CAOS Contribution to Understanding Cultural/Ethnic Differences in the Prevalence of Bipolar Affective Disorder in New Zealand

Abstract

Objective: To examine whether data on the clinical profiles of psychiatric service users can inform the claim that Maori in the New Zealand community have an increased rate of bipolar disorder.

Method: The standardized recordings of a variety of clinical phenomena previously collected in a New Zealand study of psychiatric service users was extracted for those persons who were diagnosed with a bipolar mental disorder. The individual clinical phenomena were then compared by ethnicity.

Results: The most dramatic result was the increased recording of high levels of overactivity or disruptive aggressive behaviour in Maori compared with European in psychiatric service users, despite which, the rates of the disorder did not significantly differ by ethnicity.

Conclusion: In the community survey lay raters were deciding on the presence or absence of phenomena including hyperactivity without the contextual knowledge that clinicians use. It is possible that the Composite International Diagnostic Interview decision tree was being fed an excessive amount of that criterion.

One of the most interesting intercultural findings of the recently published Te Rau Hinengaro: the New Zealand Mental Health Survey was the apparent high rate of bipolar disorder in Maori and Pacific people [1,2]. This information was based on a survey of the general population using the Composite International Diagnostic Interview (CIDI) with hierarchical diagnosis allocating rules except for substance abuse disorder [3]. That study specifically excluded those who were receiving hospital-based or any other institutional care at the time.

Apart from substance use/abuse-related disorders, the single diagnostic category that had the greatest cultural variation was bipolar disorder, for which the apparent rate for Maori was twice that for Pakeha.

The other large New Zealand mental health project funded by the National Mental Health Research and Development Strategy in recent years was the Casemix Outcome Study (CAOS) [4, 5]. That project, based on current service users, included particularly those with episodes of hospitalization during a 6 month period in 2002. The contribution of the CAOS project and other refereed publications reflecting on Maori psychiatric health issues have recently been comprehensively reviewed [6]. The interpretation of this new, community prevalence, information from Te Rau Hinengaro may be able to be informed by more detailed analysis of some of the data collected on Maori and Pakeha during the CAOS study.

In particular, comparison by culture/ethnicity of those treated for bipolar disorder in the inpatient services during 2002 may give information to inform the debate necessary to understand the findings of Te Rau Hinengaro with regard to the apparent community prevalence of bipolar disorder. By comparing, between the different cultural/ethnic groups, more detailed, standardized, descriptions of their clinical phenomena during treated illness episodes we may better understand those apparent community prevalence findings. Particularly to elucidate whether the results do indeed represent a reliable and valid difference in the rates in different ethnic groups or reflect a lack of cultural sophistication in assigning diagnosis by the CIDI, or an even more conceptually global inadequacy in the definitions of the phenomenon underpinning psychiatric diagnoses in the context of different cultures [3, 6].

Our aim in this brief paper is to compare aspects of the clinical profiles of Maori and Pakeha presenting for care and diagnosed with bipolar disorder. The Health of Nation Outcome Scales (HoNOS) [7, 8] ratings provide data on 12 different items, the first nine of which directly reflect on the psychopathology profile.

Method

In the CAOS project, information was collected on the costs of episodes of care, service user characteristics and aspects of services provided. A total of 19 239 episodes of care were studied. These were for both inpatients and community patients.

Incidence and prevalence figures cannot be extrapolated from these findings and applied to the general population because the participating services were simply those that volunteered, representing approximately one-third of those in New Zealand.

Among the patient characteristics reliably recorded was the profile of clinical psychopathology as reflected in the HoNOS [7, 8]. These HoNOS data give a reliable recording of the observations made on 12 particular items, for 96% of the service users who had an episode of care during that 6 month period.

HoNOS profiles are compared between two self-identified cultural groups: Maori and European (‘Pakeha’). The numbers of Pacific people in the CAOS data were too small to consider.

We looked at two categories: bipolar/manic and bipolar/depressed episodes. There were 407 episodes for the first and 3954 for the second.

Results

Data for the HoNOS items dealing with clinical psychiatric phenomena of the diagnosed manic patients are presented in Table 1, for the two ethnic groups. The mean ratings for self-harm and cognition in both ethnic groups do not reach clinically significant levels given that they are >1. Item 1, which summarizes behavioural features concerned particularly with overactivity, shows a clear difference between European and Maori: the latter scoring more highly. Similar direction changes were apparent for the use of alcohol and drugs and to a lesser extent the presence of hallucinations or delusions.

Table 1.

Selected HoNOS items for mania

HoNOS items	Maori Mean rating (95% CI)	European Mean rating (95% CI)	p
Overactive/disruptive etc.	2.21 (1.90–2.52)	1.52 (1.33–1.72)	<0.001
Self-harm	0.61 (0.34–0.88)	0.50 (0.35–0.66)	0.48 (n.s.)
Alcohol/drug overuse	1.75 (1.37–2.13)	0.71 (0.53–0.90)	<0.001
Cognition	0.90 (0.66–1.14)	1.07 (0.91–1.22)	0.26 (n.s.)
Hallucinations/delusions	1.81 (1.39–2.23)	1.06 (0.86–1.26)	0.002
Depression	0.82 (0.50–1.14)	0.79 (0.64–0.95)	0.87 (n.s.)
Other mental/behavioural problem	1.67 (1.23–2.11)	1.62 (1.35–1.90)	0.86 (n.s.)

CI, confidence interval; HoNOS, Health of Nation Outcome Scales.

The ethnic differences for item 1 (overactivity) in either bipolar phases are illustrated in Figure 1. The same higher ratings for overactivity for Maori is apparent in patients presenting in a depressive phase of bipolar disorder (see Table 2). On average, Maori patients were rated as significantly less depressed. Ratings of cognition symptomatology and self-harm did not differ significantly between the two ethnic/cultural groups.

Figure 1.

Mean Health of Nation Outcome Scales ratings for ‘overactive/disruptive’ in the two bipolar diagnostic groups.

Table 2.

Selected HoNOS items for bipolar depression

HoNOS items	Maori Mean rating (95% CI)	European Mean rating (95% CI)	p
Overactive/disruptive	1.17 (1.03–1.30)	0.71 (0.67–0.76)	<0.001
Self-harm	0.85 (0.73–0.97)	0.87 (0.82–0.92)	0.82 (n.s.)
Alcohol/drug overuse	1.03 (0.88–1.19)	0.52 (0.48–0.57)	<0.001
Cognition	0.85 (0.72–0.97)	0.66 (0.62–0.70)	0.06 (n.s.)
Hallucinations/delusions	0.65 (0.52–0.78)	0.34 (0.30–0.38)	<0.001
Depression	1.43 (1.31–1.55)	1.86 (1.81–1.91)	<0.001
Other mental/behavioural problem	1.28 (1.13–1.43)	1.58 (1.52–1.64)	<0.001

CI, confidence interval; HoNOS, Health of Nation Outcome Scales.

Discussion

Concerns about the mental health of Maori and also Pacific people have been highlighted in many publications [1, 2, 5, 6, 9]. Analysis of the full CAOS data has demonstrated that Maori have higher scores on psychotic symptoms and lower scores on depressive symptoms than non-Maori male service users [4]. It also showed that among those living in areas of a similar level of socioeconomic deprivation, Maori service users had on average higher levels of severity and lower levels of functioning than non-Maori service users. Other investigations have shown that Maori attending primary health care had higher rates of common mental disorders (anxiety, depression and substance abuse) than non-Maori and exhibited more severe symptoms [9].

The findings of the more recent study Te Rau Hinengaro: the New Zealand Mental Health Survey[1, 2, 10] not only provide further information about the prevalence of mental illness in New Zealand but also seemed to confirm previous findings of higher rates of mental illness and severity among Maori and Pacific people, similar to the elevated rates in the UK ethnic minority groups [1, 2, 10, 11]. Those results, however, have attracted some debate and criticism [3, 12]. In particular, Jorm was concerned about diagnostic thresholds and proliferation of diagnoses in relation to the use of CIDI and the current categorical as compared with the dimensional diagnostic system [3]. Hickling et al. demonstrated that comparing cross-cultural and iso-cultural raters resulted in apparently similar prevalence rates but that the diagnostic groups were contributed to by different individuals [13].

Item 1 on the HoNOS examines problems with overactive, disruptive, agitated, or aggressive behaviour. Rating rules stipulate that all types of behaviour must be rated, regardless of intention, insight or awareness, and context must be considered. The rating rules are clear in terms of how specific behaviours should be rated but without considering possible causes of such behaviours.

The striking differences for item 1 (overactivity) in both bipolar phases are of interest. Maori are not rated to be as depressed as are European patients. But, as with the UK-based work on Afro-Caribbean patients, does that apparent reliability indicator hide a real difference in the chosen groups?

The findings on alcohol and drug usage are in the same directions for the service user (patient) group as in the community-based survey. The higher ratings for overactivity in the current patients could be seen as consistent with work on the use of neuroleptics in these New Zealand groups [14], and the rate of use of seclusion in a New Zealand inpatient unit [15]. This has also been noted for specific phenomena, such as hallucinations [16].

Those patients who contributed data to the CAOS study when in the depressed phase of a bipolar disorder show the same intercultural variation as for the pure manic patients, in particular, higher rates of overactivity and substance use/abuse and less display (or recognition) of depression.

Whether these are cultural variations in how depression or dysphoria are experienced or displayed lacks currently available proof, but must at least support the view of Jorm [3], and challenge the cross-cultural robustness of a nomothetic instrument such as the CIDI.

The CAOS data did not find any difference in the rates of diagnosis of bipolar disorder when examined by ethnicity/culture. Yet it is clear that the clinical profile of the Maori and Pacific groups who received that diagnosis did differ. It would be of interest to examine the data utilized in the CIDI decision-tree in reaching the prevalence conclusions of the New Zealand epidemiological study [1, 2]. The ethnic differences in prevalence rates raise the possibility of cultural differences in the experience and reporting of key symptomatology and are in accordance with previous reports of cultural variation in hallucinatory experiences and reported prevalence of schizophrenia [16–18]. It may be seriously postulated that in the presence of prominent overactivity behaviour the threshold for the diagnosis of bipolar disorder, particularly mania, is more readily reached. Similarly for agitated depression. It is also possible that there are some cultural differences in the experience and reporting of hyperactivity. This is relevant to surveys using CIDI in which lay interviewers rather than psychiatrists use information to generate diagnoses. In the CAOS data ICD-10 diagnoses were made by psychiatrists who would most of the time take contextual issues into account when making such diagnoses. This is reflected in the finding that despite differences in the clinical profile of the Maori and Pacific groups there was no difference in the rates of bipolar. We argue that lay interviewers, for example, may be impaired by the lack of access to or knowledge (sometimes familiarity) of contextual issues when making judgements on clinical phenomena when using CIDI. It is also possible that lay interviewers’ interpretation of hyperactivity is distorted by preconceived assumptions about Maori patients as being ‘overactive, over-talkative, loud’, leading to the reported inflated rates of bipolar disorder in this group.

Conclusions

Clinical profiles indicate that Maori patients in both categories are considered/perceived to be overactive. This is consistent with previous findings regarding use of seclusion and neuroleptic medications. It is possible that lay interviewers are picking up ethnically or culturally based activities, which allows the CIDI decision tree to generate the diagnoses of bipolar disorder. We believe that CIDI will not replace the experience and wisdom of psychiatrists who would normally consider the bigger picture when making such diagnoses. It is reasonable to conclude that there is a need to develop a culturally sensitive and robust assessment tool that can take into account cultural variation in symptom reporting, and the use of phenomenological constructs as would psychiatrists. When the distribution of diagnoses in mental health service users is less culturally idiosyncratic than in community research, we need to be confident we are comparing like with like, before accepting the findings as knowledge.