Abstract
Rising suicide rates among young people are an increasing source of national concern [1]. While overall rates of suicide have remained fairly constant over the last 70 years, suicide rates in Australians aged from 65 to 75 have dropped markedly while rates among young people have increased. A recent report from the Australian Bureau of Statistics [2] found that in 1998 people in the 25–44 years age group had the highest rate of suicide (23 suicides per 100 000), followed by young people in the 15–24 years age group (17 suicides per 100 000). Among Australians over 65 years of age, the rate was 15 suicides per 100 000 people.
The report also highlighted the dramatic differences in suicide rates between the sexes and between states and territories. For example, suicide rates for men throughout Australia were 23.1 per 100 000 in 1998 while for women, they were 5.6 per 100 000 [2]. In 1998 the Northern Territory had the highest overall state based death rate from suicide of 21 per 100 000 compared to the next highest of 16 per 100 000 in Queensland and South Australia [2]. The Australian Bureau of Statistics data also showed that people living in rural areas had a higher rate of suicide with 17 per 100 000 compared to 13 per 100 000 in capital cities.
It was against this background that, in 1998, an apparent epidemic of suicide occurred over a four month period among young people in the Tiwi islands – Bathurst and Melville islands, situated above Darwin. This cluster of suicides highlighted the urgent need for a more detailed analysis of issues specifically relating to suicide in the Top End of the Northern Territory than had previously been available.
The Top End of the Northern Territory covers approximately two-thirds of the Northern Territory's land mass and contains 79% of its population. A high proportion of the population is rural with 36% of people living in regions outside Darwin. The population characteristics of the Top End follow that of the Northern Territory as a whole. The Northern Territory has a younger population than the rest of Australia, 64% of the Territory's population of 190 000 people being 34 years or younger compared to 51% of the general population of Australia. The Territory's Aboriginal population comprises an important minority of Territorians, 28% of the population as a whole, compared with 2% of all Australians [3, 4]. The Northern Territory Aboriginal population is younger still than the Northern Territory as a whole, with 77% of Aboriginal males and 75% of Aboriginal females being younger than 35 [4–6].
Coroner's records are currently the main source of information about suicide in the Northern Territory. In the Territory, the Coroner does not formally presume that suicide has occurred and the term suicide is not used in any formal context. Coronial findings refer to the deceased having ‘intentionally contributed to his own death’ [7]. In this study we equate such a conclusion with suicide. Suicide is one category of ‘reportable’ deaths that are investigated by Coroner's constables – police attached to the Coroner's court. The majority of suicide inquests are conducted by the Deputy Coroner, a senior public servant with legal qualifications in the Coroner's Department. If there are uncertainties about the manner of death, or there are special circumstances such as a death in custody, the matter is determined through a full coronial inquiry.
A review of Top End individuals whom the Coroner considered had ‘intentionally contributed’ to their own death was undertaken after the Northern Territory Coroner granted access to relevant coronial records. Coroner's files for the Top End are maintained at the Coroner's office in Darwin while files for the remainder of the Northern Territory are kept at Alice Springs. The review of the Top End files was for the eight year period from 1991 to 1998.
Given the unusual population make up of the Northern Territory and concern over the events in the Tiwi Islands, we undertook a descriptive study, but also tested the hypotheses that the means of death of Aboriginal Territorians would differ from those of non-Aboriginal Territorians. We also investigated whether there were any links between individual suicides, or between suicides that occurred in particular locations.
Materials and methods
The study was conducted from 1991 to late 1998 and included all persons in the Top End of the Northern Territory deemed by the Coroner to have committed suicide. The Coroner considered that 187 people had committed suicide during this period. Six files from the end of 1998 were not available for the study because they were still being finalized for coronial determination by the coroner's constables. The study was therefore conducted on 181 deaths by suicide. The study protocol was reviewed and approved by the Aboriginal Ethics sub committee and Joint Institutional Ethics Committee of Royal Darwin Hospital and the Menzies School of Health Research. No attempt was made to review coronial files in which deaths by other means were recorded. Thus no attempt has been made to assess the ‘accuracy’ of the coronial designation of cause of death.
Study method
The study was carried out at the Coroner's Court from June to August 1999. All files were examined by one of us (RP) and the contents coded using the Coding Instrument from the West Australian Coroner's Database on Suicide. This research instrument was developed by the West Australian Youth Suicide Advisory Committee as a source of information on the nature and extent of suicidal behaviour within Western Australia [8].
The West Australian coding instrument, as part of a linkage to the Western Australia Mental Health Information System, contained a further page where data from existing mental health records for the person who committed suicide was incorporated. This item was deleted from the current study due to the impracticality of obtaining the data during the time frame for the study. Instead, two further items were added to the West Australian Instrument. Those items sought to ascertain whether there was any mention in the coroner's record of behaviour suggestive of mental illness in the person who committed suicide, and whether there was any possibility of linkage. These questions were specifically intended to address whether there was different symptomatology preceding suicide in Aboriginal people and to attempt to establish linkage of factors in this population group.
All files for the period were examined and factors for Aboriginal and other suicides compared using data categories contained in the survey instrument. The establishment of the ‘Aboriginality’ of the person who committed suicide is sometimes a sensitive and difficult task for the Coroner's officers. Information is sought from the family, community sources and medical files of the deceased and a decision eventually made on the best available information by the Deputy Coroner. However, it is recognized by the Coroner's office that, given the sensitive nature of such investigations, there may be rare errors in identification.
Formal recognition of mental illness was confined to evidence of treatment for mental illness by a general practitioner or mental health worker. However, statements of presumed mental illness by people close to the deceased were also tabulated. The statements were usually not about mental illness per se but comments on abnormal behaviour that was out of character. These statements might also include an impression about whether the victim was ‘depressed’ using a broad, lay view of the depressive state.
Results
The sample were designated in relation to Aboriginality as ‘Aboriginal’ or ‘Other’. One hundred and forty-nine of the 181 individuals whose deaths were studied were male (28 Aboriginal and 121 Other). Thirty-two of the suicide population were female (7 Aboriginal, 25 Other). The Darwin Statistical District contains 64% of the Top End's population and accounted for 73% of total suicides. However, only 33% (n = 11) of total suicides of Aboriginal people occurred within the Darwin Statistical District.
The mean age of the Aboriginal males who suicided was 27.4 years (range 11–44 years) while the mean age of Aboriginal females who suicided was 24.1 years (range 16–39 years). The mean age of ‘other’ male and female's who suicided was 37.7 years (range 15–73 years) and 43 years (range 22–81 years), respectively.
Suicide methods
Tables 1 and 2 contain detailed information about the methods used to commit suicide in the Top End. The tables also contain the overall figures for Australia as a whole during the same period, taken from a recent publication from the Australian Bureau of Statistics [2].
Comparison of suicide methods for males in the Top End for the period 1991–1998 compared for averaged figures for Australian males for the same period
Comparison of suicide methods for females in the Top End for the period 1991–1998 compared for averaged figures for Australian females for the same period
Hanging was prominent as a suicide method in Top End Aboriginal males and females compared to Australia as a whole. In the Top End Other female population, death by toxic fumes was slightly greater than that for Australia as a whole, while death from poisoning was lower.
There appeared to be a greater number of deaths by firearm in Top End males generally. However, there also appeared to be a substantial reduction in suicide by firearm over time. For the years 1991–1996, firearms represented 29% of suicide methods in the Top End. In 1997–1998 this rate dropped to 14%. It may be of relevance that 1996 saw the introduction of national uniform firearms legislation and the firearms ‘buy back’ scheme.
Information was available about the origin and legal status of a firearm in 29 out of 44 cases. Where information was available, the majority of the weapons 62% used to commit suicide were legally registered and maintained by their owner at home. The remainder were obtained illegally or obtained elsewhere, for example, borrowed from a friend or relative.
Recognition of mental illness
In 80 (or 44%) of the Top End cases studied, the person who committed suicide was identified by their family or friends as being ‘depressed’ prior to their suicide. A chi-squared analysis of the expression of depression between the Aboriginal and ‘other ‘population revealed no significant difference (χ 2 = 1.09, df = 1, p < 0.3). However, when a similar comparison was carried out with the total male and female populations, a significant difference (χ 2 = 9.48, df = 1, p < 0.01) was observed.
Sixty-four (or 80%) of these 80 cases were noted to have received some form of assistance prior to suicide. General practitioners were the main source of help (n = 26), with a lesser number of people consulting psychiatrists (n = 17). It was noted that males were equally likely to consult a psychiatrist (n = 15) or general practitioner (n = 16), whereas more women consulted general practitioners (n = 10) than psychiatrists (n = 2).
There was a formal recognition of mental illness in the deceased from their hospital and attendant medical records or supplementary information in 36 (35 Other, 1 Aboriginal) of the cases. The most common diagnosis among this group was major depression (14 male, 8 female). Six were noted to have schizophrenia. Other diagnoses include anxiety disorders, adjustment disorders, chronic substance abuse disorders and organic illness such as frontal lobe disorder.
In 55 (or 30% of the total cases), it was noted that the deceased had expressed ideas of self-harm prior to their suicide. A comparison of expressed self-harm prior to suicide between Aboriginal and Other populations revealed no significant difference (χ 2 = 0.45, df = 1, p < 0.7). However, with respect to attempted self-harm prior to suicide (34 cases), there was a significant difference between the proportion of Aboriginal and Other populations with Aboriginal people more likely to show this behaviour (Aboriginal, n = 10 Other, n = 24, χ 2 = 2.72, df = 1, p < 0.05).
At a more detailed level, Tables 3 and 4 show the number and nature of the signs of abnormal behaviour observed by family and others in individuals prior to their suicide. Approximately 70 per cent of Aboriginal males and Other females were recognized by others as having one or more features of behaviour which may have been characteristic of mental illness prior to their deaths (see Table 3). These features were less recognizable in Aboriginal females and Other males, approximately 55 per cent of whom showed some sign of abnormal behaviour that may have been the result of mental illness. For both the Aboriginal and Other male populations, the most common signs were behaviour that was out of character, aggression, depressed mood and anxiety or agitation. Depressed mood was the most commonly observed sign in both female populations (see Table 4).
The number of signs of abnormal behaviour that may have represented mental illness in individuals prior to suicide observed in each population group
The signs of abnormal behaviour observed by others prior to the suicide of an individual
All of the Aboriginal females had experienced one or more stresses prior to their suicide. Of the remaining groups, approximately 85 per cent of each group had one or more identified stresses. The most common stresses across all the groups were relationship breakdown, trouble with family and friends and medical illness (which ranged from transient conditions such as an upper respiratory infection and mild diarrhoea to serious conditions such as terminal cancer and HIV). The death of someone close was a stress for Aboriginal males, Aboriginal females and Other males. Financial problems, legal problems and issues related to unemployment or employment were also identified as significant stresses for the Aboriginal and Other males.
Substance abuse
Among the Top End suicides, a greater proportion of males in both the Aboriginal and Other populations had a history of alcohol abuse prior to suicide. Sixty per cent of Aboriginal males and 45% of Other males had this history compared to 29% of Aboriginal females and 28% of Other females. These differences were not statistically significant.
The evidence at autopsy of alcohol intoxication prior to suicide was inconsistent with a history of previous alcohol abuse. For example, only 25 per cent of Other males had evidence of alcohol intoxication at post-mortem, though, as a notable point, 48 per cent of Other females were intoxicated at the time of their death. There was a marked difference in blood alcohol level between the general Aboriginal and Other populations at autopsy. The Aboriginal population (n = 15) had a mean blood alcohol level of 0.211 mg/100 mLs whereas the Other population (n = 54) had a mean blood alcohol level of 0.113 mg/100 mLs.
Linkages
There were a number of apparent linkages with some of the suicides. On two occasions, there was a paired suicide that appeared to involve a suicide pact. Other paired suicides involved a married couple who committed suicide by a similar method within a day of each other and two young male cousins who also committed suicide by a similar method within a short time interval. From the examination of the suicides in the Tiwi region we can conclude that there were some resemblance's between them, but that a more detailed investigation would be required before any links could be established with certainty.
Regional issues in Aboriginal suicides
Despite the fact that there were only a small number of suicides among Aboriginal people in the Top End during the eight years of the study, some important trends were still visible.
Suicide risk in various regions of the Top End was variable. Certain regions such as Port Keats on the Western Coast of the Top End had no recorded suicides for the period despite high rates of alcohol related trauma and death from other causes. Suicides in adjacent regions, such as Daly River were also low. However, there appeared to be a generally higher risk for suicide in Aboriginal people resident in the East Arnhem Region (the area around Gove on the Eastern tip of the Top End) where there were four suicides of Aboriginal people in the period 1991–1994 and a further three suicides in the period 1995–1998. The methods of suicide were either by hanging or firearm. However, the number of suicides in the Tiwi islands present a different picture. A disturbing trend within the figures was a marked increase in suicide among the 2000 Tiwi people resident on Bathurst and Melville Islands during the second half of the study period. From 1991 to 1994, there had been only two suicides in this tribal group. However, this jumped to six suicides between 1995 and 1998. Three of the suicides, which occurred between 1995 and 1998, took place over a four-month period, from August to November 1998. All of the suicides were by hanging. This increase in suicide occurred against a background of widespread self-harm behaviour, in the Tiwi communities, ranging from falling from power poles to self-injury through shooting or attempted hanging [9].
Discussion
Suicide data for the Top End reflect important recent trends in youth suicide and rural suicide in Australia. Thus, recent studies [10, 11] have pointed to the increasing use of hanging by young people in Australia. The suicide methods used by individuals in the Top End are consistent with this national trend.
The use of hanging as a method of suicide in Aboriginal people living in isolated rural communities is of particular concern. The prominence of hanging as a method of suicide in the Aboriginal people of the Top End is striking and is consistent with a similar increase in suicide by hanging in Aboriginal people in Western Australia [8]. Hunter and his colleague's [12] have identified important psychological factors related to hanging in remote Aboriginal communities in Queensland which may be relevant in the Top End context.
However, there are also some local factors that are important in the consideration of suicide in Aboriginal people in the Top End. In connection with the apparently higher risk of suicide for Aboriginal people in East Arnhem, commentators such as Eastwell [13] have noted a predisposition for Aboriginal people in this region to suffer from psychosis and with a higher prevalence of psychosis in these families generally. This may increase the risk of suicide in affected individuals. Cawte [14] also notes suicide ‘fits’ associated to alcohol withdrawal and ‘shame’ suicides which may also be relevant to suicides in this area.
With respect to the suicide rates among the Tiwi, however, there appears to be different contributory factors. It is noteworthy that the increase in suicide occurred roughly 20–30 years after the widespread introduction of alcohol to the Tiwi community. This is a similar picture to that which Hunter (personal communication) obtained in the Kimberleys where a rise in suicide also occurred 20–30 years after the introduction of alcohol. The emotional trauma of young people exposed to alcohol related violence and social disruption in their childhood, combined with individual and community ‘at risk’ factors related to continuing alcohol use in the community elucidated by Hunter et al. [12] may be very relevant to the Tiwi situation. The additional burden of mental illness related to the widespread use of marijuana in the community may be a further contributing factor. The close geographical proximity, contemporaneous nature of suicide and similarity of method in the Tiwi people appears to suggest that links between suicides may be of importance.
The greater rate of suicide by firearm in the Top End may be a reflection of the greater proportion of the population who live in rural areas or the popularity of hunting as a recreational activity among the Top End population. The reduction in suicide rates in the years following national uniform firearm legislation is also worthy of mention. Prior to the legislation and associated firearms ‘buy back’ in May 1996, the Northern Territory Police Department estimates that there were 22 500 licensed shooters and 48 000–50 000 licensed weapons in the Northern Territory. They further estimate that these numbers had fallen to 16 500 licensed shooters and 37 000 licensed weapons in the latter part of 1996 following the ‘buy back’ scheme. The virtual halving of the rate of suicide by firearm from 29% of total suicides prior to the ‘buy back’ to 14% for the period following the ‘buy back’ appears to be associated with the reduction of available firearms and is consistent with other studies [15]. It is necessary, however, to be cautious when examining short-term influences, such as the ‘buy back’, when suicide rates may also be influenced by longer term trends.
The fact that only 64 out of 181 (or 35%) of people who committed suicide in the Top End appeared to have sought help prior to their suicide may be a concern in relation to the potential underrecognition of severe mental illness in the region. However, data from other communities would need to be examined before any conclusions about the Top End could be drawn. The Top End figures support the observation that men appear to be further disadvantaged in this regard, while women tend to be more expressive in their presentation of mental illness and thus more prone to recognition [16]. Furthermore, there is concern that 109 (or 60%) out of the total cases may have given some sign of mental illness prior to their suicide without the benefit of effective intervention. However, often this distress occurred within a short time period prior to the suicide, reducing the window of opportunity to provide help to the individual. It is hoped that attempts to improve recognized deficits in mental health literacy in Australia [17], such as the recent introduction of the ‘Mind Matters’ [18, 19] curriculum into high schools in Australia, will increase recognition of mental illness and thereby increase the potential for intervention.
The rate of alcohol abuse prior to suicide in the ‘Other males’ population is similar to data for this population from Queensland [20] and overseas [21]. The rate of alcohol abuse in the Aboriginal male population prior to suicide is consistent with data for substance abuse and Aboriginal suicide in North Queensland and the Kimberleys [22, 23]. The higher mean blood alcohol levels in the Aboriginal population probably reflect recognized issues with respect to Aboriginal alcohol abuse where there is less frequent but heavier drinking [24, 25]. There are a variety of explanations to explain inconsistencies between a history of alcohol abuse and the evidence of blood alcohol at post-mortem. Results of alcohol assays from autopsy were not always available. In addition, there was also evidence that some of the ‘Other females’ with no prior history of alcohol abuse had used alcohol to aid the process of their suicide. The inconsistency in investigation of substance abuse may also account for the limited number of other drugs recorded in suicide investigations. Cannabis was rarely mentioned in the police reports or coronial determinations. It was detected at autopsy in only six cases. Opiates and amphetamines were detected in only nine cases. Factors that appeared to limit the collection of this data were: a consideration by the investigating police that cannabis and other substance abuse was not a significant factor in suicide, and also the cost to the coroner's office of completing wide ranging drug screens where drug use was not thought to be important in an investigation.
This study has a number of limitations. Due to the small population and limited number of suicides in the Top End, it is difficult to generalize the findings to populations outside the area. The comparative youth of the Top End population also limits the generalizability of findings. The issues surrounding the identification of ‘Aboriginality’ following a suicide make the establishment of absolute numbers in each group open to question. The limited number of Aboriginal female victims in the study means that statistical evaluation of any data from this group is of limited value. Resource limitations meant that data extraction was performed by one of us only, and no formal analysis of test-retest or interrater reliability of the data extraction was therefore possible. Nevertheless, the timeliness and comprehensive nature of the study may make it of wider interest.
In summary, we have extracted information from the available Coronial records of all individuals who committed suicide in the Top End of the Northern Territory during the years 1991–1998. We found important differences in the methods of suicide used by Aboriginal residents who had committed suicide over that period, and high levels of identifiable but untreated mental illness and substance abuse among many of the individuals who had taken their own lives. There were marked regional differences in the numbers of suicides among the Aboriginal and Torres Strait communities of the Top End, and some indication of linked suicides in a number of instances.