Annual Report on Deaths in Custody 2016-2017

The Correctional Service of Canada (CSC) acknowledges, accepts, and takes seriously its obligation to ensure the safety and security of all offenders in its custody. Deaths in custody, particularly non-natural deaths, remain a complex and difficult issue. Providing the appropriate treatment for offenders with life-threatening, non-curable illnesses and the prevention of non-natural deaths remain a fundamental priority of the organization. The current report builds on the previous in-depth review of all deaths in custody across a 16-year period by extending analysis to the 2016/2017 fiscal year. The annual reports will continue to contribute to the organization's ability to quickly identify trends, areas for opportunity, and initiatives to reduce deaths in custody, particularly non-natural deaths.

In 2016/2017, the majority of deaths were the result of natural causes (81%). There were also three deaths by suicide, four by overdose, and two by homicide. There were no accidental deaths or deaths where staff were involved during this fiscal year. In general, this makeup of deaths in custody is consistent with trends over the last 16 years.

Offenders who died in custody varied considerably, but tended to be White, serving sentences for a homicide related offence, and serving indeterminate sentences. Those who died from natural causes tended to be older (M = 59.83). More specifically, 71% of offenders who died from natural causes were over the age of 55 years, while those who died of non-natural causes were younger (M = 41.46 years; 81% were between 18 and 54 years of age).

Deaths in the 2009/2010 to 2016/2017 period ^{Footnote 1} were examined more closely. Among natural deaths, cancer was the leading cause, followed by cardiovascular-related conditions. Chronic health issues were common among offenders who died by natural cause, with at least one chronic health condition unrelated to the cause of death identified in 96% of cases. Mental health issues were also quite prevalent, identified in 51% of cases. Other common elements in natural deaths included the presence of cigarette smoking as a contributing factor (identified in 48% of cases) and offender non-compliance with medication (34% of cases).

Suicide was the most common type of non-natural death, accounting for 70 deaths. Most suicide deaths occurred by hanging. Those who died by suicide tended to have life-histories marked by substance misuse (89%) and mental health concerns (i.e., 89% had at least one mental health condition, and nearly 60% had at least one prior suicide attempt). Changes in mental health medication regimens appeared in 38% of cases and in nearly 70% of such cases, the change occurred within three weeks of the incident. Over the 8-year period analyzed, 35 offenders died in custody from an overdose, with fentanyl being identified in 37% of the cases (i.e., either as a standalone substance or in combination with others).

CSC has prioritized learning from all deaths that occur in custody in an effort to prevent future non-natural deaths and improve on the treatment of offenders with life-threatening, non-curable illnesses. The investigations and reviews conducted following deaths in custody allow for the identification of areas of need in the service. CSC works proactively to implement recommendations and consider policy and practice in light of findings, thereby contributing to the safety and well-being of offenders, as well as staff and the public.

Introduction

The Correctional Service of Canada (CSC) takes seriously all offender deaths that occur while in custody. In an effort to enhance transparency, understand the nature of and circumstances leading to deaths in custody, and to inform policy, its 2015/2016 Annual Report on Deaths in Custody provided an in-depth review of all deaths in custody that occurred between 2009/2010 and 2015/2016. This detailed report was conducted in response to recommendations of the Third Independent Review Committee on Deaths in Custody and other inquiries. The current report builds on the previous report by extending analysis to the 2016/2017 fiscal year.

This year’s report examines all deaths in custody that occurred in 2016/2017 and explores trends since 2000/2001, with a more detailed analysis for the 2009/2010 - 2016/2017 period. Both natural deaths (i.e., where the cause of death is the result of an illness or internal breakdown of the body) and non-natural deaths (i.e., all deaths that do not qualify as a natural death or where the cause is initially undetermined) are explored in detail, with consideration paid to the manner of death, the circumstances surrounding the death, and the profile of offenders who died. Regional and yearly variation is analyzed where appropriate. For non-natural deaths, suicide and overdose deaths are also examined more closely. It is important to note that due to low base numbers, yearly fluctuations, variations, or trend analysis should be interpreted with caution; this is particularly so in the case of sub-categories of non-natural deaths.

All deaths in custody are subject to review and investigation under CSC policy. As per Commissioner’s Directive (CD) 041, Incident Investigations ^{Footnote 2}, CSC’s Incident Investigation Branch investigates all non-natural deaths ^{Footnote 3} and CSC’s Health Services Sector conducts a Mortality Review for all natural deaths, except in the rare instances where the circumstances warrant no further investigation. This report examines investigations thematically in an effort to discern trends and identify areas for improvement.

The underlying goal of the Annual Report on Deaths in Custody is to help inform organizational policy and practices in an effort to reduce and prevent all non-natural offender deaths and provide appropriate person-centered and compassionate health care to offenders with life-threatening, non-curable illnesses ^{Footnote 4}. This report also promotes accountability and transparency by providing detailed information on all deaths in custody. The CSC Research Branch (Policy Sector) will continue to track all deaths that occur in custody to advance these goals.

Data Source and Methodology

The information used to conduct the analysis presented in this report was gathered from a variety of sources. Coroner’s reports, toxicology reports and Warden’s situation reports (where available) were systematically examined for relevant information pertaining to each death. For all deaths from fiscal year 2009/2010 onwards^{Footnote 5}, either a Mortality Review (for natural cause deaths) or Board of Investigation (for non-natural deaths) was obtained and systematically coded to extract relevant information. In order to ensure the quality and consistency of the information gathered, each death was audited by a different coder and any discrepancies were reconciled. The Offender Management System (OMS) was used to gather demographic and other information about the deceased offender, as well as to retrieve incident reports on the death.

Previously developed and tested coding forms were used by multiple coders in order to capture information from Board of Investigation reports and Mortality Reviews in a reliable manner^{Footnote 6}. Once coded, each death was entered into a database maintained using survey software and subsequently extracted to statistical software for analysis. The database that was previously developed will continue to be maintained to track all deaths in custody and used for analysis in subsequent annual reports and research projects.

The information extracted from these reports was broad-based and wide-reaching, and there were some types of information that were not consistently found. It was impossible to ascertain whether the information had simply not been recorded, or non-reporting indicated that the term was irrelevant or non-applicable. In these cases, the item was coded as not indicated. When information was not available for 25% or more of cases, items were not included in the analyses. For example, in coding for the item ‘Primary source of support in year prior to death,’ information was not available in 54% of non-natural deaths; hence, this item was not included in the report for non-natural deaths. An overview is provided for the entire 17-year period, while a more in-depth analysis is presented for an 8-year period (2009/2010 - 2016/2017).

Looking at trends over low base-rate phenomenon, such as deaths in custody, can be challenging. To overcome the natural fluctuations that can occur year-to-year, comparisons were normally carried out by comparing 2016/2017 to an amalgamation of previous 7-years. In some cases, the full 8-year period was analyzed as a whole in order to preserve anonymity due to small numbers. Not all percentage totals in the tables will add to 100% due to rounding and this is more pronounced due to the small numbers in some tables.

Changes to the Annual Report

This annual report provides greater definitional clarity regarding the types of deaths that occur in federal custody (i.e., the originating incident occurred within a federal institution or at a hospital while in CSC’s custody). Natural cause reflects incidents where the cause of death is the result of an illness or internal breakdown of the body. All other types of deaths that occur in custody that do not meet this definition would fall into the category of non-natural deaths. There are a number of types of non-natural deaths that fall within this category:

1. Suicide reflects incidents where the offender intentionally ended their own life (e.g., hanging/ligature, cutting, blunt force trauma, overdose, and other manners).
2. Overdose reflects incidents where the offender died as a result of ingesting licit or illicit drug(s); overdoses will only be classified as a suicide if there is evidence of intent (e.g. a suicide note).
3. Homicide reflects incidents where an offender dies as a result of harm caused by another offender.
4. Accidental reflects incidents that are non-natural in nature but are the result of an accident (e.g., slip/fall).
5. Staff Involved reflects incidents where staff were directly involved in an incident that lead to the death of an offender (regardless of intent; e.g., use of force during a security incident such as an attempted escape). Please note that there has been a change in terminology from ‘Staff Intervention’ to ‘Staff Involved’ in order to remove ambiguity and be more inclusive of the variety of incidents involving staff.
6. Undetermined reflects incidents where the specific cause of death could not be determined, as well as deaths where the investigation process is ongoing at the time of this report.

The format and presentation have also been updated in this report. Text boxes are used to provide definitions and contextual information, as well as to highlight notable findings, throughout the report. Graphical representation of data is also presented through the report to complement the text. Finally, the majority of the tables referred to within this document have been included as appendices in order to make this document more accessible to readers, while still providing the level of detail as the previous annual report.

The Research Branch gratefully acknowledges the assistance of the Incident Investigation Branch, Health Services and the Parole Board of Canada in gathering and assisting with the interpretation of the varied documents and reports used in the analysis. This report would not have been possible without their assistance and co-operation.

Overview of Deaths in Federal Custody

Deaths in Custody over a 17-year period: Manner of Death and Regional Differences

In 2016/2017 there were 38 natural cause deaths, accounting for 81% of all deaths during this fiscal year. There were also three deaths by suicide, four by overdose, and two by homicide. There were no accidental deaths or deaths where staff were involved during this fiscal year. The overall rate of deaths in custody in 2016/2017 was 3.32 per 1000 offenders. The rate of natural deaths in 2016/2017 was 2.68 per 1000 offenders, while the rate of non-natural deaths in 2016/2017 was 0.64 per 1000 offenders.

Across 17-year period (2000/2001 and 2016/2017), there were a total of 904 deaths in federal custody. Of these deaths, 67% were natural cause deaths (see Table 1). Non-natural deaths account for 33%; however, this percentage has fluctuated over this 17-year period, ranging from a high of 44% in 2002/2003 to a low of 19% in 2016/2017. It is important to note that the proportion of natural versus non-natural deaths in custody can fluctuate greatly, producing a magnified effect on percentages due to low base numbers.

In 2016/2017, the Quebec region recorded the highest number of natural deaths in terms of raw numbers (12), followed by the Ontario region (10). The Atlantic region had the lowest number (1). Non-natural deaths were highest in raw numbers in the Prairie region (i.e., 5 compared to 2 in both the Quebec and Ontario regions). Regional breakdowns were not examined within the single fiscal year given the small numbers.

Likewise, the Quebec and Ontario regions had the highest number of natural deaths across the 17-year period examined (2000/2001-2016/2017; 176 in both cases), while the Atlantic region had the lowest (48; see Table 2). Over this period, the Ontario region had the highest number of non-natural deaths in custody (80), followed by the Prairie region (76) and the Quebec region (63). An examination of the breakdowns within each region indicate that the Quebec and Ontario regions had the largest percentage of natural deaths (i.e., 73% and 68%, respectively). In terms of non-natural deaths, the Atlantic region had the largest percentage of suicide deaths (27%) and the smallest percentage of overdose deaths (4%). The Prairie region had the largest percentage of homicide deaths (11%) and the second largest percentage of both overdose deaths (9%) and suicide deaths (21%). The Ontario region had the largest percentage of overdose deaths (10%) and the smallest percentage of suicide deaths (14%).

Profiles of Offenders Who Died in Federal Custody

Offenders who died in custody varied considerably, but tended to be White, 55 years of age or older, serving a sentence for a homicide related offence, and serving indeterminate sentences. Women accounted for two of the deaths that occurred in 2016/2017 and nine of the deaths that occurred in the previous seven years. Results by gender were not presented due to low numbers and the possibility of revealing identifying information about individual offenders.

Offenders who died of natural causes in 2016/2017 were more likely to be White, 55 years of age or older, and were typically serving an indeterminate sentence (see Table 3 in appendix). These offenders tended to be serving time for a homicide related offence and classified as either medium or minimum security. At the time of death, offenders who died of natural causes had typically served either a relatively short amount of time (less than five years) or a considerable amount of time (20+ years).

Offenders who died of non-natural causes tended to be White (67%, n = 6), under the age of 44 (67%, n = 6), serving time for a homicide related offence (44%, n = 4), and classified as medium security (67%, n = 6; see Table 3 in appendix). At the time of death, they had typically served a relatively short amount of time (less than five years of their sentence; 67%, n = 6). The most common index offence among both natural and non-natural deaths was homicide, in both 2016/2017 (61% and 44%).

When considering the 8-year period (2009/2010 - 2016/2017), differences were observed between the offenders who died of natural causes and those who died by non-natural manner. For example, those who died from natural causes were more likely to be White (71% versus 29% for non-natural causes), serving an indeterminate sentence (55% versus 39% for non-natural causes), be classified as minimum security (30% versus 9% for non-natural causes), and be older (M = 59.83). More specifically, 71% of offenders who died from natural causes were over the age of 55 years. Those who died of non-natural causes were more likely to be classified as maximum security (28% versus 10% for natural causes) security and were more likely to be younger (M = 41.46 years). Over three quarters of offenders who died of non-natural causes were between the ages of 18 and 54 years (81%). Among Indigenous offenders in 2016/2017, a greater number of deaths were natural (56%) compared to non-natural (44%).

Natural Deaths in Federal Custody

Manner of Death and Regional Differences

In 2016/2017, there were 38 natural cause deaths in custody, accounting for 81% of all deaths during this fiscal year. The three most common subtypes of natural cause deaths in 2016/2017 were cancer, cardiovascular related, and other (24% each; see Table 5 in appendix). Given the relatively low number of deaths overall in this fiscal year, comparisons with the previous 7-year period should be made with caution. Therefore, while this report will describe natural cause deaths for 2016/2017, the full 8-year period (2009/2010 - 2016/2017) will also be discussed. When natural cause deaths are examined across this time period, the two most common subtypes of natural cause death remained both cancer (34%) and cardiovascular related (28%; see Figure 1).

Figure 1. Subtype of Natural Cause Deaths, 2009/2010 - 2016/2017

Given the small numbers in the current fiscal year, regional variation was examined across the 8-year period (2009/2010 - 2016/2017). Cancer related natural deaths were highest in raw numbers in the Ontario (37) and Quebec (26) regions, and lowest in the Atlantic (6) region across this time period (see Table 6 in appendix). Cardiovascular related deaths were highest in the Quebec region (30).

When the percentages within each region were examined, cardiovascular related deaths were the most common in the Atlantic, Quebec and Prairie regions (i.e., 37%, 38%, and 33%, respectively), with cancer related deaths being the next most common subtype in these regions (i.e., 32%. 33%, and 24%; see Figure 2). In the Ontario and Pacific regions, cancer related deaths were the most common subtype of natural cause death (i.e., 41% and 38%), with cardiovascular related deaths the next most common subtype in these regions (i.e., 14% and 30%). Liver related deaths were the most common in the Prairie (20%) and Ontario (11%) regions.

Figure 2. Subtype of Natural Cause Deaths by Region, 2009/2010 - 2016/2017

Further Examination of Natural Deaths

In 2016/2017, 15 (65%) offenders who died of natural causes had a Do not Resuscitate order (DNR) on file and 20 (87%) were receiving palliative care^{Footnote 7} (see Table 7 in appendix). Nearly half of offenders had their next-of-kin involved (i.e., contact, visits, decision-making, and/or consent). These proportions exclude the 14 offenders who died unexpectedly from natural causes and one whose investigation is ongoing. In 2016/2017, the final cause of death was related to substance misuse in 19% of natural deaths and to cigarette smoking in 27% of natural deaths. During this fiscal year, 95% of offenders who died of natural causes had at least one chronic health condition^{Footnote 8} identified, 62% had at least one mental health condition identified, and 35% had noncompliance issues with their medication.

When the full 8-year period was examined, 136 (81%) offenders who died of natural causes had a DNR on file and 147 (88%) were receiving palliative care (see Table 7 in appendix). Fifty-five percent had their next-of-kin involved (i.e., contact, visits, decision-making, and/or consent). Again, these proportions exclude offenders who died unexpectedly from natural causes. During this time period, 96% of offenders who died of natural causes had at least one chronic health condition identified, 51% had at least one mental health condition identified, and 34% were noncompliant with their medication. In 24% of cases, the final cause of death was related to substance misuse and in 48% of cases the final cause of death was related to cigarette smoking.

There was minimal regional variation when it came to the presence of health conditions and palliative care receipt. Over the 8-year period, over 90% of individuals who died of natural causes had at least one chronic health condition identified, regardless of the region (see Table 8 in appendix). Between 70% and 84% of individuals who died expectedly of natural cause deaths were receiving palliative care across the regions. Those in the Prairie region were the most likely to have a DNR on file (94%), though the percentage was over 60% in all regions. The Pacific region had the smallest percentage of individuals who had their next of kin involved (37%), compared to over 50% of individuals in the remaining regions. The Quebec region had the smallest percentage of individuals who had noncompliance issues with their medication (23%), compared to 34-37% in all other regions. They also had the smallest percentage of individuals who had at least one mental health condition identified (30%), compared to 54-70% in all other regions. The Atlantic region had the largest percentage of individuals who had noncompliance issues with their medication (47%), compared to 23-42% in all other regions. The Pacific region had the largest percentage of individuals who had at least one mental health condition identified (70%), compared to 30-58% in all other regions.

In 2016/2017, in terms of chronic health conditions identified in offenders who died by natural cause, cardiovascular conditions were the most common, noted in 76% of offenders (i.e., at least one condition was identified; see Table 9 in appendix). Other common issues included chronic gastrointestinal conditions (i.e., 65% of offenders had at least one condition identified), blood-borne viruses/infections, and endocrine conditions (within each category, 49% of offenders had at least one condition identified). When the full 8-year period was examined (2009/2010 - 2016/2017), cardiovascular conditions remained the most common type of chronic health conditions identified (71%; see Figure 3). However, respiratory conditions (46%) and gastrointestinal conditions (43%) were the next most common types of chronic health conditions. Reproductive conditions were the least commonly identified conditions for both 2016/2017 and the full 8-year period (i.e., 8% and 10%, respectively).

Figure 3. Types of Chronic Health Conditions, 2009/2010 - 2016/2017

There was minimal regional variation when it came to the most common types of health conditions identified across the 8-year period (2009/2010 - 2016/2017). Cardiovascular conditions were the most commonly identified and reproductive conditions were the least commonly identified type of chronic health conditions, regardless of region (see Figure 4 and Table 10 in appendix). However, there was some regional variation in the prevalence of conditions. For example, while there was no regional variation in the most common and least common type of chronic conditions identified, the percentages of offenders with these conditions did vary. The percentage of offenders with at least one chronic cardiovascular condition identified ranged from 65% in the Ontario region to 83% in the Quebec region and ranged from 5% in the Atlantic region and 12% in the Pacific region for chronic reproductive conditions. In the Atlantic and Ontario regions, chronic gastrointestinal conditions were the second most common type of condition identified (i.e., 58% and 50% respectively), while respiratory conditions were more common in the Quebec, Pacific and Prairie regions (i.e., 54%, 52%, and 47%).

Figure 4. Types of Chronic Health Conditions Identified by Region, 2009/2010 - 2016/2017

In 2016/2017, only 5% of offenders who died of natural causes had no chronic health conditions identified, which is consistent across the full 8-year period (see Figure 5 and Table 11 in appendix). In 2016/2017, 76% of offenders had between two and eight chronic health conditions identified, while 16% had nine or more identified. Similarly, 79% of offenders in the full 8-year period had between two and eight chronic health conditions identified and 10% had nine or more identified.

Figure 5. Number of Chronic Health Conditions Identified, 2009/2010 - 2016/2017

In 2016/2017, natural cause deaths most commonly occurred on Wednesday (27%) and were less likely to occur on Monday and Saturday (8% each; see Figure 6 and Table 12 in appendix). Across the full 8-year period, natural cause deaths most commonly occurred on Sunday (19%) and were less likely to occur on Tuesday (11%). In comparison to the full 8-year period, where 29% of natural cause deaths occurred in the morning (i.e., between 6am and 12pm), natural cause deaths were least likely to occur during this time in 2016/2017 (16%; (see Figure 7 and Table 13 in Appendix).

Figure 6. Natural Deaths by Day of the Week, 2009/2010 -2016/2017

Figure 7. Natural Deaths by Time of Day, 2009/2010 - 2016/2017

Compliance Issues in Relation to Natural Deaths

Reports produced by the Incident Investigation and Health Services branches in response to deaths outline if and where compliance issues with policy were discovered. In 2016/2017, the most common compliance issue identified in Mortality Reviews for natural cause deaths related to reporting requirements according to policy (24%), which is similar to the previous 7-year period (29%). Examples of such compliance issues identified over the 8-year period (2009/2010 - 2016/20017) include:

• incomplete or missing documents within files,
• exclusion of follow-up tests in files,
• confusion around where the documents were located, and
• failing to complete documents within the specified timeframe.

Across the 8-year period (2009/2010-2016/2017), the majority of compliance issues were similar across regions, with a few exceptions. The Quebec region had the highest percentage of compliance issues noted regarding exposure protocols (23%). The Atlantic region had the highest percentage of compliance issues around reporting requirements as per policy (37%), and the Quebec region had the lowest (20%). The Ontario region had the highest percentage of non-compliance when it came to providing support (Critical Incident Stress Management) to staff (32%) and offenders (33%) noted.

Non-Natural Deaths in Federal Custody

Manner of Death and Regional Differences

In 2016/2017, there were nine non-natural deaths in custody, including three suicide deaths, four overdose deaths, and two homicide deaths (see Table 14 in appendix). Overall, non-natural deaths were lowest in raw numbers in 2016/2017 compared to the seven fiscal years prior. Given low base numbers, fluctuations are likely to occur and should be interpreted with caution.

Over the 8-year period analyzed (2009/2010 - 2016/2017), suicide deaths were the most common cause of non-natural death (49%, n = 70), followed by overdose deaths (25%, n = 35), and homicide deaths (13%, n = 19; see Table 14 in appendix). Other types of non-natural deaths were relatively uncommon. As a percentage of non-natural deaths, suicide deaths have been decreasing, from 64% (9) in 2014/2015 to 33% (3) in 2016/2017. At the same time, the percentage of overdose deaths has been increasing, from 9% (2) in 2012/2013, to 44% (4) in 2016/2017.

Figure 8. Non-Natural Deaths in Custody Over an 8-year Period, 2009/2010 - 2016/2017

In terms of raw numbers, non-natural deaths over the 8-year period were highest in the Ontario region (37) followed by the Prairie region (36), and lowest in the Atlantic region (18; see Figure 9). In terms of the sub-types of non-natural deaths, suicide deaths were highest in the Prairie region (17), while overdose deaths were highest in the Ontario region (12; see Table 15 in appendix). As a percentage of non-natural deaths, suicide deaths were most common in the Atlantic region (61%), while overdose deaths were most common in Ontario (32%) and Quebec (30%).

Figure 9. Number of Non-Natural Deaths in Custody by Region, 2009/2010 - 2016/2017

Compliance Issues in Relation to Non-Natural Deaths

Over the 8-year period examined, the most common compliance issues noted in investigations pertaining to non-natural deaths related to reporting requirements (45%), exposure protocol (25%) and search logs/documentation of items found (22%; see Table 18 in appendix).

Recommendations from Board of Investigation (BOI) Reports

Investigations into non-natural deaths contain recommendations in cases where a need for a policy or other type of change has been identified. A thematic analysis of BOI recommendations was previously conducted for non-natural deaths that occurred between 2009/2010 and 2015/2016 and was continued for the 2016/2017 fiscal year. The recommendations put forth in the BOI reports were previously organized into eight general themes:

1. Review a current process or implement a new process,
2. Change or merge reports/forms,
3. Review or change the guidelines, roles, responsibilities, and/or training of staff,
4. Change the type of equipment used in medical emergencies,
5. Changes to improve communication and sharing of information between services and departments,
6. Review current policy or implement policy change to prevent a similar situation,
7. Change institutional infrastructure, and
8. Best practices.

In 2016/2017, the most common recommendations related to improving the communication and sharing of information between services and departments, and reviewing/changing the guidelines, roles, responsibilities, and/or training of staff.

As with the previous 7-year period, many of these recommendations were to clarify particular aspects of a staff member’s position or policies/procedures. Such recommendations were intended to eliminate confusion during medical emergencies or improve the training of all staff members who come into contact with offenders, especially contract employees. Overall, the recommendations across each category aimed to improve the Services’ overall care of individuals with mental health or substance misuse issues in order to prevent similar situations (i.e., non-natural deaths) from occurring, whether through policy changes, more effective training for staff or improved communication between services and departments.

Suicide Deaths in Federal Custody

Profile of Offenders who Died by Suicide

The three offenders who died by suicide in custody in 2016/2017 between the ages of 25 and 44, were classified as either medium security or maximum security, and all were serving indeterminate sentences for homicide-related offences (see Table 19 in appendix). This profile is consistent with that of offenders who died by suicide in custody in the 7-year period prior (2009/2010 - 2015/2016). In the previous 7-year period, the 67 offenders who died by suicide in custody tended to be White (64%) or Indigenous (27%), with an average age of 41. They tended to be medium security (49%) or maximum security (30%) offenders. Many were serving indeterminate sentences (42%), and serving time for homicide-related offences (39%).

Across the full 8-year period examined (2009/2010 - 2016/2017), those who died by suicide tended to have life-histories marked by substance misuse and mental health concerns. For example, 80% had a history of substance misuse, 89% had at least one mental health condition, and nearly 60% had at least one prior suicide attempt (see Table 23 in appendix). It was also common for those who died by suicide to have had previous mental health interventions; for example, 60% were on medication for mental illness at the time of death, 87% had regularly accessed mental health services, and 31% had been previously placed in a regional treatment centre (see Table 25 in appendix). Of note, 38% had a medication change prior to the suicide; in nearly 70% of such cases, the change occurred within three weeks of the incident.

Details Surrounding Suicide Deaths in Custody and Compliance Issues Identified

Over the last 8 years, 39% (n = 27) of suicide deaths have occurred in segregation (see Table 23 in appendix). The percentage was highest in the Atlantic and Pacific regions (46%, n = 5 in both cases), and lowest in the Ontario region (25%, n = 4; see Table 24 in appendix). Hanging was the primary method utilised in suicide deaths (see Table 20 appendix). Suicide deaths were most common in the middle of the week (between Tuesday and Thursday), peaking on Wednesdays (21%), and in the afternoon period (40%; see Table 21 in appendix).

Over the 8-year period analyzed, the most common compliance issue in relation to suicide deaths related to reporting requirements (46%, n = 32; see Table 27 in appendix). Other common compliance issues related to exposure protocol (noted in 23% of cases, n = 16) and search logs/documentation of items found (noted in 19% of cases, n = 13).

Overdose Deaths in Federal Custody

Profile of Offenders who died by Overdose

Over the 8-year period analyzed, 35 offenders died in custody from an overdose, with four deaths occurring in 2016/2017. Offenders involved in overdose deaths tended to be White (86%), medium security (71%), serving an indeterminate sentence (26%) or short (less than four year) sentence (23%) for a homicide-related offence (26%) or robbery (23%; see Table 28 in appendix). The average age of offenders who died by overdose was 39. Those who died by overdose tended to have documented histories of substance misuse (86%; see Table 34 in appendix).

Details Surrounding and Compliance Issues Related to Overdose Deaths

Of the four overdose deaths that occurred in 2016/2017, two involved prescription medications and two involved illegal substances (see Table 29 in appendix). All four cases involved opioids as a standalone or contributing substance. Across the 8-year period examined, illegal substances were most common, evident in 74% (n = 26) of cases, with opioids being present as a standalone or contributing substance in 89% (n = 31) of cases. In terms of specific substances, fentanyl was involved in 37% (n = 13) of cases, heroin was involved in 20% (n = 7), and morphine was involved in 23% (n = 8; see Tables 30 and 31 in appendix). Overdose incidents were most common between Wednesday and Saturday and in the morning period (see Tables 32 and 33 in appendix).

Over the 8-year period analyzed, the most common compliance issues in relation to overdose deaths related to reporting requirements, noted in 43% (n = 15) of cases. Other common compliance issues related to search logs/documentation of items found (31%, n = 11) and exposure protocol (29%, n = 10; see Table 36 in appendix)

Non-Natural Deaths in Federal Custody by Other Manner

Profile of Offenders who Died by Other Manner and Compliance Issues Identified

Homicides, accidental deaths, and undetermined cause deaths tend to occur less frequently than other non-natural deaths. In 2016/2017, deaths by other non-natural manner included two homicide deaths (see Table 38 in appendix). Over the 8-year period analyzed, 3 5 other types of deaths occurred (i.e., nine accidental deaths, 19 homicide deaths, and seven deaths that were undetermined).^{Footnote 9} Offenders who died by other non-natural manner in the 8-year period tended to be White (63%) or Indigenous (26%), medium (57%) or maximum security (29%), serving an indeterminate sentence (46%), and serving time for a homicide-related offence (40%). The average age of these offenders was 41 years.

Offenders who died by other non-natural manner over the 8-year period^{Footnote 10} often had histories involving substance misuse, health issues, or mental health issues (see Table 41 in appendix). For example, 56% (n = 18) had histories of substance abuse, 44% (n = 14) had prior drug-related offences/incidents, 47% (n = 15) had at least one chronic health issue, and 50% (n = 16) had at least one mental health issue. The most common compliance issue in relation to non-natural death by other manner related to reporting requirements (44%, n = 14; see Table 42 in appendix). Other common compliance issues related to exposure protocols, search protocols, and search logs/documentation of items found.

Conclusion

Deaths in custody, particularly non-natural deaths, remain a complex and difficult issue for all correctional jurisdictions. CSC recognizes that it is vital to continuously work to enhance relevant prevention and intervention strategies. CSC has prioritized learning from all deaths that occur in custody in an effort to improve treatment for offenders with a life-threatening, non-curable illness and prevent future non-natural deaths in custody.

As previously mentioned, CSC strives to provide compassionate, innovative, patient- and family-centered hospice palliative care (HPC)^{Footnote 11} to offenders with a life-threatening, non-curable illness^{Footnote 12}. The Hospice Palliative Care Guidelines for Correctional Service Canada (2009) is informed by the principles and standards of the Canadian Hospice Palliative Care Association (Canadian Hospice Palliative Care Association, 2013; Ferris et al., 2002). HPC aims to provide self-directed care by an Interdisciplinary HPC Team, who are both knowledgeable and skilled in all aspects of the caring process related to their discipline of practice, in order to meet each offender’s unique health and social care needs. HPC is available across a continuum of care, from diagnosis to end-of-life and may be provided in different environments, such as an offender’s parent institution, at a CSC Regional Hospital where 24-hour nursing care is available, or in community agencies and facilities such as Haley House.

CSC also continues to expand strategies in place to prevent non-natural deaths in custody. Current measures in place to combat overdose deaths, for example, include drug prevention and detection strategies (e.g., use of ion scanners, urinalysis testing, intelligence gathering to lead searches, and drug detection dogs; Correctional Service Canada, 2017; Johnson, Cheverie, & Moser 2010), harm reduction measures, including opiate agonist therapy (Cheverie 2014) and the availability of naloxone to both medical and non-medical staff working in CSC institutions, as well as the availability of correctional programming targeting substance misuse (Ternes, Doherty & Matheson 2014; Doherty, Ternes & Matheson, 2014; Kunic & Varis 2009). To reduce the likelihood of suicide deaths, CSC utilizes tools to screen individuals for mental illness and suicide risk (Mills and Kroner 2010; Archambault et al. 2010; Stewart et al. 2009), and has procedures in place for enhanced observation and care of individuals at risk (Correctional Service Canada, 2012).

Though the presentation of information may differ across each annual report as the number of deaths that occur each fiscal year can fluctuate greatly, the goal of these reports remains the same. This annual report, and future reports, will continue to contribute to the organization's ability to quickly identify trends, areas for opportunity, and initiatives to reduce deaths in custody. Furthermore, this report contributes to transparency and open communication regarding both natural and non-natural deaths, thereby facilitating the continual collaboration with stakeholders and experts who may contribute to the important goal of preventing non-natural deaths and improving the health care of offenders with life-threatening, non-curable illnesses while in federal custody.

This report also highlights the need for CSC to maintain its efforts for the reduction of deaths in custody, particularly non-natural deaths. These efforts must include the acquisition of knowledge, the development of capacity, and the introduction of measures, protocols, and initiatives to increase our efficiencies. CSC must also self-reflect on how it manages its in-custody population to assist offenders in their own path to reintegration, in a safe and secure environment. Finally, CSC must support its professional workforce, without which CSC will not be able to fulfil its mission. Training, collaboration, and better communication must be maintained or increased in order to build cohesion and expertise in managing potentially tragic situations that can result in a loss of life.

References

Aday, R. (2006). Aging prisoners’ concerns toward dying in prison. OMEGA: Journal of Death and Dying, 52(3), 199-216.

Aday, R. & Wahidin, A. (2016). Older prisoners’ experiences of death, dying, and grief behind bars. The Howard Journal of Crime and Justice, 55(3), 312-327. doi: 10.1111/hojo.12172

Archambault, K., Stewart, L., Wilton, G., & Cousineau, C. (2010). Initial results of the Computerized Mental Health Intake Screening System (CoMHISS) for Federally Sentenced Women. Research Report R-230. Ottawa, ON: Correctional Service Canada.

Canadian Hospice Palliative Care Association (2013). A Model to Guide Hospice Palliative Care: Based on National Principles and Norms of Practice (Revised and Condensed Edition). Ottawa, ON: Canadian Hospice Palliative Care Association. Retrieved from http://www.chpca.net/media/319547/norms-of-practice-eng-web.pdf

Cheverie, M., MacSwain, M., Farrell MacDonald, S., & Johnson, S. (2014). Institutional Adjustment of Methadone Maintenance Treatment Program (MMTP) Participants. Research Report R-323. Ottawa, ON: Correctional Service Canada.

Correctional Service Canada (2009). Hospice Palliative Care Guidelines for Correctional Services Canada. July 2009.

Correctional Service Canada (2010). Commissioner’s Directive (CD) Number 041: Incident Investigations.

Correctional Service Canada (2012). Commissioner’s Directive (CD) Number 843: Interventions to Preserve Life and Prevent Serious Bodily Harm.

Correctional Service Canada (2017). Commissioner’s Directive (CD) Number 566-8-1: Use of Non-Intrusive Search Tools.

Corrections and Conditional Release Act (S.C 1992, c. 20), section 121.

Crawley, E. & Sparks, R. (2006). Is there life after imprisonment? How elderly men talk about imprisonment and release. Criminology and Criminal Justice, 6(1), 63-82. doi: 10.1177/1748895806060667

Doherty, S., Ternes, M., & Matheson, F.I. (2014). An Examination of the Effectiveness of the National Substance Abuse Program High Intensity (NSAP-H) on Institutional Adjustment and Post-Release Outcomes. Research Report R-290. Ottawa, ON: Correctional Service of Canada.

Ferris, F. D., Balfour, H. M., Bowen, K., Farley, J., Hardwick, M., Lamontagne, C., Lundy, M., Syme, A., & West, P. (2002). A Model to Guide Hospice Palliative Care: Based on National Principles and Norms of Practice. Ottawa, ON: Canadian Hospice Palliative Care Association.

Johnson, S., Cheverie, M., & Moser, A. (2010). Assessing the Impact of Enhanced Drug Interdiction Activities at Kingston Penitentiary: A Pilot Study. Research Report R-232. Ottawa: Correctional Service Canada.

McKendy, L., Biro, S., & Keown, L.A. (2019). Overdose Incidents in Federal Custody, 2012/2013 - 2016/2017. Special Report SR-18-02. Ottawa, ON: Correctional Service of Canada

Mills, J. and Kroner, D. (2010). Concurrent Validity and Normative Data of the Depression Hopelessness and Suicide Screening Form with Women Offenders. Research Brief B-47. Ottawa, ON: Correctional Service Canada.

Morton, J. B. & Anderson, J. C. (1991). Elderly offenders: the forgotten minority. Corrections Today, 44(7), 14-17.

Sanders, M. & Stensland, S. (2018). Living a life full of pain: Older pain clinic patients’ experience of living with chronic low back pain. Qualitative Health research, 28(9), 1434-1448. doi: 10.1177/1049732318765712

Stewart, L. A., Harris, A., Wilton, G., Archambault, K.Y. Cousineau, C., Varrette, S., & Power, J. (2009). An Initial Report on the Results of the Pilot of the Computerized Mental Health Intake Screening System (CoMHISS). Research Report R-218. Ottawa, ON: Correctional Service Canada.

Stewart, L. A., Nolan, A., Sapers, J., Power, J., Panaro, L., & Smith, J. (2015). Chronic health conditions reported by male inmates newly admitted to Canadian federal penitentiaries. Canadian Medical Association Journal Open, 3(1), E97-E102. doi: 10.9778/cmajo.20140025

Ternes, M., Doherty, S., & Matheson, F.I. (2014) An Examination of the Effectiveness of the National Substance Abuse Program Moderate Intensity (NSAP-M) on Institutional Adjustment and Post-Release Outcomes. Research Report R-291. Ottawa, ON: Correctional Service Canada.

Appendix A - Tables