Abstract
Studies conducted in Australia and overseas have consistently found that ethnic groups from non-English-speaking backgrounds (NESB) underutilise mental health services [1–3]. Language problems, lack of information about services, and concerns about stigma and shame have been identified as contributing to the under-representation of the elderly from NESB in statistics of mental healthcare utilisation [4].
Limited proficiency in English is frequently cited as the most problematic factor for the ethnic elderly in accessing mental health and other aged care services [5–6]. This is not only a problem for recently arrived elderly migrants, but also for many who came to Australia in the post-World War II migration boom. This group, which is generally poorly literate, has struggled to acquire English proficiency and a significant number have never achieved this, despite many decades of residence in Australia [7–8]. This is particularly the case for women from NESB whose proficiency in English is generally lower than men of the same ethnic group [9].
The underutilisation of mental health services by the ethnic elderly has also been attributed to a greater reliance on family networks [4]. Factors which may influence a family's decision not to pursue professional help for a psychiatrically unwell elderly relative include the stigma attached to mental illness, and a sense of shame associated with abandoning a loved one to an impersonal and unfamiliar environment. This can place a significant burden on families from NESB, particularly if there are strong cultural expectations with regard to caring for ageing family members. If such carers have little contact outside their ethnic community, then they are less likely to know about, or be able to access, home-based nursing or respite services.
Attempts are being made to address the inequitable access of elderly persons from NESB to the range of aged health and support services in the community. These have been outlined in The Ethnic Older Persons Strategy launched by the Commonwealth Department of Human Services and Health in 1995 [10]. Two epidemiological trends have particularly highlighted the need for greater culture-sensitivity in the planning and development of new mental health services for the aged and the remodelling of existing services. One is the increasing proportion of elderly in most communities from NESB. This is a particularly pertinent issue for those ethnic groups whose peak period of immigration was in the immediate post-war period, as a significant proportion of these communities are now elderly [8]. Also, the rate of increase in the proportion of aged people is much greater among communities from NESB than among the Australian-born. A second trend is reflected in studies of the prevalence of mental disorder in Australia which have found that, while this is stable for Australian born after 60 years (about 16%), for the elderly from NESB the rate increases from 14% at ages 60–64 years, up to 26% of individuals aged 75 years and over [1]. It is, therefore, not surprising that suicide rates are greater for older migrants (both males and females), than the Australian-born elderly [1].
In evaluating the extent to which existing mental health services are meeting or not meeting the needs of the elderly from NESB, it is important that service utilisation data not be interpreted as a proxy for estimates of mental illness prevalence or relative need for psychiatric services [8–11]. However, examination of the patterns of service utilisation by elderly persons from NESB can assist in identifying whether particular ethnic groups are failing to access psychiatric services, and whether variables relating to their experience of the service differ significantly from Australian-born elderly. In a retrospective study of the utilisation of adult psychiatric services by patients of ESB and NESB, Trauer [12] found longer median lengths of stay for NESB inpatients, and lower rates of voluntary admission for NESB patients. He also found a variable and generally low use of interpreters for this group, although English fluency was not routinely recorded in the files. Trauer was unable to determine if interpreters were being underutilised or not required for the majority of patients from NESB. However, he did conclude that there was a need to improve the recording of ethnicity-relevant factors in patient's histories, access to interpreters and the ‘user-friendliness’ of psychiatric settings for patients from NESB.
The aim of the current study was to examine the admission patterns for elderly persons from NESB and ESB to a psychogeriatric service, the catchment area of which covers the ethnically diverse northern and western regions of metropolitan Melbourne. Although most studies of mental health service utilisation have focused on younger adult populations from NESB, there is an increasing recognition that service-utilisation patterns for the elderly in these communities need to be examined separately [1]. Sociodemographic, clinical and service usage variables were used to compare an inpatient population from NESB and ESB. These data were examined for differences in admission patterns, as has been found in adult psychiatric services [12]. A further aim was to identify areas of service provision which could be improved for elderly patients from NESB who require psychiatric hospitalisation.
Methodology
Subjects
All admissions over a 12-month period (January to December 1996) to the acute admission unit of the aged psychiatry service for the north-west and western catchment population of Metropolitan Melbourne were examined. This 30-bed unit is the main acute inpatient facility in this region for elderly individuals with psychiatric illness. Elderly patients admitted to private psychiatric hospitals were not included; however, there are very few such facilities in this region and most elderly are of working-class background and do not have private health insurance. This unit is also the only facility in this catchment area for elderly patients who require compulsory admission. As a consequence, the more severe end of the spectrum of psychiatric illness is overly represented in this sample.
Measures
The following variables were collated from the files of all patients admitted to the unit during the specified 12-month period: date of birth, sex, country of birth, preferred first language, marital status, living arrangement, use of interpreter, diagnosis, legal admission status, duration of stay and previous psychiatric treatment.
In order to compare whether the ethnic mix of the sample was representative of the different cultural groups in the catchment area, sociodemographic data were obtained from the Australian Bureau of Statistics 1996 Census [13].
For purposes of analysis, the sample was divided into subgroups from ESB (Australia and United Kingdom) and NESB on the basis of the individual's preferred first language.
Procedure
The majority of patients were assessed prior to admission by one of the two community teams which are the initial points of contact with the service and provide a ‘gatekeeping’ role for the inpatient unit (a minority of patients were transferred directly from general hospital accident and emergency departments or inpatient services). Only those patients who were assessed as being unsuitable for management in their home environment were admitted, usually because of concern about self-neglect or self-harm, the complexity of required investigations and treatment, or lack of appropriate social support.
Data analysis
Analysis was undertaken to examine the characteristics of the sample subgroups using a range of descriptive and inferential statistics. The extent to which the ethnic composition of the first-admission sample reflects the proportion of different cultural groups in the catchment area population was assessed using exact binomial tests. This procedure tests the likelihood that the observed proportion in the sample equals the proportion in the catchment area. Comparisons between subgroups with NESB and ESB on sociodemographic and clinical variables were made using a range of parametric and non-parametric tests. Continuous data were compared using equal variances t-tests, while highly skewed data such as length of stay were transformed to normality using a log-transformation procedure before analysis. The Mann-Whitney U-test, a non-parametric procedure that also reduces the influence of outliers, produced a similar result. Chi-squared tests were carried out on cross-tabulated data, and exact tests of significance carried out where appropriate.
Missing data
There was a small amount of missing data on a number of variables. It was not possible to determine the language status for two patients due to missing data on preferred first language. Details on the marital status of a further six patients were not available, nor the need for an interpreter for eight patients. The amount of missing data on other variables was negligible. Each analysis was conducted using all available data, hence the number of cases across the analyses ranged from 243 to 251, with most analyses based on a sample size of 251. The amount of missing data was not considered sufficient to present a threat to the validity of the findings of this study.
Methods
There were 253 admissions to the psychogeriatric unit during the specified 12-month period. Of these, 221 were first admissions, and the remaining 32 admissions involved 25 patients who were re-admitted one or more times. Apart from the comparison with Census data, for which only first admissions were included, the unit of analysis was taken to be an admission to the psychogeriatric unit. The gender distribution of the total sample was 164 (64.8%) females and 89 (35.2%) males. 128 (50.6%) patients were born in Australia, 23 (9.1%) patients were born in the United Kingdom and 102 (40.3%) patients were born in countries in which English was not their original language. The mean age of patients from ESB was 74.7 years (SD = 7.6) and the mean age for patients from NESB was 74.2 years (SD = 7.1).
Initially, admissions by country of birth were compared with 1996 Census data [13] for all persons over the age of 65 years living in the catchment area serviced by the unit. Re-admissions were not included in this analysis. Table 1 shows that elderly persons from ESB and those from European countries with post-World War II immigration to Australia were represented in the inpatient admissions in similar proportion to the catchment area figures. Of note, elderly persons from Vietnam were under-represented in the inpatient population, although the inpatient numbers are small for this group. In contrast, elderly patients from European backgrounds, apart from Italy, Greece and Yugoslavia, were over-represented in the inpatient population.
Comparison of first admission rates with 1996 Census data over 65 years
The sociodemographic characteristics of inpatients from NESB and ESB are shown in Table 2. There was a very strong trend for more males from NESB to be admitted than elderly females (X 2 = 4.04; df = 1; p = 0.045). There was a statistically significant association between NESB status and marital status (X 2 = 12.24; df = 4; p = 0.02), with a greater proportion of the NESB sample currently married than patients from an ESB (48% vs 32.7%), and fewer NESB patients separated or divorced (4% vs 14.6%). Although NESB patients were more likely to live with their spouse, children or other family (33.8% vs 21.3%), and less likely to live in hostel or similar supported accommodation (6.5% vs 14.5%), these differences were not statistically significant (p = 0.15). The proportion of each group in nursing homes was very similar (35.1% vs 34.1%), suggesting that, when this level of care is required, the presence of advanced physical dependency or behavioural disturbance takes precedence over differing cultural approaches to caring for the elderly.
Means (SD) or percentage frequencies for sociodemographic variables for non-English-speaking background (NESB) and English-speaking background (ESB) groups
Table 3 shows that there was a strong trend for elderly patients from NESB to be admitted involuntarily (X 2 = 3.79; df= 1; p =0.05), although there was no significant difference in the mean length of stay. Previous contact with psychiatric services and previous admissions to the psychogeriatric unit were similar across both groups. When admission diagnoses were compared across the two groups, a significant difference was noted in the dementia and affective disorder categories. Agreater proportion of elderly NESB patients were diagnosed with dementia and fewer were diagnosed with affective disorders (X 2 = 15.9; df = 4; p = 0.003). This was most marked for males from NESB; only 8.8% were diagnosed with affective disorder, while 37.0% males from ESB were diagnosed with affective disorder. The difference was less apparent for females; 37.2% were diagnosed with affective disorder in contrast to 53.3% females from ESB. The figures are reversed for the diagnosis of dementia; 64.7% males from NESB and 42.6% males from ESB received this diagnosis. The difference was, once again, less marked for females; 30.2% from NESB and 20.8% from ESB received this diagnosis. No statistical association was detected between diagnosis and NESB status for females (X 2 = 3.9; df =4; p = 0.42); however, this association was statistically significant for males (X 2 = 12.8; p = 0.009).
Means (SD) or percentage frequencies of admission variables for non-English-speaking background (NESB) and English-speaking background (ESB) groups
When country of birth status was compared with preferred first language, 60.2% patients were born in countries where English was the first language while 69.3% patients nominated English as their preferred language. This discrepancy resulted from a small number of patients from countries other than Australia and the United Kingdom who also nominated English as their preferred language. Interpreters were used in 62.3% of those patients for whom English was not their preferred first language. Information was not available about the threshold of English proficiency which determined if an interpreter was used, nor the frequency of interpreter use during the admission period. Involuntary admissions were significantly higher for patients from NESB (60.5%), who required an interpreter, than for those NESB patients who did not require an interpreter (X 2 =4.3; df = 1; p = 0.03).
Discussion
This 12-month review of admissions to a geriatric psychiatry inpatient unit found significant differences in the utilisation of the service by patients of NESB as compared with those from ESB. Before discussing the relevance of these findings for future service planning and provision, the limitations inherent in this study need to be addressed.
First and most fundamentally, utilisation rates of inpatient psychiatric services do not necessarily reflect the need or prevalence of psychiatric morbidity within the community [8]. This is particularly the case for elderly persons from ethnic minority groups whose accessibility to such services may be influenced by cultural attitudes to mental illness and a preference for alternative patterns of help-seeking behaviour [4]. Second, the term ‘NESB’ is a crude categorisation which does not take into account the heterogeneity of communities given this designation, nor factors such as the length of time individuals have lived in Australia, their acculturation into the society, nor fluency with the English language.
Third, the methodology of this study involved a retrospective file audit which utilised clinician diagnoses rather than standardised research assessment instruments. There was, however, significant diagnostic consistency provided by the four senior consultant psychiatrists who worked in the inpatient unit during the period of the study. Finally, it must be noted that the sample numbers for some ethnic groups were small and this affected the stability of statistical findings.
Despite these limitations, the findings of this study revealed both similarities and differences in admission patterns between elderly patients from NESB and ESB. In general, admission rates for elderly patients from NESB reflected the representation of that ethnic group in the catchment area population figures, although admissions from European countries, apart from Italy, Greece and Yugoslavia, were over-represented. The admissions from individual European countries were too small in number to analyse individually; however, this finding is consistent with the relatively greater rate of psychiatric admissions in New South Wales for certain European countries noted by Minas et al. [8]. These authors also found that the admission rates for most other NESB groups were consistently lower than for the Australian-born, but that there was significant variation between ethnic groups. In particular, the lowest psychiatric admission rates were for individuals from South-East Asia. During the 12-month period of this clinical file review only two elderly Asian-born patients were admitted to the psychogeriatric unit. Although the Asian community in this region of Melbourne is steadily growing, their overall numbers are still relatively small, and it is not clear at this stage if they significantly under-utilise the psychogeriatric service. This group also constitute the most recent elderly immigrants to Victoria, have the lowest rates of proficiency in English and are most likely to be heavily dependent on relatives and feel culturally isolated. It is, therefore, likely that they do not readily access psychiatric services, even when these are needed. As a consequence, their mental health care needs require careful monitoring by service providers. The admission rates for the two largest ethnic communities in the catchment area, Italian and Greek, were similar to their percentage representation in the catchment area population. This finding suggests that the elderly from these well-established European communities, many of whom would have migrated in the post-World War II period, utilise psychiatric services in a similar manner to the Australian-born elderly.
Differences in rates of compulsory admission between elderly patients from NESB and ESB were a significant finding in this study. Other studies, examining adult, rather than specifically aged inpatient samples, have found that the rate of involuntary admission is similar across ethnic groups, but that patients from NESB significantly under-utilise the option for voluntary psychiatric admission [8],[12]. Minas et al. [8] have attributed the consistency in involuntary admission rates across cultural and language groups to two major factors. One factor is the inherent lack of choice in compulsory detention, so that the decision about hospitalisation is taken out of the hands of the patient and his/her family. The second contributing factor is the relatively uniform prevalence of severe psychiatric illness, regardless of ethnic background. These findings would suggest that the relatively greater use of compulsory admission for elderly psychiatric inpatients from NESB in this study is likely to result from an under-utilisation of voluntary admission in this group rather than an actual increase in involuntary admissions. For patients from NESB, who also required an interpreter, the relative rate for involuntary admission was even greater. Cultural attitudes to psychiatric hospitalisation may be one factor contributing to resistance by elderly migrant patients, or their families, in accepting voluntary admission.
With regard to gender, the greater female representation in both NESB and ESB groups is to expected on the basis of the greater longevity of women and, hence, their need and use of old age services in general. Within the NESB group, however, this difference was less marked, so that relatively more elderly males from NESB were admitted than from ESB. There were also relatively more elderly males from NESB who were admitted with a diagnosis of dementia. Patients with dementia are generally only admitted to a psychogeriatric unit if there are behavioural complications present which cannot be managed, either in the community, or in a mainstream aged care facility. As the rate of dementia in males from NESB is unlikely to be greater than in other elderly, this finding suggests that the threshold for admission for this group may be lower. Alternatively, presentation for treatment may be delayed beyond the time when community-based intervention strategies could be instituted. A relatively greater increase in admissions for dementia was also seen in elderly females from NESB, but not to the same extent as males. This may reflect the greater community tolerance of dementia-related behavioural disturbance in females than males.
In contrast to the above finding, affective disturbance was diagnosed significantly less in patients from NESB, as compared with ESB. This difference was more marked for males than females and warrants exploration, given the higher prevalence of depressive and anxiety disorders and suicide in older NESB migrants as compared with the Australian born elderly [14],[15]. Poor English proficiency and cultural variations in the expression of depression are obvious factors that may have contributed to the low diagnosis rate of affective disturbance in patients from NESB. There may also be cultural factors which contribute to depression being more ‘masked’ in males in this group [5].
Although this study has focused on admissions to a psychogeriatric unit, only a relatively small number of elderly individuals who are referred to the service eventually require admission. A much greater number are managed in the community, often intensively for a period of time, and admission is avoided as much as possible. Even when a patient has required an inpatient admission, this often represents only a small fraction of the time that they are being managed by the service. It would, therefore, not be appropriate to extrapolate from the findings of this inpatient study to the service as a whole. Differences between inpatients from NESB and ESB were noted in sociodemographic variables, patterns of diagnosis and relative rates of involuntary admission. These findings have been important in assisting the psychogeriatric service to be more aware of particular obstacles that patients from NESB may experience in accessing psychiatric treatment; however, these inpatient findings now need to be examined in the context of all referrals to the service. Patterns of community-based utilisation of the service by patients of NESB and ESB are being examined in order to undertake a comparison with the inpatient data and to facilitate further improvement in service provision for elderly patients from all ethnic backgrounds.
Footnotes
Acknowledgements
We would like to acknowledge Associate Professor Henry Jackson and Associate Professor David Ames for their helpful comments in the preparation of this paper.