DEATHS IN CUSTODY

Final Report

Submitted to the Office of the Correctional Investigator by:
Thomas Gabor, Ph.D.
Professor of Criminology
University of Ottawa

February 28, 2007

TABLE OF CONTENTS

ACKNOWLEDGMENTS

3

1.0 INTRODUCTION

4

2.0 METHODOLOGY

6

3.0 BASIC VICTIM AND INCIDENT CHARACTERISTICS

7

4.0 FINDINGS

9

5.0 SUMMARY AND CONCLUSIONS

21

APPENDICES

22

2

ACKNOWLEDGMENTS
The author would like to express his profound gratitude for the guidance and support of
staff members at the Office of the Correctional Investigator. He is especially grateful for the
efforts and support of Dr. Ivan Zinger, Mr. John Reid, and Mr. Ed McIsaac.

3

1.0 INTRODUCTION
The Office of the Correctional Investigator (OCI) investigates and attempts to resolve
complaints from individual offenders under federal jurisdiction. In addition, it has a
responsibility to review and make recommendations on the policies and procedures of
Correctional Services Canada (CSC) that relate to individual complaints. In this way, systemic
areas of concern can be identified and appropriately addressed.
Over the last decade, OCI has become increasingly concerned about the high number of
deaths and self-inflicted injuries in federal institutions. In his last Annual Report (2005-06), Mr.
Howard Sapers, Correctional Investigator of Canada, stated that his office was especially
concerned about the number of similar recommendations made year after year by CSC’s national
investigations, provincial coroners, and medical examiners. He also expressed concern about the
ability of the Correctional Service to implement these recommendations on a national level. In
his report, the Correctional Investigator undertook to conduct a comprehensive review of reports
and recommendations dealing with deaths in custody and other matters. Mr. Sapers noted that,
in order to reduce the number of fatalities, a timely and systematic follow-up on corrective
actions was required to ensure that preventive measures are implemented.
It is in this context that the present project was undertaken. Specifically, this project
examined all reported deaths, due to other than natural causes, occurring over a five-year period
in Canadian federal correctional institutions. Thus, all deaths determined by CSC to be
homicides, suicides, and accidents have been included, with the aim of identifying areas in which
improvements might enhance CSC’s ability to prevent or respond to assaults and attempted selfinjury in the future. The goal was to move from the analysis of incidents, on a case-by-case
basis, to an overall assessment of trends and patterns.
Key questions posed by this study included:
1. Do the Boards of Investigation (BOIs) and Coroners’ reports reveal a pattern of
shortcomings on the part of CSC staff or policies?
2. Do more recent incidents indicate that CSC has improved its capacity to prevent and
respond to assaults, acts of self-injury, and accidents?
3. How does CSC respond to the findings and recommendations of BOIs and Coroners?
4. How expeditiously does CSC respond to deaths in convening investigative boards and in
acting upon their recommendations?
5. Could some of the fatalities have been foreseen and possibly prevented?

4

It is important to note that some investigative reports commend CSC staff for exemplary
behaviour in highly stressful circumstances. These BOI reports occasionally note that the
manner in which institutional personnel responded to an incident should constitute a “Best
Practice” that should be emulated nationwide. The focus of this report, however, was on the
identification of systemic issues that have compromised the prevention of, or response to, fatal
assaults and acts of self-injury.
It is also important to acknowledge that a study confined to fatal incidents may possess
an inherent bias. As this is a study of people who succeeded in ending their lives, in harming
others, or who were victims of tragic accidents, this report does not include those incidents in
which lives may have been saved through actions by CSC personnel that were in full compliance
with existing policies and procedures. Thus, it may be that those cases resulting in deaths
involve a disproportionate number of compliance issues.
Notwithstanding the possibility that incidents resulting in fatalities may reveal more
issues with existing practices than non-fatal ones, the number of non-natural fatalities—which is
far in excess of the rate in the civilian population--deserves attention in its own right.
Institutional homicide and suicide rates, when combined, are calculated here to be nearly eight
times the rates found in the population as a whole. 1 This situation underscores the urgency of
identifying those areas in which actions have fallen short of the optimal across a variety of
institutional settings.

1

According to Statistics Canada, the national suicide rate for 2003 was 11.9 per 100,000 Canadians. In 2003-2005,
the homicide rate in Canada was in the range of 2 incidents per 100,000, resulting in a combined rate of
approximately 14 incidents per 100,000. The present study reveals an annual average of 13.6 homicides and
suicides in federal institutions from 2001-2005. According to CSC’s Research Branch, there were 12,561 men and
women in custody in Canadian federal institutions on any given day in 2004/2005, yielding a rate of 108 deaths by
suicide or homicide per 100,000 inmates—nearly 8 times the rate in the civilian population.

5

2.0 METHODOLOGY
This study examined 82 reported suicides, homicides, and accidental deaths in custody from
2001 to 2005 (inclusive). The cause of death in each case was determined by CSC; however,
where a Coroner’s Office subsequently concluded that a death was due to some other factor, this
was then deemed to be the more definitive cause of death. The study reviewed Board of
Investigation reports, CSC Action Plans, Coroner’s Reports, correspondence between CSC and
both OCI and Coroners’Offices, and other documents pertaining to each fatality. The cut-off
date for inclusion of documents in the study was November 1, 2006. Incidents occurring during
the study period were not included if BOI reports were not available by that date. Incidents were
included, however, where Action Plans or Coroner’s reports were unavailable.
A coding instrument was developed to guide the process of recording information from each
file (see Appendix A). Information drawn from each file included:
•
•
•
•
•

The cause of death;
The institution in which the incident occurred;
Basic information about the victim (age, gender, Aboriginal status);
Current offences and criminal history;
Relevant dates, including those of the incident, admission to institution, parole eligibility,
completion of the BOI report, and CSC’s national or regional response to the report;
• Pre-indicators of the incident and risk factors (e.g., substance abuse, previous suicide
attempts, mental health issues, family support, institutional history); and,
• Relevant findings and recommendations contained in BOI and Coroners’ Reports, as well
as CSC’s response to these.
The issues raised in the findings and recommendations of BOI and Coroners’ Reports
were placed in one of 15 categories (Appendix B). These categories included: 1) post-incident
emergency care and resources; 2) counts and patrols; 3) mental health issues, programming, and
suicide prevention; and 4) security issues. The categories were developed on the basis of a
preliminary review of 15 files and through consultation with OCI personnel. The study also
noted whether a recommendation was directed to CSC Headquarters, regional authorities, or
institutional officials. CSC’s responses to findings of non-compliance and recommendations
were placed in one of five categories:
1.
2.
3.
4.
5.

Agree: No Action
Agree: Action Consistent
Agree: Action Inconsistent
Disagree
Recommendation Ignored

Some interpretation was required in those instances in which CSC Action Plans did not explicitly
state whether the Service agreed or disagreed with a finding of non-compliance or with a
recommendation.

6

3.0 BASIC VICTIM AND INCIDENT CHARACTERISTICS
Table 1 shows that over 60 percent of the deaths in custody examined in this project were
suicides and that the remainder was almost equally split among homicides and accidents.
Almost nine of every ten suicides were by hanging, whereas stabbings accounted for over half of
all homicides and drug overdoses accounted for 80 percent of the unintentional deaths. All but
one incident involved men and three of every ten victims were under 30 years of age. Aboriginal
persons accounted for more than a fifth of the victims.
Inmates who had been involuntarily transferred seemed to be especially at risk. In fact,
one-fifth of the victims had spent less than 30 days at the institution at which their death
occurred. There were other cases in which an impending transfer, the denial of parole, an
unsuccessful appeal, or the loss of a significant other played a role, reinforcing the idea that
transitions and other critical events in their lives can elevate the risk of self-injury and other
types of harm.
More than 90 percent of the victims possessed a criminal record for a prior offence as an
adult and/or juvenile, and over 90 percent were serving their last sentence for violent crimes.
Almost a third of the victims were serving a life sentence and over half had passed their full
parole eligibility date at the time of their death.

7

TABLE 1 – Key Characteristics of the Victims and Incidents (N=82)
VICTIM/INCIDENT
CHARACTERISTIC

% OF ALL CASES

Homicides

20.7

Suicides

61.0

Accidents

18.3

Males

98.8

Under 30 years of age

29.3

Aboriginal descent

22.0

Incident occurred within 30 days of
admission to that institution
Possessed a criminal record prior to
most recent offence(s)
Most recent offence involved
violence
Was serving life sentence

20.7

Passed full parole eligibility date

51.2

93.9
92.7
32.9

8

4.0 FINDINGS
Finding #1 - Several Concerns Are Raised Repeatedly by Investigative
Boards and Coroners in a Significant Number of Death in Custody Cases
Table 2 displays the number and proportion of cases in which various concerns have been
raised by BOIs or Coroners in their findings of non-compliance with existing practices or
policies, or in their recommendations.
a) Post-Incident Medical/Emergency Care, Resources, & Decontamination
In almost two-thirds of the cases, some shortcoming was noted by a BOI or Coroner in the
response of personnel to the emergency, in the adequacy of emergency resources, and/or in the
decontamination of the area surrounding the victim in the immediate aftermath of the incident.
The concern that was perhaps raised in the largest number of cases related to the failure of
officers to perform Cardio-Pulmonary Resuscitation (CPR), or a delay in so doing, upon the
discovery of an inmate without any apparent vital signs. Delays in taking other actions (e.g.,
notifying health personnel or emergency responders) were also noted in a number of cases.
Questions were raised about the adequacy of the training of officers in the administration of CPR
and in the prevention of contamination through body fluids during its administration. In many
cases, for example, officers did not wear protective masks. Overall, officers often appeared
uncertain as to what to do when a body was discovered.
Additional shortcomings noted in the post-incident emergency response included:
• The inadequate decontamination of cells or other areas in which fatalities occurred;
• The absence of on-site defibrillators;
• Concerns about the quality of emergency care and nursing staff in several institutions,
especially on the night shift;
• The inaccessibility of emergency supplies in institutions.

9

Table 2: Issues Raised in the Findings and Recommendations of Boards of Investigation
and Coroners Following Deaths in Custody (N=82)

ISSUE

# OF CASES

% OF CASES

Post-Incident Medical/Emergency Care,
Resources, & Decontamination

54

65.9

Recordkeeping and Information Sharing
Among Staff Within Institutions

43

52.4

Security Practices, Video Surveillance,
and Evidence Gathering

42

51.2

Patrols, Counts, and Live Body
Verification

36

43.9

Mental Health Issues, Programming, and
Suicide Prevention

36

43.9

Availability of Illicit Drugs &
Paraphernalia, & Monitoring of
Prescription Drugs
Post-Incident Stress Management
Services for Staff & Inmates

21

25.6

18

22.0

Post-Incident Family Concerns—
Notification, Personal Effects, Funeral
Arrangements
Information Sharing Between
Institutions

11

13.4

8

9.8

Inmates’ Institutional Placement and
Security Classification

6

7.3

Pre-Incident Medical Care and
Resources

3

3.7

Private Family Visits—Screening of
Visitors and Security Procedures

3

3.7

Expeditious Resolution of High Priority
Grievances

3

3.7

Prevention of and Response to Prison
Disturbances

3

3.7

Other Issues

21

25.6

10

b) Recordkeeping and Information Sharing Within Institutions
In more than half the cases, issues were raised about the failure of institutional staff to
record relevant medical or mental health information on the inmate’s file or to otherwise share
such information with others working with an inmate. In a number of files, BOIs and Coroners
noted that poor communications existed between health care or psychological personnel and
those involved daily with inmates (correctional officers, members of the case management team).
Information on stresses experienced by the inmates or threats against them often were not shared
with other personnel. Mental health interventions and previous suicide attempts, which might
have resulted in closer monitoring, were sometimes not noted in the inmate’s file. In one case,
the failure of an inmate who committed suicide to pick up his anti-depressant medication for
three days was not shared with his case management team. In another case in which an inmate
committed suicide following an unsuccessful appeal of his conviction, there was no alert in the
Offender Management System (OMS) despite a history of self-harm, suicide attempts (including
one following a previous unsuccessful appeal), and a history of substance abuse.
Recordkeeping issues also arose in relation to officers’ rounds. The recording of these
rounds appeared to be inconsistent, making it difficult to ascertain their precise timing and
frequency in the post-incident investigation. In one homicide, officers sensed that an incident
was going to occur and that they were being observed by inmates; however, they failed to record
this in the log book. With regard to contraband, there was no alert in one case that a deceased
inmate’s wife had previously triggered a drug detection device. In a subsequent visit, the inmate
choked on a bag of drugs he had swallowed to avoid detection. In addition, health care
personnel were not always informed when searches of an inmate’s cell revealed substantial
quantities of prescription drugs that had been dispensed by health staff.
c) Security Practices, Video Surveillance, and Evidence Gathering
In over half the cases, BOIs and/or Coroners raised some security concern(s) or issue(s)
pertaining to the collection or preservation of evidence. Concerns were raised repeatedly about
the quality and coverage provided by video cameras. Evidence was sometimes lost or of low
quality and video surveillance often was found to be inadequate in ranges, living units, and
recreational areas. In a number of cases, cells were obscured by privacy panels, curtains, and
mesh, making it more difficult to verify the condition of an inmate during patrols.
Inmate movements were sometimes found to be poorly controlled, especially during
recreational activities. A number of homicides occurred in a gym area and, in one case, a metal
bar was removed from a weight room and used as the murder weapon. In another case, an
inmate was so intoxicated he fell to his death over a railing. The Coroner’s report noted that
closer supervision during free time was required so that a state of intoxication of such severity
could not go unnoticed.

11

A number of killings and suicides were gang-related. Killings could be due to inter-gang
rivalries or to discipline within a gang. Suicides had occurred due to gang-related pressures on
inmates. The lack of anti-gang strategies or of trained security intelligence analysts was
mentioned in several files.

d) Patrols, Counts, and Live Body Verification
In just under half the cases, issues were raised regarding some aspect of patrols or counts,
including the failure of correctional officers to ensure that inmates were still alive in their cells.
The main concerns related to range walks that may not have been done or counts that were done
improperly. In several cases, questions were raised as to whether national protocols were needed
with regard to rounds and counts in Native centers. Issues concerning the proper functioning of
the Silverguard guard monitoring system were also mentioned in several files.

e) Mental Health Issues, Programming, and Suicide Prevention
In nearly half the cases, BOIs and/or Coroners raised concerns about the services
available to inmates with mental health issues and to those with a history of self-injury. In a
number of cases, there were suggestions that more could have been done to assist individuals
with well documented records of self-harm and suicide attempts. In a number of files, questions
were asked about the competence of clinical personnel and about the quality of assessments as to
the mental state of an inmate or the degree to which he was at risk of committing suicide. In
several cases, psychologists or psychiatrists wrongfully believed that suicide threats, suicidal
ideation reported by the inmate, or abnormal behaviour were nothing more than malingering or
manipulation on the part of the inmate. In one case, previous suicide attempts, declining mental
health, and knowledge that the inmate was giving away his possessions did not lead the
psychologist who assessed him to view the inmate as one at an elevated risk to commit suicide.
Some institutions reportedly do not have a multidisciplinary mental health team to assess
inmates during intake, to deal with suicidal inmates, and to provide input into placements into,
and releases from, mental health ranges. In several cases, it was noted that suicide watch was de
facto segregation and did little to respond to the inmate’s mental health needs. There were also
many references to structural modifications in cells and shower rooms that would make it more
difficult for inmates to commit suicide by hanging, the suicide method in close to 90 percent of
the cases.

f) Availability of Contraband
In a quarter of the cases, BOIs and/or Coroners expressed concerns about the availability
of illicit drugs and related paraphernalia, as well as about the administration of prescription
drugs. References were made to the ease of bringing drugs into institutions and, in several cases,
to the role played by spouses and girlfriends during family visits. It was noted in several files
that an inmate had played a role in the institutional drug trade or “culture”. The diversion of

12

methadone, following administration, was mentioned as one problem. Several victims had been
found to be in possession of a large quantity of illegal substances, prescription drugs, and drug
paraphernalia. One individual was reported to have consumed $1,000 worth of heroin daily.
One Coroner lamented that it is difficult to talk about the rehabilitation of inmates when inmates
have access to illicit substances.

g) Post-Incident Stress Management
Following more than a fifth of the incidents, officers and/or inmates were not offered
services to deal with the stresses associated with a death in their midst. In some of these cases,
stress management services were offered but they were not offered promptly.

h) Other Issues
A smaller number of cases involved such issues as post-incident family concerns, the failure
to transfer critical information about an inmate who was transferring to another institution, the
inappropriate placement of an inmate, pre-incident medical care, security procedures during
private family visits, the failure to resolve an inmate’s grievance in a timely fashion, and issues
in the prevention of, and response to, prison disturbances.

Finding #2 – There is No Evidence that Correctional Services Canada has
Improved its Overall Capacity to Prevent or Respond to Deaths in Custody
During the Five-Year Study Period
Finding #1 indicates that some issues continue to surface in death in custody cases. It is
plausible that CSC has improved its capacity to prevent and respond to these cases over time.
Two indicators were used to determine whether this was the case.
First, the number of non-natural deaths was compared over the five-year study period.
Table 3 shows that there was a spike in deaths in 2003. This spike was followed by a number of
fatalities in the final two years of the study period (2004-2005) that actually exceeded the
number of deaths in the first two years (2001-2002). Furthermore, there was at least one
additional case in 2005 that was excluded from the study, as the BOI report was unavailable as of
November 1, 2006, the present study’s cut-off date for including documents.
Table 3 – Deaths in Federal Institutions,
2001-2005 (N=82)
YEAR
# OF DEATHS
2001
2002
2003
2004
2005

15
14
23
14
16

13

A second measure of whether CSC’s responses to deaths in custody were becoming more
effective over time involved a comparison of incidents occurring in the last two years of the
study period, with those taking place during the entire study period. Thus, did the BOI and
Coroners’ Reports raise fewer concerns following the incidents in 2004-2005 than from 20012003?
Table 4 indicates that there is no reason to believe that these concerns have diminished
over the five-year study period. In fact, five of the six main categories of concerns were raised
more often in the last two years of the study period than in the entire study period. Thus, in 2004
and 2005, post-incident emergency care and related issues were raised in three-quarters of the
cases. Concerns with security practices were raised in 60 percent of the cases, while
recordkeeping and information sharing, as well as patrols and counts, were mentioned as
concerns in over half the cases. Issues relating to the availability of contraband were raised in a
third of the cases in 2004-2005, as opposed to a quarter of the cases over the entire study period.
It was only in relation to mental health issues that concerns were not raised more frequently in
the last two years than in the study period as a whole.
Therefore, neither the analysis of the number of deaths from 2001-2005, nor the study of
concerns raised by BOIs and Coroners, supports the assertion that fatalities have diminished over
time or that CSC has responded more effectively, in 2004-2005, in the core areas frequently
identified by BOIs and Coroners.

Table 4 – A Comparison of Issues Raised by Boards of Investigation and Coroners in
Fatalities Occurring in 2004-2005 with Cases Occurring During the Entire Study Period
ISSUE
% OF CASES
% OF CASES
(2004-2005)
(2001-2005)
Post-Incident Medical/Emergency Care,
76.7
65.9
Resources, and Decontamination
Security Practices, Video Surveillance,
and Evidence Gathering

60.0

51.2

Recordkeeping and Information Sharing
Among Staff Within Institutions

56.7

52.4

Patrols, Counts,
Verification

Body

53.3

43.9

Mental Health Issues, Programming, and
Suicide Prevention

43.3

43.9

Availability of Illicit Drugs
Paraphernalia,
&
Monitoring
Prescription Drugs

33.3

25.6

and

Live

&
of

14

Finding #3 – Correctional Services Canada Tended to Act on the Findings and
Recommendations of Boards of Investigation, but Often Disagreed With, or
Took No Action on, Coroners’ Recommendations
Table 5 illustrates that, in nearly three-quarters of the cases in which a finding of noncompliance or recommendation was made by an investigative board, CSC, either nationally or at
the regional level, responded in a manner consistent with that finding or recommendation. Thus,
it can be said that BOI findings and recommendations were treated seriously. In some cases, an
Action Plan reported that the appropriate action had already been taken, while in other cases it
was reported that a directive had or would be issued. It is beyond the scope of this project to
verify whether concrete actions, consistent with a finding or recommendation, had actually been
taken and whether such actions were sustained over time and applied regionally or nationally,
when the recommendation urged that this be done. This analysis is based on the assumption that
statements made on CSC Action Plans and regional responses are accurate.
In another 8 percent of the cases, the Service agreed with a BOI finding or
recommendation but either took no action or, in a few rare cases, took an action that was
inconsistent with a finding or recommendation. CSC ignored 11 percent and disagreed with 8
percent of BOI findings and recommendations.
CSC was more likely to resist or to fail to act on Coroners’ recommendations. The
Service complied with just over a third of the recommendations. CSC asserted that it disagreed
with nearly another third of the recommendations. The Service agreed with, but took no action
on, another 27 percent of the Coroners’ recommendations.
When BOI and Coroners’ findings and recommendations are combined, CSC agreed with
and acted consistently on these in over two-thirds of the cases. This finding begs the question as
to why similar issues continue to arise in fatal incidents in federal institutions if, in most
instances, the Service takes concrete measures to improve its capacity to prevent and respond to
acts of self-injury and assaults. Further investigation is required to assess the implementation of
BOI and Coroners’ recommendations.

15

Table 5 - Correctional Services Canada’s Responses to Boards of Investigation and
Coroners’ Findings and Recommendations Following Deaths in Custody
RESPONSE

TO BOI
REPORT

%

%

ALL
RESPONSES

%

6.5

TO
CORONERS’
REPORTS
18

Agree: No
Action

24

27.3

42

9.6

Agree: Action
Consistent

270

72.6

25

37.9

295

67.4

Agree: Action
Inconsistent

7

1.9

0

0

7

1.6

Disagree

30

8.1

21

31.8

51

11.6

Finding/Recommendation
Ignored
Total

41

11.0

2

3.0

43

9.8

372

100.1*

66

100.0

438

100.0

* Column does not equal 100% due to rounding error

Finding #4 – Typically, A Significant Period of Time Elapses Between An
Institutional Fatality and the Adoption, by Correctional Services Canada, of
Formal Measures to Address Issues Arising From It
Table 6 displays the amount of time elapsing from the date a fatality occurred to the
completion of the BOI Report and the formal approval of remedies at the national or regional
level. A small number of cases were excluded from these analyses where the inmate’s file was
unclear as to the date of the submission of a BOI report or the date of the adoption of an Action
Plan by CSC’s Executive Committee.
The table shows that an average of 165 days (over 5 months) elapsed between an incident
and the completion of the BOI report. An average of another 10 months elapsed between the
completion of the BOI report and the review of remedies by CSC’s Executive Committee or at
the regional level. Thus, it took an average of nearly 16 months following fatalities to formally
adopt measures to address issues arising from the incidents. The 16-month figure underestimates
the true length of time. As there was no Action Plan in 14 of the 75 cases in which clear dates
were available from the files, the study cut-off date of November 1, 2006 was used to calculate
the promptness of the Service’s response. It is possible that, in some of these 14 cases, it may be
many additional months before an Action Plan is drawn up and approved.

16

Averages are deceptive in that they may obscure extreme values. In 15 of the 75 cases
with clear dates available in the files, Action Plans were approved within six months of the
incident. However, on the other extreme, 10 cases were not resolved until at least two years after
the incident and 7 of these cases did not receive the approbation of senior officials for at least
three years following the incident.
Table 6 – Average Number of Days Elapsing Between Deaths in Custody and
Key Phases of Correctional Services Canada’s Responses
FROM INCIDENT TO FROM BOI REPORT FROM INCIDENT TO
SIGN-OFF OF THE
TO APROVAL OF
APPROVAL OF
BOI REPORT
ACTION PLANS
ACTION PLANS
165 days

310 days

475 days

Finding #5 – It is Likely that Some of the Deaths in Custody Could Have Been
Averted Through Improved Risk Assessments, More Vigorous Preventive
Measures, and More Competent and Timely Responses by Institutional Staff
It is difficult to say, with any certainty, that a particular fatality could have been
prevented had institutional staff performed some action differently. The life of an inmate bent on
suicide, for example, may be saved as a result of the actions of vigilant and competent staff, only
to be extinguished in subsequent attempts. Thus, in some cases, staff acting in an optimal
manner and in compliance with all existing procedures, cannot avert that which appears to be
inevitable.
Notwithstanding this point, to suggest that none of the deaths could have been prevented
would reflect a fatalism that would be an enormous impediment to an improvement of practices
in any system. It would also ignore the fact that many people may have a history of suicide
attempts and eventually desist from actions that are self-injurious. Also, the fatalistic notion that
nothing can be done to prevent suicides and homicides ignores the impulsive nature of many of
these acts. There is much evidence in the behavioural sciences that rage and despair leading to
these extreme acts are often transient and may quickly dissipate following unsuccessful suicide
or homicide attempts. Therefore, those responsible for the care of individuals at high risk of
harm ought to strive to prevent as many incidents as possible, to analyze them, and to implement
constructive remedies.
It is beyond the scope of this project to arrive at a figure representing the number of
incidents that might have been prevented had institutional staff, in each case, responded in an
exemplary manner, had all resources been in place, and had all preventive actions possible been
undertaken. In many cases, however, it is clear that the Service fell short in implementing its
own policies and practices, and in doing everything possible to avert a fatality. Table 2 indicates
that BOIs and Coroners raised a number of concerns in the majority of cases. One might infer
from this that a number of deaths might not have occurred had CPR been administered more

17

promptly, had officers received better training in First Aid, rounds and counts been done
properly and on time, crucial information shared with mental health or front-line staff, mental
health assessments and placements done more prudently, and so on.
In fact, in a few cases, a BOI or Coroner/Medical Examiner indicated that a particular
action or omission on the part of institutional staff may well have contributed to a fatality. In
one suicide case, the Board noted that there was no evidence the inmate should have remained in
segregation after the fifth day review and noted that this may have contributed to the suicide.
The inmate had a history of suicidal behaviour, substance abuse, impulsivity, and had no stable
emotional relationships. At the time of his suicide, he was hearing voices, behaving strangely,
and pacing in his cell. Segregation was thought to exacerbate his condition.
In a case involving a drug overdose, the BOI noted that the staff did not ensure that a live
body was counted during the 12:30 and 4:30 counts. The Board further asserted that if the
proper live body verification had been done, the outcome might have been different.
In one homicide case, the institution was at heightened readiness for a disturbance due to
escalating tension. The disturbance nevertheless ensued and the murder, which was enabled by
the disturbance, occurred. In another homicide, staff viewed the assault but failed to intervene
effectively. A spray (likely pepper spray) was used to neutralize the assailant but was
ineffective. The BOI noted that there is no standard national training in managing these types of
situations. It further noted that there was also a delay in calling the ambulance as none of the
staff on duty at the time knew the emergency number.
In a number of suicide and drug overdose cases, Boards and Coroners have commented
on the ease with which inmates can access drugs and alcohol. In several files, it was mentioned
that the diversion of methadone and prescription drugs was a chronic issue at some institutions.
One Coroner’s report dealing with an inquest into a methadone overdose mentioned that inmates
were adept at transferring contraband to one another, even in segregation. In addition, several
accidental deaths were related to the smuggling and trafficking of contraband. Two inmates
choked to death on a bag of drugs brought in during a family visit. In one of these cases, the
spouse had previously triggered a drug detection device, yet no alerts were present on the
Offender Management System. One highly intoxicated inmate fell to his death over a railing.
Several other suicides and homicides were linked to the institutional drug trade and the
accumulation of drug-related debts.
A number of inmates who committed suicide were deemed to be malingering or, for other
reasons, were not viewed as being at risk, despite previous attempts, substance abuse problems,
and mental health issues. In one case, the inmate had this profile, in addition to giving away
some of his possessions. Nevertheless, the psychologist who assessed him did not recommend
that monitoring of the inmate be increased. In other cases, inmates with a history of self-injury
made statements about their intention to commit suicide or were distraught about some matter
and no additional monitoring was ordered. In several suicide cases, BOIs and Coroners
commented on structural elements in cells that may have facilitated a suicide.

18

Several cases were marked by a virtual comedy of errors. In one case, an inmate who
died of an accidental overdose was known to have overdosed on two previous occasions. His
death was discovered slowly due to the failure of officers to conduct a proper count and the
failure of staff to notify officers that he did not report to work. Upon discovery of the body,
officers did not conduct CPR. In another example, this time a homicide case, officers sensed
something was wrong and that they were being observed by inmates, while conducting their
rounds. They took no action and made no record of this event. The homicide ensued and there
was a delay in discovering the body and in notifying the police. In another homicide case,
outdated health care facilities precluded the treatment of an inmate who had been assaulted. The
inmate did not appear to have life-threatening injuries. In addition, officers failed to call an
ambulance promptly and also did not deliver first aid.
Gang activities were considered to have played a role in a number of homicides and
suicides. In eleven of the cases (13.4%) and close to half of the homicides, the file clearly
indicated that the incident was gang-related. Several gang-related killings occurred just hours
after a transfer to an institution with obvious incompatibles. In one case of a gang member who
was murdered within hours of an institutional transfer, there were no notes on the transfer
documents to the effect that there were two incompatible inmates in the receiving institution.
Apart from the case material discussed above, there is additional evidence that the
victims of homicides, suicides, and fatal accidents in custody may constitute a high-risk subpopulation within the federal correctional system. Overall, more than half of the files indicated
that there were proximal or more long-term pre-indicators of the event.
Although the information on federal inmates available to this project is limited, some
comparisons and observations can be made in relation to the largest group, the suicide victims.
The vast majority of suicide victims displayed a history of substance abuse (91.8%); had
previously attempted suicide (82.2%); and had previous institution infractions, escapes or
violations of conditional releases (68.0%).
In addition, mental health issues pertaining to suicide victims were discussed in almost all
files reviewed. A complete breakdown and assessment of the types of mental health issues was
not undertaken in this study, as the files reviewed consisted of a combination of observations,
symptoms, and/or mental health diagnoses. The CSC conceded in its recent Mental Health
Strategy that its intake assessment of the mental health of offenders upon admission is
inadequate (e.g., it consists of a few questions on such matters as previously psychiatric
hospitalizations and prescriptions for psychotropic medications). Only offenders with evident,
very serious mental health issues or symptoms are referred for a more comprehensive
psychological assessment by either a psychologist or a psychiatrist.

19

The lack of comprehensive mental health assessment at intake hinders the ability of CSC
to better identify those at risk of committing suicide, as well as those at risk of attempted suicide
or self-injury. A sound and comprehensive mental health intake assessment is required for CSC
to implement a more effective suicide and self-injury prevention strategy. The CSC’s Mental
Health Strategy also calls for significant investments in the care, treatment and support of
offenders in custody with mental health issues. Without a comprehensive intake assessment and
adequate mental health services, care and support, some offenders will continue to fall through
the cracks.

20

5.0 SUMMARY AND CONCLUSIONS
This project reviewed all deaths from 2001-2005, occurring in federal custody and
deemed to be due to other than natural causes. During that period, all cases in which a Board of
Investigation had been convened and submitted its report were included. Therefore, in statistical
terms, this was a study of a “population” rather than a “sample”. Just over 60 percent of the 82
cases examined were suicides. Homicides and accidents (usually drug overdoses) each
accounted for approximately one-fifth of the total. All but one of the cases involved men. Many
of the victims were young men—30 percent were under the age of 30—and over a fifth was of
Aboriginal descent. A disproportionate number of incidents occurred following an institutional
transfer. Over 90 percent of the victims had a criminal record prior to the offence for which they
were incarcerated at the time of their death. Over nine out of ten were serving sentences for
violent offences. Close to a third were serving life sentences and more than half were past their
full parole eligibility dates.
The present study provides strong support for the belief that certain issues continue to
arise in fatalities occurring in federal institutions. Concerns relating to post-incident emergency
care, recordkeeping and information sharing within institutions, and various security matters
were raised by BOIs and/or Coroners in more than half the cases. Concerns relating to mental
health programming and suicide prevention, as well as with counts or patrols, were raised in over
40 percent of the cases. Issues relating to the control of illicit or prescription drugs and those
dealing with post-incident stress management also surfaced in a significant number of cases.
There was no indication that these problems have abated over time. Specifically, this analysis
showed that they arose as frequently, and sometimes more so, over the past two years. In
addition, the annual number of fatalities is not declining.
These findings support the concern that, overall, Correctional Services Canada is not
incorporating into current practices, the lessons that can be learned from previous incidents.
Further investigation is needed to understand the impediments to reform in order to minimize the
number of fatalities occurring in custody.
A review of CSC’s Action Plans, the Service’s formal response to Board of Investigation
findings and recommendations, does suggest that BOI reports are taken seriously, as CSC’s
Executive Committee and Regional officials usually agree with and issue directives that are
consistent with BOI recommendations. However, further investigation is required to determine
whether corrective action is actually implemented nationally or regionally and whether it is
sustained, as similar problems continue to persist.
The analysis of CSC’s responses to Coroners’ recommendations suggested that the
Service disagreed with, ignored, or failed to take any action in relation to more than 60 percent
of these recommendations. Further investigation is required to understand the rejection of such a
high proportion of Coroners’ recommendations. The development of a dialogue between CSC
and Coroners’ offices appears to be warranted. It should be noted that Coroners’ reports often do
not contain recommendations and, when recommendations are made, they are usually few in
number. Therefore, these offices cannot be accused of dispensing an excessive amount of advice
to correctional officials.

21

The present study also found that the time elapsing between a fatality and the formal
response of the correctional system is considerable. On average, this figure is 16 months,
although some cases take more than three years to resolve. This matter has been the subject of
many correspondences between the Service and the Office of the Correctional Investigator.
Every effort needs to be made to expedite the investigation of incidents and the response to them.
There are indications that some of the fatalities occurring in the past five years might
have been prevented. Some BOI reports suggest that the outcome of several cases may have
been different had institutional staff discharged their duties as required. In some cases, staff
failed to comply at a number of levels. There were serious errors made in assessing the suicide
risk of several victims and gross errors on the part of medical staff in responding to emergencies.
First responders (often correctional officers) often did not know what was expected of them and
frequently failed to administer first aid. Emergency medical resources were often unavailable,
especially at night. One major concern has been the absence of Automatic External
Defibrillators as standard equipment in institutions. Vital information pertaining to an inmate’s
propensity to self-injure or risk of assault often went unrecorded or unshared, both within and
among institutions. More than half the files indicated that there were pre-indicators to the
incident. Furthermore, many of the victims appeared to be at higher risk than the norm by virtue
of their mental health issues and previous suicide attempts.
Further study is required as to how risk assessments can be undertaken on a more routine
basis. Ideally, this should be the work of a multidisciplinary team, rather than the responsibility
of one professional. There should be more attention to the recording and sharing of information
on the risks to which inmates are exposed and careful monitoring to ensure that a comprehensive
risk management strategy is implemented across the federal system. Such a strategy might
include a more formal examination of all suicide attempts, as these are often harbingers of
successful suicides.

22

APPENDIX A
CODING FORM
OCI FILE #
INSTITUTION

FPS #

Name

Sex

D.O.B.

Aboriginal

Age
Yes _____

Cause of Death

No ____

Date of Incident
Date of CSC’s Investigation Report
Date of EXCOM’s Response or Region’s 1st Re
Length of Sentence
Sentence Commencement
Admission to Institution
Day Parole Eligibility (passed or date)
Full Parole Eligibility (passed or date)
Statutory Release Date
Warrant Expiry Date

CRIMINAL HISTORY
None ____

Juvenile ____

OFFENCES

VIOL

Adult ____

VSEX

Juvenile & Adult ____

SEX

PROP

DRUG

WEAP

OTH

Previous Offences
Index Offences
Viol = violent i.e. (murder, manslaughter, robbery, assault, kidnapping, hostage , threats)
Vsex = violent sex offences (i.e. sexual assault levels 1-3)
Sex = non violent sex offences (i.e. exhibitionism, invitation to sexual touching, exploitation)
Prop =property (i.e. break & enter, theft, forgery, possession of stolen goods)
Drug =drugs (i.e. possession or trafficking of illegal or scheduled substance)
Weap =weapons (i.e. possession of prohibited / restricted weapon)
Oth =other (i.e. prostitution)

23

RISK FACTORS

YES

NO

DK/ NA

Proximal Pre-Indicators
Long-Term Pre-Indicators
Substance Abuse
Previous Suicide Attempts
Mental Health Issues
Active Intervention for Mental Health Issues
Family Support
Conditional Release Violations
Institutional History
•

Cooperative

•

Program Participation

•

Violations / Escapes

(dk=don’t know, n/a=not applicable)

24

ISSUES RAISED IN THE FINDINGS AND RECOMMENDATIONS
IN CSC’S AND CORONER’S INVESTIGATIONS

CSC Board of Investigation Recommendations:
Recom.
#

Code
#

Directed To
Inst.

Region

CSC Response

National

Agree:
No
Action

Agree:
Action
Consistent

Agree:
Action
Inconsist.

Disagree

Recom.
Ignored

1
2
3
4

Other________________________________________________________________________________
CSC Board of Investigation’s Key Findings Not Appearing in Recommendations:
Finding
#

Code
#

CSC Response
Agree:
No Action

Agree:
Action
Consistent

Agree:
Action
Inconsistent

Disagree

Finding
Ignored

1
2
3
4

Other________________________________________________________________________________
Coroner’s Recommendations:
Recom.
#

Code
#

CSC Response
Agree:
No Action

Agree:
Action
Consistent

Agree:
Action
Inconsistent

Disagree

Recom.
Ignored

1
2
3
4

Other_________________________________________________________________
Noteworthy Quotes and Source:

25

Appendix B
Issued Raised in the Recommendations and Findings of CSC and Coroners
1. Patrols, Counts, and Live Body Verification
2. Pre-Incident Medical Care and Resources
3. Post-Incident Medical/Emergency Care and Resources, as well as Decontamination
4. Mental Health Issues, Programming, and Suicide Prevention
5. Security Practices, Video Surveillance, and Evidence Gathering
6. Custody and Care Issues--Availability of Illicit Drugs, Drug Paraphernalia, Weapons, and
Monitoring of Prescription Drugs
7. Inmate’s Institutional Placement and Security Classification
8. Private Family Visits—Screening of Visitors and Security Procedures
9. Sensitivity to Family Concerns (Post-Incident)—Notification, Personal Effects, Arranging
Funerals
10. Provision of Post-Incident Stress Management Services to Staff and Inmates
11. Expeditious Resolution of High Priority Grievances
12. Prevention of and Response to Prison Disturbances
13. Information Sharing Between Institutions
14. Recordkeeping and Information Sharing Among Staff Inside Institutions
15. Other___________________________________

26