Australian Institute of Criminology

Skip to content

Self-inflicted deaths in Australian prisons

Trends & issues in crime and criminal justice no. 513

Matthew Willis, Ashleigh Baker, Tracy Cussen and Eileen Patterson
ISSN 1836-2206
Canberra: Australian Institute of Criminology, August 2016

Abstract | The AIC has operated the National Deaths in Custody Program since 1991, following a recommendation by the Royal Commission into Aboriginal Deaths in Custody. This study updates and extends an earlier AIC study, which examined the important issue of self-inflicted deaths in prison custody using deaths in custody data for the period 1999–2013. Regrettably, suicide remains a common cause of death in prison; however, it is no longer the most common cause of death. While self-inflicted deaths constituted half of all prison deaths between 1980 and 1998, they have declined to the point that between 2004 and 2013—the most recent decade for which data are available—they represented around a quarter of prison deaths. This decrease reflects the considerable progress made by corrective services administrators in developing policies and practices, and implementing cell designs, that address suicide risk factors. Nonetheless, these data and coronial findings suggest there is still room for improvement, particularly in responding to prisoners with psychiatric needs and in the management of unsentenced prisoners.

Reducing the rate of self-inflicted deaths in prison has long been a priority for correctional agencies across Australia and internationally. The rate of self-inflicted deaths in the prison population greatly exceeds that of self-inflicted deaths in the community; prisoners represent a particularly vulnerable and high-risk group for suicide. This paper updates Dalton’s (1999) national overview of self-inflicted deaths, which analysed data from the Australian Institute of Criminology’s (AIC) National Deaths in Custody Program (NDICP) for the period 1980–1998. The current paper analyses data for the period 1999–2013, allowing changes in the prevalence and nature of prison suicides to be examined over a 34-year period.

International research shows prison inmates have a higher rate of suicide than their counterparts in the general community, with prison suicide rates typically three to five times those of the general community (Corben 2006; Fazel et al. 2011; Jenkins et al. 2005; Johnston 1991; WHO 2007; Wobeser et al. 2002). Offenders enter the prison system with more risk factors for suicide than those that apply to members of the general community (Konrad et al. 2007; Larney et al. 2012), and remain at elevated risk of suicide following their release (Daigle & Naud 2012; Pratt et al. 2006).

Studies examining risk factors for self-inflicted deaths in prisons have found many prisoners who suicided entered the prison system with histories of prior suicide attempts (Fazel et al. 2008; WHO 2007). Across all Australian prisons, 16 percent of those entering prison report intentionally having harmed themselves, while 11 percent had thought of harming themselves in the preceding 12 months (AIHW 2012). Both histories of actual self-harm and thoughts of self-harm were more common among female than male prison entrants, and among non-Indigenous than Indigenous prison entrants (AIHW 2012). A New Zealand study found broadly comparable results, with lower proportions of Maōri than European prisoners experiencing thoughts of suicide; Maōri prisoners who attempted suicide were, however, more likely to be successful (Simpson 2012).

A study of suicidal ideation and suicide attempts in a NSW prison found 34 percent of inmates reported lifetime suicidal ideation and 21 percent had attempted suicide (Larney et al. 2012). Among prisoners in this NSW study who attempted suicide, 58 percent reported a lifetime history of suicidal ideation and suicide attempts. Another NSW study found 68 percent of subjects had a documented history of intentional self-harm (O’Driscoll, Samuels & Zacka 2007), while in South Australia 40 percent of prisoners who suicided had a documented history of previous attempts (Austin, van den Heuvel & Byard 2014). A study in Western Australian prisons found a little over half of all prisoners (56.4% female and 47.3% male) had contemplated suicide at some point in their life (Fleming, Gately & Kraemer 2012). Some 39.5 percent of females had previously attempted suicide, while the same was true for 29.6 percent of males. Female prisoners tended to report their suicidal thoughts had increased while in custody, while males reported a decrease in suicidal thoughts (Fleming, Gately & Kraemer 2012).

A similar pattern has been seen overseas. Some 82 percent of prisoners in the UK who suicided had a history of self-harm and suicide attempts while in custody (Shaw, Appleby & Baker 2003). A Dutch prison study found 54 percent of suicide victims in their sample had a history of attempted suicide, including in the community (Blaauw, Kerkhov & Hayes 2005).

It is apparent that the life experiences of prisoners, including the experience of imprisonment, increase the risk of suicide both in and out of prison. A study of over 40,000 people released from Queensland prisons found that released women were 14.2 times, and released men 4.8 times, more likely to die from suicide than the general population (Spittal et al. 2014). While the overall rate of self-inflicted death after release was comparable with that observed for drug-related deaths, drug-related deaths spiked in the two weeks after release whereas suicides did not. Being married was a strong protective factor against suicide, whereas multiple prior imprisonment was a risk factor (Spittal et al. 2014).

Psychiatric illness, particularly depression, and drug use are risk factors for suicide (Fazel et al. 2008; Marzano et al. 2003; Shaw, Appleby & Baker 2003). Blaauw, Kerkhov & Hayes (2005) found 78 percent of their Dutch sample had a history of psychiatric care and 58 percent a history of illicit drug use. A study using NDICP data examined the case histories of the 75 female prisoners who died in prison or police custody between 1980 and 2000. The largest proportion of these deaths (32%; n=24) were the result of hanging (Collins & Mouzos 2002). While the examination of case histories did not separate self-inflicted from other deaths, the authors noted that many of the deceased women had experienced alcohol and other substance misuse, psychiatric illness, domestic violence and self-harm (Collins & Mouzos 2002). Studies focused on NSW and South Australian prison populations have similarly found strong correlations between psychiatric illness and substance use among prisoners who attempt suicide or who suicide (Austin, van den Heuvel & Byard 2014; Larney et al. 2012; O’Driscoll, Samuels & Zacka 2007).

Several studies have identified higher rates of suicide among unsentenced than sentenced prisoners. Studies in both NSW and South Australia found more than 50 percent of suicide victims were on remand at the time of their death (Austin, van den Heuvel & Byard 2014; O’Driscoll, Samuels & Zacka 2007). A UK study found that, while unconvicted remandees constituted only 19 percent of the total prison population, they represented almost half of those committing suicide, while another 20 percent of suicide victims had been convicted but not yet sentenced (Shaw, Appleby & Baker 2003). Another UK study of near-lethal self-harm in female prisoners found that 35 percent of those attempting suicide were on remand (Marzano et al. 2010).

Prison suicide risk is generally highest early in the detention period. In South Australia, 26 percent of suicides were found to occur in the first week of custody and 39 percent in the first month (Austin, van den Heuvel & Byard 2014). In the UK, 11 percent of inmates who suicided did so within 24 hours of entering custody, and 32 percent took their lives within seven days (Shaw, Appleby & Baker 2003). In a New Zealand study by the Independent Police Conduct Authority (ICPA 2012), half of those who suicided in police custody did so less than four hours after being detained.

Being accommodated in single cells, compared with sharing a cell, can contribute to suicide risk (Austin, van den Heuvel & Byard 2014; Fazel et al. 2008; IPCA 2012; O’Driscoll, Samuels & Zacka 2007; Shaw et al. 2003; WHO 2007). One UK study found that, when inmates suicided in shared accommodation, 52 percent of them did so when their cellmate was absent (Shaw, Appleby & Baker 2003).

Defining self-inflicted death

Within the NDICP any death that is the result of a prisoner’s own actions, regardless of the underlying intention behind those actions and whether the prisoner was intending to take their own life, is classified as a self-inflicted death. Self-inflicted deaths are subcategorised within the monitoring program according to intent—that is, whether determined by a coroner to be:

  • intentional;
  • not intentional;
  • accidental hanging—where a coroner has found a self-inflicted hanging to be accidental or the product of misadventure; or
  • where intention was not stated or was unable to be determined by a coroner.

Non-self-inflicted deaths incorporate all other causes of death, including death by natural causes and unlawful homicide.

The current study

This paper draws on data from the NDICP for the calendar years 1999 to 2013. This period was chosen to ensure the complete dataset was accounted for when comparing these results with the previously published results (Dalton 1999). A review of the NDICP in 2011 resulted in a change from calendar year reporting to reporting by financial year. While the use of calendar-year data for this paper does not align with the current reporting cycle, it does allow comparison with the earlier analysis (Dalton 1999) and also with data on causes of death in the general community, as reported by the Australian Bureau of Statistics (ABS; ABS 1999, 2000a–2012a). However, there are some limitations to this comparability. The 19-year data period analysed by Dalton (1999) is four years longer than the 15-year period analysed for this paper, which should be taken into account when considering the overall number of deaths.

A 2011 review of NDICP data identified a number of cases previously omitted from the dataset (see Lyneham & Chan 2013). This resulted in the inclusion of additional (n=18) deaths in prison custody not previously included in the dataset. Dalton’s earlier (1999) analysis of suicide in prison reported that a total of 787 people died in prison custody in Australia between 1980 and 1998, and that 46.6 percent of those deaths (n=367) were the result of suicide. With the addition of cases following the 2011 review, and the finalisation of some coronial matters still open at the time of the earlier report, the NDICP database now shows that between 1980 and 1998 there were 794 deaths, of which 47.4 percent (n=376) were self-inflicted. The authors do not consider the small change to the number and proportion of self-inflicted deaths occurring between 1980 and 1998 affects Dalton’s (1999) overall findings and conclusions. However, where comparisons are made in this paper across the whole dataset, the amended data will be used, which will in some cases differ from the data reported by Dalton (1999).

Since 2000, the AIC has had access to the data in the National Coronial Information System, which includes information from coroners’ reports. As noted above, deaths are classified as self-inflicted for the NDICP if they are determined, through a coronial process, to be self-inflicted, regardless of the intention of the detainee. The counting rules used by the ABS in coding deaths in the general community due to intentional self-harm or suicide have changed over time and may not be directly comparable with those used in the NDICP. Nonetheless, the data are considered sufficiently comparable for the type of analysis used in this paper.

Comparisons with self-inflicted deaths in the general community are restricted to the period 1999 to 2012, as cause of death data for the Australian population in 2013 were not available at the time of writing.

Self-inflicted deaths since 1999

From the period 1 January 1999 through 31 December 2013, 693 deaths occurred in Australian prisons (Figure 1). Of these, 240 deaths were self-inflicted (34.6%), 447 deaths (64.5%) were not self-inflicted and the nature of six deaths (0.9%) was unable to be determined as no relevant coronial finding was available at the time of writing. Two of these indeterminate deaths occurred in 2011, one in 2012 and three in 2013. The proportion of self-inflicted deaths between 1999 and 2013 (34.6%) was substantially lower than the proportion of self-inflicted deaths between 1980 and 1998 (47.4%).

Between 1980 and 2013, the average number of prison custody deaths each year from all causes increased from 41.8 deaths per year in 1980–1998 to 46.2 in 1999–2013, indicating growth in the overall prison population. At the same time, the average number of self-inflicted deaths decreased from 19.8 to 16 deaths per year. Accordingly, while 47.4 percent of prison deaths from 1980 to 1998 were self-inflicted, this decreased to 34.6 percent of deaths in the later period.

Figure 1: Deaths in prison custody, self-inflicted and other causes, Australia, 1980–2013

Source: National Deaths in Custody Program, AIC data file

As shown in Figure 1, while there has been some fluctuation, both the number and proportion of self-inflicted deaths have decreased since the earlier analysis (Dalton 1999). This was particularly so in the last decade (2004–2013) with an average of 11.7 self-inflicted deaths per year, representing 27.5 percent of deaths over this period. This compares with an annual average of 27.2 self-inflicted deaths, representing 46.8 percent of all deaths in prison custody between 1994 and 2003. Since 2006, self-inflicted deaths have consistently represented less than a third of deaths in prison custody.

Demographic characteristics

The vast majority of offenders in custody are male; at the most recent national prison census on 30 June 2014, 92.3 percent of Australian prisoners were male, and this proportion has remained essentially stable throughout the study period (ABS 2000b–2012b, 2013–14). It is therefore to be expected that most of those who die in prison custody will be male, and NDICP data shows this is the case. During the period 1999–2013, 92.5 percent (n=222) of self-inflicted deaths were of males.

Between 1999 and 2013, the proportion of Indigenous people in Australian prison populations gradually increased from 20 percent in 1999 and 18.9 percent in 2000, to 27.4 percent in 2013 (ABS 1999, 2000a–2012a). Over the same period, 18.9 percent of all deaths (n=131) were of Indigenous prisoners, as were 16.3 percent of self-inflicted deaths (n=39). In comparison, between 1980 and 1998 Indigenous prisoners accounted for 14.4 percent of self-inflicted deaths. While there has been—in the context of very small numbers—considerable fluctuation in the rate from year to year, the overall proportion of Indigenous suicides has not increased in line with the increased representation of Indigenous persons in the Australian prison population during the study period.

Figure 2 shows the relative proportions of Indigenous and non Indigenous deaths across Australian jurisdictions. The data depicted in this figure show there is considerable variation between jurisdictions in the proportion of deaths, by Indigenous status, that are self-inflicted.

During the period from 1998 to 2013 there were no self-inflicted Indigenous prisoner deaths in the ACT, Tasmania or Victoria. These are jurisdictions with relatively low numbers of Indigenous prisoners. In Western Australia, 80.0 percent of deaths of Indigenous prisoners were self-inflicted. This compared with 41.7 percent of Indigenous prisoner deaths in NSW being self-inflicted, 21.4 percent in Queensland and 13.6 percent in the Northern Territory.

Data available through the NDICP does not support a deeper examination of the factors underlying these inter-jurisdictional variations. The variations may be due to differences in the overall health of Indigenous prisoners entering the prison system, degrees of separation and distance from family and home communities, and differences in aspects of correctional management such as visit protocols.

Figure 2: Australian prison populations 2013 by Indigenous status, and self-inflicted deaths in prison custody by Indigenous status, 1998–2013

Source: National Deaths in Custody Program, AIC data file; Australian Bureau of Statistics 2013

Figure 3: Self-inflicted deaths in prison custody by age group and Indigenous status, Australia, 1999–2013

Source: National Deaths in Custody Program, AIC data file

When comparing self-inflicted deaths from 1999 to 2013 by age group and Indigenous status (Figure 3), the largest proportion of self-inflicted deaths among non-Indigenous prisoners was among 30–34 year olds (21.9%; n=44), followed by among 25–29 year olds (16.9%;n=34) and 35–39 year olds (14.9%;n=30). Indigenous prisoners showed a different age distribution, with 28.2 percent (n=11) of self-inflicted deaths occurring among 20–24 year olds, 23.1 percent (n=9) among 35–39 year olds and 18 percent (n=7) among 30–34 year olds. While only 4.2 percent (n=10) of all prisoners who took their own lives were aged between 17 and 19 years, 10.3 percent (n=4) of Indigenous prisoners were in this youngest group, compared with three percent (n=6) of non-Indigenous prisoners. There were no self-inflicted deaths of Indigenous people aged over 50 years.

Differences in the age distribution of Indigenous and non-Indigenous self-inflicted deaths reflect differences in age distributions in the overall population, with a greater relative proportion of Indigenous people being younger. Within the prison population, Indigenous prisoners are nearly twice as likely as Indigenous prisoners to be aged 24 years or younger (ABS 2015). The large proportion of Indigenous self-inflicted deaths of 35–39 year olds may reflect the relatively small proportions of older Indigenous prisoners. However, consideration of these differences should take into account the small numbers of Indigenous people in each age group.

Mean age at death for all self-inflicted deaths in the 1999–2013 period was 34.5 years (SD=11.2), which was markedly lower than the average age for non-self-inflicted prison deaths (51.1 years; SD=16.2).

Cause of death

Consistent with the earlier analysis, between 1999 and 2013 prisoners were much more likely to take their own lives by hanging (n=206; 86%) than by any other means. Of the remainder, 20 prisoners (8%) inflicted external/multiple trauma upon themselves which led to their death, six (3%) died from other or multiple causes (for example asphyxia), five (2%) died after ingesting drugs and the remaining three (1%) died from head injuries.

Almost all Indigenous prisoners who took their own lives did so by hanging (n=37; 95%). The remaining Indigenous self-inflicted deaths were the result of external/multiple trauma or other/multiple causes.

Across both non-Indigenous and Indigenous prisoners, a range of fittings and materials were used in hanging deaths. Hanging points included ‘other fitting in cell’ (35%), cell bars (32%), shower fixtures (13%), and bunk beds (12%). Prisoners were most likely to hang themselves using bed sheets (59%). Other items used included rope or cord (10%), shoelaces (10%) and other clothing, including belts (10%).

Following the report of the Royal Commission into Aboriginal Deaths in Custody in 1991, considerable attention was drawn to the need to reduce hanging points in cells. Between 1999 and 2013, coronial recommendations were recorded in the NDICP in 86 cases (73 hangings, 13 other causes). Of the recommendations made concerning hanging deaths, 33 referred to removing hanging points and/or issues such as the need to ensure compliance with ‘safe cell’ principles and the need for custodial staff to carry Hoffman knives for cutting hanging materials. Coronial recommendations also frequently referred to the need for improved responses to prisoners with mental health issues, including through improved identification and notification of risk, and better access for prisoners to mental health services. Findings based on coronial recommendations suggest that, while prison authorities have made substantial progress in eliminating hanging points, they remain in some environments. More recent coronial findings indicate a need for further improvements in managing prisoners at risk of suicide, particularly those with mental health issues.

Legal status, time in custody and offence type

While the early stages of custody appear to be a risky time for all prisoners who may be suicidal, the risk is heightened for Indigenous and unsentenced prisoners, both separately and cumulatively.

Legal status

Unsentenced prisoners were relatively more likely to take their own lives than sentenced prisoners (see Table 1). From 1999 to 2013, unsentenced prisoners accounted for 29.3 percent of all deaths in custody (n=203), but 53.3 percent of self-inflicted deaths (n=128). Non-Indigenous prisoners were slightly more likely (54.2%; n=109) to take their own lives while unsentenced compared with Indigenous prisoners, of whom 48.7 percent (n=19) were unsentenced at the time of their death.

The rate of self-inflicted deaths of unsentenced prisoners ranged from 0.54 to 3.96 deaths per 1,000 unsentenced prisoners, with an average of 1.61 deaths per 1,000 unsentenced prisoners per year. The rate has decreased noticeably in recent years. From 2006 to 2013 the annual average was 0.91 deaths per 1,000 unsentenced prisoners. In contrast, the rate of self-inflicted deaths among sentenced prisoners ranged from 0.09 to 0.89 deaths per 1,000 sentenced prisoners between 1999 and 2013, at an average of 0.35 deaths per 1,000 prisoners. Between 2006 and 2013, the average self-inflicted death rate fell to 0.22 per 1,000 sentenced prisoners.

Time in custody and sentence length

Consistent with broader research, NDICP data show the first one to three months of imprisonment are a time of heightened risk of, and vulnerability to, self-inflicted death. Of the 233 self-inflicted deaths occurring between 1999 and 2013 where time in custody was recorded, 117 (50.2%) occurred during the first three months of imprisonment. This included 68 deaths (29.2% of all self-inflicted deaths) that occurred during the first month of imprisonment.

Unsentenced prisoners are particularly vulnerable to suicide risks early in their imprisonment (see Table 1). A little over half (56.4%; n=22) of Indigenous prisoners who took their own life did so in the first three months of being in custody, including six (15.4%) in the first week. Each of those six was unsentenced, and 73.6 percent (n=14) of all unsentenced Indigenous prisoners who took their own lives did so in the first three months.

Similarly, 50.8 percent (n=102) of non-Indigenous prisoners who suicided did so during the first three months in custody. A little under two thirds (62.4%; n=68) of non-Indigenous unsentenced prisoners who suicided did so in the first three months, and 22.9 percent (n=25) within the first week. Among sentenced non-Indigenous prisoners who suicided, 29.3 percent (n=27) did so within the first three months, including 12 percent (n=11) within the first seven days.

Table 1: Self-inflicted deaths in prison and the Australian population, legal status and time in custody, by Indigenous status, 1999–2012
Time in custody Indigenous Non-Indigenous All prisoners
Unsentenced Sentenced Unsentenced Sentenced Unsentenced Sentenced
n % n % n % n % n % n %
<24 hrs - - - - 3 2.8 1 1.1 3 2.3 1 0.9
1 to <2 days 2 10.5 - - 7 6.4 1 1.1 9 7.0 1 0.9
2 to <7 days 4 21.1 - - 15 13.8 9 9.8 19 14.8 9 8.0
7 to <14 days - - - - 7 6.4 5 5.4 7 5.5 5 4.4
14 days to <1 mth 1 5.3 2 10.0 9 8.3 2 2.2 10 7.8 4 3.5
1 to <3 mths 7 36.8 6 30.0 27 24.8 9 9.8 34 26.6 15 13.3
3 to <6 mths 1 5.3 6 30.0 15 13.8 16 17.4 16 12.5 22 19.5
6 to <12 mths 3 15.8 3 15.0 14 12.8 8 8.7 17 13.3 11 9.7
1 to <2 yrs 1 5.3 2 10.0 4 3.7 14 15.2 5 3.9 16 14.2
2 to <5 yrs - - - - 3 2.8 12 13.0 3 2.3 12 10.6
5 yrs+ - - 1 5.0 1 0.9 12 13.0 1 0.8 14 12.4
Missing - - - - 4 3.7 3 3.3 4 3.1 3 2.7
Total 19 100 20 100 109 100 92 100 128 100 113 100

Source: National Deaths in Custody Program, AIC data file

Most serious offence

Prisoner census data indicates that at 30 June 2015 a little over half of all sentenced prisoners in Australian prisons (54.0%, n=19,503) had been imprisoned for a most serious offence involving violence (ABS 2015). Violent offenders are somewhat over-represented in self-inflicted deaths. From 1999 to 2013, 65 percent (n=157) of self-inflicted deaths occurred among detainees, both sentenced and unsentenced, incarcerated for a violent offence including homicide (n=51; 21%), assault (n=46; 19%); robbery (n=28; 12%) and sex offences (n=26; 11%). Unlawful entry with intent was the most serious offence of 11 percent (n=26) of detainees who suicided and theft offences accounted for six percent (n=14).

Unsentenced prisoners charged with homicide accounted for the largest proportion of self-inflicted deaths (26%; n=33). This group is over-represented in the unsentenced prisoner self-inflicted death statistics, given that those charged with homicide accounted for just 6.1 percent of unsentenced prisoners in the most recent prisoner census (ABS 2015). The next-highest proportion of self-inflicted deaths among unsentenced prisoners was of those charged with acts intended to cause injury (17%; n=22). However, those charged with this offence made up 30.5 percent of unsentenced prisoners (ABS 2015). Conversely, among sentenced prisoners, self-inflicted deaths occurred most frequently among those with a most serious offence involving acts intended to cause injury (21.4%; n=24). In the overall prison population, 17.7 percent of all sentenced prisoners were serving sentences for acts intended to cause injury (ABS 2015).

Community comparisons

Deaths in prison custody are much more likely to be self-inflicted than deaths in the general community. While suicide is a relatively rare cause of death in Australia overall, it is a relatively common cause of death in prison.

Figure 4 shows the proportion of self-inflicted deaths in prison and in the overall Australian population, as a percentage of deaths from all causes, for the period 1999 to 2012. During this time, the proportion of self-inflicted deaths in the overall population remained largely stable, accounting for between 1.3 percent and 1.9 percent of deaths (mean=1.6%). In contrast, while the proportion of self-inflicted deaths in prison generally trended downward, they accounted for between 17.9 and 51.2 percent of all deaths (mean=34.6%).

Rates of self-inflicted deaths in prison are far higher than those in the community. Figure 5 shows the rate of self-inflicted deaths in the general community between 1999 and 2013 ranged between 8.6 and 13.2 per 100,000 persons in the total population (mean=10.8 per 100,000). Among prison populations, self-inflicted deaths over the same period occurred at a rate of between 23.8 and 142.8 per 100,000 persons in the total prison population (mean=67.8 per 100,000). There has been an overall decline in the rate of prison suicides over the period, with the rate in 2012 being 2.2 times that of rate in the general population rate in the same year, compared with 11.6 times in 2000.

Figure 4: Self-inflicted deaths in prison and in the Australian population, percentage of all deaths in relevant population, 1999–2012

Source: National Deaths in Custody Program, AIC data file; Australian Bureau of Statistics, Causes of Death, 1999–2012

Figure 5: Self-inflicted deaths in prison and in the Australian population, rate per 100,000 relevant population, 1999–2012

Source: National Deaths in Custody Program, AIC data file; Australian Bureau of Statistics, Causes of Death, 1999–2012

The earlier analysis of self-inflicted deaths (Dalton 1999) also examined the differences between prison and community suicide rates. It found community suicide rates had increased from 10.9 deaths per 100,000 in 1980 to 15.1 deaths per 100,000 in 1997. Suicides in prison tended to fluctuate more widely, but ranged from 102.6 per 100,000 prisoners in 1980 (9.4 times the community rate) to 179.7 per 100,000 prisoners in 1998 (11.9 times the community rate in the previous year). Therefore, both prison and community suicide rates have decreased since Dalton’s (1999) analysis, with a generally greater decrease in prison deaths.

While there are clear differences between self-inflicted deaths in prison and in the community, those differences narrow when sex and age characteristics are taken into account. As noted, prisoners are overwhelmingly likely to be male, and males accounted for three-quarters (75%) of suicides in the overall Australian population in 2012 (ABS 2012a). Just under half (49.9%) of prisoners were aged from 20 to 34 years (ABS 2015), and this same group is over-represented in suicides in the wider community; more than a quarter of deaths of 20 to 34 year old males in Australia are the result of suicide (ABS 2012a). Prisoners tend to fall into the highest-risk sex and age categories for suicide.


While some issues remain, policy and procedural changes to prison administration, the management of offenders and institutional design have had a major impact on the likelihood of prisoners taking their own lives. Each of Australia’s eight correctional services departments was invited to contribute to this research by completing a short questionnaire to inform a better understanding of the policy and operational changes that may have contributed to the decline in self-inflicted deaths.

A collated response received from the Correctional Services Administrators Council (CSAC) identified a number of strategies used across jurisdictions to identify, manage and respond to risks of self-harm and suicide. These strategies include:

  • screening and assessment— undertaken upon reception of a detainee at a detention facility and updated as required throughout the detainee’s length of stay. Screening and assessment identifies episodes of prior attempted suicide, self-harm and current ideation, as well as concerns for safety and wellbeing that detainees may have upon entering custody;
  • mandatory notification—cross-jurisdictional mandatory notification of prisoner-at-risk status is made to relevant parties including other corrective services staff and healthcare providers;
  • documenting risks detainees present to themselves and/or others—often referred to as alerts, risks are flagged to ensure information about risk is clearly evident to anyone responsible for providing care to, or managing the risk of, a prisoner;
  • staff training—including in suicide and self-harm awareness and prevention, responding to incidents, and incident reporting and associated records management;
  • offender management—including reviews of detainee classification, restrictions on movement and/or access to areas and/or other people (based on assessments of risk), as well as providing access to programs and services, and visits from family and/or friends and external service providers (based on assessment of needs);
  • policies and procedures governing responses to suicide, self-harm attempts and completed suicides—generally described as a whole-of-prison response comprising the provision of medical assistance, oversight and review, including internal investigation, notification to relevant persons (for example police and coroners); and
  • alteration of the physical environment—including the removal of hanging points and the installation of security cameras, duress alarms etc.

Many of these strategies reflect coronial recommendations in cases of self-inflicted prison deaths. It is not possible to say from the available information whether the strategies were implemented in response to coronial recommendations or whether the recommendations relate to individual jurisdictions which had not implemented strategies adopted by other jurisdictions. Nonetheless, it is apparent there is substantial alignment between the strategies adopted by correctional agencies and the approaches recommended by coroners.


The proportion of self-inflicted deaths in prison, as a proportion of all deaths, has markedly decreased in the period following Dalton’s (1999) analysis. While close to half of all deaths in prison custody from 1980 to 1998 were self-inflicted, between 1999 and 2013 this reduced to approximately a third, and since 2004 to approximately a quarter. This change has occurred alongside an increase in the number and relative proportion of natural-cause deaths, which during the latter period consistently outnumbered all other causes of death (Lyneham & Chan 2013).

Although deaths in Australian prisons are rare, it is important to understand the circumstances surrounding any death, to inform the development of evidence-based policies and procedures designed to prevent deaths and improve responses when deaths occur.

Data from the NDICP indicates that self-inflicted deaths are no longer the most common type of death in Australian prisons, but their incidence still remains relatively high. At the same time, the data shows that unsentenced prisoners, particularly those facing lengthy terms in prison on homicide charges, are at greater risk of suicide than sentenced prisoners, and this risk is heightened in the first weeks of imprisonment. What could not be determined through existing data sources is the extent to which changes in legal or other circumstances contributed to self-inflicted deaths. For now, it is the task of correctional agencies, coroners and independent oversight bodies to determine how a number of factors may have contributed to any individual’s death.

Finally, this paper highlights the potential for improving NDICP data collection so future reporting can consider additional information that may assist in providing a better understanding of the confluence of factors that lead to self-inflicted deaths in prison. Further, it demonstrates the need for ongoing attention to policy and operational developments—particularly in the context of the psychological distress and mental health issues confronting many prisoners—to further reduce the incidence of prison suicide.

Additional information could be particularly valuable in light of findings about the increased risk of suicide for unsentenced prisoners, particularly early in their custody. NDICP data does not provide information on events occurring during the prisoner’s sentence that may contribute to suicidal ideation, such as refusal of parole, further charges or disciplinary action. Nor does the NDICP provide insights into personal events, such as relationship issues, the death of loved ones or conflicts within the prison, that may have contributed to an individual’s decision to take their own life.

While such information is outside the scope of the NDICP, further study into the justice-related and other factors precipitating suicidal intentions would be valuable—for instance, understanding whether changes in status, such as being found guilty after a period on remand or being sentenced, increase suicide risk. For Indigenous prisoners there may be certain times during their custody when they are at increased risk, such as when being in custody prevents a prisoner carrying out cultural obligations. A greater understanding of potential risk factors could inform better responses to prisoners facing particularly difficult circumstances.


URLs correct as at October 2015

  • Austin AE, van den Heuvel C & Byard RW 2014. Prison suicides in South Australia: 1996-2010. Journal of Forensic Sciences 59(5): 1–3
  • Australian Bureau of Statistics 1999, 2000a, 2001a, 2002a, 2003a, 2004a, 2005a, 2006a, 2007a, 2008a, 2009a, 2010a, 2011a, 2012a. Causes of death, Australia. Catalogue no. 3303.0.
  • Australian Bureau of Statistics 2000b, 2001b, 2002b, 2003b, 2004b, 2005b, 2006b, 2007b, 2008b, 2009b, 2010b, 2011b, 2012b, 2013, 2014, 2015. Prisoners in Australia. Catalogue no. 3303.0.
  • Australian Bureau of Statistics 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012. Causes of death, Australia. Catalogue no. 4517.0.
  • Australian Institute of Health and Welfare 2012. The health of Australia’s prisoners. Canberra: AIHW
  • Blaauw E, Kerkhov JFM & Hayes LM 2005. Demographic, Criminal, and Psychiatric Factors Related to Inmate Suicide. Suicide and Life-Threatening Behaviour 35(1): 63–75
  • Collins L & Mouzos J 2002. Deaths in Custody: A Gender-specific Analysis. Trends & Issues in Crime and Criminal Justice no. 238.
  • Corben S 2006. NSW Inmate Census 2005: Summary of Characteristics. Statistical Publication no. 27. Sydney: NSW Department of Corrective Services.
  • Daigle MS & Naud H 2012. Risk of dying by suicide inside or outside prison: The shortened lives of male offenders. Canadian Journal of Criminology & Criminal Justice 54(4): 511–528
  • Dalton V 1999. Suicide in prison 1980 to 1998: national overview. Trends & Issues in Crime and Criminal Justice Series no. 126. Canberra: Australian Institute of Criminology.
  • Dear G 1999. Preventing self-harm in prison: Do we need different strategies for Indigenous and non-Indigenous prisoners? Paper presented to the Best Practice Interventions in Corrections for Indigenous People Conference. Adelaide: Australian Institute of Criminology.
  • Fazel S, Grann M, Kling B & Hawton K 2011. Prison suicide in 12 countries: An ecological study of 861 suicides during 2003–2007. Social Psychiatry and Psychiatric Epidemiology 46(3): 191–195
  • Fleming J, Gately N & Kraemer S 2012. Creating HoPE: Mental health in Western Australian maximum security prisons. Prisons, Psychiatry, Psychology and Law 19(1): 60–74
  • Independent Police Conduct Authority 2012. Thematic Report: Deaths in custody—Ten Year Review. Wellington, NZ: IPCA.
  • Jenkins R, Bhugra D, Meltzer H, Singleton N, Bebbington P et al. 2005. Psychiatric and social aspects of suicidal behaviour in prisons. Psychological Medicine 35(2): 257–269
  • Johnston E 1991. Royal Commission into Aboriginal Deaths in Custody. Canberra: AGPS.
  • Konrad N, Daigle MS, Daniel AE, Dear GE, Frottier P et al. 2007. Preventing suicides in prisons, part 1: Recommendations from the International Association for Suicide Prevention Task Force on Suicide in Prisons. Crisis 28(3): 113–121
  • Larney S, Topp L, Indig D, O’Driscoll C & Greenberg D 2012. A cross-sectional survey of prevalence and correlates of suicidal ideation and suicide attempts among prisoners in New South Wales, Australia. BMC Public Health 12(14).
  • Lyneham M & Chan A 2013. Deaths in custody in Australia to 30 June 2011: twenty years of monitoring by the National Deaths in Custody Program since the Royal Commission into Aboriginal Deaths in Custody. Monitoring Report no. 20. Canberra: Australian Institute of Criminology
  • Marzano L, Fazel S, Rivlin A & Hawton K 2010. Psychiatric disorders in women prisoners who have engaged in near-lethal self-harm: case-control study. British Journal of Psychiatry 197(3): 219–226
  • O’Driscoll C, Samuels A & Zacka M 2007. Suicide in New South Wales prisons, 1995-2005: towards a better understanding. Australian and New Zealand Journal of Psychiatry 41(6): 519–524
  • Pratt D, Piper M, Appleby L, Webb R & Shaw J 2006. Suicide in recently released prisoners: a population-based cohort study. Lancet 368(9,530): 119–123
  • Shaw J, Appleby L & Baker D 2003. Safer prisons: A national study of prison suicides 1999-2000 by the National Confidential Inquiry into Suicides and Homicides by People with Mental Illness. Preston, UK: National Confidential Inquiry into Suicides and Homicides by People with Mental Illness.
  • Shaw J, Baker D, Hunt IM, Moloney A & Appleby L 2004. Suicide by prisoners: national clinical survey. British Journal of Psychiatry 184: 263–267
  • Simpson S 2012. Suicide and suicidal behaviour in New Zealand prisons.
  • Spittal MJ, Forsyth S, Pirkis J, Alati R & Kinner SA 2014. Suicide in adults released from prison in Queensland, Australia: a cohort study. Journal of Epidemiology and Community Health, 68(10): 993–998
  • Wobeser W, Datema J, Bateman B & Ford P 2002. Causes of death among people in custody in Ontario, 1990–1999. Canadian Medical Association Journal 167(10): 1,109–1,113
  • World Health Organization 2007. Preventing suicides in jails and prisons. World Health Organization and the International Association for Suicide Prevention.

About the authors

Matthew Willis is a Research Manager with the AIC.

Ashleigh Baker was formerly a Research Officer with the AIC.

Tracy Cussen was formerly a Senior Research Analyst with the AIC.