of 144 studies has found that mental health related stigma is associated with avoidance of health seeking [34]. Indeed, it may be that individuals with psychosocial disabilities in particular, may be at greater risk of discrimination and avoidance due to mental health stigma. Stigma is based on negative attitudes and misunderstanding that leads to prejudice and discrimination. Stigma can result in a lack of support and empathy for individuals living with a mental illness, which can leave sufferers feeling embarrassed, misunderstood, and marginalised. Stigma may affect self-esteem, leading to avoiding seeking treatment, and the social withdrawal from the rest of the community through avoidance [35].

Our study further shows that respondents with a physical disability were more likely to cite both discrimination and avoidance, relative to those without this disability. We speculate that physical disability is more ‘visible’ relative to other health conditions, creating a point of difference to the norm and consequently a visible target for discrimination. Several studies have shown that individuals with high visible disabilities are treated less favourably than people with less visible or invisible disabilities. For example, Gouvier et al. 1991 showed that job applicants with visible disabilities received the most unfavourable ratings by potential employers. This was particularly the case in recruiting for jobs requiring high levels of public contact [36]. Similarly, Crocker and Major, 1994 noted individuals with invisible disabilities had less problematic or anxiety-provoking social interactions than people with more visible disabilities [37].

More generally, people with physical disabilities may avoid social, familial and economic situations due to physical and organisational obstacles such as inadequate access to transportation and a lack of communication assistance as well as discriminatory attitudes [38]. A recent international study shows that the association of discrimination with healthcare avoidance was particularly pronounced amongst people with physical disabilities, relative to those with other disability types [24]. This is important as studies show that avoidance behaviours, particularly when occurring in contexts important to human capital such as healthcare, is strongly associated with experiencing psychological distress [11]. Moreover, there is new evidence that the ‘effect’ of disability discrimination and avoidance on poor mental health outcomes is exaserbated for respondents living with physical disabilities including incomplete use of arms, fingers or legs and feet or those with experiences of blackouts/seizures [39].

Context of Disabilitiy discrimination and avoidance

Our final research objective was to understand the experience of discrimination and avoidance in different settings and contexts across an individual’s lifecycle. Importantly, we demonstrated that disability discrimination and avoidance occurs in contexts that are vital to human capital investments. For example, in healthcare settings, education settings and within the labour force. Consistent with a lifecycle approach, discrimination and avoidance in the workplace and education settings were prominent among younger and middle age Australians with a disability.

Indeed, both avoidance and discrimination in the workplace specifically due to physical barriers has been cited by the Australian Human Rights Commission, who note mobility and access issues combined with an unwillingness of employers to implement reasonable adjustments to the physical environment [21]. Labour force participation of people with a disability is 54% compared with 82% for people without a disability and this ratio has been stable over the past decade [40]. The Australian Human Rights Commission’s 2016 National Inquiry report cited several brief case studies indicating specific failings of reasonable workplace adjustments and the barriers that such failure makes to on-going employment for people with disabilities [32], pp. 187–191]. This is despite the fact that many benefits of employing people with disability were identified in an earlier report [41].

Further work by Darcy, Taylor and Green (2016) using natural data from the Australian Human Rights Commission determined that disability discrimination complaints comprised the largest proportion of cases compared with all other discrimination types in Australia [42]. Within the disability discrimination complaints data, employment discrimination made up the greatest proportion in these cases. They found that people of employment age with physical mobility disability reported higher rates of discrimination, particularly in relation to access barriers in the workplace. They speculate that employers incorrectly assume expensive hiring costs associated with potential employees who have a disability and may also be unaware of government subsidy programs to assist in making workplace adjustments.

We further identified experiences of discrimination and avoidance due to disability in education settings, particularly among younger Australians. The existing literature notes that students living with disability are often excluded and/or teachers refuse to or are reluctant to make reasonable adjustments [43]; amounting to institutional discrimination [44]. A Senate Inquiry in 2016 received evidence of systemic barriers experienced by students with disability and their families including “difficulties enrolling, failure of schools to provide the reasonable adjustments required by students, exclusion from school activities, a shortage of services in rural and remote areas of Australia and low expectations of students with disability from school staff and others, leading to a failure to take seriously the educational needs of students” [45]. This was compounded by the family’s financial means, geographical location and indigenous status.

Interestingly, whereas education and workplace discrimination and avoidance was clustered among younger and middle aged Australians, the same was not the case for discrimination in healthcare settings. The likelihood of discrimination in healthcare was relatively flat, or invariant, by age. Discrimination in healthcare settings is particulary important as recent research shows disability discrimination specifically, may be associated with deteriorating health as well as avoidance of healthcare services [24, 46]. Experiencing access barriers to healthcare, such as discrimination, can demotivate people to attend to current illnesses posing further issues for individuals’ re-engagement into effective healthcare treatment. Delaying treatment for deleterious health conditions may lead to an escalation in illness severity, contributing to a greater health burden [47, 48]. This may contribute to poorer quality of life, unnecessary hospitalisations and eventually placing greater strain on acute care systems [49–51].

Limitations

In interpreting the results from this study, it is important to note the limitations. Firstly, the data are cross-sectional and as noted earlier, there is a possibility of a selectivity effect in survival. As individuals with higher economic and social resources are more likely to exhibit higher survival prospects relative to their financially disadvantaged peers; in cross-sectional data we may be observing these individuals [52]. Moreover, the SDAC module on discrimination was only completed by repondents living in private households in the community. Further data collections would be necessary to generalise these findings to persons with a disability living in non-private dwellings such as those in long-term care residences. Longitudinal data would be required to disentangle these effects, unfortunately there is a dearth of information on the experiences of disability discrimination in Australia. In addition, the measures utilised by the ABS rely upon recall over a 12-month period, and thus may be subject to recall bias. The measures on discrimination are also self-reported and may be subject to further bias. For example, some individuals may feel uncomfortable disclosing or discussing such experiences, biasing the prevalence downwards.

Recommendations for further research

Noting these limitations, the levels of disability discrimination and avoidance that we observe pose questions about the implications for the individual. In the broader field of research in social inequalities in health, pathways between exposure to discrimination and poor health outcomes (both physical and mental), have been extensively investigated specifically in the field of racism research [53]. Racism has been shown to be associated with negative health impacts through several key pathways including by: (1) increasing stress; (2) decreasing health promoting behaviours (e.g. physical activity); (3) increasing health-damaging behaviours (e.g. alcohol and drug use); (4) reducing access to key health-promoting resources (e.g. employment, education, health and aged care services etc.); (5) increasing dysregulation (e.g. sleep disruption); and (6) increased health-damaging exposures (e.g. toxic substances) [53]. The results may also potentially relate to the increased levels of exposure to other types of discrimination, including for people with disabilities and the variability of these exposures over the life span and with demographic characteristics [54, 55]. Recent Australian studies have provided some evidence that experiencing discrimination is associated with poor mental health outcomes and lower levels of self-rated health [11, 12, 39]. However, an important area of future research is to validate the pathways between exposure to disability discrimination and deleterious health outcomes, as has occurred in racism research.

Related to that above, the second key research priority is the use of longitudinal data with validated measures of disability discrimination and avoidance. The cross-sectional results from SDAC, although informative, are limited to measuring a statistical association, rather than enabling a determination of more complex causal pathways between disability exclusion and health outcomes. In the Australian case, further data collections are necessary.

Although not the focus of this paper, the high prevalence of discrimination and avoidance reported by some groups of people living with disabilities and in specific contexts, suggests the time is ripe for policy overhaul and immediate attention. In particular, through the continued roll out of the National Disability Insurance Scheme (NDIS), it is important to consider the role of disability exclusion in acting as a barrier to accessing mainstream services for people living with a disability.

Many policy solutions have been offered in the literature. For example, positive media campaigns and general education by way of interventions are an obvious way to address physical and psychosocial disability stigma and discrimination. Beyondblue, an Australian NGO, suggests two effective approaches to reduce stigma: (i) educational approaches which involves the dissemination of factual information through media, social media, books, flyers, movies, websites etc. and (ii) contact approaches which involves direct interpersonal contact with people with a psychosocial disability [35]. However, consideration must be given to strategies that are delivered in a collaborative, sustainable and multi-sectoral way and be supported by reform and policies that influence national attitudes and behaviours [35].

Apart from these general recommendations, the heterogeneity in discrimination and avoidance that we observe in contexts critical to human capital (eg., healthcare, education, workplace) strongly underscores the need for sectoral solutions. For example, the 2016 Senate Inquiry received evidence of systemic barriers experienced by students with disability and their families [45]. They concluded with ten comprehensive recommendations which included the establishment of a national strategy around driving a cultural change, particularly at leadership level, to recognise all students with disability as learners; increase school participation and access rates for students with disability; ensure students with disability can access adjustments and interdisciplinary support to maximise learning; improve accountability; and establish an independent review and complaints mechanisms so all stakeholders can have full confidence in the system.

Strong support for addressing discrimination in employment have also become a national priority as a results of the Australian Human Rights Commission’s Willing to Work Inquiry [32]. Consequently, recommendations from this Inquiry included supporting increasing the workforce participation of people with disabilities by (i) forming a national priority around the issue, (ii) which is closely linked with policy planning for an ageing workforce and (iii) in conjunction with goals to reduce disability stigma and avoidance.

Although sectoral responses to disability discrimination and avoidance are now being considered by governments as indicated above, there is considerable opportunity for research to help direct implementation and evaluate proposed strategies.