The Deaths of Offenders Serving Community Corrections Orders (Trends and Issues in Crime and Criminal Justice No 107)

Research into deaths in custody has shown that people in prison or police custody are much more likely to die than are people of the same age and gender in the broader community. This study finds that offenders serving community corrections orders - particularly parole - have an even higher probability of death than those in prison.

The authors analysed data from Victoria, and found that among an average of around 7000 persons serving community corrections orders there were between 50 and 70 deaths per year since 1991. Between 1995 and 1998, there were 198 deaths - 62 from drug/alcohol causes and 29 from suicide.

This paper identifies the risks faced by these offenders, especially high risk drug and alcohol behaviour. But whose responsibility is the appropriate care? There are significant issues related to duty of care, given that one-third have orders requiring no supervision.

The study focuses in detail on Victoria, but points out that we know very little about the issue, or about the situation in other States and Territories. A case is made for further research in other jurisdictions.

This Trends and Issues is essentially a synopsis of the report of a study undertaken by the authors for CORE¾ the Public Correctional Enterprise, a service agency of the Department of Justice in Victoria. That study was prompted by a Victorian government task force which investigated the nature and extent of suicide, particularly youth suicide, and suggested strategies to prevent or to minimise its incidence. The Suicide Task Force’s report did not specifically refer to the risk of suicide among offenders serving community corrections orders, but did refer to the high incidence of suicide among prisoners.

The task force also identified a number of factors associated with a high risk of suicide in the community generally. The factors included: mental illness, drug and alcohol abuse, antisocial behaviour, unemployment and homelessness. To the extent that these factors may be found in any population of persons serving community corrections orders, the recommendations of the task force are relevant to that population. At all events, CORE successfully sought funding to undertake research to investigate deaths among the client group managed by Community Correctional Services (CCS).

It is to be noted that, even though the primary motivation for the research was suicide prevention, the focus was to be on deaths from all causes among offenders serving community corrections orders.

The methodology followed in the project incorporated both quantitative and qualitative approaches. The data gathering and analysis started with a relatively large statistical review of the total community corrections client population, followed by a number of analyses of populations decreasing in size down to the study of the files of the relatively small number of cases where the cause of death was clearly suicide.

Visits to community corrections centres and interviews with community corrections officers were also undertaken; data were obtained from a number of other jurisdictions; and a literature review was conducted.

The latter revealed that very little had been published anywhere in the world on the subject of deaths in community corrections. Only two such previous research projects have been undertaken, both in Australia. The first of these was by Haege in New South Wales. The report of this project was not formally published but copies have been widely circulated to interested persons. The study is described as "the first accurate count of probation and parole clients who died under supervision", and was based on 128 case histories of offenders who died between July 1982 and December 1983. At that time there was an average of 11,794 offenders under supervision in New South Wales.

Comparing these figures with deaths in the general community, Haege calculated that "the number of deaths in the supervision population is at least...6.4 times greater than one would expect for this same age group in the general community". She also found that parolees seemed to be more at risk than offenders serving other orders.

The only other empirical study was that of Fleming, McDonald and Biles (1992)¾ one of a series undertaken by the Research Unit of the Royal Commission into Aboriginal Deaths in Custody, at the request of the conference of Australian and New Zealand correctional administrators. Its focus was the two-year period, 1987 and 1988. Data collection forms were circulated to all Australian jurisdictions and New Zealand. A total of 394 forms were returned (31 for Aboriginal offenders and 363 for non-Aboriginal offenders) with the largest number coming from New South Wales (206) and Victoria (89). Queensland and Western Australia were unable to supply all of the information requested. It was seen as complementary to other studies completed by the research unit on deaths in prisons.

The rate of deaths for community corrections at that time was 5.6 per 1000 offenders serving orders, per year. The rate for Aboriginal offenders was 5.3 and for non-Aboriginal offenders it was 5.7. (The rates were compared with the rate for all deaths in prisons, then 2.6 per 1000 prisoners per year.) On this comparison, the report stated:

As with the study by Haege, it was found that the rate of death for parolees (15.1 per 1000) was higher than for probationers (10.0 per 1000) and for offenders serving community service orders (2.2 per 1000).

It was established that, on a daily average basis, over the years 1996- 97 and 1997- 98 there were just over 7000 offenders being managed by Community Correctional Services, of whom some 1100 were women. A further 3000 cases were, at any time, either in custody or the subject of breach proceedings.

Of the 7000 plus persons actively serving orders in June 1998, 8.4 per cent were serving intensive corrections orders, 39.7 per cent community based orders, 22.7 per cent fine default orders, 16.3 per cent community work orders, and 12.8 per cent post-prison or parole orders (see Figure 1). (Note that offenders serving fine default and community work orders are not required to report to a community corrections officer for supervision.)

Figure 1: Persons Serving Community Corrections Orders in Victoria, by Order Type, June 1998

The community corrections population of Victoria is significantly different from the general Victorian population, with young males being noticeably over represented (see Figure 2).

Figure 2: Demography of Offender Population Compared to Victoria’s Population aged 17+

There was some difficulty in ascertaining the number of offenders who died whilst serving community corrections orders. CORE documents contain several differing, though overlapping, lists of such persons. A master database was created by the researchers and this showed that, over the period 1991- 92 to 1997- 98 a total of 418 offenders had died from all causes. Of these, 357 were male and 60 female. The annual totals varied between around 50 and 70, with little or no indication of trends over that period (see Figure 3).

Reasonably accurate data on the causes of death for the deceased offenders were available only for 1995- 96 to 1997- 98, and even for this limited period about the accuracy were expressed some doubts. This was particularly the case with suicide, where errors of classification are more likely, but also applied to the latter part of the period when coroners' verdicts were not always known. Nevertheless, as can be seen from Table 1, "drugs/alcohol" was the most common cause of death for both males and females.

The age group with the highest number of deaths was 30 to 39 years, nearly one half of which were due to "drugs/alcohol". For offenders 50 years and over, however, the most common cause of death was "natural causes". There was little difference between males and females in the cause of death for different age groups.

Table 1: Recorded Deaths by Gender and Cause of Death, 1995- 96 to 1997- 98
Cause of Death	Male	Female	Total
Accidents	19	3	22
Natural Causes	24	5	29
Drugs/Alcohol	53	9	62
Violent Incidents	6	1	7
Suicide	25	4	29
Unknown	43	5	49*
Grand Total	170	27	198*

An intensive study of the files of the offender deaths that were classified as suicide over the period August 1995 to July 1997 was undertaken. The researchers aimed to identify any consistent factors or patterns associated with suicide with a view to developing preventive strategies.

A total of 26 files was produced for examination by the research team. After close scrutiny, three of these were excluded as, even though the coroner had found that the offender had caused his or her own death, there was no finding in relation to intention and thus these cases could not lawfully be seen as suicide. For the 23 remaining cases (see Table 2) the most common mode of suicide was hanging and the most common type of order being served was a community based order (CBO). One case of suicide by a parolee was found.

Because of the relatively small number of suicide cases, the search for reliable predictors of suicide was not particularly successful. However, it was observed that some suicides occurred when a period of satisfactory compliance was suddenly followed by a pattern of non-compliance. This may be an early warning sign of risk, particularly where the client has psychiatric or drug problems or where there is an imminent possibility of imprisonment for breach.

These observations could be developed into a hypothesis to be tested against a larger sample of cases in the future.

A number of individual telephone interviews were conducted and a number of group discussions were held with the staff during visits to community corrections centres, to obtain some indication of the views of community corrections officers on preventing the deaths of their clients, and on appropriate responses if death occurs. No attempt was made to obtain a representative sample of all CCS staff, but those interviewed had all had the experience of at least one of their clients committing suicide. Some tentative conclusions were drawn from this exercise.

In the first place, it was apparent that there is deep concern among community corrections officers about their responsibilities with regard to duty of care. The level of concern, which may be based on some degree of misunderstanding, amply justifies this subject being the focus of specific scrutiny later in the project.

Secondly, on a closely related subject, it was clear from these inquiries that community corrections officers need specific training in the role and functions of the coroner's court. The level of ignorance of this subject revealed by these discussions surprised the researchers, but it should be capable of being corrected within a relatively short time.

Thirdly, a theme that emerged clearly from both the interviews and group discussions was a degree of dissatisfaction with the support services available to community corrections clients, especially with regard to psychiatric treatment and the treatment for drug and alcohol abuse. [It is understood that recently these latter services have received increased government funding.] A number of references were made to waiting lists for admission to specialist services, which were seen by the supervising officers as unacceptable when a client was facing a crisis and in urgent need of treatment. Reference was also made to the particular difficulties faced by clients who simultaneously needed both psychiatric treatment and treatment for drug or alcohol abuse.

The fourth and probably most important finding from these visits and interviews was the severe level of distress experienced by the overwhelming majority of community corrections officers when confronted with the news that one of their clients had committed suicide. At one level this could be seen as a welcome finding as it demonstrates that the staff working in this field have a commendable level of personal commitment to their work and to their clients. They are certainly not detached from the emotional aspects of working with people and, in the interests of their own wellbeing, it is essential that the distress which officers experience be recognised and responded to appropriately. [CORE management advises that CCS staff are routinely offered post-trauma counselling, which is optional, following any traumatic event, including an offender death.]

In order to compare the level of all deaths (not only suicides) of offenders in community corrections in Victoria with other jurisdictions, letters were sent to all other Australian and New Zealand correctional authorities requesting basic information. It was recognised that some work would be involved in gathering the requested information as few jurisdictions have offender record systems which readily yield the numbers of offenders serving community corrections orders who die in any specified period.

In the event, statistical information was received from only Queensland, Western Australia, South Australia and the Northern Territory. This was supplemented by the data for New South Wales collected by Haege some 15 years earlier (see Table 3).

Data in this table are not strictly comparable across jurisdictions as different definitions of who is serving a community corrections order may have been applied. In particular, it can be seen that Queensland and South Australia have relatively high numbers of offenders serving orders which must have an impact on the rates of death per 1000 offenders.

Nevertheless, with the possible exception of the Northern Territory, the death rates in all cases are higher than one would expect for offenders in prison.

The data available, together with that obtained in the study by Fleming et al. (1992), enabled a comparison to be made between the death rates of offenders serving post-court orders with those serving post-prison, or parole, orders. The data shown in Figure 4 strongly support the hypothesis that persons serving parole orders, in all of the jurisdictions covered, were notably at greater risk of death from all causes than are offenders serving other forms of community corrections orders. It may be speculated that this difference is due to a more serious level of criminality and a more greatly disturbed lifestyle being associated with imprisonment and its aftermath.

Figure 4: Comparison of Post-Court and Post-Prison Death Rates, by Jurisdiction

Figure 5 uses data from the Australian Bureau of Statistics and the National Deaths in Custody Monitoring and Research Unit of the Australian Institute of Criminology, together with the central findings of this study, to show the death rates from all causes for the general population, for prisoners and for offenders serving community corrections orders by age group.

Figure 5: Death Rates in Victoria, All Causes, Comparison between General Population, Prison Population and CCS Clients, by Age.

It can be seen that for each of the younger age groups, CCS clients are at a much greater risk of death than prisoners or members of the general community. For the age group 55 to 64 years, the differences between the three groups are not great, but prisoners have the highest rate. For persons 65 years and over, however, death rates are much higher in the general community than they are in either of the two correctional populations. A possible explanation for the latter finding is the fact that, comparatively, very few persons aged 65 years or more are in prison or are serving community corrections orders, but all will die at some time and their deaths are more likely to be in the general community which includes, for these purposes, hospitals and nursing homes.

In the complete absence of reported case law on the duty of care in relation to the community corrections population, the principles must be deduced by analogy to those applicable to the custodial population.

At common law the assertion and exercise of custodial control creates a duty of care to the detained person (Howard v. Jarvis (1958) 98 CLR 177). Whether that duty is breached depends on the facts in the particular case (Dixon v. Western Australia (1974) WAR 65; L v. Commonwealth (1976) 10 ALR 269).

The extent of the duty, and thus the question of whether it has been breached, "is tailored to the act and function to be performed" (Knight v. Home Office (1990) 3 All ER 237). For example, a prison's duty in relation to a psychiatrically disordered inmate would be less than that of a psychiatric hospital if that same person were a patient.

The governing principles thus seem to be that: the point on the control continuum will influence whether, and to what extent, a duty of care exists; and the question of breach will be determined in the context of the function to be performed by the correctional authority.

Clearly, then, there is no basis for a duty of care in relation to the 39 per cent (see Figure 1) of the community corrections population who in Victoria receive no supervision. With regard to the remaining 61 per cent¾ serving supervised community based orders, intensive corrections orders or parole¾ a duty of care might exist depending on the nature and extent of the supervision required and undertaken. For example, an ICO requiring participation in an intensive drug rehabilitation pro-gram might in some circumstances create a duty of care going beyond immediate program delivery protocols. But the more intermittent and less intensive the supervision, the less likely it is that the element of control necessary to create such a duty would be held to exist.

In summary, bearing in mind concern about client suicides and overdoses, one can say that a duty of care is exceptional (rather than standard, as with prisoners) and depends on the particular circumstances.

An organisation can assume additional risk and thus extend its potential liability (Pyrenees Shire Council v. Eskimo Amber Pty. Ltd. (1998) 151 ALR 147). The administrative practices of CCS in some cases seem to have come close to bringing about such an outcome.

One factor pushing CCS in this direction has been coronial over-sight. The authors have observed in other circumstances that, increasingly, there is a tendency for inquests to go beyond the formal issue of ascertaining the cause of death to the stage of evaluating professional standards or working practices of organisations involved with the deceased.

CORE was mindful of this trend in introducing the requirement that a death report be compiled for all cases involving community corrections clients.

The other factor is that of public administration standards generally. Under its statute CORE is required "to supervise and manage persons under its control in a safe, secure, humane and just manner", and is subject to audit according to this criterion. This contributes to a tendency to go beyond what common law principles of liability would require.

Deaths in the community corrections population predictably continue to exceed those amongst both the prison and the comparable general populations. Preliminary analysis enables identification of some higher risk groups and situations. Duty of care issues are complex, and have not yet been fully worked through. Public administration standards are muddying the waters of strategy and responsibility. The capacity of community corrections regimes to become an effective source of diagnosis and prevention is limited.

Nevertheless the issue is one of legitimate concern. There would be benefit in corrections authorities collaborating to produce a definitive national database which, after analysis, would enable a considered debate to take place as to their appropriate role in this area. The likelihood is that, with the correctional dollar increasingly being allocated to prisons, their capacity to absorb additional responsibility will be very limited.

This paper is published with the approval and encouragement of CORE¾ the Public Correctional Enterprise, a service agency of the Department of Justice in Victoria. The authors would like to thank the management and staff of CORE, particularly Ann Reilly and Barbara Thorne, for their advice and assistance with this project.

Professor David Biles is a Consultant Criminologist and Professorial Associate of Charles Sturt University. Professor Richard Harding is the Director of the Crime Research Centre of the University of Western Australia. John Walker is a Consultant Criminologist and Associate of the Australian Centre for Security Research, University of Western Sydney. All three are former senior members of staff of the AIC.

Mode	CBO	CW	FD	Parole	Total
Hanging	4	2	3	1	10
Carbon monoxide	3	1	0	0	4
Other poisoning	1	0	1	0	2
Gunshot	1	0	1	0	2
Jumping	1	0	0	0	1
Cut Wrist	0	1	0	0	1
Unknown	0	2	1	0	3
Total	10	6	6	1	23

State/Territory	Years Covered	No. of Deaths	Average Client No.	Rate/1000 Clients
New South Wales	1982-83	85	11,794	7.2
Victoria	1997-98	59	7,081	8.3
Queensland	1997-98	95	14,776	6.4
Western Australia	1997	32	5,900	5.4
South Australia	1997-98	36	8,375	4.3
Northern Territory	1996-98	7	560	2.9