Indigenous Australians (Aboriginal and Torres Strait Islander people), who represent approximately 2.5% of the total Australian population, are disadvantaged relative to other Australians across a broad range of social and health indicators and are at greater risk of morbidity and mortality from mental health-related conditions [1]. For example, in 2005-06, Indigenous Australians were 2-3 times more likely than other Australians to be admitted to hospital for intentional self-harm and 1.9 times more likely to be admitted for mental and behavioural disorders [1]. Indigenous Australians are also more likely to be exposed to a range of life stressors. In 2002, Indigenous people were about twice as likely as non-Indigenous Australians to report the death of a family member/close friend in the previous year and about 3.5 times more likely to report being affected by someone's alcohol/drug problems or abuse/violent crime [1]. Mental disorders accounted for an estimated 15.5% of the total burden of disease and injury (as measured by Disability Adjusted Life Years) among Indigenous Australians in 2003, second only to cardiovascular disease (17.5%) [1].

In keeping with Indigenous understandings of health [2], the emphasis in Australia in recent years has been on the broader concept of social and emotional wellbeing, rather than the more narrow--and arguably more negative--concept of mental ill-health. However, progress has been limited by the paucity of data, with most available information to date relating to the use of health services for mental illness. Social and emotional wellbeing is one of nine key result areas in the current National Strategic Framework for Aboriginal and Torres Strait Islander Health, and one of the nominated actions for governments is to improve the evidence base about social and emotional wellbeing in culturally appropriate and sensitive ways [3].

The collection of data on the social and emotional wellbeing of Indigenous Australians has long been a contested and sensitive area, with slow progress over many years. Following extensive consultation (including a stakeholder workshop in 2003), negotiation and field testing, an "interim" social and emotional wellbeing module was developed through a partnership involving the Australian Bureau of Statistics (ABS), the Australian Institute of Health and Welfare and the National Aboriginal Community Controlled Health Organisation [4]. This interim module was first used in the ABS's 2004-05 National Aboriginal and Torres Strait Islander Health Survey (NATSIHS). It included eight domains: 1) psychological distress; 2) impact of psychological distress; 3) life stressors; 4) discrimination; 5) anger; 6) removal from natural family; 7) cultural identification; and 8) positive wellbeing. Although it was recognised that other domains were of interest, it was agreed by the key stakeholders that no satisfactory measures were available at that time [4]. For most of the measures included in the interim module, there were no comparable data for the non-Indigenous population. One exception was psychological distress, for which data for the non-Indigenous population were available from the ABS's 2004-05 National Health Survey (NHS).

Some basic data on psychological distress from the NATSIHS have been presented previously [1, 4, 5], but more detailed analysis is required to adequately understand the socio-demographic patterns of psychological distress within the Indigenous population, as well as to assess the utility and robustness of the measure. In other populations, low socioeconomic status (SES) has been associated with psychological distress [6–8], but it is not known whether, or to what extent, this holds true in the Indigenous population. Traditional measures of SES are not necessarily equivalent across social groups, and they may not adequately measure all relevant aspects of what they purport to measure [9, 10]. Although exposure to interpersonal racism occurs across the SES spectrum, increased self-reported interpersonal racism has been associated with high SES, at least in some settings [11–13]. Such exposure could potentially reduce the SES gradient in psychological distress. In addition, there are recognised differences in the cultural identity and SES of Indigenous Australians living in remote and non-remote areas [14], and these may impact on the size and nature of the gradient for people living in different areas.

The aims of the current study are: 1) to present more detailed data on psychological distress among Indigenous Australians; 2) to examine the relationships between socio-demographic measures and an indicator of very high psychological distress (VHPD) among a nationally representative sample of Indigenous Australian adults; and 3) to compare these relationships with corresponding patterns in the non-Indigenous population.

Data for Indigenous and non-Indigenous adults aged 18-64 years were taken from the NATSIHS and the NHS, two national surveys conducted in parallel by the ABS in 2004-05. These two surveys had very similar content and in most cases the wording of questions on particular topics was identical [15]. This analysis is limited to responses to questions deemed by the ABS to be comparable across the two surveys [16].

Extensive details on survey methodology have been published elsewhere [5, 15–19]. Briefly, both surveys were conducted using multi-stage sampling strategies; the first stage involved random selection of either communities or census collection districts (CD), and subsequent stages involved selection of dwellings and individuals within households [17, 19]. Both the NHS and NATSIHS samples were designed to provide reliable estimates for Australia as a whole as well as for selected sub-national areas, such as State/Territory, capital city versus balance of state within each state (NHS) and remote versus non-remote areas (NATSIHS). Indigenous respondents from the NHS were included with NATSIHS data to provide Indigenous population estimates [17]. Both surveys were limited to usual residents of private dwellings and conducted by trained ABS interviewers. Very remote areas were out of scope in the NHS but in scope for the NATSIHS. In the NHS and in non-remote areas in the NATSIHS, data were collected using Computer Assisted Interviews. In remote areas of the NATSIHS, pen and paper interview forms were used and some questions were simplified or deleted. After accounting for sample loss (e.g. dwellings out of scope or vacant, households with no adults, etc.), 89% of selected households in the NHS were classified as fully/adequately responding [19]. Corresponding figures in the NATSIHS were 85.5% in remote Indigenous communities and 83.4% in other areas [17]. More details about the design, conduct and results of the surveys are available elsewhere [5, 15–19].

To allow data access to interested researchers, the ABS created a Confidentialised Unit Record File (CURF) for the NATSIHS. This file includes unit records for Indigenous respondents of the 2004-05 NATSIHS and the 2004-05 NHS, as well as unit records for non-Indigenous respondents from the 2004-05 NHS [16]. Although the CURF contains unit records for participants of all ages, this analysis is limited to data from the 20,849 adult respondents (5,417 Indigenous and 15,432 non-Indigenous) aged 18-64 years. Children less than 18 years were not asked questions about psychological distress. The exclusion of those aged ≥ 65 years was due to uncertainty about the applicability of socioeconomic indicators among older people, as well as the relatively small size of this group in the Indigenous population [1].

The socio-demographic profile of the Indigenous population was significantly different from that of the non-Indigenous population, with a younger age distribution, lower educational attainment, and greater levels of disadvantage across a range of indicators (Table 1). There were also marked differences within the Indigenous population according to remoteness of residence.

Association of VHPD with SES indicators

VHPD was more commonly reported for Indigenous than non-Indigenous people in every age group (Figures 1 and 2). After adjusting for age, most of the independent variables examined were significantly associated with VHPD among non-Indigenous males and females and among Indigenous males and females in non-remote areas, although the relative odds were generally larger for non-Indigenous males and females than for their Indigenous counterparts (Tables 3 and 4).

	Remote Indigenous males	Non-remote Indigenous males	All Indigenous males	Non-Indigenous males
	OR (95% CI)†	OR (95% CI)†	OR (95% CI)†	OR (95% CI)†
Marital status
Married	0.4 (0.2-0.9)	1.1 (0.6-1.8)	0.8 (0.5-1.2)	0.4 (0.3-0.6)
Not married	1.0	1.0	1.0	1.0
Main language
English	1.0	---	1.0	1.0
Not English	0.4 (0.2-0.9)	---	0.4 (0.2-0.8)	1.0 (0.6-1.7)
Highest year of school completed
Year 10 or more	1.0	1.0	1.0	1.0
Less than Year 10‡	0.7 (0.3-1.5)	2.7 (1.7-4.3)	1.9 (1.3-2.8)	3.2 (2.3-4.6)
Highest non-school qualification
Bachelor/post-graduate degree	---	0.1 (0.0-0.5)	0.1 (0.0-0.6)	0.2 (0.1-0.4)
Diploma	---	0.1 (0.0-0.7)	0.2 (0.0-0.8)	0.5 (0.3-1.0)
Certificate	1.0 (0.4-2.8)	0.6 (0.3-1.1)	0.7 (0.4-1.2)	0.6 (0.4-0.8)
No qualifications	1.0	1.0	1.0	1.0
Employment status
Employed	1.0	1.0	1.0	1.0
Unemployed	2.6 (0.8-8.6)	5.0 (2.4-10.4)	4.3 (2.3-7.9)	7.2 (4.1-12.7)
Not in the labour force	1.7 (0.7-3.8)	5.0 (2.9-8.8)	3.8 (2.4-6.0)	7.9 (5.5-11.3)
Housing tenure
Owner/purchaser	0.8 (0.1-4.5)	0.4 (0.2-0.7)	0.5 (0.3-0.8)	---
Renter/other tenure	1.0	1.0	1.0	---
Ran out of food and couldn't afford to buy more (last 12 months)
Yes	2.3 (1.1-4.7)	4.1 (2.5-6.7)	3.1 (2.0-4.6)	6.2 (4.2-9.1)
No	1.0	1.0	1.0	1.0
Equivalised household income quintile§
1 (lowest)	1.5 (0.2-11.5)	3.9 (0.7-22.0)	3.0 (0.9-10.5)	18.9 (10.4-34.2)
2	0.7 (0.1-5.2)	3.0 (0.4-20.4)	2.0 (0.5-8.3)	8.6 (4.5-16.4)
3	0.6 (0.0-10.8)	1.0 (0.1-7.2)	0.9 (0.2-4.0)	3.8 (2.1-7.2)
4	---	1.0 (0.1-8.0)	0.8 (0.1-4.0)	2.1 (1.0-4.3)
5 (highest)	1.0	1.0	1.0	1.0
Not known/not stated	1.2 (0.2-9.6)	2.2 (0.3-14.4)	1.8 (0.5-7.1)	5.2 (2.9-9.3)
	Trend‖: p = 0.16	Trend‖: p < 0.001	Trend‖: p < 0.001	Trend‖: p < 0.001
SEIFA quintile**
1 (most disadvantaged)	---	1.3 (0.3-5.2)	1.0 (0.3-3.9)	7.1 (4.0-12.5)
2	---	0.7 (0.1-3.1)	0.7 (0.2-3.2)	4.6 (2.4-8.7)
3	---	0.4 (0.1-1.9)	0.4 (0.1-1.8)	4.3 (2.3-7.8)
4	---	0.8 (0.2-2.7)	0.7 (0.2-2.4)	4.4 (2.4-8.0)
5 (least disadvantaged)	---	1.0	1.0	1.0
	Trend: n/a	Trend: p = 0.16	Trend: p = 0.24	Trend: p < 0.001

* Source: Weighted data from the National Aboriginal and Torres Strait Islander Health Survey 2004-05 confidentialised unit record file (CURF) [16]. Very high psychological distress is defined as a K5 score of ≥ 15

† OR, odds ratio; CI, confidence interval. All estimates are adjusted for age group (18-24, 25-34, 35-44, 45-54 and 55-64 years). Figures in bold are statistically significant at p < 0.05

‡ Includes those who never went to school

§ Gross weekly equivalised cash income of household, using the OECD scale [17]. Quintiles are based on national figures

‖ Trend analysis excludes those with equivalised household income quintile 'not known/not stated'

** SEIFA, Socioeconomic Index for Areas, Index of Relative Disadvantage. Quintiles are based on national figures

	Remote Indigenous females	Non-remote Indigenous females	All Indigenous females	Non-Indigenous females
	OR (95% CI)†	OR (95% CI)†	OR (95% CI)†	OR (95% CI)†
Marital status
Married	0.4 (0.3-0.7)	0.5 (0.3-0.7)	0.5 (0.3-0.6)	0.5 (0.4-0.6)
Not married	1.0	1.0	1.0	1.0
Main language
English	1.0	1.0	1.0	1.0
Not English	0.4 (0.3-0.7)	2.4 (0.7-8.9)	0.6 (0.4-0.9)	1.6 (1.1-2.3)
Highest year of school completed
Year 10 or more	1.0	1.0	1.0	1.0
Less than Year 10‡	1.1 (0.7-1.7)	2.2 (1.5-3.2)	1.8 (1.3-2.4)	2.8 (2.1-3.6)
Highest non-school qualification
Bachelor/post-grad degree	2.2 (0.8-5.7)	0.5 (0.2-1.4)	0.7 (0.3-1.4)	0.4 (0.3-0.6)
Diploma	1.2 (0.3-4.8)	0.8 (0.3-2.1)	0.9 (0.4-2.0)	0.4 (0.2-0.7)
Certificate	1.1 (0.7-1.8)	1.1 (0.7-1.5)	1.1 (0.8-1.5)	1.0 (0.8-1.3)
No qualifications	1.0	1.0	1.0	1.0
Employment status
Employed	1.0	1.0	1.0	1.0
Unemployed	2.0 (1.0-4.0)	2.3 (1.3-4.0)	2.2 (1.4-3.5)	4.6 (2.8-7.5)
Not in the labour force	1.3 (0.9-2.0)	2.1 (1.4-3.1)	1.9 (1.4-2.6)	2.7 (2.3-3.3)
Housing tenure
Owner/purchaser	1.1 (0.5-2.7)	0.4 (0.3-0.7)	0.5 (0.4-0.8)	---
Renter/other tenure	1.0	1.0	1.0	---
Ran out of food and couldn't afford to buy more (last 12 months)
Yes	1.8 (1.2-2.7)	3.5 (2.4-4.9)	2.8 (2.1-3.6)	6.3 (4.8-8.2)
No	1.0	1.0	1.0	1.0
Equivalised household income quintile§
1 (lowest)	1.8 (0.5-6.7)	2.1 (0.7-6.0)	1.9 (0.8-4.9)	7.0 (4.5-10.8)
2	1.5 (0.3-7.1)	1.1 (0.4-3.4)	1.2 (0.5-3.1)	3.4 (2.1-5.7)
3	1.0 (0.2-5.6)	1.0 (0.3-3.1)	1.0 (0.4-2.6)	2.8 (1.8-4.3)
4	1.0 (0.2-5.7)	0.7 (0.2-2.6)	0.7 (0.2-2.3)	2.0 (1.2-3.3)
5 (highest)	1.0	1.0	1.0	1.0
Not known/not stated	1.1 (0.3-4.9)	1.4 (0.5-4.3)	1.3 (0.5-3.3)	2.7 (1.6-4.4)
	Trend‖: p = 0.07	Trend‖: p = 0.001	Trend‖: p < 0.001	Trend‖: p < 0.001
SEIFA quintile**
1 (most disadvantaged)	---	2.3 (0.7-7.8)	2.1 (0.6-7.0)	2.9 (1.9-4.2)
2	---	1.8 (0.5-6.2)	1.9 (0.6-6.8)	1.6 (1.1-2.4)
3	---	1.9 (0.5-6.9)	1.9 (0.6-6.8)	1.4 (0.9-2.1)
4	---	1.1 (0.3-4.3)	1.1 (0.3-4.3)	1.2 (0.8-1.8)
5 (least disadvantaged)	---	1.0	1.0	1.0
	Trend: n/a	Trend: p = 0.02	Trend: p = 0.04	Trend: p < 0.001

* Source: Weighted data from the National Aboriginal and Torres Strait Islander Health Survey 2004-05 confidentialised unit record file (CURF) [16]. Very high psychological distress is defined as a K5 score of ≥ 15

† OR, odds ratio; CI, confidence interval. All estimates are adjusted for age group (18-24, 25-34, 35-44, 45-54 and 55-64 years). Figures in bold are statistically significant at p < 0.05

‡ Includes those who never went to school.

§ Gross weekly equivalised cash income of household, using the OECD scale [17]. Quintiles are based on national figures

‖ Trend analysis excludes those with equivalised household income quintile 'not known/not stated'

** SEIFA, Socioeconomic Index for Areas, Index of Relative Disadvantage. Quintiles are based on national figures

For example, the relative odds of VHPD among those who did not complete Year 10 (compared with those who completed Year 10 or more) was 2.7 (95% CI 1.7-4.3) for Indigenous males in non-remote areas and 3.2 (95% CI 2.3-4.6) for non-Indigenous males; for females, the corresponding figures were 2.2 (95% CI 1.5-3.2) and 2.8 (95% CI 2.1-3.6). There was a significant inverse trend in VHPD with equivalised household income quintile for non-remote Indigenous males and females and non-Indigenous males and females (p ≤ 0.001 in each of the four groups), but the gradient was steeper for non-Indigenous males (OR for a one-unit change in income quintile = 0.47) and females (OR = 0.64) than for their non-remote Indigenous counterparts (0.62 and 0.73, respectively). People who were unemployed were significantly more likely to report VHPD than those who were employed, with relative odds of 5.0 (95% CI 2.4-10.4) for non-remote Indigenous males, 7.2 (95% CI 4.1-12.7) for non-Indigenous males, 2.3 (95% CI 1.3-4.0) for non-remote Indigenous females and 4.6 (2.8-7.5) for non-Indigenous females.

Among Indigenous males and females in remote areas, by contrast, only marital status, main language, and running out of food (and unemployment among females) were significantly associated with VHPD (Tables 3 and 4). Although running out of food was significantly associated with VHPD in all groups examined, the relative odds were smallest among Indigenous people in remote areas (males: OR = 2.3, 95% CI 1.1-4.7; females: OR = 1.8, 95% CI 1.2-2.7) and largest among non-Indigenous people (males: OR = 6.2, 95% CI 4.2-9.1; females: OR = 6.3, 95% CI 4.8-8.2), with Indigenous people in non-remote areas showing intermediate values (males: OR = 4.1, 95% CI 2.5-6.7; females: OR = 3.5, 95% CI 2.4-4.9).

Very high psychological distress was more commonly reported by Indigenous than non-Indigenous Australians in this nationally representative study. The prevalence was associated with most traditional indicators of SES--education, employment, income, home ownership and area-level disadvantage--in both the Indigenous and non-Indigenous populations in non-remote areas, although the relative odds were generally larger in the non-Indigenous population. By contrast, these traditional SES indicators were not significantly associated with VHPD in Indigenous people in remote areas. Other factors, including marital status, main language and food insecurity, were generally significant in all groups, although the relationship between main language and VHPD was not always in the same direction.

The lower prevalence of VHPD among Indigenous people in remote areas, and the lack of association with many socio-demographic variables in this group, suggest that living in an Indigenous-majority environment (such as exists in most remote communities in Australia) may mitigate the risk of psychological distress to some degree. There is evidence from the UK that increased ethnic density is beneficial for minority communities and that this is partially mediated by reduced exposure to racism as well as attenuated impact of such exposure [28, 29].

The higher prevalence of VHPD in Indigenous Australians is consistent with other data indicating that Indigenous Australians have a much higher burden of hospitalisation for intentional self-harm as well as mental and behavioural disorders [1]. For Indigenous Australians, 16% of general practitioner visits in 2005-2010 related to mental health problems, compared with 11% for all Australians [30].

Life stressors such as the death of a family member/close friend and alcohol/drug problems or abuse/violent crime among friends or relatives are more common for Indigenous people compared with non-Indigenous people [1] as are higher rates of suicide [31], disability and chronic disease [1]. Indigenous people are also exposed to high levels of racism, trauma and grief [13, 32–34].

In addition to its morbid effects, psychological distress is also related to mortality. Using data from the 1997 to 2000 US National Health Interview Survey (NHIS) linked to the US National Death Index through 2002, Pratt found a significantly higher risk of death (during a mean follow-up time of nearly four years) among those with high K6 scores, even after adjusting for a range of variables relating to socio-demographics, health behaviours and physical illness. There was a dose-response relationship between K6 score and risk of mortality [35].

There is evidence that psychological distress is associated with chronic health conditions. For example, in the 2002 NHIS, the prevalence of psychological distress (K6 score ≥ 13) in adults aged 40+ years was higher among those with self-reported congestive heart failure (CHF) (10.0%), myocardial infarction (MI) (6.4%) and coronary heart disease (4.1%) compared with those with no cardiovascular disease (2.8%), with elevated odds of psychological distress among those with MI and CHF remaining after adjustment for a range of socio-demographic and health risk factors [36]. In Australia, in cross-sectional data from almost 78,000 adults aged 18-70 years as part of the Australian Work Outcomes Research Cost-benefit Study, psychological distress (K6 score ≥ 13) was significantly associated with age, sex, marital status, education and income. After adjusting for these factors, psychological distress was associated with a range of health conditions and health risk factors, and the relative odds increased with multiple morbidity [37].

These findings may, in part, explain the increased prevalence of VHPD among Indigenous Australians, who have a greater burden of physical illness and an adverse risk profile. However, in the present study, only about a third of Indigenous participants with VHPD indicated that physical problems were the cause of their feelings most or all of the time. More work is needed to understand Indigenous people's perspectives regarding the causes of VHPD and the role played by physical illness.

For Indigenous males and females as well as for the Indigenous population as a whole, mean scores on the K5 were lower in remote than non-remote areas, and differences between males and females were more pronounced in remote compared to non-remote areas. Other studies have also found regional differences, but the direction of the difference has not always been consistent. In a study of non-metropolitan areas of New South Wales, Kelly et al. found lower mean K10 scores in remote areas than in very remote or inner or outer regional areas; mean scores were similar for males and females [38]. Although non-significant, a similar trend was evident in South Australia with lower mean K10 scores in accessible/moderately accessible areas compared to highly accessible and remote/very remote areas [39]. In the US, Dhingra and colleagues found that, among those in the 2007 Behavioral Risk Factor Surveillance System (BRFSS) (with a landline telephone), urban residents were more likely than rural residents to have mild (K6 7-12 out of 24) or serious (K6 13+) psychological distress after adjustment for socio-demographic characteristics [40].

The higher prevalence of VHPD among Indigenous Australians and the relationship with SES indicators (at least in non-remote areas) are consistent with data on ethnic minority groups in other countries. In New Zealand, K10 scores were significantly associated with sex, age group, education, equivalised household income and area-level deprivation. Maori and Pacific people had higher mean K10 scores than other ethnic groups, even after adjusting for age, sex, educational qualifications and equivalised household income [8]. In the 2002 Canadian Community Health Survey, high K10 scores (> 9 out of 40) were associated with being female, having low education, low income, younger age, and being unmarried; Aboriginal Canadians were more likely to have high psychological distress, but this was only apparent among those of low income [7]. In the 2001-04 NHIS in the US, the prevalence of serious psychological distress in the last 30 days (as indicated by a K6 score ≥ 13 corresponding to K5 ≥ 16) was reported as 3.1% overall. Higher prevalence of serious psychological distress was seen in females, those not married, living in poverty, and who did not complete high school. The relationship between living in poverty and serious psychological distress was observed among Hispanics, Non-Hispanic Whites and Non-Hispanic Blacks alike [6]. In the 2004-08 NHIS, American Indian or Alaska Native (AIAN) adults had a similar prevalence of serious psychological distress (K6 ≥ 13) to that seen in Black and Hispanic adults (3.4% versus 3.4% and 3.5%, respectively), but higher than that in White (2.9%) or Asian (1.4%) adults. However, there were marked differences by sex, with AIAN males having the highest prevalence and AIAN females having the second lowest prevalence among the five ethnic groups examined [41].

These international studies suggest a much lower prevalence of very high psychological distress than that found among Indigenous Australians. However, similar figures were found in a 2003-4 American study, in which 15.4% of 1,202 low-income multi-ethnic workers had a K6 score ≥ 13. This study also found that after adjusting for poverty, psychological distress was significantly associated with workplace abuse and high exposure to racial discrimination [42].

Significant associations among Indigenous Australians have previously been reported for SES and diabetes [43, 44], renal disease [45, 46] and cardiovascular disease [47] (but not for asthma [48]). The present study indicates that there are significant associations between SES indicators and psychological distress, but these relationships appear to be attenuated, especially in remote areas. This finding is somewhat consistent with data from a study of 963 Indigenous people from a socio-economically disadvantaged coastal region in Australia, in which socio-demographic characteristics were largely non-significant in explaining psychological distress as measured by the K10 [49]. This suggests that risk factors such as racism, with common and/or cross-cutting effects across the socioeconomic spectrum for Indigenous Australians [13, 32, 33, 50], may contribute to psychological distress. In the 2004-05 NATSIHS, those who reported they had been treated badly because they were Aboriginal or Torres Strait Islander were more likely to have high or very high psychological distress (K5 ≥ 12) than those who said they had not been treated badly (39.5% versus 25.0%) [4].

Over half of Indigenous respondents with VHPD in the present study indicated they were not always able to carry out their normal roles/work due to their distress, and about a third had consulted a health professional about their feelings in the last 4 weeks. Such findings accord with international studies demonstrating that psychological distress leads to absence from work both in the short and long term [51, 52]. These data serve to further highlight the importance of psychological distress as a health risk factor for Indigenous Australians.

The main strengths of the current study are the use of nationally representative data, comparisons between Indigenous and non-Indigenous populations, and identical socio-demographic measures with comparable scales in the two populations. Although bias is always possible in any survey with less than complete participation, the high response rates in both the NHS and the NATSIHS suggest that any such bias is unlikely to be large. The main limitations relate to the cross-sectional nature of the study and the potential misclassification of VHPD and socio-demographic factors.

Kessler and colleagues have noted that the K6 has minimal bias with respect to age, sex and education [53]. Although bias relating to other factors such as culture or language is possible, the instrument has been used successfully in a wide range of settings. Most notably, the K6 has been validated and used in a diverse group of 14 countries taking part in the World Mental Health survey initiative (Brazil, Bulgaria, Colombia, India, Japan, Lebanon, Mexico, New Zealand, Nigeria, China, Romania, South Africa, Ukraine, and the United States) [53], with additional work in the United States [20, 21], Japan [54] and Australia [22]. The K6 has also recently been used in two US Native American populations living on or near reservations and found to be an appropriate screening tool for psychological disorders as well as a good indicator of severity [55].

Although the measure used to assess psychological distress in the NATSIHS was thus based on a widely validated instrument, it has not been widely used before in this population, and one of the questions comprising the K6 was omitted. Previously published data from the 2004-5 NATSIHS suggest that the measure has some validity, based on its association with a range of factors such as positive wellbeing, anger, number of life stressors, mental illness stressor, racial discrimination, and removal from natural family [4].

The data used in this analysis are derived from two different surveys, the NATSIHS and the NHS. Although the ABS planned these two surveys to be run in parallel, with the methodology and question wording matched as closely as possible to allow the data to be compared, some important differences may remain. In particular, NHS participants were administered the full K10, while NATSIHS participants were administered only the K5 (with slightly altered wording), and this may have affected participants' responses. Previous work in the US National Survey on Drug Use and Health has shown variations in K6 scores depending on question order and context [56]. Although it is possible that differences between the NATSIHS and NHS could have affected the results, the extent and direction of any such bias is unclear. We were not able to undertake sensitivity analysis (for example by comparing K5 responses for Indigenous NHS respondents who completed the full K10 and NATSIS respondents who completed only the K5) because the necessary data were not included in the dataset.

Aside from its use in the ABS survey program, we are aware of only one published study using the K5 [57]. In this study of 298 Aboriginal adults aged 15-54 years living in a remote area of Northern Australia, the mean (s.d.) K5 score was 6.58 (2.12). This is substantially lower than the mean K5 score for remote Indigenous males in the NATSIHS (8.27 (0.45)). A study of 184 Indigenous Australians between 2007 to 2009 used a version of the K6 with two additional items (focused on happiness and anger) to validate the Growth and Empowerment Measure (GEM). Findings suggest that, in comparison to the K6 alone, the 'K6 + 2' had slightly higher internal consistency and stronger correlations with the GEM (sub) scales. As such, it may constitute a more valid measure of psychological distress for Indigenous Australians [58].

Psychological distress is only one aspect of mental health and, in turn, mental health only one aspect of social and emotional wellbeing for Indigenous Australians. Previously published data from the 2004 to 2005 NATSIHS indicate that psychological distress does not necessarily correspond to a lack of wellbeing and vice versa. As such, it is important to assess both psychological distress and wellbeing in surveys relating to Indigenous Australians [4].

Information used to determine SES may have been incorrectly reported by (or on behalf of) some participants, and only limited detail was available on the SES indicators examined here. Data on housing tenure was not available in the NATSIHS CURF for the non-Indigenous population. Despite the use of comparable scales, the equivalence of a given level of SES may not be guaranteed across individuals or population groups. For example, the meaning of a certain level of education may vary over time and place, and years of education do not necessarily reflect the quality of education received, nor its social or economic value [59, 60].

Similarly, the use of SEIFA quintiles based on the whole population may not adequately capture the socioeconomic position of population subgroups such as Indigenous Australians [61]. No information was available about other potentially important SES measures, such as total household assets or childhood SES. An area-based measure of disadvantage was included, but no other information was available about neighbourhood/area characteristics. Although equivalised household income is intended to adjust for household size and economies of scale, the relatively high mobility of Indigenous people, including movement of individuals across households [62], can make it difficult to assess both Indigenous household income and household size, both of which are required to calculate equivalised income.

Because information on socio-demographic factors and psychological distress were collected at the same time, the temporal relationships between socio-demographic variables and psychological distress are not certain. For example, employment status may change as a result of having psychological distress. This may explain in part the observed relationship between psychological distress and being unemployed or out of the labour force. Similarly, data on physical health problems were collected at the same time as data on psychological distress and were based on self-report. These factors could have influenced the estimate of the proportion of psychological distress explained by physical health problems, although the direction of any such bias is uncertain. Moreover, it is not possible in a national survey to collect information on the full range of the factors that may relate to both psychological distress and SES. Hence, there may be unmeasured confounders which could account for all or part of the findings presented here.

Despite these limitations, the NATSIHS data provide the best available information on psychological distress in Indigenous Australian adults that can be compared directly to the non-Indigenous population.

This paper demonstrates for the first time in a national survey, the higher absolute levels of VHPD combined with smaller socio-demographic gradients in the Australian Indigenous population. This suggests the relevance of risk factors with cross-cutting effects across the socioeconomic spectrum, such as racism. The lower prevalence of VHPD among Indigenous people in remote areas compared to those in non-remote areas, and the lack of association with many socio-demographic variables in this group, suggest either that the instrument may be less valid for remote-living Indigenous people or that living in an Indigenous-majority environment (such as exists in most remote communities in Australia) may mitigate the risk of psychological distress to some degree.

These findings highlight a clear need for services to meet the needs of Indigenous people with VHPD as well as interventions to address factors which may act as determinants of psychological distress for Indigenous people. The services and interventions that are most appropriate may vary according to factors such as location, language, gender and socioeconomic status. The study also demonstrates the need for further research into the most valid approach to measuring and monitoring levels of psychological distress among Indigenous people as an important aspect of social and emotional wellbeing.