Abstract
As in many other countries, Australia has increased its emphasis on access to mental health care for people from non-English-speaking backgrounds over the last decade or so. At a national level, government initiatives have included the National Agenda for a Multicultural Australia [1], the Access and Equity Strategy [2], the National Health Strategy [3] and the National Mental Health Strategy [4, 5]. The National Mental Health Strategy, for example, established the Australian Transcultural Mental Health Network in June 1995 to improve the quality and accessibility of mental health services for people from non-English-speaking backgrounds.
At a State and Territory level, initiatives have included the development of specific ethnic mental health policies in most States and Territories and the establishment of transcultural mental health services in Victoria, New South Wales, Western Australia and Queensland. Each of these services is concerned with enhancing the provision of mental health services to people from non-English-speaking backgrounds (through research, training, policy direction, provision of forums for consumer and carer participation, mental health promotion, provision of clinical back-up, advocacy and brokerage).
These initiatives have been underpinned by data from several studies suggesting that people from non-English-speaking backgrounds have reduced access to mental health services relative to their English speaking counterparts. Studies in Victoria [6–8], New South Wales [9] and Western Australia [10] have reported that people from non-English-speaking backgrounds as a collective group, or people from specific birthplace groupings (e.g. those from South-East Asian communities), are underrepresented in both inpatient and community specialist mental health services (although they may be proportionally represented among those consulting general practitioners about mental health problems) [7]. As well as reporting variability across birthplace groups [6–10], these studies have suggested that access may be particularly problematic for those with English language difficulties [7]. Other factors that may interact with birthplace to impact upon access are diagnosis and duration of residence in Australia [10].
The difficulty with most of these studies is that their points of comparison tend to be population-based census data, which do not take into account health status. So, for example, one Victorian study reported that 20% of Victorians were born in a non-English-speaking country and that 16% of adults admitted to psychiatric inpatient services were born in a non-English-speaking country, and concluded that this group were underrepresented in inpatient services [6]. However, if this group had better mental health status than those from English-speaking countries, then it might be reasonable to expect their rate of inpatient admissions to be lower.
The relative rates of mental health problems among people from non-English-speaking and English-speaking backgrounds is itself is a much-debated issue [3]. Many studies that purport to examine rates in fact examine utilization. For example, Stuart et al. report high rates of mental illness among migrants from Greece, and low rates among those from South-East Asia [11]. However, their rates are based on consultations with mental health or primary care providers, or other evidence of treatment for mental health problems (e.g. use of psychotropic medication). In other words, they provide estimates of treated prevalence, rather than community prevalence.
Some community-based estimates of prevalence have been derived, based on survey data [12–14]. As a general rule, these studies have tended to be less likely than the studies of treated prevalence, described above, to find differences between people from Englishspeaking and non-English-speaking backgrounds. However, it should be noted that the majority of these are at least 20 years old, are based on relatively small samples, define people from non-English-speaking backgrounds in different ways and exclude people who do not speak English. This raises questions about their current generalizability.
In isolation, the estimates of prevalence based on service utilization figures or on community-based surveys do not allow conclusions to be drawn about access to services, because they fail to consider ‘unmet need’. Andrews defines ‘unmet need’ as ‘the proportion of people who meet criteria for a disorder and do not see a clinician’ [15]. This is akin to the rate of untreated disorders, which can provide a reference point from which to determine whether some groups underutilize (or, for that matter, overutilize) services. Meadows et al. take a more complex view of ‘unmet need’, noting that there are people who satisfy diagnostic criteria but have no need for services, and people who do not meet diagnostic criteria who do have need for services. They suggest that ‘perceived need… can be seen as being closely linked with perceptions of mental health care, barriers to care, and the prevalence of untreated disorders’ [16]. By either definition, there is a paucity of work examining relative levels of ‘unmet need’ for people from non-Englishspeaking and English-speaking backgrounds, particularly in the Australian context.
Recently, a major epidemiological survey of the mental health of the general Australian community was conducted (the National Survey of Mental Health and Wellbeing, NSMHWB), that provided the potential to explore the issues described above. Early reports of the findings of this survey have indicated that the 12-month prevalence of mental disorders among people from non-Englishspeaking backgrounds is 14.5% (12.5% for males, 16.9% for females). This compares with 18.6% (18.7%, 18.4%) for those born in Australia, and 15.7% (15.1%, 16.4%) for those born in other main English-speaking countries [17, 18].
In addition to providing for relative prevalence estimates, the NSMHWB also allows rates of service use to be calculated, and permits an explicit examination of unmet need. The current study utilizes the NSMHWB to consider the access to mental health care of people from non-English-speaking backgrounds relative to that of people from English-speaking backgrounds, doing so in the context of the mental health status of both groups. It also considers whether, if they perceive that they have needs for care, these needs are met.
Method
Data sources
All data were taken from the NSMHWB [17, 19]. Data were provided to the authors in a confidentialized unit record file, and analyzed using SPSS (Version 10; SPSS, Chicago, IL, USA) [20] and SUDAAN (Version 7.5.3; SUDAAN Research Triangle, Research Triangle Park, NC, USA) [21].
The NSMHWB was a population-based survey conducted by the Australian Bureau of Statistics between May and August 1997, designed to provide information on the prevalence of a range of mental disorders for Australian adults and to examine health service use for given mental health problems. Australian households were randomly selected, and the adult with the next birthday from each was invited to participate. If this individual declined, no further attempt was made to recruit from that household. This sampling method yielded 10 641 participants aged 18 or over, from whom trained interviewers collected self-report data.
Simple cross-tabulations were conducted, and odds ratios calculated. The statistical technique of jackknife replicate weighting provided for estimation of all statistics taking into account the design error arising from the clustered probability sampling methodology [22].
Specific data available from the NSMHWB are described below.
Mental health status
The NSMHWB incorporated a modified version of the Composite International Diagnostic Interview (CIDI) [23], which permitted mental health status to be determined. The CIDI translated ICD-10 diagnostic criteria into symptom-based questions that could be administered by non-clinical interviewers and scored in a manner that enabled a diagnosis to be assigned. The CIDI identified the presence or absence of three types of mental disorder during the previous 12 months, namely anxiety disorders (including panic disorder, agoraphobia, social phobia, generalized anxiety disorder, obsessive–compulsive disorder and posttraumatic stress disorder), affective disorders (including depression, dysthymia, mania, hypomania and bipolar disorder) and/or substance abuse disorders (including harmful use/abuse and dependence). These disorders were selected on the grounds that they were likely to affect at least 1% of the general adult population, and could be detected within the limitations of a household survey. An individual could experience more than one disorder. The survey also included ‘screeners’ for psychosis and personality disorders, as well as questions on neurasthenia, but these are not considered here on the grounds that their validity is less well tested than that of the CIDI.
Mental health service use
The NSMHWB provided data on each individual's use of any of the following services for mental health problems: inpatient unit, general practitioner, psychiatrist, psychologist, other mental health professional, and/or other non-mental health professional.
Unmet need
All individuals who used services were asked whether they felt they had needs in any of the following areas: information, medication, counselling, skills training and/or social interventions. Perceived needs were considered: (i) fully met when individuals received as much of a particular type of help as they felt they needed; (ii) partially met when they received a particular type of help, but less than they felt they needed; and (iii) not met when individuals felt that they needed a particular type of help, but did not receive any. Responses were aggregated in the following way: (i) all perceived needs fully met; and (ii) at least one instance where perceived needs were partially met or not met.
Country of birth
Country of birth data were collected in the NSMHWB, aggregated to the level of: (i) born in an English-speaking country (defined as Australia, New Zealand, UK, Ireland, Canada, USA and South Africa); and (ii) born in a non-English-speaking country (defined as all other countries). These definitions were taken directly from the Australian Bureau of Statistics [17], and the confidentialized unit record file provided to the authors did not permit any disaggregation of those born in non-English-speaking countries.
For simplicity, the remainder of this paper refers to those born in an English speaking country as ‘people from an English-speakingbackground’ and those born in a non-English-speaking country as ‘people from a non-English-speaking background’. It is acknowledged, however, that the notion of ethnic background is more complex than this.
It should be noted that the NSMHWB excluded potential respondents who did not speak English. The explanatory notes relating to the conduct of the NSMHWB state that ‘proxy, interpreted or foreign language interviews were not conducted’ [17]. They do not describe how English language ability was measured, or how the decision was made that an individual spoke English well enough to participate. No information is available regarding potential respondents who were excluded in this way, either in terms of their absolute numbers or their specific ethnic make-up [Rawson M: personal communication].
Results
Anxiety disorders
Almost 10% of all respondents from non-English-speaking backgrounds reported experiencing an anxiety disorder during the 12 months prior to the survey. This 12-month prevalence figure was not significantly different from that for people from English-speaking backgrounds (OR = 1.00, 95% CI = 0.78–1.27) (see Table 1a).
Mental health status, service use and unmet need by country of birth: anxiety disorders (a) Anxiety disorders by country of birth grouping (all survey participants)
People from non-English-speaking backgrounds with anxiety disorders were not significantly different from their English-speaking counterparts in terms of their likelihood of reporting receiving some form of mental health care. In total, 37.68% of the former group and 46.84% of the latter did so (OR = 0.69, 95% CI = 0.44–1.06). In terms of individual services, people from non-English-speaking backgrounds were significantly less likely to seek care for a mental health problem from ‘other health professionals’, with 6.30% doing so compared with 13.22% (OR = 0.44, 95% CI = 0.22–0.90). There were no significant differences in terms of likelihood of using any other type of individual service (see Table 1b).
When those with anxiety disorders and perceived needs for care were considered, 24.50% of those from non-English-speaking backgrounds reported that their needs were fully met, and 29.04% of those from English-speaking backgrounds did so. Again, this difference was not statistically significant (OR = 0.79, 95% CI = 0.44–1.42) (see Table 1c).
Affective disorders
In total, 5.21% of people from non-English-speaking backgrounds reported experiencing an affective disorder in the 12 months prior to the NSMHWB. This 12-month prevalence figure did not differ significantly from that for those from English-speaking backgrounds (5.90%) (OR = 0.88, 95% CI = 0.63–1.22) (see Table 2a).
Mental health status, service use and unmet need by country of birth: affective disorders (a) Affective disorders by country of birth grouping (all survey participants)
Among those with an affective disorder who were from non-English-speaking backgrounds, 50.24% received some form of mental health care according to the NSMHWB. The corresponding figure for those from English-speaking backgrounds was not statistically significantly different at 66.03% (OR = 0.52, 95% CI = 0.19–1.39). The same finding held true for the use of each individual service for mental health problems (see Table 2b).
Those with an affective disorder who expressed needs for care were considered in isolation, and the degree to which they perceived their needs were met were examined. There was no significant difference between those from non-English-speaking backgrounds and those from English-speaking backgrounds in terms of their likelihood of stating that their needs were fully met, with 28.04% of the former group doing so, and 35.60% of the latter (OR = 0.70, 95% CI = 0.36–1.39) (see Table 2c).
Substance-use disorders
People from non-English-speaking backgrounds were significantly less likely to experience substance-use disorders than those from English-speaking backgrounds. Just under 4% of those from the former group reported that they had experienced a substance-use disorder in the 12 months leading up to the survey, compared with over 8% of those from the latter (OR = 0.45, 95% CI = 0.31–0.65) (see Table 3a).
Mental health status, service use and unmet need by country of birth: substance-use disorders (a) Substance-use disorders by country of birth grouping (all survey participants)
Taking self-reported service use data from the NSMHWB, it can be seen that 31.07% of those with substance-use disorders who were from a non-English-speaking background received mental health care, compared with 28.66% of their English-speaking background counterparts. This difference was not significant (OR = 1.12, 95% CI = 0.61–2.06). There was no significant difference between people from non-English speaking and English-speaking backgrounds in terms of use of any individual service for a mental health problem (see Table 3b).
When those with substance-use disorders and perceived needs for care were considered alone, just under 19% of those from a non-English-speaking background had their needs fully met, and just over 26% of those from an English-speaking background did so. This difference was not statistically significant (OR = 0.66, 95% CI = 0.17–2.63) (see Table 3c).
Any mental disorder
In total, 14.53% of people from non-English-speaking backgrounds who responded to the NSMHWB reported experiencing at least one mental disorder in the 12 months prior to the survey. The corresponding prevalence figure for people from English-speaking backgrounds was significantly higher at 18.21% (OR = 0.76, 95% CI = 0.64–0.91) (see Table 4a).
Mental health status, service use and unmet need by country of birth: any mental disorder (a) Any mental disorder by country of birth grouping (all survey participants)
Around 40% of people from both non-English-speaking backgrounds and English-speaking backgrounds with at least one mental disorder reported receiving services for a mental health problem (OR = 0.91, 95% CI = 0.62–1.35). Proportions of service users were also similar among the two groups in terms of use of individual services for mental health problems (see Table 4b).
Likewise, the proportions of those with perceived needs who felt that their needs were fully met were similar for both groups at around 30% (OR = 0.81, 95% CI = 0.54–1.23) (see Table 4c).
Discussion
The above findings suggest that people from non-English-speaking backgrounds and English-speaking backgrounds were equally likely to experience anxiety disorders and affective disorders. The former were significantly less likely to experience substance-use disorders. The strong impact of substance-use disorders meant that they were also less likely to experience any disorder, when all disorders were considered together. It is worth exploring the possibility that age may have had a significant impact here. There is a strong age-related risk for substance- use disorders, with young people more likely to experience such disorders [17]. Age could potentially have acted as a confounder, with those from non-Englishspeaking backgrounds being more likely to fall into older age categories (as younger, second generation individuals who were born in Australia would be classified as from English-speaking backgrounds). Supplementary analysis that only considered young people (aged less than 45) found that those from non-English-speaking backgrounds were still less likely to experience substanceuse disorders than their English speaking counterparts (5.81% compared with 12.04%; OR = 0.45, 95% CI = 0.28–0.73). When those aged 45 or over were considered in isolation, there was no significant difference between those from non-English and English-speaking backgrounds (1.98% compared with 3.52%; OR = 0.55, 95% CI = 0.26–1.18).
When those with each disorder type were considered alone, there was no significant difference between people from non-English-speaking backgrounds and people from English-speaking backgrounds in terms of their likelihood of using services for mental health problems. In the case of anxiety disorders and affective disorders, however, there was a tendency for access to services for mental health problems to be lower for people from non-English-speaking backgrounds.
Within each disorder grouping, when those who perceived that they had needs for care were considered, there was also a tendency for people from non-Englishspeaking backgrounds to be less likely to report that their needs were fully met. Again, these differences did not reach statistical significance, possibly because of small numbers.
In addition to the issues about small numbers, there are several limitations of the current study that should be considered when interpreting the above results. First, the data were provided to the authors in aggregate form, thus making it impossible to examine differences in morbidity, service utilization and levels of unmet need for individual non-English-speaking background groups. Earlier literature suggests that there may be considerable variation across country of birth groupings in terms of both prevalence of psychiatric disorders and service utilization [6–10], and these differences may be masked by considering all groups in aggregate. In addition, this lack of detail makes it impossible to determine how generalizable these findings are to the total Australian population.
Second, the NSMHWB excluded potential respondents who could not speak English. Given the complexity of psychiatric interviewing, it may have been that even those with reasonable ‘everyday’ conversational English were excluded. This clearly creates a potential selection bias. There are some studies that suggest that the prevalence of mental health problems may vary inversely with language ability (which may be a proxy variable for a cluster of factors, including recency of arrival in Australia and lower acquisition of cultural knowledge and skills). For example, higher psychological distress has been found in those with poor English relative to those with greater competency in English [24]. There are also studies that suggest that those with the greatest difficulty with English have the poorest access to mental health care among all those from non-Englishspeaking backgrounds [7]. This problem has been noted elsewhere in the psychiatric epidemiology literature [25].
If these limitations are ignored and the findings are interpreted in the light of strict tests of statistical significance, it would appear to be true to say that at an aggregate level, when people from non-Englishspeaking backgrounds are compared with people from English-speaking backgrounds, they are not statistically significantly different in terms of their likelihood of accessing services, both in terms of overall access and in terms of access to individual services. This finding is generally at variance with Australian literature that suggests that overall levels of utilization, and levels of utilization of individual specialist mental health services, are lower for people from non-English-speaking backgrounds [6–10]. It is, however, consistent with some Australian and international literature that suggests that use of general practitioners for mental health problems may be equally common among those from non-English speaking and English-speaking backgrounds [7, 26]. This may be related to a number of factors, including the relatively greater supply of bilingual general practitioners, and the fact that there may be less stigma attached to seeking help from a general practitioner than from a specialist mental health service [27].
Given the limitations, however, caution should be exercised in interpreting the findings too literally. Close examination of the data reveals consistent trends that did not reach significance, but generally moved in the same direction, and suggested that people from non-Englishspeaking backgrounds may have reduced access to care and be less likely to have their needs met. This interpretation is more consistent with other Australian quantitative studies that have considered access to mental health care for people from non-English-speaking backgrounds, which have tended to suggest that access to care for this group is poor [6–10].
Clearly there is a need to build on this population-based work, by oversampling people from non-English-speaking backgrounds generally and people from particularly small community groups in particular. Efforts must also be made to ensure that those who do not speak English are included in population samples. A more detailed picture of this nature would allow an examination of groups for whom access to mental health care is particularly poor, which in turn would enable strategies to be put in place to redress imbalances. This work should further explore the reasons for needs being unmet by services, in recognition of the fact that access to mental health care is only part of the equation, and does not say anything about the nature or quality of mental health care that people from non-English-speaking backgrounds (or, for that matter, English-speaking backgrounds) receive once they ‘get through the door’. Such work will further clarify the extent to which the various policy initiatives and associated research and advocacy efforts that began during the 1990s are having an impact.
Footnotes
Acknowledgements
The National Survey of Mental Health and Wellbeing was funded by the Mental Health Branch of the Commonwealth Department of Health and Aged Care, under the National Mental Health Strategy. It was conducted by the Australian Bureau of Statistics. This study is an extension of work made possible by funding from the National Health Priorities and Quality Branch, Commonwealth Department of Health and Aged Care.