Summary
This web report builds on two earlier reports by the AIHW investigating the health of people from culturally and linguistically diverse (CALD) backgrounds:
- Reporting on the health of culturally and linguistically diverse populations in Australia: An exploratory paper, and
- Chronic health conditions among culturally and linguistically diverse Australians, 2021.
Analysis of the Australian Bureau of Statistics’ 2021 Census of Population and Housing (2021 Census) has been used to investigate the key social determinants of health that explain the differences in the reporting of long-term health conditions across CALD populations. Statistical modelling was used to understand how social determinants of health (education, labour force status, income, housing tenure, housing suitability, occupation, marital status, citizenship and remoteness) are associated with cultural and linguistic diversity (proficiency in spoken English, time since arrival in Australia, age at arrival in Australia, country of birth and languages used at home) and long-term health conditions reported in the 2021 Census.
Results
Statistical modelling indicates the strength of association between long-term health conditions and CALD variables is affected by age and the social determinants of health. For each CALD variable, the odds of reporting long-term health conditions generally approached those of a reference (or comparison) group, after adjusting for the effects of age and the social determinants of health.
However, this was not the case for all long-term health conditions as there were notable differences for asthma and mental health conditions. After adjusting for age and social determinants of health, the odds of reporting asthma or mental health conditions changed little compared with the reference group. The proportion of people reporting asthma or a mental health condition is similar from the age of 15 years, unlike the other health conditions investigated which become more common with age. This indicates factors beyond the scope of these analyses contribute to the development of these conditions.
Binomial logistic regression (statistical) modelling
Binomial logistic regression was used to estimate the probability of an event occurring (such as a long-term health condition). The statistical modelling shows association between variables but does not explain what is causing the association. For more information about the statistical modelling including use of reference groups, see the Technical notes and refer to the Methods section. For detailed findings, see Detailed results.
Proficiency in spoken English
People with low English proficiency (i.e. they indicated in the 2021 Census they spoke English ‘not well’ or ‘not at all’) are more likely to report most of the long-term health conditions examined than those with high English proficiency (i.e. speak English ‘well’ or ‘very well’). But when adjusting for the effects of age and the social determinants of health, there is stronger alignment between both groups meaning that age and the social determinants of health influence people reporting these long-term health conditions.
Time since arrival
Similar patterns were found using the time since people arrived in Australia to analyse long-term health conditions; age was associated with much of the difference but sometimes the social determinants of health played a role. For example, males who have been in Australia more than 20 years have odds ten times higher of reporting heart disease than males who arrived in Australia in the last ten years. When accounting for the effects of age, and then age combined with the social determinants of health (fully adjusted models), the odds drop to a similar level to those who were born in Australia.
Age at arrival
After adjusting for the effects of age and the social determinants of health, people who arrived in Australia when they were 65 years or older had similar odds of reporting a long-term health condition to people who arrived in Australia when they were younger. Figure 1 demonstrates how the statistical models did not greatly change the odds ratios for females who came to Australia as children (0 to 14 years) reporting a long-term health condition, and this young age group was closest to people born in Australia.
Figure 1: Odds ratios of females reporting a long-term health condition by Age at arrival, compared with those born in Australia, 2021
The figure shows the odds ratios of females reporting a long-term health conditions by age at arrival, compared with those born in Australia. In 2021, those aged 15 to 24 years when they arrived were the least likely to report a long-term health condition.
| Age at arrival | Unadjusted odds | Adjusted for age only | Full model |
|---|---|---|---|
| Born in Australia | 1.00 | 1.00 | 1.00 |
| 0 to 14 years | 0.70 | 0.70 | 0.75 |
| 15 to 24 yrs | 0.45 | 0.43 | 0.52 |
| 25 to 64 yrs | 0.56 | 0.45 | 0.54 |
| 65 years and over | 1.81 | 0.49 | 0.71 |
“Born in Australia” is the reference group for these models
Source:
Source: AIHW analysis of PLIDA, 2022
|
Data source overview
For more details about interpreting the modelling results see the Detailed results and the Technical notes.
Figure 1 also highlights that while females who were 65 years or older when they came to Australia were more likely to report any long-term health condition(s) compared with those born in Australia, they were less likely to report any long-term health condition(s) compared with Australian-born people when adjusted for the effects of age and social determinants of health.
Country of birth
Country of birth is one of the more commonly used CALD variables in health datasets, so it is useful to understand how strongly age and the social determinants of health are associated with the odds of reporting long-term health conditions.
While people born in some countries were more likely to report conditions like diabetes and heart disease than people born in Australia, these differences disappeared when they were adjusted for the effects of age and the social determinants of health. Similarly, people born in other countries who were less likely to report particular diseases compared with the Australian population, became more likely to report those conditions after adjusting for the effects of age and social determinants of health.
Figure 2 shows how adjusting for the effects of age and the social determinants of health impacts on the odds of males born in different countries reporting diabetes in different ways. While the odds drop for males born in Italy and England, for males born in India and the Philippines the odds of reporting diabetes increase to more than twice the odds of Australian-born males reporting diabetes.
Figure 2: Odds ratios of males reporting diabetes by Country of birth, compared with males born in Australia, 2021
The figure shows the odds ratio of males reporting diabetes by country of birth in 2021. The fully adjusted model shows those born in the Philippines and India were most likely to report diabetes.
| Country of birth | Unadjusted odds | Adjusted for age only | Full model |
|---|---|---|---|
| Australia | 1.00 | 1.00 | 1.00 |
| China | 0.77 | 0.88 | 0.93 |
| England | 1.40 | 0.87 | 0.95 |
| India | 1.06 | 1.91 | 2.24 |
| Italy | 3.77 | 1.60 | 1.66 |
| Malaysia | 1.06 | 1.07 | 1.32 |
| New Zealand | 0.94 | 0.92 | 1.01 |
| Philippines | 1.80 | 2.14 | 2.29 |
| South Africa | 0.80 | 0.82 | 0.98 |
| Vietnam | 1.39 | 1.38 | 1.19 |
“Born in Australia” is the reference group for these models
Source:
Source: AIHW analysis of PLIDA, 2022
|
Data source overview
For more details about interpreting the modelling results see the Detailed results and Technical notes.
Language used at home
After adjusting for the effects of age and the social determinants of health, the odds of people who spoke a language other than English at home reporting a long-term health condition more closely aligned with that of people who spoke ‘English only’. There were exceptions however, as shown in Figure 3 below. For example, after adjusting for the effects of age and the social determinants of health, males who spoke Hindi, Punjabi or Arabic at home were more likely to report diabetes compared with males who spoke ‘English only’.
Figure 3: Odds ratios of males reporting diabetes by language used at home, compared with those who speak English (only), 2021
The figure shows the odds ratios of males reporting diabetes by language spoken at home in 2021. The fully adjusted model shows those speaking Arabic, Punjabi or Hindi were the most likely to report diabetes.
| Language spoken at home | Unadjusted odds | Adjusted for age only | Full model |
|---|---|---|---|
| English (only) | 1.00 | 1.00 | 1.00 |
| Mandarin | 0.64 | 0.98 | 1.04 |
| Arabic | 1.55 | 2.2 | 1.87 |
| Vietnamese | 1.00 | 1.3 | 1.15 |
| Cantonese | 1.08 | 1.06 | 1.07 |
| Italian | 2.10 | 1.34 | 1.33 |
| Greek | 1.66 | 1.21 | 1.14 |
| Punjabi | 0.58 | 1.75 | 1.82 |
| Hindi | 1.30 | 2.78 | 3.21 |
| Spanish | 0.88 | 1.14 | 1.11 |
“English (only)” is the reference group for these models
Source:
Source: AIHW analysis of PLIDA, 2022
|
Data source overview
For more details about interpreting the modelling results see the Detailed results and the Technical notes.
Moving forward
People from CALD groups are varied in their migration experience, countries of birth, ethnic and cultural backgrounds, languages, traditions, religions and beliefs. But the analyses found many of the differences observed in the odds of reporting long-term health conditions are associated with age and the social determinants of health. However, the situation is far from uniform across the different CALD groups, indicating that other determinants of health (e.g. biology, environment, health service accessibility) are playing a role in whether people are reporting specific long-term health conditions.
The analyses also showed there are some groups within the CALD population who appear to have increased odds of reporting long-term health conditions that are currently being masked at the unadjusted level by their younger age and higher social determinants of health profile. This report provides valuable insights for additional research, policies, and future directions for these diverse populations.