Abstract
The majority of the costs of psychiatric consultation in Australia are met from taxpayers contributions, supported either through the Commonwealth-funded Medicare universal insurance scheme, or through the state-funded public mental health services. Training for psychiatrists is almost exclusively within publicly funded services. Hence psychiatric care is a heavily government-subsidized aspect of mental health care delivery, whether ‘private’ or ‘public’ sector. For such welfare funded services, principles of social equity (see, for instance, Bebbington [1]) would suggest that variables such as place of residence, age, and sex should not independently influence service utilization. However, studies in recent decades have described substantial disparities in the use of private and public sector psychiatric services in Australia. Within the private sector, and before the introduction of universal coverage through Medicare, Opit and Gadiel described use of psychiatric services by clients of a major private health insurance fund [2]. They found that females received more services than males; that service use rose with age, reaching a peak in the 31– 45-year-old age group, and then fell; that psychiatric service use had risen over the period 1977–1979; and that the distribution of services was correlated with socioeconomic rating. Following the introduction of universal insurance, Jorm and Henderson examined usage of private psychiatric services across Australia using Medicare data [3]. Private psychiatrists saw 1.25% of the female population and 0.91% of the male population in the year under study. Among children, boys received more services than girls, but from the mid-teens onwards, females were greater consumers than males. The age distribution of rates of service use showed substantial variation, with both children and the elderly being low users. The peak users of psychiatric services were the 35– 44-year-old group. The difference between the sexes was also greatest in this age group. Geographic variation in rate of use of private sector services was considerable, with a similar effect of socioeconomic status as found by Opit and Gadiel [2]. This variation was found to be greater for psychiatric services than for physician services.
Andrews and Hadzi-Pavlovic surveyed a random sample of psychiatrists working in both public and private sector settings about their last 20 patients [4]. They found that 61% of patients were female and that 77% of consultations were in the private sector. Psychotic and neurotic disorders made up most of the caseload, accounting for around 40% each. Treatment time per patient was highly skewed in distribution, with a small number of patients (in particular, those with personality disorders) receiving very prolonged treatment. Public and private service use was also surveyed in the national health survey (NHS) carried out by the Australian Bureau of Statistics in 1989–1990 [5]. This survey involved interviewing a sample of over 22 000 households involving over 54 000 respondents, for selfreport information about medical consultations and other health care actions in the previous two weeks. The survey found that 0.06% reported attending as an outpatient of a hospital for a mental disorder, 0.16% reported consulting a specialist for a mental disorder and 0.39% consulting a general practitioner (GP) for a mental disorder. People receiving specialist mental health services tended to be younger, better educated and more likely to be in high-income groups compared to those receiving GP mental health services. However, those receiving specialist services were also less likely to be employed.
Another part of the picture is provided by systematic examination of the distribution of the psychiatrist workforce in public and private mental health services in Australia, which has shown ‘maldistribution of the workforce, both by state/territory and by geographical location’ [6]. This report also suggested that waiting times for consumers to see a psychiatrist were in many areas unacceptably long.
The Australian National Survey of Mental Health and Wellbeing (NSMHWB) provides us with a particular opportunity to examine the role played by psychiatrists in the delivery of mental health care in Australia, at a time point towards the end of the span of the first national mental health plan. The NSMHWB enables us to pose questions regarding social equity, and also locate the work of psychiatrists within the context of other service utilization and the meeting of needs for services in the community.
Aims
This report presents analyses of the data set from the exploring of:
– The relationship between age, sex, area of residence, diagnosis, and disability, on the likelihood of consulting a psychiatrist;
– Patterns of specialist and generalist consultation for different disorder groups;
– Inequalities in utilization of psychiatric services, including whether any inequalities found are different in degree or magnitude from inequalities in utilization of medical and surgical specialists;
– The meeting of perceived needs by psychiatrists, and by other service providers.
Method
The household survey [7–9] employed clustered probability sampling, with delivery of a computerized field questionnaire including sections of the composite international diagnostic interview (CIDI) [10], to assess for affective, anxiety and substance misuse disorders according to ICD-10 [11]. The survey also collected data on selfreported service utilization and perceived needs for care [12, 13]. A 78% response rate was achieved and data from 10 641 cases were included in the confidentialized unit record files (CURFs) released by the ABS. Analyses reported here were performed with the SUDAAN statistical package [14] with jackknife replication based on replicate weights provided by the ABS in the 1998 release of the CURF.
Results
Service utilization: descriptive analyses
Proportions of people seen by psychiatrists
One point eight per cent of the population reported consulting a psychiatrist in the last year (males 1.6% (Standard Error [SE] 0.2), females 2.0% [SE 0.3]). This represented 7.3% (0.9) of all people identified by the CIDI with an ICD-10 disorder (males 6.4% 1.[1] females 8.2% 1.[1]). Age specific rates for consultation among those with disorder were: 18–24 years 4.4% (1.1), 25–44 years 7.7% (1.1), 45–64 years 9.9% (2.1), and 65 + years 2.0% (2.1). In the age group 18–24, significantly higher percentages of younger females with disorder than males with disorder reported consulting (7.0% 1.[9] against 2.0% 1.[4]; difference + 5.0% [95]% CI 0.[4–9].6, p < 0.05).
There were no substantial differences in proportions of low or high consulters between the sexes. Of all consulting males 17.5% (3.0) were only seen once, while 15.7% (6.8) of females were seen only once. For those with over 10 consultations reported in the year, the figures here were, for males 27.0% (4.9) and for females 26.6% (6.4).
Area of residence and psychiatric consultation
We explored the effect of living in ‘remote’ areas (population centres size under 10 000), ‘rural centres’ (non-metropolitan areas with population sizes between 10 000 and 100 000) and metropolitan areas. Within the metropolitan areas we examined the areas defined by the five quintiles of the index of relative socioeconomic disadvantage (IRSED) [15] providing for comparison of areas with different socioeconomic characteristics. This seven-fold division (‘remote’, ‘rural centres’ and five IRSED quintiles in the metropolitan areas) is helpful in that it provides for seven roughly equivalent subgroups of the total population, each with between 1.5 million and 2 million inhabitants. The findings are summarized as follows, concentrating on the extreme values. The lowest percentage of people consulting psychiatrists overall is in remote areas (1.0% [0.3]) this represents 5.95% (1.9) of people with ICD-10 disorder identified with the CIDI in those areas. In rural centres this latter rate is 6.3% (1.6). The percentage of people with disorder consulting psychiatrists is lowest in the most disadvantaged areas of the cities (5.4% 1.[5]). Through the fourth, third and second IRSED quintiles, as disadvantage decreases, the rates are 7.2% (1.5), 6.1% (1.4) and 7.0% (2.1). The highest rates of utilization are in the least disadvantaged areas of the cities, both as a percentage of all with disorder (10.8% 2.[6]), and as a percentage of the population seen (2.7% [0.6]). Very few males with disorder in the most disadvantaged areas of the cities were found to be accessing psychiatric services (1.8% 1.[4] against 8.5% 2.[7] for females), whereas in the least disadvantaged areas, utilization appears comparable between gender groups (males 10.2% 2.[7] females 11.4% 3.[1]).
The service providers consulted for broad categories of mental health problems
Table 1 presents some description of the consultation patterns found overall for different disorder groups. Many individuals consulted multiple providers, and the survey instrument provided for 15 categories of provider to be potentially endorsed by the participant. Theoretically this yields a huge number of possible combinations of service provider (2 15 = 32768), some three times the total survey sample. Here this large set of potential combinations of providers is reduced to seven categories, concentrating on combinations including a psychiatrist, with some other common groups for comparison. This table also provides a further perspective on the role of the psychiatrist by examining the patterns of care involved in the response to needs of different diagnostic groups.
Diagnostic group and consultation
Service utilization: inferential analyses
We may consider next whether the area based differences in utilization of the services of psychiatrists are due to the action of confounders including age, sex and disability, and also whether this is an effect specific to this type of specialist service provision. Table 2 presents this analysis. We control through logistic regression for the effect of age and sex, also for disability, using as an index the presence of days out of role as assessed in the brief disability questionnaire (BDQ). We examine the extent to which these variables impact on primary care utilization for mental or physical health problems, to provide for comparison with the situation as it applies to specialist services. The disparity in specialist mental health utilization is confirmed through this analysis, as can be seen in column 2 of the table, with significantly lower utilization of psychiatrists services in more disadvantaged areas of the cities and remote areas compared to less disadvantaged areas of the cities. This analysis does not reveal any corresponding difference in utilization of general practice services for mental health problems, as is apparent in column 1.
Demographics, disability and consultation, odds ratios from simultaneous logistic regression
Presence of physical health problems was assessed through selfreport of having a diagnosed mental physical health disorder as guided by a series of prompts for major groups of complaints. For specialist care in respect of physical health problems there is one significant finding which is that in the middle IRSED quintile there is less utilization than in the highest quintile. However, the most disadvantaged quintile has almost exactly the same likelihood of specialist utilization as the least disadvantaged, so there is no convincing progressive trend of association with this variable. For GP utilization in respect of physical health problems there is again no convincing trend though if anything the suggestion is towards greater utilization in the more disadvantaged areas.
Perceived need and people consulting psychiatrists
Perceived need and the meeting of it for people consulting psychiatrists
As introduced above, the NSMHWB provided for some estimation of the views of consulters regarding their mental health needs and the meeting of them [12, 16]. Table 3 presents a summary of these consumer perceptions as sampled for all people seeing a psychiatrist. This analysis includes people seeing a psychiatrist whether or not they were seeing other providers as well.
Perceived needs and the meeting of them, persons seeing psychiatrists
Almost all consumers seeing a psychiatrist did endorse some need for mental health care, and only very rarely were all the identified needs unmet. The most frequently endorsed needs were counselling, followed by medication, then information, social interventions and skills training. In most categories of perceived need, met need was of greater frequency than unmet need. The exception here was ‘skills training’.
Satisfaction and types of service providers
Satisfaction is assessed with a percentage derived with ‘met need’ (see for instance column 4, Table 3) as the numerator and with the sum of ‘unmet need’ + ‘partially met need’ + ‘met need’ (being all perceived need, as for instance in the sum of columns 2–4 in Table 3) as the denominator.
The comparison is made for all those consulting a psychiatrist, a GP, a psychologist, an ‘other mental health professional’, and an ‘other health professional’. For this analysis we treat each of these provider categories separately, acknowledging that this does not allow for direct attribution of effect to the nominated provider. However, it does lead to the retention of more substantial cell sizes in the relevant subgroups. Hence the categories present ratings of all care received where single or multiple providers provide the care, but in any case including the specific provider identified at the head of the column.
Of all the types of provider examined this table shows that for this estimator of satisfaction, care that includes a psychiatrist singly or in combination with others is consistently rated as the most satisfactory across all need categories. For all perceived needs, where there is psychiatrist involved in care, the proportion of need rated as met was 61.2% (5.0). The performance rated poorest by consumers is the performance of all care packages including a GP in meeting perceived need for skills training, where the percentage is only 15.1% (3.1).
Discussion
Explanatory models for inequalities
How could the area-based disparities in utilization of psychiatric services arise? A useful way of thinking about the problem is in terms of Goldberg and Huxley's model of the pathways to psychiatric care [17]. The first filter involves cases in the community consulting in primary care. In this paper, specifically in the findings in Table 2, we have not found a significant bias in utilization of GP services, so the first filter does not seem to be the major influence operating here. The second filter involves the detection of mental disorder by the GP, and the third involves referral to specialist services. The data examined here do not enable us to separate these filters. However, it does seem to be within the second and third filters that the discretion to refer must be operating in order to explain these findings.
One possible influence on the functioning of the first to the third filters is the severity of disorder, as more severe conditions may be more recognized by the patient and doctor. However, this explanation seems unlikely in this case, since the differences between areas persist when we control for disability.
Another possibility is that patients in some areas are unwilling to accept referral, perhaps because of stigma or because of cost. The possibility of stigma cannot be dismissed using our data. The finding that GPs are no less likely to be consulted for mental disorders suggests no significant bias in help-seeking for mental health problems generally, though there could be a greater effect in relation to specialist consultation. Cost could be a factor because the majority of private psychiatric services involve some cost to the patient for what will often be a recurring service.
Another possibility is that GPs are less likely to refer patients from particular areas. The most immediately obvious reason why this might be the case is that there are few, or no psychiatrists in the area, or they have waiting lists. In this level of the filters, the psychiatrists themselves could be influencing the distribution by locating their practices in less socioeconomically disadvantaged areas, and avoiding remote areas of the country.
It is, of course, well known that the private sector workforce in psychiatry is not equitably distributed in respect of the needs of the community, and the reasons advanced for the well-established geographical maldistribution of psychiatrists in Australia [6] have included both work and personal family considerations [18].
What is evident from this study is that the maldistribution of the private sector workforce is not effectively counteracted in any preferential access to public sector psychiatrists in the areas with less private sector provision.
Types of disorder and types of provider
The great majority of all care involving psychiatrists is de facto shared care. Only about one in five people seen by a psychiatrist report the psychiatrist as the only mental health care provider. Rather the most common combinations are ones including a GP as another practitioner providing care. This is particularly evident for people with affective disorders and with comorbid disorders. The findings emphasize the need for attention to be paid to communication and coordination of care. The recent innovations regarding case conferencing involving both GPs and psychiatrists gives the potential for private practitioners generally to do better in coordinating their work with that of others, and it is to be hoped that communication will be improved with access to these facilitating measures.
Meeting needs
Table 4 shows that people seeing psychiatrists are the most satisfied of all consumers. Before any members of the profession see too much cause for self-congratulation in this, we should remind ourselves of the extent to which care is shared, and the psychiatrist may not be the one meeting the need. A limitation of the perceived need for care questionaire, constrained as it was by time within the survey instrument, is that the meeting of needs is not attributed to specific providers. Rather than being a judgement on the potency of psychiatrists, this finding of increased satisfaction may be largely a dose–response phenomenon in relation to receipt of care. Most people seeing psychiatrists are seeing other providers as well, and probably getting more care than people who do not see psychiatrists. In contrast, GPs see many more people as solitary providers. The others contributing to the care in which the psychiatrist is involved may be the ones delivering the active ingredients. The most useful thing to do here may be to look for areas for improvement. Pointers for this can be found in Table 3. This Table 3 can be read as a report card from the Nation as to how psychiatrists are doing when the patients concerned consider their complete packages of care. Overall most people rate their needs as met, and provision of medication and counselling seem well regarded, with a predominance of met need. There is only one item where unmet need predominates over met need, the item of skill training. This is in line with the findings of the survey overall [16], but this paper shows that even people getting care packages of the relatively intensive nature that tends to go with seeing a psychiatrist, still tend to report this area of need as unmet.
Comparisons with others consulted, meeting of need
Limitations
The use of the NSMHWB provides a perspective on the activities of the psychiatric workforce in Australia. We cannot describe the work profile of psychiatrists in general. A few very high usage patients might heavily influence this. Also we cannot address the work associated with psychotic disorders, which was better estimated in the low prevalence survey [19]. Findings related to the elderly must be treated with caution since the survey excluded residents in institutions. The self-report nature of the data collection is a further limitation.
Despite these limitations, no other data source at present provides us with an alternative methodology for this kind of analysis. What this examination does is set the input of psychiatrists as perceived by patients in the context of community morbidity from anxiety, substance misuse and affective disorders, and in the context of other providers seen.
Conclusions
Australia has a nationally funded insurance scheme, including specialist consultation. Despite this, there are substantial inequities in utilization of psychiatric services which act to the detriment of people living in remote areas and disadvantaged areas of cities. If government, and the profession, desire to correct these inequities, health care reform is indicated, including changes in the incentives for psychiatrists.
Since most care is shared, psychiatrists need to communicate with others who deliver interventions, most frequently GPs.
Most patients report their needs as met. But this is not uniform. Help with skills development is less well met, an observation that underscores the need for collaboration by a range of providers.
Footnotes
Acknowledgements
We acknowledge the Australian Commonwealth Department of Health and Aged Care, Mental Health Branch, who funded and otherwise supported the NSMHW, and the Australian Bureau of Statistics for data collection and preparation. The Commonwealth Department of Health and Aged Care provided funding to support the analyses carried out here through general practice evaluation program grant 661.