Abstract
The mental health experience of Maori has been a source of concern for some years. Durie [1] has described how the undermining of traditional supports of health in Maori society has lead to poorer health in general, and mental ill health in particular. Bridgman and Dyall [2] found that while Maori were not noted to have higher rates for first admission for psychotic disorder, they did experience a higher rate of readmission and lower rates of admission for affective disorder. Later presentation for mental health care raised concerns regarding the appropriateness of the services to meet the needs of Maori [3]. In Otago, Edmonds et al. [4] a cohort of Maori and non-Maori patients admitted to the inpatient psychiatric services and found the Maori admission rate to be four times higher than non-Maori. Rates of family psychiatric history did not differ although Maori were found to have more social disadvantages. The sources of referral for Maori admissions were more likely to be from the law. The most common diagnosis did not differ between the groups.
The Census of New Zealand Prisoners, 1999 [5] found that approximately half are Maori, about 10% of Pacific Island ethnicity and the remainder largely Pakeha (European New Zealanders). This is an over-representation by the former two groups, with Maori being 14.1% of the general population and Pacific peoples 6.2% [6]. The Maori and Pacific Island communities are much younger in their age structure [6], contributing to this over-representation, but much more predominantly so for Maori [7].
Maori are over-represented among prisoners who commit suicide, both in relation to other prisoners and in relation to the non-imprisoned populations [2, 8–12] and suicide is more common among younger Maori than in non-imprisoned populations [3, 13].
These findings suggest there is likely to be a major issue of psychiatric morbidity and unmet need for assessment and treatment for Maori prisoners. In our study of psychiatric morbidity in New Zealand prisoners, we detected significant morbidity among all prisoners, a considerable proportion of which was unrecognized [14, 15]. We present further analysis on this data to address whether the rates of mental disorder, suicidality or treatment experience differ between the ethnic groups.
Method
The methodology for this study is fully described elsewhere [14, 15]. We approached all female prisoners, all male remand prisoners and an 18% random sample of sentenced male prisoners from every prison in New Zealand. For sentenced male prisoners, selection of names was performed on a rolling muster through the alphabet. Interviewing occurred from October – December 1997, and February – June 1998.
Assessment tools
The interview process was comprised of five sections: informed consent; demographic data; diagnostic interview for mental illness (CIDI-A); and a screening diagnostic interview for relevant personality disorders (PDQ-4 +). The presence of specific information about suicidal thoughts or actions was sought by adapting the questions regarding suicidal thoughts and actions in the CIDI-A, and asking specific questions regarding these symptoms since entry to prison. Demographic data included a variety of variables including information regarding past psychiatric history, past psychiatric treatment and history of previous psychiatric hospitalization. Ethnicity was self-identified according to New Zealand Census categories.
The CIDI–Auto 2.1 data were scored for DSM-IV diagnosis using the WHO SPSS scoring algorithms (Release 2.1, May, 1998). The version of the CIDI-A was chosen which comprised sections examining: anxiety disorders; depression; mania; psychosis; eating disorders; alcohol; drugs; obsessive-compulsive disorder; post-traumatic stress disorder; and the use of the Mini-Mental State Examination. These sections were chosen as they identified the most common and clinically significant disorders in the prison setting.
The PDQ-4 + was selected and computerized to screen for relevant personality disorders. Several studies attest to the high sensitivity and only modest specificity of the PDQ. The PDQ questionnaire seems particularly over-inclusive, with high rates of false positives [16–19]. We screened for the five most relevant personality disorders from DSM-IV, namely borderline, antisocial, narcissistic, histrionic and paranoid personality disorders.
Statistical analysis was performed on the Statistics Package for the Social Sciences (SPSS).
Procedure
The interviews were performed by four regional teams. Each team had a team leader, an experienced clinician, who supervised 8–14 interviewers and their data collections from 220 to 350 prisoners. Interviewers had a range of skills from those with mental health knowledge such as clinical psychology students and psychiatric nurses, to lay people with interviewing skills. One team were Maori. All were instructed to raise any concerns about the wellbeing of any prisoner they interviewed with the team leader or the clinical co-ordinator of the team. Referral to the regional forensic psychiatric services was offered to prisoners where appropriate. If the prisoner could not speak English, an interpreter was arranged. This was used four times, on each occasion by older men of Pacific Island ethnicity. Kaumatua and Kuia (male and female Maori Elders) worked with the research teams and were available to be with subjects if needed.
Interviews took approximately 1 h and 15 min (range 35 min to over 3 h). As the PDQ4+ interview was performed last; the completion rate was lower than for the CIDI-A. If an interview was interrupted, the CIDI-A was closed and was restarted from there if it could be recommenced.
Acceptance and completion rates
Eligible for inclusion were 200 remand and sentenced women prisoners, 540 remand male prisoners and 4447 sentenced male prisoners. All female prisoners were approached. One hundred and seventy (or 85%) consented to interview, 162 completed the CIDI-A (81%) and 158 completed the PDQ4 + (79%). All 540 remand male prisoners were approached. Four hundred and fifty-two (or 83.7%) consented to interview, 441 completed the CIDI-A (81.7%) and 405 completed the PDQ4 + (75%). Of the 4447 sentenced male prisoners, 18% (798) were randomly selected for interview, having been stratified by security rating. Of these, 660 (82.7%) consented to interview, 645 completed the CIDI-A (80.2%) and 592 completed the PDQ4 + (74%).
Results
The self-identified ethnicity of prisoners is presented in Table 1. Maori are similarly over-represented in each component of the prison population, being 49.8% of the remand male, 46.2% of sentenced male and 52.4% of female prisoners. Pacific peoples show a somewhat different pattern. Of remand men, 10.2% identified as Pacific Island origin, 8.5% of sentenced men but only 2.9% of female prisoners.
Self-identified ethnic affiliation of New Zealand prisoners
Maori prisoners are significantly younger than others. Their mean age is 27.99 years (SD 8.84), compared to 31.33 (SD 11.54) for Pakeha/others (t-test p < 0.000). There is no difference in the gender ratio between Maori and non-Maori, although there are proportionately fewer Pacific Island women (Wilcoxon two-tailed test, p < 0.000).
Maori and Pacific Island prisoners are significantly less well educated than other prisoners (Wilcoxon two-tailed test, p < 0.000). Fewer than 20% have completed secondary education and less than 5% tertiary study of any sort, approximately the same as in the non-prison Maori population [6]. Prior to entry to prison, Maori and Pakeha prisoners reported very similar living circumstances (primarily living with a partner, flatting or living alone) while the most common living situation of Pacific Island prisoners was with their parents. All prisoner groups showed a similar distribution of offences, except that Pakeha showed more property and drug offences.
Lifetime prevalence of mental disorder by ethnicity is presented in Table 2. All the differences between ethnic groups in Table 2 were assessed for statistical significance using Pearson χ 2-test. No significant differences in lifetime rates of disorder were found by ethnic group.
Lifetime prevalence of mental disorders by ethnicity
Current prevalence of mental disorder (within the last month) by self-identified ethnic group is presented in Table 3. Major depression was more common among Pakeha/other, than among Maori or Pacific Island people, the only significant difference between the groups.
Current prevalence of mental disorders by ethnicity
We asked five questions regarding self-harming ideation during this current period of imprisonment. We defined the suicidal as those prisoners who reported to ‘have thought a lot about suicide’ at some time during the current period of imprisonment. Their ethnic profile is presented in Table 4. Maori reported significantly lower rates of thoughts of suicide (χ 2-test, p = 0.007, two-tail) than other ethnic groups.
Suicidal thoughts by ethnic group
Given that Maori are more frequently represented among those who suicide in prison than expected by their proportion in the prison population [9], this finding of a lower mean number of suicidal thoughts requires further consideration. Of the 30 prisoners who committed acts of self-harm in prison, 17 (56.7%) were Maori, 13 (43.3%) Pakeha/other and none of Pacific Island ethnicity. This difference was not statistically significant from the total prisoner population. (Mann–Whitney Test, NS).
As the mean age of Maori prisoners is significantly younger than for non-Maori, we separated the younger prisoners, defined as those less than 20 years of age and compared them with older prisoners. Prisoners younger than 18 years are also of interest as they overlap with youth services in terms of jurisdiction.
There were 40 prisoners less than 18 years. By ethnicity, 25 (or 62.5%) were Maori, eight were of Pacific Island ethnicity (20%) and seven were ‘other’ (17.5%). The rate of young Maori and Pacific Island offenders is greater than in the total sample. The size of these groups was too small to detect differences in rates of mental disorder in relation to the general prison population. When analyzed as a total group (that is prisoners less than 18 years relative to prisoners 18 years and over), no significant differences were detected in the rates of disorder.
There were 173 prisoners less than 20 years of age. By ethnicity, 88 (or 50.9%) were Maori, 29 (16.8%) were of Pacific Island ethnicity and 56 (32.4%) Pakeha/other. These proportions are similar to the total population, unlike the under 18 group, with a trend only towards a higher representation of Pacific Island prisoners. We compared the prevalence of current mental disorders and lifetime diagnosis of substance misuse of this group by ethnicity. The only disorder that was significantly more frequent in one ethnic group was a lifetime diagnosis of substance abuse or dependence which was more common in Pakeha/other than in Maori or Pacific Island young prisoners (Pearson χ 2-test = 10.076, p = 0.006, two-tailed). There were no other differences in the prevalence of current mental disorders or lifetime substance misuse between the various ethnic groups of prisoners aged less than 20 years.
Treatment experience
All prisoners were asked whether they had received treatment for mental health problems since being in prison. Self report of treatment received in prison is listed in Table 5.
Self report of treatment prior to and since imprisonment, by ethnicity of prisoners
It appears that despite experiencing the same prevalence of current and lifetime mental disorders, Maori and Pacific Island prisoners receive fewer services both prior to and during imprisonment. This is reflected both in their rates of prescription of psychiatric medication and in their report of being seen by a psychiatrist or psychologist. Notably, their rates of contact with mental health services in the community prior to entry to prison are similarly lower than the Pakeha/ other rates, both for attendance at a community mental health facility or inpatient care.
Discussion
The research procedure was an adaptation of a WHO study methodology that has not been specifically validated within New Zealand, nor specifically culturally evaluated. Such validation was beyond the scope of this study. The CIDI has however, been used in many countries and appears cross-culturally robust. Therefore, although the design lacked cultural specificity and is at risk of committing cross-cultural error [7], its general acceptability and robust international design counteracts some of these difficulties. Self-identification of ethnicity is the most widely used means for ethnic analysis in New Zealand studies, but may still be associated with underreporting of Maori ethnicity [20]. Some prisoners preferred to refer to themselves as ‘Kiwi’ rather than reply as Maori or European.
Response rates for Maori cannot be specifically derived as refusers did not have their ethnicity recorded. The ethnic breakdown of the research subjects was no different to the prison census data [5] suggesting that there was no ethnicity bias among those refusing to participate. Anecdotally, we did not note greater resistance from Maori. In one region we used only Maori interview staff with similar response rates to the other teams.
As noted in our previous paper [15], mental disorder is common among prisoners, but we lack appropriate population norms against which to compare the data. The only major study of the prevalence of psychiatric disorders in the community in New Zealand [21, 22] was performed in Canterbury and had too few Maori or Pacific Island subjects to allow separate comparison. Male and female ethnicity over-representation in prison was similar for Maori, but why Pacific Island women should be under-represented is unclear. This might suggest protective factors of extended family and church affiliation acting for them, but not for Pacific Island men. There is little overall difference between the ethnic groups in the prevalence of mental disorders, either lifetime or current. Indeed only one disorder is different between the groups on current prevalence and that (major depression) is more common in non-Maori groups.
The rate of suicidal thoughts is also lower among Maori but their tendency to act upon such ideas is equivalent to non-Maori, resulting in the same rate of actual suicidal behaviour. This may be because Maori prisoners are more reluctant to admit to suicidal feelings. Alternatively, when such ideas are present they may carry greater lethality because rate of successful prisoner suicide has been found to be similar to non-Maori in the time period 1973–1988 [9], but elevated at 65% of all prison suicides in the time period 1980–1995 [12]. Further, it may be that the language of the questions or the manner in which they were asked did not access psychological attitudes relevant to suicide, namely shame (whakamaa) and humiliation (whakamomori) [7].
Despite having equal need for treatment Maori and Pacific Island prisoners report receiving fewer services within prison and prior to coming to prison. Maori and Pacific Island prisoners have fewer prior admissions, fewer community outpatient appointments and less treatment in prison for the same degree of psychiatric morbidity. This regrettably confirms other findings of later engagement with services and greater likelihood of Maori presentations being compulsory and more severe [2, 4]. Service responses therefore need to cater for half the population being Maori, and must have screening processes that adequately detect the presence of disorder and offer acceptable and effective care and treatment for the detected disorders. Intervention must be able to overcome the under-provision of services to Maori that persists from the community to prison treatment experience, and appears to persist despite attempts of forensic mental health services to implement culturally appropriate services to inmates in prison [23]. Overcoming the health inequities for Maori has been frustrated by poor integration of policy, purchasing and provision of mental health services [24], deficiencies that need to be addressed within the provision of mental health services to inmates.
The problems identified in the study demand treatment responses from the health sector for major mental disorders. Substance abuse and dependence, important in relation to reoffending and as a contributing factor for other mental disorders, is currently the responsibility of the Department of Corrections. Any service response must involve both screening and intervention services that are acceptable and effective for those needing care. Pathological gambling has a lifetime prevalence of approximately 33% of male and 45% of female inmates in New Zealand, with a high comorbidity with substance misuse among Maori gamblers [25, 26]. These needs must be considered alongside treatment interventions for substance misuse.
Why are mental disorders not more common among Maori and Pacific Island groups than among Pakeha/ other prisoners? Offenders commonly have a pattern of sociodemographic vulnerability present in their developmental history (poverty, marginalization, poor educational achievement, broken homes, large family size, male gender, history of abuse). Such factors also are commonly associated with mental disorders. If inmates are vulnerable to developing mental disorders, such stressors appear to be giving rise to mental disorder in this study at the same rate, regardless of ethnicity, but at an increased rate relative to the general population. Such a finding is noted internationally [27, 28]. The problem is that these sociodemographic factors are much more common in the lives of Maori and Pacific Island people, resulting in their over-representation in prison, and similarly elevated rate of mental disorder. Once in prison, the ethnic groups manifest very similar rates of mental disorder.
These findings do not mean the same treatment response is necessarily appropriate for the same disorder in all people. While there is not a particular sensitivity to mental disorder among any particular ethnic group, service responses must aim at both increasing the detection of all mental disorders, and particularly asking if the service response is meeting the needs of Maori, given they are the largest ethnic group present in NZ prisons.
Footnotes
Acknowledgements
We thank Lorna Dyall and Joanne Baxter. The Study was funded by the New Zealand Department of Corrections, and Ministries of Health and Justice.