Mental health service use and ethnicity: An analysis of service use and time to access treatment by South East Asian‐, Middle Eastern‐, and Australian‐born patients within Sydney, Australia

Abstract

Objective

The current research aimed to assess the association between country of birth and use of a specialised mental health service in Sydney, Australia.

Methods

Patient file data were analysed from individuals who accessed the Clinic for Anxiety and Traumatic Stress in Western Sydney between 1996 and 2010. Patients had undergone a clinical assessment and research interview prior to receiving treatment. Data on demographic information and health history were extracted from these files. South East (SE) Asian‐ and Middle Eastern‐born minority groups were compared with an Australian‐born majority group, using country of birth as a proxy measure of ethnicity. Ratios of service use by group were compared with data on ethnicities residing within the local government area health district.

Results

Relative to the local population, country of birth minority status was associated with fewer patients accessing the service, with SE Asian‐born patients reporting low service use across all cohorts studied. However, Middle Eastern‐born patients' service utilisation increased over time, becoming commensurate with the local population. Middle Eastern‐born patients reported a significantly shorter delay to seek treatment compared with Australian‐born patients, although no significant differences were reported between ethnic minority groups.

Conclusions

Differences between SE Asian‐ and Middle Eastern‐born groups in service utilisation patterns over time and treatment delay relative to an Australian‐born group highlight the importance of better understanding the impact of ethnicity on service use.

Key words:

• clinical/counselling psychology
• cross‐cultural psychology
• ethnic and race issues
• health attitudes and behaviours

What is already known about the topic

• Although consideration of ethnic diversity is critical in understanding health service use among different ethnic minority groups, there is little evidence that focuses on such differences.

• Low levels of access to mental health services are reported in SE Asian‐ and Middle Eastern‐born migrant populations within Australia, with these two ethnic minorities representing the largest ethnic minority groups residing within the Bankstown area of Western Sydney, Australia.

• International literature purports migrants experience a delay to access health treatment; however, research within an Australian context suggests that this may not be the case for ethnic minorities accessing specialised mental health services within Sydney, Australia.

What this topic adds

• The current study provides a direct comparison between two main ethnic minority populations living within Sydney who have previously reported service utilisation difficulties, SE Asian‐ and Middle Eastern‐born patients, as well as a comparison with an Australian‐born patient sample.

• Country of birth, as a proxy measure of ethnic minority status, is associated with significant differences in patterns of service utilisation. SE Asian‐born patients consistently underutilised services, while Middle Eastern‐born patients showed improved service utilisation across time.

• Significant differences existed between Middle Eastern‐born patients and Australian‐born patients on duration of time to access a specialised mental health service, with Middle Eastern‐born patients likely to access treatment 2 years faster.

International research reports that migrants, refugees, and ethnic minority groups are less likely to, and report greater difficulty in, accessing or receiving appropriate health care (Armstrong & Swartzman, 2001; Betancourt & Cervantes, 2009; Cowan, 2001; Cummings & Druss, 2011; Garrison, Roy, & Azar, 1999; Murray & Skull, 2004; Pirkis, Burgess, Meadows, & Dunt, 2000; Whitley & Lawson, 2010). As these difficulties result in poorer health treatment and health outcomes, focus should be placed on improving timely access to mental health services (Snowden, Masland, Peng, Wei‐Mien Lou, & Wallace, 2011). Given the degree of multicultural heritage now present within many countries, it is also important to consider cultural and linguistic diversity when evaluating health needs (Rehzaho, 2008). Caution must be taken against treating all ethnic minority groups as homogenous, when considering service utilisation (Dogra, Singh, Svirydzenka, & Vostanis, 2012). It is only through analysing individual ethnic minority groups (Hsu, Davies, & Hansen, 2004) that differences in the presentation of psychological distress and health service need can be accurately captured. Yet many national health surveys, such as those conducted within Australia, frequently group all persons from non‐English‐speaking backgrounds into one category when investigating mental health (McEvoy, Grove, & Slade, 2011; Pirkis et al., 2000; Sharma, 2012).

Many immigrants report difficulties understanding or accessing health services within the Australian health‐care system (Blignault, Ponzio, Rong, & Eisenbruch, 2008; Renzaho, Polonsky, McQuilten, & Waters, 2013), with lower levels of access to mental health treatment reported across migrants groups, including those of South and South East (SE) Asian or African and Middle Eastern country of birth specifically (Anikeeva et al., 2010; Boufous, Silove, Bauman, & Steel, 2005; Wong, Lam, & Poon, 2010). SE Asian refugees within Australia tend to not access mental health treatment, despite documented need (Hart, 2002). High levels of family burden are reported by Vietnamese families of patients with schizophrenia for example (Anikeeva et al., 2010), where there is preference to look to family systems for support before turning to a medical system. Given that Vietnamese migrants are one of the largest foreign‐born groups residing within Australia (Anikeeva et al., 2010), inadequate engagement and underrepresentation within health services are a significant concern.

Even though ethnic minority populations residing in other English‐speaking countries have previously reported longer health treatment delays relative to the ethnic majority, in particular Asian immigrants (Lam & Kavanagh, 1996), the literature appears to be mixed in the Australian context. Patients from an Indochinese background report a longer delay to access psychiatric treatment for the first illness episode compared with Australian‐born patients (Lam & Kavanagh, 1996). However, although Chinese outpatients with anxiety disorders report lengthy treatment delays of 7 years on average (Ho, Hunt, & Li, 2008), other research indicates this may be commensurate with the general population. For instance, a specialist anxiety clinic within a metropolitan area of Australia has reported treatment delays of up to 11 years for the general population, which sits roughly in line with data from the USA citing 6–14 years on average (Thompson, Issakidis, & Hunt, 2008). One study reporting on separate ethnic minority groups found that those of Asian or Middle Eastern ethnicity reported no significant differences compared with the Australian‐born population in duration of time taken to reach a specialist mental health service (Steel et al., 2006). Moreover, this same research reported that the median treatment delay to a specialist mental health service was as little as 6 months, significantly shorter than previous research has indicated. Further research is needed to clarify treatment delay among ethnic minorities who access a specialist outpatient anxiety or mental health service within the Australian population.

One approach that can determine whether differences in mental health service use or duration to seek treatment are present between ethnicity groups is to study particular local areas where there is a high degree of cultural diversity. Considering the Bankstown local area within Western Sydney, 55% of the population speak languages other than English at home, with Arabic speakers from Lebanon and Vietnamese the most common ethnic minority populations (Australian Bureau of Statistics, 2011a). Between 2006 and 2011, these migrant groups increased while the local Australian‐born populations decreased. Given SE Asian and Middle Eastern ethnic groups have previously reported underutilisation of mental health services (Boufous et al., 2005) but commensurate treatment delays (Steel et al., 2006) within Australia, understanding the factors that may impact the mental health service engagement of these specific ethnic minority populations will improve the ability of services to meet their health needs.

The Clinic for Anxiety and Traumatic Stress at Bankstown Hospital in Sydney, Australia (Bankstown Anxiety Clinic; BAC) is a free local area mental health service founded in 1996, designed to cater for patients of differing ethnicity backgrounds with anxiety‐ and mood‐based disorders. Historically, BAC has provided specialist group therapy for the treatment of anxiety and stress in Arabic‐speaking women, and individual treatment for Vietnamese women experiencing panic attacks (Wagner & Joukhador, 2001). In 1996 soon after establishment, both Middle Eastern (Arabic‐speaking) and SE Asian (Vietnamese‐speaking) patients were underrepresented at the BAC (Wagner & Joukhador, 2001). Due to this underrepresentation, the Arabic‐speaking community was targeted specifically in order to increase levels of participation, by contacting local bilingual counsellors and Arabic‐speaking general practitioners. This contact resulted in the running of two Arabic‐speaking women's groups for stress management. At this time the clinic employed an Arabic‐speaking psychologist, and later a second Arabic‐speaking psychologist was also employed.

However, no commensurate intervention was targeted towards the local Vietnamese‐speaking population, nor had the clinic employed a Vietnamese‐speaking clinician. When service utilisation was assessed 10 years later in 2006, underutilisation of the service by Vietnamese patients persisted, relative to the ratio of Vietnamese speakers residing in the local population (Wagner, Manicavasagar, Silove, Marnane, & Tran, 2006). Moreover, the same research indicated there was no treatment delay difference between this sample of Vietnamese patients and Australian‐born patients. Given previous research indicating low rates of service utilisation by Asian communities, it is likely that BAC continues to be characterised by lower rates of service utilisation by SE Asian‐born and Vietnamese‐speaking patients, compared with Middle Eastern‐born or Arabic‐speaking patients, or Australian‐born ethnic majority populations living within the same local area.

Through an analysis of patient service use and clinical information, the current research aims to assess the numbers of patients who accessed BAC relative to the general population within the same local area, to determine levels of service utilisation across SE Asian‐, Middle Eastern‐, and Australian‐born patients across time. Second, the time taken to access BAC by country of birth will be assessed to determine duration of time to access treatment by country of birth to determine whether differences exist between these ethnicity groups. Overall it is predicted that:

1. SE Asian‐ and Middle Eastern‐born patients' service use will be lower than service use by Australian‐born patients, based on proportions of these ethnicities within the local population.

2. Middle Eastern‐born patients will exhibit a higher rate of service use relative to SE Asian‐born patients across time.

3. Country of birth status will not be associated with a delay in accessing treatment.

Methods

Participants and design

Patient files of individuals who were assessed at BAC in Western Sydney between 1996 and 2010 for an anxiety or mood disorder were analysed. In order to be eligible for treatment, all patients were required to currently reside within the local area health district. The Sydney South West Area Health Service Human Research Ethics Committee granted approval for the collection of the current data. Country of birth groupings were used in line with previous research as a means of categorising ethnic identity (Boufous et al., 2005). Of the total 1,487 patients who were assessed during this period, 1,198 were included for further analysis as they were either Australian, North African/Middle Eastern, or SE Asian, using country of birth groupings as defined by the Australian Standard Classification of Cultural and Ethnic Groups (Australian Bureau of Statistics, 2011b). Although crude, this method of grouping allows for a direct comparison with Australian national health data.

Of the Australian‐born patients, 97.4% spoke English at home (n = 1,058; age in years: X = 36.2, SD = 13.6; 67.7% female). Middle Eastern‐born patients (n = 102; age in years: X = 36.0, SD = 10.1; 66.7% female; 71% speaking Arabic at home) were from a total of 10 countries: Algeria (n = 2), Egypt (n = 6), Iran (n = 2), Iraq (n = 3), Jordan (n = 4), Lebanon (n = 70), Palestine (n = 3), Saudi Arabia (n = 1), Syria (n = 7), and Turkey (n = 5). SE Asian‐born patients (n = 37; age in years: X = 37.6, SD = 10.7; 73% female; 32.4% speaking Vietnamese at home) comprised nine countries: Cambodia (n = 3), East Timor (n = 1), Indonesia (n = 2), Malaysia (n = 2), Myanmar (n = 1), Laos (n = 1), Philippines (n = 8), Thailand (n = 2), and Vietnam (n = 17).

Measures

In order to assess service utilisation, the number and the proportion of new patients from Australian, Middle Eastern, and SE Asian countries of birth was compared with the population residing within the local government area. Population data were extracted from the Census for Population and Housing, collected by the Australian Bureau of Statistics across time intervals of 5 years, ranging from 1996 to 2011 (Australian Bureau of Statistics, 2011c).

Time of symptom onset was assessed using the Structured Clinical Interview for Diagnosis (SCID; First, Spitzer, Gibbon, & Williams, 1997). The SCID is a semi‐structured clinical interview designed to identify the disorders listed in the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 1997). Modules A (mood disorders) and F (anxiety disorders) were administered. Treatment delay in years was calculated by subtracting the age of first symptomatic expression for an anxiety or mood disorder from patients' age at presentation to BAC.

Procedure

Data were extracted from individual files of patients who had undergone an individual clinical assessment and research interview by a clinical psychologist or psychologist prior to receiving subsequent individual or group cognitive behavioural treatment for the management of an anxiety or mood disorder. Patients from a non‐English‐speaking background had been assessed with an interpreter or by an Arabic‐speaking psychologist when needed. Demographic and health information was extracted from files, including age, country of birth, language spoken at home, referral source, and age of onset of symptoms.

Planned analyses

Chi‐square test for goodness‐of‐fit analyses were used to assess the statistical significance between the proportions of patients seen at the clinic, compared with the expected proportion, based on the ethnicity profile of the local government area generated from the census survey, by ethnicity group across time. Across all groups, duration to reach treatment ranged from less than one to 65 years with an interquartile range (IQR) of 11. Due to significant skew on duration to seek treatment, the median time in years was estimated. Differences between groups were calculated using the Mann–Whitney test.

Results

Preliminary analyses

There were no significant differences between groups (Australian‐born, Middle Eastern‐born, SE Asian‐born) on gender ratio (χ² (2, n = 1,197) = .52, p = .771) or age (F(2, 1,193) = .21, p = .808). There was also no significant difference between referral source (χ² (2, n = 1,040) = 2.99, p = .224), with all groups most likely to be referred to BAC by a general practitioner compared with any other health professional or department (frequency and percentile information for referral type by group is displayed in Table 1).

Table 1. Frequency and percentile information for referral type by ethnicity

	Australian‐born (n = 1,053)		Middle Eastern‐born (n = 99)		South East Asian‐born (n = 36)
	Freq.	%	Freq.	%	Freq.	%
General practitioner	682	65.6	58	58.6	21	56.8
Hospital department	13	1.3	4	4.0	—	—
Community health	31	3.0	5	5.1	3	8.1
Psychologist	42	4.0	3	3.0	2	5.4
Psychiatrist	65	6.3	7	7.1	—	—
Counsellor	3	.3	—	—	—	—
Other health professional	106	10.2	14	14.1	7	18.9
Self‐referred	89	8.6	6	6.1	4	10.8
Government agency	9	.9	2	2.0	—	—

Tests of hypotheses

Hypotheses 1 and 2: Australian‐born patients were significantly overrepresented within the clinic across all time points, while SE Asian‐born patients were significantly underrepresented (refer to Table 2). Middle Eastern‐born patients were significantly underrepresented between years 1996 and 2001; however, at later time points, the proportion of new patients of Middle Eastern background was not statistically significantly different from proportions of this ethnicity group within the local population.

Hypothesis 3: Analyses indicated no significant differences between Australian‐born (4 years, n = 733, IQR = 11) and SE Asian‐born patients (3 years, n = 30, IQR = 15.5; U = 10491.00, z = −.43, p = .668), or between Middle Eastern‐born (2 years, n = 69, IQR = 7) and SE Asian‐born patients (U = 926.0, z = −.84, p = .403) as predicted. However, a significant difference in treatment delay was reported between Australian‐born and Middle Eastern‐born groups (U = 21306.5, z = −2.18, p = .029), with Middle Eastern‐born patients reporting a shorter median delay to treatment.

Table 2. Chi‐square goodness‐of‐fit analyses for observed number of patients at Bankstown Anxiety Clinic (BAC) compared with the expected number of patients, based on local population data (Australian Bureau of Statistics, 2011c), by ethnicity and year interval

		1996–2001				2002–2006				2007–2011
		Observed, n (%)	Expected, n	χ^2	p	Observed, n	Expected, n	χ^2	p	Observed, n (%)	Expected, n	χ^2	p
Total inhabitants	Population	165,070	—	—	—	170,488	—	—	—	182,234	—	—	—
New patients	BAC	940	—	—	—	387	—	—	—	159	—	—	—
Australian‐born	Population	96,976 (58.7)	—	—	—	96,655 (56.7)	—	—	—	102,515 (56.2)	—	—	—
	BAC	671 (71.4)	534.0	83.37	<.000	280 (72.4)	217.7	41.13	<.001	120 (75.5)	87.7	27.22	<.001
Middle Eastern‐born	Population	13,923 (8.4)	—	—	—	16,250 (9.5)	—	—	—	18,745 (10.3)	—	—	—
	BAC	56 (6.0)	77.5	6.53	.011	30 (7.8)	36.5	1.27	.259	16 (10.1)	16.1	.01	.986
SE Asian‐born	Population	12,312 (7.5)	—	—	—	13,623 (8.0)	—	—	—	16,289 (8.9)	—	—	—
	BAC	28 (3.0)	69.2	26.54	<.001	8 (2.1)	30.7	18.26	<.001	2 (1.3)	13.9	11.17	.001

Discussion

The current research aimed to assess rates of service utilisation relative to the local population and duration of treatment delay for Middle Eastern‐, SE Asian‐, and Australian‐born patients who had accessed services at a specialised mental health service within Western Sydney, Australia. As predicted, a difference in proportion of service use between the local population and BAC patients by country of birth grouping revealed that Australian‐born patients, as the Australian ethnic majority, were more likely to have been seen at the clinic across time. While ethnic minority groups, Middle Eastern‐ and SE Asian‐born patients, reported lower than expected service utilisation on at least one time point between 1996 and 2011. The current results are consistent with previous findings that migrants of ethnic minority status report poorer mental health service utilisation compared with ethnic majority patients living within the same geographical region. Similar findings are reported globally, with poorer access reported in Chinese migrants living within Britain (Cowan, 2001), to African Americans in the USA (Whitley & Lawson, 2010).

The current study uniquely assessed two concurrent ethnic minority populations in order to highlight differences in service utilisation, which may occur within the same geographical area. As predicted, both at baseline and over time, rates of service use differed by country of birth grouping, with Middle Eastern‐born patients showing improved rates of service across time, becoming commensurate with the rates of access expected from the local population, compared with SE Asian‐born patients, whose rates of service use was consistently low. This finding may in part be due to BAC fostering early engagement with the Arabic‐speaking community. However, it is interesting to note that although proportionately greater numbers of Middle Eastern patients accessed the service over time, despite similar referral pathways used, the service use of SE Asian patients was consistently low.

It must be noted that the prevalence of disorders within the community was not assessed, nor the difference in clinical presentations between ethnic groups, to account for cross‐cultural differences in illness expression, or the difference in stressors related to migration experiences. However, despite these, the finding of lower service utilisation by SE Asian groups does sit in line with previous literature, highlighting the difficulties with engaging Asian communities in health services both in Australia (Anikeeva et al., 2010; Blignault et al., 2008) and overseas (Hsu et al., 2004) despite a need for services being present (Hart, 2002). The finding of a difference in service utilisation patterns across time also reiterates the importance of analysing ethnic minority groups separately in order to better understand service utilisation (Dogra et al., 2012).

Although research across global settings has anticipated a longer duration of time to seek treatment for ethnic minority groups compared with the ethnic majority, research within the Australian context has indicated that this difficulty may not be present (Steel et al., 2006; Wagner et al., 2006). Therefore, despite differences in the degree of service utilisation, there was no predicted difference between groups on duration to seek treatment. Although results indicated that there was no significant difference between SE Asian‐ and Middle Eastern‐born patients, in line with previous findings of no reported differences between ethnic minority groups (Steel et al., 2006). significant differences were reported between those of Middle Eastern and Australian background. Unexpectedly, those of Middle Eastern country of birth reported a significantly shorter treatment delay compared with Australian‐born patients.

It appears therefore that those Middle Eastern patients who were able to access this particular mental health service were able to do so within a relatively shorter time period from the illness onset, relative to both the Australian‐born and SE Asian‐born patients. This result may be partly explained by the practice of contacting Arabic‐speaking local health professionals in order to foster engagement with the Arabic‐speaking and wider Middle Eastern community, and the availability of both an Arabic‐speaking psychologist and translator services within BAC. Given all groups were as likely to be referred via a general practitioner, it is likely that targeting general practitioners within the Middle Eastern community may have resulted in a decrease in duration of time to seek treatment at this specialist service. Second, this finding may also be explained by the help‐seeking patterns of those with Middle Eastern ethnicity. Help‐seeking patterns appear to differ both between ethnic minority groups, and ethnic minority and majority groups (Rudell, Bhui, & Priebe, 2008). Regardless, the finding of a difference between ethnic groups reiterates the importance of further understanding engagement to services.

Despite providing insight into the association between country of birth and service use, the current study is limited by the inherent difficulties in utilising pre‐existing data. For example, the time taken to access treatment was calculated using the time between the onset of symptoms and being seen in the clinic; however, this analysis assumes that no other health treatment has occurred, which would not necessarily be the case. A better understanding of health history, and particular illness expression, may assist in furthering the understanding of how different ethnic minority groups engage with mental health services. Moreover, the limited information about previous contact with health services means it is not possible to accurately predict the direct impact that early engagement with other Arabic health services had, and how this may have translated to a difference in findings for both service utilisation and time to access treatment. Additionally, the time spent in Australia was not recorded, meaning that the impact of acculturation and ethnic identity was not formerly measured.

Given the use of a naturalistic sample in the current research, the sample sizes for groups, SE Asian‐born in particular, were comparatively low, reducing statistical power. Caution must therefore be taken when interpreting negative findings. Moreover, in order to compare the current data directly with available health data, the statistical analyses were not able to control for factors such as age or gender. While these factors were consistent across country of birth groupings within the current sample, simple statistical analyses were used with the intention to highlight whether differences exist within a community sample. Given the significant differences found, future research could endeavour to account for ethnicity and other individual factors when assessing ethnic minority differences in service utilisation patterns across time and the duration to seek treatment.

The significant differences found between groups in service utilisation patterns over time and time taken to seek treatment do beg the question as to what makes one particular cultural group more or less able and/or willing to engage with health services compared with another. It should be acknowledged that there may be a difference in prevalence of mental health disorders between ethnic populations, which has not been tested, and may indicate that for some populations health services may not be culturally appropriate or warranted. However, attitudes towards help seeking (Hamid, Simmonds, & Bowles, 2009; Rudell et al., 2008), stigma associated with mental illness (St Louis & Roberts, 2013), and cultural expressions of distress (Dura‐Vila & Hodes, 2012) are all additional factors that may impact on barriers to engage with local mental health services by ethnic minority patients and could be further investigated in the future. Based on the current findings, it may be expected that community engagement with local practitioners of a culturally diverse background has a direct impact on service utilisation and duration of time to access appropriate treatment. It could also be argued that pre‐existing stigma or negative health belief attitudes would be directly correlated with lower rates of mental health access. Future research could be undertaken, perhaps through the recruitment of a specific community group or ethnic minority patient population, in order to test these predictions with the aim to further understand barriers to service use and the reasons for patients' engagement.

Conclusions

Ethnic minority status, as assessed by an individual's country of birth, is related to lower service utilisation relative to the Australian‐born population. This finding was reflected in the current study with continued lower service use by SE Asian‐born patients relative to Australian‐born patients living within Sydney, Australia. Differences in service use over time by country of birth status highlight the importance of investigating ethnic minority groups individually in order to better understand the way culture and ethnic identity may impact on service use. In line with previous Australian literature, the current study found no significant differences between ethnic minority groups, SE Asian‐ and Middle Eastern‐born patients, on the length of time taken to seek treatment. Surprisingly however, those of Middle Eastern background reported a significantly shorter treatment delay compared with ethnic majority Australian‐born patients. Further research is needed to better understand the impact of demographic factors such as treatment and illness history, cultural factors such as stigma or illness expression, or the community engagement practices of health service on service utilisation.