Australian Institute of Criminology

Skip to content

Deaths in custody in Australia 1990-2004

Trends & issues in crime and criminal justice no. 309

Jacqueline Joudo
ISBN 1 921185 00 7 ; ISSN 0817-8542
Canberra: Australian Institute of Criminology, June 2006

Foreword | The National Deaths in Custody Program contains a dataset spanning 25 years with details of every death occurring in prison custody, police custody and custody-related operations, and in juvenile detention facilities across Australia. This report summarises the trends in these deaths between 1990 and 2004. Compared with the period 1980-1989 in which the majority of deaths in custody occurred in police custody, prison custody deaths have accounted for the overall majority of deaths since 1990. Deaths in police custody have generally decreased during this time while the number of deaths in police operations has increased. An increase in the number of deaths of non-Indigenous persons accounted for the bulk of the increase in the number of deaths in police operations. The NDICP is a valuable program which enables continued monitoring of all deaths in custody across Australia and identification of changing trends.

Toni Makkai
Director

Background

The National Deaths in Custody Program (NDICP) was established at the Australian Institute of Criminology (AIC) in 1992 in response to a recommendation of the Royal Commission into Aboriginal Deaths in Custody (RCIADIC) that an ongoing program monitor all deaths in custody in Australia. The RCIADIC outlined the types of deaths that would require notification to the NDICP as deaths of persons in prison custody, police custody or juvenile detention; deaths due to lack of proper care, or injuries sustained whilst in such custody; deaths of persons who die or are fatally injured in the process of being detained by police or prison officers; and deaths of persons attempting to escape custody (RCIADIC, 1991: 190).

The NDICP gathers information on these types of deaths from all state and territory police departments and correctional services with a dataset dating back to 1980. Williams (2001) examined Australian deaths in custody between 1980 and 1989 and compared them with those between 1990 and 1999. This paper examines the general trends in deaths in custody over the 15 year period following the RCIADIC, from 1990 to 2004. The main findings are presented in three sections: deaths in prison custody; deaths in police custody (Category 1) and deaths in police custody-related operations (Category 2). Further statistical information on the findings presented here can be found in the NDICP annual report (Joudo & Veld 2005).

It is widely recognised that Indigenous people are over-represented across the criminal justice system and this trend is also present in the deaths in custody statistics. Various underlying issues have been raised in recent government reports as contributing to Indigenous disadvantage generally, and in increasing the likelihood of contact with the criminal justice system. Behaviours such as substance and solvent abuse and harmful alcohol use are among the health issues that have been identified as contributing to poorer health and shorter life expectancies among Indigenous persons than the wider Australian population (SCRGSP 2005). It has been suggested that these behaviours may cause long-term health problems and increase the need and opportunity for some Indigenous persons to commit offences due to 'enforced idleness and poverty' (Hunter 2001: 3). In an AIC study of the impact of substance abuse on offending by Indigenous males it was found that 69 percent of Indigenous adult male prisoners had used alcohol at the time of arrest or commission of the offence, whereas only 27 percent of non-Indigenous prisoners reported alcohol use at either arrest or commission of offence (Putt, Payne & Milner 2005). Indigenous prisoners were also more likely than non-Indigenous prisoners to attribute their most recent serious offence to drugs.

Prison custody, 1990–2004

The deaths reported in this section refer only to deaths in adult correctional facilities, including those occurring during transfers and in medical facilities (see Box 1). Nine deaths have been recorded in the custody of juvenile justice or welfare authorities since 1990, none since 2000. Due to this small number, deaths in juvenile justice/welfare custody have been excluded from analysis.

Box 1: Definition of deaths in prison custody

Deaths in prison custody include those deaths that occur in prison or juvenile detention facilities. This also includes the deaths that occur during transfer to or from prison or juvenile detention centres, or in medical facilities following transfer from adult and juvenile detention centres (RCIADIC 1991: 189-190)

During the decade investigated by the RCIADIC (1980-1989) the majority (61%) of deaths of Indigenous persons occurred in police custody (Williams 2001), while in the following 15 years the majority of deaths of Indigenous persons have occurred in prison custody (63%). Of the 772 prison deaths since 1990, almost 19 percent (n=145) were of Indigenous prisoners (Table 1). While the number of deaths of non-Indigenous prisoners has consistently exceeded deaths of Indigenous prisoners, the rate of Indigenous prisoner deaths exceeded the rate of non-Indigenous prisoner deaths in just over half of the 15 years since RCIADIC (Figure 1), reflecting the general over-representation of Indigenous persons in the prison population. In fact, the most recent prison census data reveal that the rate of imprisonment of Indigenous persons was 16 times higher than the non-Indigenous imprisonment rate (ABS 2005). Overall, the rates of death between 1990 and 1999 for both Indigenous and non-Indigenous prisoners fluctuated between two and six deaths per 1000 prisoners, with the exception of 1992, and have become more similar since 1999, at which time they both began to trend downward (Figure 1).

Table 1 : Deaths in prison custody by Indigenous status, 1990-2004
Prison
Indigenous Non-Indigenous Total persons
n % n % N %
Note: Some totals will not sum to column total due to missing data. Column percentages may not sum to 100 due to rounding errors and missing data.
Source: AIC, NDICP 1990-2004 [computer file]
Sex
Male 137 94.5 602 96.0 739 95.7
Female 8 5.5 25 4.0 33 4.3
Age
47 32.4 118 18.8 165 21.4
25-39 years 67 46.2 255 40.7 322 41.7
40-54 years 26 17.9 138 22.0 164 21.2
55 years + 5 3.4 116 18.5 121 15.7
Mean age (years) 32.3 39.0 37.7
Cause of death
Hanging 64 44.1 264 42.1 328 42.5
Natural causes 66 45.5 208 33.2 274 35.5
Head injury 0 0.0 9 1.4 9 1.2
Gunshot 0 0.0 3 0.5 3 0.4
External/multiple trauma 7 4.8 57 9.1 64 8.3
Drugs/alcohol 6 4.1 72 11.5 78 10.1
Other 0 0.0 10 1.6 10 1.3
Manner of death
Self-inflicted 65 44.8 291 46.4 356 46.1
Natural causes 66 45.5 208 33.2 274 35.5
Justifiable homicide 0 0.0 4 0.6 4 0.5
Unlawful homicide 6 4.1 42 6.7 48 6.2
Accident 7 4.8 69 11.0 76 9.8
Other 0 0.0 9 1.4 9 1.2
Most serious offence
Violent 88 60.7 307 49.0 395 51.2
Theft related 36 24.8 188 30.0 224 29.0
Drug related 4 2.8 57 9.1 61 7.9
Traffic 3 2.1 18 2.9 21 2.7
Good order 8 5.5 35 5.6 43 5.6
Other 6 4.1 21 3.3 27 3.5
Total number 145 627 772

Figure 1 : Prison custody deaths by Indigenous status, 1990-2004(a)

(a) Rate per 1000 prisoners

Source: AIC, NDICP 1990-2004 [computer file]

AIC studies have found that Indigenous prisoners are more likely to begin offending regularly at younger ages than their non-Indigenous counterparts (Makkai & Payne 2003) and are more likely to be younger than non-Indigenous prisoners when they first commit a property or violent offence (Putt, Payne & Milner 2005), consistent with previous findings in this area. Young Indigenous male offenders are also more likely to enter the adult criminal justice system than their non-Indigenous counterparts which may be directly related to the number of prior arrests and convictions and therefore a lower likelihood of avoiding incarceration (Chen et al. 2005; Weatherburn, Fitzgerald & Hua 2003). Non-Indigenous persons who died in prison were generally older (mean=39 years), than their Indigenous counterparts (32 years), which is consistent with prison census data (ABS 2005).

The high rates of death among Indigenous prisoners reflect a greater vulnerability of Indigenous persons in terms of both poor health and self-harm as suggested in recent reports (SCRGSP 2005). A NSW survey of prisoner health found that Indigenous prisoners were six times more likely to have diabetes than non-Indigenous prisoners and that prevalence of hypertension among Indigenous prisoners increased between 1996 and 2001 (D'Souza, Butler & Petrovsky 2005). The comparative and continued poor health of Indigenous prisoners has been reported elsewhere (SCRGSP 2005) and provides a useful backdrop against which the statistics on cause of death can be interpreted. Hanging deaths and those due to natural causes have consistently been the most common causes of death in prison since 1990. For Indigenous prisoners, deaths due to natural causes are only slightly higher than hanging deaths and these two causes account for almost 90 percent of all Indigenous prisoner deaths (Table 1). A lack of financial resources to seek treatment for long-term health problems, combined with a scarcity of such services for Indigenous persons prior to incarceration may also play a role in the poor health of Indigenous prisoners (SCRGSP 2005) and exacerbate the number of Indigenous prisoner deaths due to natural causes.

For non-Indigenous persons, hanging deaths accounted for 42 percent of all deaths and natural causes for 33 percent of deaths. Although these two categories combined account for the majority of non-Indigenous prisoner deaths, a sizeable portion of non-Indigenous deaths (22%) are also due to external trauma, drug and/or alcohol toxicity and head injuries combined.

Indigenous rates of arrest have been found to be higher than non-Indigenous rates for offences which are likely to result in incarceration (i.e. violent and serious theft offences: Harding et al. 1995; Weatherburn, Fitzgerald & Hua 2003). While numbers of deaths of persons whose most serious offence was a violent offence peaked in 2000 at 37, accounting for 60 percent of deaths that year, the proportion of deaths of persons incarcerated for violent offences increased to 72 percent in 2003 (n=28) and 74 percent in 2004 (n=29). Recent prisoner census data reveal that just over half the Indigenous prisoners were incarcerated for violent offences compared with 37 percent of non-Indigenous prisoners (ABS 2005). A similar pattern emerges in the deaths in custody data, with 61 percent of Indigenous and 49 percent of non-Indigenous prisoners who died in custody having been incarcerated for violent offences (Table 1). The next most prevalent offence group in both Indigenous and non-Indigenous prisoners who died was theft related crimes. Almost one quarter of the Indigenous prisoners and 30 percent of non-Indigenous prisoners who died in custody between 1990 and 2004 had been incarcerated for theft-related offences.

Police custody (Category 1), 1990–2004

A resolution of the Australasian Police Ministers Council in May 1994 expanded the definition of police custody into two categories (see Box 2) to more clearly identify the level of contact and therefore an officer's ability to influence or control a person's behaviour. Prior to 1994 only deaths in police custody (Category 1a) were being recorded by the NDICP. As a result, data on deaths occurring during police operations (Categories 1b and 2) between 1990 and 1994 were collected retrospectively and the NDICP now holds information on all deaths in police custody or custody-related police operations spanning 15 years.

Deaths in police custody are divided into two main categories:

Category 1 (Police custody)

  1. Deaths in institutional settings, for example, police stations or lockups, police vehicles, during transfer to or from such an institution, or in hospitals, following transfer from an institution.
  2. Other deaths in police operations where officers were in close contact with the deceased. This would include most raids and shootings by police but would not include most sieges where a perimeter was established around a premise but officers did not have such close contact with the person to be able to significantly influence or control the person's behaviour.

Category 2 (Police-related custody)

Other deaths during custody-related police operations, including situations where officers did not have such close contact with the person to be able to significantly influence or control the person's behaviour. It would include most sieges, as described above and most cases where officers were attempting to detain a person, for example, during a pursuit.

This definition of a 'death in police custody' is based on a resolution of the Australasian Police Ministers Council in 1994.

During this time period, a total of 452 deaths occurred in both police custody and custody-related operations. Deaths in police custody have decreased since 1990 when they peaked at 25, falling to nine in 2004 (Figure 2).

Figure 2 : Trends in deaths in police custody and custody-related operations, 1990-2004

Source: AIC, NDICP 1990-2004 [computer file]

Indigenous persons account for 18 percent of deaths in police custody (Table 2). The 2002 National Police Custody Survey (Taylor & Bareja 2005) reported that Indigenous persons were 17 times more likely than non-Indigenous persons to be in police custody. Based on the custody incident data reported in 2002, the rate of death of Indigenous persons in police custody (Category 1) that year was just over 56 deaths per 100,000 relevant population. This was considerably greater than the rate of 30 deaths per 100,000 relevant population for non-Indigenous persons. Despite these findings the raw number of deaths of Indigenous persons has remained low and relatively stable since 1990, while the trend for non-Indigenous deaths shows greater fluctuation and an overall downward trend (Figure 3).

Table 2 : Deaths in police custody (Category 1) and custody-related operations (Category 2), 1990-2004
Police custody Police operations
Indigenous Non-Indigenous Total persons Indigenous Non-Indigenous Total persons
n % n % N % n % n % N %
Note: Some totals will not sum to column total due to missing data. Column percentages may not sum to 100 due to rounding errors and missing data.
Source: AIC, NDICP 1990-2004 [computer file]
Sex
Male 31 79.5 168 93.3 199 90.9 40 88.9 179 95.2 219 94.0
Female 8 20.5 12 6.7 20 9.1 5 11.1 9 4.8 14 6.0
Age
9 23.1 30 16.7 39 17.8 29 64.4 76 40.4 105 45.1
25-39 years 18 46.2 92 51.1 110 50.2 11 24.4 70 37.2 81 34.8
40-54 years 11 28.2 43 23.9 54 24.7 4 8.9 30 16.0 34 14.6
55 years + 1 2.6 15 8.3 16 7.3 1 2.2 12 6.4 13 5.6
Mean age (years) 33.2 35.2 34.8 23.4 31.0 29.5
Cause of death
Hanging 8 20.5 28 15.6 36 16.4 0 0.0 3 1.6 3 1.3
Natural causes 15 38.5 23 12.8 38 17.4 2 4.4 3 1.6 5 2.1
Head injury 2 5.1 9 5.0 11 5.0 4 8.9 19 10.1 23 9.9
Gunshot 6 15.4 70 38.9 76 34.7 3 6.7 53 28.2 56 24
External/multiple trauma 4 10.3 6 3.3 10 4.6 33 73.3 101 53.7 134 57.5
Drugs/alcohol 3 7.7 32 17.8 35 16.0 0 0.0 1 0.5 1 0.4
Other 1 2.6 10 5.6 11 5.0 3 6.7 6 3.2 9 3.9
Manner of death
Self-inflicted 8 20.5 44 24.4 52 23.7 6 13.3 74 39.4 80 34.3
Natural causes 15 38.5 24 13.3 39 17.8 2 4.4 3 1.6 5 2.1
Justifiable homicide 4 10.3 59 32.8 63 28.8 1 2.2 5 2.7 6 2.6
Unlawful homicide 0 0.0 0 0.0 0 0.0 0 0.0 2 1.1 2 0.9
Accident 10 25.6 44 24.4 54 24.7 36 80.0 98 52.1 134 57.5
Other 2 5.1 7 3.9 9 4.1 0 0.0 4 2.1 4 1.7
Most serious offence
Violent 10 25.6 69 38.3 79 36.1 8 17.8 54 28.7 62 26.6
Theft related 5 12.8 37 20.6 42 19.2 25 55.6 53 28.2 78 33.5
Drug related 1 2.6 3 1.7 4 1.8 0 0.0 3 1.6 3 1.3
Traffic 1 2.6 5 2.8 6 2.7 6 13.3 47 25.0 53 22.7
Good order 20 51.3 49 27.2 69 31.5 3 6.7 10 5.3 13 5.6
Other 2 5.1 16 8.9 18 8.2 3 6.7 18 9.6 21 9.0
Total number 39 180 219 45 188 233

Figure 3: Trends in deaths in police custody and custody-related operations by Indigenous status, 1990-2004

Source: AIC, NDICP 1990-2004 [computer file]

Deaths in police custody are more likely to involve males (91%) and persons aged between 25 and 39 years than other age groups, with 46 percent of Indigenous deaths and 51 percent of non-Indigenous deaths in this age group.

Between 1980 and 1989 most deaths in police custody were due to hangings, but since the RCIADIC deaths due to hanging have decreased, averaging 2.4 deaths per year. This is likely to be a result of efforts to reduce hanging points in police cells and lockups and identify vulnerable detainees. In a study monitoring injuries in police custody in NSW between November 2001 and June 2003, Sallybanks (2005) found that attempted hangings comprised 30 percent of all injury incidents and that most took place in police vans rather than cells, attributed by the author to the reduction of hanging points in cells. Other strategies implemented in NSW to increase the safety of police and detainees include: recording of medical conditions and risk factors, observation, safe custody training for police officers, prompt release of bailed detainees and encouraging alternatives to charging and custody (Sallybanks 2005). The implementation of such initiatives in NSW and elsewhere is likely to be behind the decrease in police custody (Category 1) deaths since 1990. A reduction in rates of Indigenous custody was reported in the 2002 National Police Custody Survey. While Indigenous persons were still over-represented, Indigenous custody rates have declined in Western Australia since 1992, and in Queensland, South Australia and the Northern Territory since 1995 (Taylor & Bareja 2005).

Since the RCIADIC, 35 percent of deaths have been caused by fatal gunshot wounds (Table 2). Of these 76 deaths, 69 were the result of police shootings and the majority were of non-Indigenous persons (n=63). Most Indigenous deaths in police custody were due to natural causes (39%).

Violent offences were the most common group of offences committed by persons who died in police custody since 1990 (Table 2). Thirty-eight percent of non-Indigenous persons and 26 percent of Indigenous persons had been detained for these offences. Over half the deaths of Indigenous persons were of those in custody for the commission of good order offences, most involving public drunkenness. Findings from the National Police Custody Survey (Taylor & Bareja 2005) showed that almost 24 percent of Indigenous persons in police custody were there for public order offences, compared with 17 percent of non-Indigenous persons. The rate (per 100,000 population) of custody incidents in the National Police Custody Survey for public order offences was higher for Indigenous persons at all ages than the rate for non-Indigenous persons.

Police custody-related operations (Category 2), 1990–2004

The trend in all deaths in police custody (Category 1) and custody-related operations (Category 2) since 1990 is relatively stable with no overall increase or decrease discernable, although a different pattern emerges when Category 1 and Category 2 deaths are separated Figure 2). Since the RCIADIC, 233 deaths have been recorded as having occurred during the course of police operations. Although the number of Category 1 deaths has generally decreased during this time, the number of deaths in police operations has increased, most notably since 1999. Police operation deaths account for almost 52 percent of all deaths in police custody since 1990.

Indigenous persons accounted for 19 percent of deaths in police operations and have remained relatively low and stable since 1990, while non-Indigenous deaths have shown greater fluctuation and an upward trend, particularly between 1998 and 2001 Figure 3). Deaths of males outnumbered deaths of females each year and comprised 94 percent (Table 2) of deaths. Seventeen percent of all deaths during custody-related operations were of Indigenous males. The mean age of deaths of Indigenous persons was 23 years, several years younger than the mean age of 31 years for non-Indigenous persons. Most deaths in this category were of persons aged younger than 25 years (45%), with 64 percent of Indigenous persons and 40 percent of non-Indigenous persons in this age group.

Fifty-eight percent of deaths in police operations were caused by external and/or multiple trauma injuries, reflecting the fact that most of these deaths were of persons involved in motor vehicle pursuits. Seventy-three percent of Indigenous deaths and 54 percent of non-Indigenous deaths were due to these types of injuries. As a result, the manner of death for most was accidental (58%).

Fifty-six percent of Indigenous persons and 28 percent of non-Indigenous persons had committed theft offences immediately prior to their death. For Indigenous persons the majority of these theft related offences involved theft or unlawful use of a motor vehicle (80%). For non-Indigenous persons there was a higher proportion (29%) of violent offences than for Indigenous persons (18%).

Conclusion

Although the majority of Australian deaths in custody occurred in police custody between 1980 and 1989, the trend in the 15 years since the RCIADIC reveals a different picture. Prison custody deaths account for the majority of deaths in custody since 1990 and have exceeded all other deaths each year. Furthermore, while deaths in police custody have decreased since 1990, deaths in police operations have been increasing. This decline in police custody deaths may be due in part to the redesign of police cells to reduce points from which detainees can hang or harm themselves. Being able to identify such a decline in police custody deaths is one of the clear benefits of the NDICP and this finding shows that progress has been made since the RCIADIC relating to deaths specifically occurring in police custody.

Deaths of Indigenous persons were lowest in police custody (Category 1) with 39 recorded since 1990, while 45 deaths during police operations (Category 2) and 145 prison deaths were recorded in the same time period. A more distinct pattern emerged when trends in the number of deaths in police custody and deaths during police operations were separated. While the number of deaths of Indigenous persons in police custody and custody-related operations remained relatively low and stable over the 15 years, deaths of non-Indigenous persons fluctuated. The increase in police operations deaths was largely due to an increase in non-Indigenous deaths.

The importance of continuing to monitor all deaths in custody is the ability to identify changing trends in custody deaths across Australia. It is clear that Indigenous persons are still over-represented in some deaths in custody statistics as they are across the wider criminal justice system. The renewed call to address socioeconomic and health issues in recent Government reports and to move away from a sole focus on systemic bias and diversionary options for Indigenous youths (Weatherburn, Fitzgerald & Hua 2003) is in keeping with the recommendations of the RCIADIC on the need to address the underlying issues that interact and seem inevitably to lead to an over-representation of Indigenous persons at all levels of the criminal justice system. The trends revealed in the NDICP could be interpreted as one indicator of the effectiveness of initiatives aimed at overcoming Indigenous disadvantage.

Acknowledgment

The AIC acknowledges the cooperation of each of Australia's police services, prison administrators and juvenile welfare/justice authorities in supplying information as the basis of this report.

All URLs were correct at April 2006

References

  • Australian Bureau of Statistics (ABS) 2005. Prisoners in Australia 2005. cat. no. 4517.0 Canberra: Australian Bureau of Statistics
  • Chen S et al. 2005. The transition from juvenile to adult criminal careers. Crime and justice bulletin: no. 86
  • D'Souza RM, Butler T & Petrovsky N 2005. Assessment of cardiovascular disease risk factors and diabetes mellitus in Australian prisoners: is the prisoner population unhealthier than the rest of the Australian population? Australian and New Zealand journal of public health 29(4): 318-323
  • Harding R et al. 1995. Aboriginal contact with the criminal justice system and the impact of recommendations of the Royal Commission into Aboriginal Deaths in Custody Sydney: Federation Press
  • Hunter BH 2001. Indigenous Australian arrest rates: economic and social factors underlying the incidence and number of arrests. Working Paper no.10. Canberra: Centre for Aboriginal Economic Policy Research, Australian National University
  • Joudo J & Veld M 2005. Deaths in custody in Australia: 2004 National Deaths in Custody Program annual report. Technical and background paper series no 19. Canberra: Australian Institute of Criminology
  • Makkai T & Payne J 2003. Drugs and crime: a study of incarcerated male offenders. Research and public policy no. 52. Canberra: Australian Institute of Criminology
  • Putt J, Payne J & Milner L 2005. Indigenous male offending and substance abuse. Trends & issues in crime and criminal justice no 293
  • Royal Commission into Aboriginal Deaths in Custody (RCIADIC) 1991. National report volume 1. Canberra: Australian Government Publishing Service
  • Sallybanks J 2005. Monitoring injuries in police custody: a feasibility and utility study. Technical and background paper no.15. Canberra: Australian Institute of Criminology
  • SCRGSP (Steering Committee for the Review of Government Service Provision) 2005. Overcoming Indigenous disadvantage: key Indicators 2005. Melbourne: Productivity Commission
  • Taylor N & Bareja M 2005. 2002 National Police Custody Survey. Technical and background paper no.13. Canberra: Australian Institute of Criminology
  • Weatherburn D, Fitzgerald J & Hua J 2003. Reducing Aboriginal over-representation in prison. Australian journal of public administration 62(3): 65-73
  • Williams P 2001. Deaths in custody: 10 years on from the Royal Commission. Trends & issues in crime and criminal justice no. 203

Jacqueline Joudo is a research assistant in the Justice and Crime Analysis Program.