Third Independent Review of Deaths in Custody,
April 1st, 2011 to March 31st, 2014

Final Report

October 2015

Independent Review Committee:

Yvette Thériault, Chairperson
Ottawa, Ontario, Canada

Alan Leschied
London, Ontario, Canada

Shelley Brown
Ottawa, Ontario, Canada

Executive Summary

The Office of the Correctional Investigator launched an investigation lead by Gabor (2007) regarding deaths in federal prisons within Correctional Service of Canada (CSC). Subsequently, CSC mandated two Independent Reviews (2011, 2012) of non-natural deaths in custody.

The two previous Independent Review Committees (IRC) identified deficiencies in the areas of: information sharing, mental health programming, risk assessment, and security monitoring.

The present IRC report is the third in this series of reviews into federal deaths in custody. The IRC reviewed 50% of non-natural deaths in custody for the fiscal years 2011 – 2014. The mandate was to advise CSC on trends, contributing factors to those deaths, adequacy of corrective measures and action plans, information sharing and best practices. The conclusions reached by the members of the present IRC concur with the general observations made in the 2012 IRC report, that overall, the performance of CSC is satisfactory, particularly, in our view, in regards to static security.

Nevertheless, there is still room for improvement especially in the area of dynamic security which encompasses case management, suicide prevention, and intervention. In essence, the current assessment procedures, i.e., the screening tools and clinical interviews, for suicide risk need to be revised considerably and the results clearly communicated to correctional staff at all levels of care. Additionally, recent developments in suicide prevention and treatments should be implemented.


A decade ago the Office of the Correctional Investigator expressed concern regarding the number of deaths in custody in federal prisons (CSC). In 2007, Gabor provided the first independent review into these deaths. The results of the Gabor study of 82 deaths in custody from 2001-2005 identified several major deficiencies that included information sharing, mental health programming, risk assessment, and security monitoring. Subsequently, CSC initiated a number of corrective actions. Chief among these was the establishment of an IRC to conduct external reviews of CSC’s actions. Subsequently to the Gabor review, two IRCs were established which covered the fiscal years 2009/2010 (Hastings, Forestell, & Graceffo, 2011), and 2010/2011 (Weinwrath, Wayte, & Arboleda-Florez, 2012). These reviews revealed that deaths in custody have declined over time. These two reviews noted improvements in CSC practices to the extent that the organisation appears to compare favourably to other countries that was reflected in a satisfactory rating (e.g., 65%-80%, Weinwrath et al., 2012). Moreover, the response time of CSC taking corrective action on the problems identified has been reduced considerably (Hastings et al., 2011). There remained, however, areas of concern, principally, case management, information sharing, security practices and the use of segregation (see Hastings et al., 2011; Weinwrath et al, 2012).

In 2015, the Commissioner of CSC, through a Convening Order (see Appendix A), mandated a third IRC review of non-natural deaths in custody for 3 fiscal years (2011/2012 to 2013/2014).

The objectives were further specified in the Statement of Work (see Appendix B) and summarized here:

  1. To review a random sample of 23 of 46 non-natural deaths in custody during this period;
  2. An analysis of the corrective measures and the management of the action plans supported at CSC’s National Investigations Meetings in order to address the identified gaps;
  3. A pre and post incident analysis of the quality of the communication and information sharing amongst CSC’s key players;
  4. The identification and analysis of any trends observed within the following areas:
    1. The contributing risk factors or precipitating events to the deaths;
    2. The physical and mental health care prior to the deaths;
    3. The case management practices prior to the deaths;
    4. The security practices such as staff presence and monitoring of the inmate activities; and
    5. The management of and response to the emergencies in regards to the deaths of the inmates.

This IRC was directed to review the relevant literature and provide a list of “best” practices with respect to deaths in custody. The reviewers updated the declining trends in deaths over time with regard to homicide and suicide as reported by Weinwrath et al., (2012). We also produced a comparison on non-natural deaths in custody by region.


This mandated review of non-natural deaths in custody for the fiscal years 2011 – 2014 was conducted by three external reviewers with experience and knowledge in corrections, mental health, and research. What follows are the definitions taken directly from Weinwrath et al., (2012) that guided the reviewers for the analysis of trends (b-f), corrective measures, management of action plans, communication and information sharing. The material provided at the beginning of the review for CSC is listed in Appendix C.

Analysis of trends

The analysis of trends was based on previously published IRC reports, current reviews of data, CSC reports on deaths in custody, annual reports from the Office of the Correctional Investigator as well as file reviews of 23 of the 46 individuals who died a non-natural death in custody. The areas included in the analysis are defined as follows:

  1. Previous data on trends and key outcomes (e.g., homicide) over time

    The data for examining these trends was derived mostly from previously published annual reports, IRCs previous reports, and additional data provided by CSC staff.

  2. Contributing risk factors

    The IRC chose to use and define contributing risk factors as such trends that might include precipitating events that immediately or proximately linked to a death, such as inmate conflict with a staff member or other inmates, bad news from family, or a security reclassification or parole denial. Trends might also include longstanding problems such as poor health, substance abuse, or negative family history.

  3. Physical and mental health care provided

    This area of trend was defined as an inmate’s mental health history reflected in past records of suicide, self-harm, depression and other conditions. We also reviewed more recent treatment, including mental health treatment plans and regimes, medication, adequacy of treatment, and follow-up.

  4. Case management practices

    Case management was also defined as the assessment of staff work on program planning, security reviews and release preparation and the adequacy of inmate contact, report preparation and overall adequacy of the timeliness, and thoroughness of documentation.

  5. Security practices

    Security practices were defined as the correctional officer as staff play a critical role in preventing inmate deaths through dynamic and static security measures.

  6. Management and response to the emergencies in regards to the deaths in custody

    This area was defined as the emergency response. Trends of concern included assessment of correctional staff response time after an inmate was found in distress and the adequacy of response by correctional officers or health care staff in the administration of medical assistance.

Review of Sample Cases 23 of 46 non-natural deaths in custody

A sample of 23 cases was randomly drawn from the 46 non-natural deaths in custody that occurred within CSC during the fiscal years 2011 – 2014. This sample was selected and provided by CSC as representative of the 46 deaths and formed the data base for the review. Individuals’ demographic information and psychological reviews that formed part of the Boards of Investigation reports were provided to the reviewers as well as a summary of individual file information drawn from the CSC system (RADAR). This information was summarized as follows:

  1. Procedure
  • The French speaking member of the IRC reviewed five francophone and two anglophone files. The other two members each examined eight anglophone files. To ensure reliability of coding across the files, each IRC member reviewed an overlapping case independently of the other two members. One hundred percent reliability in criterion was achieved through a collaborative process of comparison on those overlapping cases.
  • Each member coded the files according to an agreed set of variables. These were based on a literature review on non-natural deaths in custody, specific file information, and previous IRC reports (see Appendix D).
  • The information reviewed included: contributing risk factors, physical and mental health care provided, case management practices, security practices, management and response to the emergencies in regards to the deaths in custody, as well as noted best practices
  • An examination was conducted of the literature along with personal contacts with international experts on best practices in areas relevant to the review.
  • Boards of Investigation findings and recommendations, and corrective measures and action plans were reviewed for all individuals. Deaths in Custody Discussion Guides by the Incident Investigations Branch were reviewed.

Corrective measures and management of the action plans

The adequacy of corrective measures and management of the action plans was based on individual Boards of Investigation findings, corrective measures and management action plans found in the related grids, and recommendations of the last IRC (2012). The findings from the Boards of Investigation were categorized in six main areas: systemic, security, clinical, policy compliance, best practices and corrective measures.

Communication and information sharing

The adequacy of the communication and information sharing amongst CSC stakeholders was examined from Boards of Investigation reports. We retained the previous definition by the IRC (2012) “as knowledge of the potential risk of overdose, suicide, or conflicts leading to violence are not of much use if not shared with other correctional staff”.

Review of suicide risk levels

Two of the reviewers analysed the estimated suicide risk levels of all cases of suicide for the sample (n = 23) using a CSC comprehensive suicide risk assessment tool as stated in Commissioner’s Directive (CD) 843, Management of Inmate Self-Injurious and Suicidal Behaviour. The psychological reviews and Boards of Investigation Reports were used for this process. Forty-one risk factors and 8 protective factors were identified by numbering the items on this assessment tool, marked as present or absent, that were then collated in providing a risk level. Protective factors were subtracted from the risk factors. Risk levels were divided into three categories: 0-33% as low risk, 34%-66% as medium risk, and 67%-100% as high risk.

Best practices in suicide prevention and intervention

A review of best practices in suicide prevention and intervention was conducted based on the current clinical and empirical literature.


Analysis of trends

  1. Previous data on trends in key outcomes (e.g., homicide) over time

    The following Tables incorporated the data on various indices of deaths in custody presented by Weinwrath et al. (2012) that were compared with the time period (2011 – 2014). The data prior to fiscal year 2011 was taken from Weinwrath et al. (2012) and for 2011/14 from CSC files.

    A running mean was used to provide a view of the trends in the outcome data (Wainer, 2014). There were three periods sampled; 2 for 5 years, and the current review which encompassed 3 years.

    Reading across line 1 of Table 1 for the period 2001/02 – 2005/06 reflects there were 17 homicides during these 5 years reflected in an average of 3.4 a year. The average population per year was 13,096. The per capita rate (average deaths ÷ average population) was calculated to assess whether the trends in homicides was due to large variations in population that may have occurred over the time periods.

The data for Table 1 indicates homicide rates declined over the three periods sampled. This decline was not confounded by population as the per capita rates declined over time as well

Table 1 - Inmate homicides
  N Yearly Average Per Capita Rate
2001/02 – 2005/06 17 3.4 .0003
2006/07 – 2010/11 12 2.4 .0002
2011/12 – 2013/14 6 2.0 .0001

Data was available on suicides since 1991 in Table 2. There was a substantial decrease after the 1991/92 – 1995/96 period as confirmed by the per capita rates. The per capita rate was similar to the two periods sampled for accidental deaths in Table 3.

Table 2 - Suicides
  N Yearly Average Per Capita Rate
1991/92 – 1995/96 82 16.4 .0012
1996/97 – 2000/01 57 11.4 .0008
2001/02 – 2005/06 56 11.2 .0008
2006/07 – 2010/11 40 8.0 .0006
2011/12 – 2013/14 31 10.3 .0007
Table 3 - Accidental Deaths
  N Yearly Average Per Capita Rate
2002/03 – 2010/11 22 4.4 .0002
2011/12 – 2013/14 9 3.0 .0002

Table 4 summarises variations in non-natural deaths across the five regions for a five year period. Two regions, Ontario and Quebec, had the lowest rates of non-natural deaths per capita (i.e., .0010; .0008).

Table 4
Non-natural Deaths (Homicide, Suicide, Overdose, Accidental, Staff Intervention, Undetermined Causes) in Custody by Regions covering the period since the first IRC was convened, i.e., 2009/10 – 2013/14
  N Yearly Average Per Capita Rate
Atlantic 12 2.4 .0017
Quebec 14 2.8 .0008
Ontario 20 4.0 .0010
Prairies 27 5.2 .0014
Pacific  13 2.6 .0013

Review of Sample Cases 23 of 46 non-natural deaths in custody

  1. Contributing risk factors

  2. Social and criminal history

    The average age (all males) was 38 years. Their cultural background was: Caucasian (48%), Métis (17%), Inuit/First Nation (17%), and others (12%).

    Two of the inmates were employed prior to entering custody.

    Education level was not reported in 48% of the sample. Thirty per cent had high school education or less.

    All of the inmates had a prior criminal history.

    Twenty-two percent had a known gang affiliation. Seventeen percent had further charges pending. The nature of the index convictions was: violent offense [55%] comprised of murder [15%], manslaughter [15%], aggravated assault [5%] and sexual violence [20%]. The 45% of cases involved in non-violent offending included drug trafficking [20%], armed robbery [20%], forgery [5%] and arson [5%].

    Twenty per cent were serving a life sentence. Of the remaining, the sentence length was greater than ten years [31%], five to ten years [31%] and less than five years [38%]. The average length of sentence was 8 years, 4 months (excluding indeterminate sentences).

    Most of these men had served 1 to 5 years [50%], with 22% having served more than 10 years and 28% having served for less than one year. In the latter category some served sentences ranging from 19 days to four months.

    Institutional factors

    Capacity of Institution. In only one case was there a question as to whether the institution was at, or over capacity, at the time of the death of the inmate.

    Level of Security. The majority of these inmates had been assigned to medium security [62%], followed by maximum [28%] and then minimum [10%]. No cases had been assigned to the Special Handling Unit. There were no cases where the level of security determination was characterised as inappropriate.

    Placement in a Mental Health Facility. None of these men had been placed in a mental health unit.

    Placement in Segregation. Twenty-six percent were in administrative segregation at the time of their deaths.

    Custody Specific Stressors. In almost two thirds of these deaths [64%], there was an identifiable stressor in the offender’s file. Examples of these stressors included: transfer to another institution; notice received of a minor institutional charge as a result of an encounter with a correctional officer four days earlier; being bullied by another offender in the six months preceding his suicide; admission of a gang member who threatened and ultimately killed another offender; movement within the region that upset medication routines; pending murder charges; pending charges regarding a sexual assault and confinement of a victim; fearful about an impending transfer; being in segregation; recent death of mother and unable to attend her funeral; concerns regarding integration with other offenders that was gang related; and a perception that family supports were too far away.

  3. Physical and mental health care provided

  4.  Deaths in custody history

    Specific Cause of Death. There was one unknown cause (a likely accidental overdose), four homicides, five known overdoses and 13 suicides.

    Specific Means Used in Homicide. Of the four homicides, all were assaults using the following: sharp piece of wood, a physical assault, knife and steel shank.

    Accidental Overdose. There were five cases where the cause of death was by an accidental overdose. One was by an epinephrine overdose causing anaphylactic shock, two by heroin toxicity and two others by a combination of Methadone and Fetanyl.

    Suicide. There were thirteen suicides amongst this group. Seven (54%) inmates actually left a suicide note; in one case a note of apology to his psychologist for “letting him down” in the event of his death.  

    Means of Death in Suicide. Among the 13 suicides, 11 (85%) died of hanging and one each died of overdose and laceration.

    Mental health history

    Prior Mental Health Issues. Ten of the thirteen (77%) individuals who committed suicide had prior identified mental health disorders. All but one individual received a mental health diagnosis while in CSC care at some point while in custody. Of the thirteen cases, there were 31 diagnoses recorded in the file, reflecting that on average each offender had 2.4 mental health disorders. Three of these suicide cases had no recorded mental health disorders. Two of the cases had one diagnosis, six cases had two diagnoses, and two cases had three diagnoses or more.

    Table 5
    Overall Distribution of Mental Health Diagnoses
    Disorder Frequency Percentage
    Bipolar Disorder 3 10
    Major depression 4 13
    PTSD 1 3
    Anxiety 2 7
    Other Depressive Disorder 5 16
    Schizophrenia 2 7
    Adjustment Disorder 1 3
    Impulse Control Disorder 1 7
    Personality Disorder 6 19
    Substance Abuse Disorder 5 16
    Cognitive Impairment 1 3

    Suicide Risk Assessment.  Of the thirteen who completed suicides only eight (62%) were documented to have received a suicide risk assessment. Six (75%) were deemed to be low risk and 2 (25%) were identified as moderate risk. None were identified as high risk. It is not clear from the Boards of Investigation findings what methods were used to determine suicide risk levels.

    Prior Evidence of a Suicide Attempt or History of Self Harm. In nine (69%) of these cases there was a recording of a previous suicide attempt and in five cases (38%) there was a history of repeated self-harm that was not considered suicidal in nature.

    Access to Mental Health Services. Nine (69%) of the individuals in the sample were seen by a mental health practitioner within the previous two months prior to their deaths, three were seen on the same day they died, one was seen the day prior to his death, one was seen within three days prior to his death, two were seen within the week prior, and two were seen within the prior two months.

  1. Case management practices prior to the deaths

  2. The IRC reviewed the case management practices for all 23 cases. There was insufficient information in regards to the quality of the case management practices to determine if case management practice followed a risk, need and responsivity model of intervention, in an effort to reduce risk of suicidal behaviour.

  3. Security practices

  4. The Boards of Investigation reports offered detailed analyses of security practices, particularly static security, and when needed, recommended corrective measures which were subsequently implemented. Minimal information was available in regards to dynamic security.

  5. Management and response to the emergencies in regards to the deaths in custody

  6. The Boards of Investigation reports gathered detailed analyses of how emergencies were managed and responded to for serious incidents leading to deaths in custody. Corrective actions were recommended in situations where questionable practices were observed, such as a correctional officer leaving an inmate to get assistance rather than calling for assistance.

Corrective measures and management of the action plans

The Boards of Investigation reported in the areas of systemic, security, clinical, policy compliance, best practices and corrective measures. A general summary of these findings are noted in Appendix E. Furthermore, the Boards of Investigation reports concluded that the majority of deaths by suicide could not have been prevented.

The recommendations from these reports are presented at the National Investigations Meetings for approval and implementation. A Hierarchy of Effectiveness method is being considered in dealing with the action plans.

“The Incident Investigations Branch reported on incidents that lead to potential non-natural deaths, i.e., homicides, suicides and accidental deaths. Of note is the number of suicides, particularly in Regional Treatment Centres, that were prevented. In FY 2012-2013, the Prairie region experienced the highest number of attempted suicides (n=7), compared to all other regions, followed by Quebec (n=5). The majority of these attempted suicides occurred at the Regional Psychiatric Centre in the Prairies” (Suicide and Attempted Suicide Discussion Guide, 2012-2013).

The current review did not have the opportunity to verify the accuracy of these numbers.

Communication and information sharing

In their findings, the Boards of Investigation reported on matters of communication and information sharing that may have contributed to a death or that could have improved the quality of service delivery. See Appendix E, sections 1-g, 3-g and 4 for examples of the need to share information more effectively.

Review of suicide risk levels

The Boards of Investigation reported that of the eight individuals who completed suicides and had been assessed for risk of suicide, only two (25%) were identified as medium risk. This result struck the review team as curious given the description of suicide risk factors documented in the individual files. Thus, two of the members reviewed the files for the 13 suicides for fiscal years 2011 – 2014 using the procedures discussed in the Method section.

As Table 6 summarizes the risk levels tabulated by the review done by the IRC were significantly higher than those that were reported in the files. The IRC review reported that 23% of these individuals were considered low risk, 62% were deemed to be medium risk and 15% were deemed high risk.

Table 6
Comparison of Risk Level Assignment for Suicide Based on assessments reported by Boards of Investigation and the reviewers’ Retrospective Analysis
Source of Information Low Medium High
Boards of Investigation 75% 25% 0%
Retrospective analysis 23% 62% 15%

In addition, we note that the Boards of Investigation concluded that the suicides could not have been prevented in several cases (see Appendix E, section 3- a, b, c, g, j, and k).

Best practices in suicide prevention and intervention

Preventing suicide in prisons: a brief review of the latest evidence

The results in the suicide risk level analysis (section e) led to a review of the suicide literature conducted by a member of the review team. What follows is a succinct review of the current literature.

Risk Assessment. While there has been a longstanding belief that predicting and preventing suicide and non-suicidal injury has been extremely difficult (Bolton, 2004; Nock, 2014; Pokorny, 1993) there have been recent developments that offer some optimism in this regard. First, empirically based structured suicide risk assessment tools such as the Columbia Suicide Severity Rating Scale (Posner et al., 2014), and which is currently used by nurses in CSC, have been generated which has been endorsed by organisations such as the U.S. Food and Drug Administration and the U.S. Substance Abuse and Mental Health Services Administration. Relatedly, leading scholars have persuasively debunked the longstanding belief that the development of empirically based suicide risk assessment tools is not possible. Niculescu et al., (2015) recently demonstrated that suicide and suicidal behaviours could be reliably predicted (with 70% accuracy) based solely on indirect markers of suicide risk (e.g., hopelessness, extreme anxiety, acute and chronic stressors, history of suicide attempts, and acute depressive symptoms). Of note, is that this study excluded all indices of self-reported suicide ideation. While some of the most novel and methodologically impressive studies have been conducted with non-correctional, psychiatric samples (e.g., Nock, 1999; Niculescu et al., 2015), corrections-based studies, albeit less methodologically sophisticated, are emerging (see Frottier, Koenig, Seyringer, Matschnig, & Fuehwald, 2009; Horon, McManus, Schmollinger, Barr, & Jimenez, 2013).

A special note on suicidal ideation. A paradigm shift in the realm of suicide ideation has occurred. The commonly accepted practice of placing immense weight on client self-report, particularly suicide ideation, is being questioned both within and outside of the correctional context. Researchers continue to gain new insights and increased specificity in terms of how suicidal ideation should be measured to maximize predictive accuracy (Posner, Subramany, Amira, & Mann, 2014). Busch, Fawcett, and Jacobs (2003) reviewed the charts for 76 patients who committed suicide while in hospital or shortly following discharge. Seventy-eight percent of patients (n = 59) verbalized no suicide ideation during their last communication with a mental health professional prior to their suicide. Similarly, the Isometsa et al., (1995) study of 571 suicides in Finland revealed that only 22% of individuals self-reported suicide ideation to a mental health professional within one month of death. More generally, the World Health Organization (WHO, 2014) acknowledges that self-reported past suicidal behaviours (attempts and ideation) are highly unreliable and will vary as a function of age, sex, religion, ethnicity and various other factors. The need to minimize the role that clinicians place on verbal expressions of suicide ideation was recently captured by a prison suicidologist.

“We should not rely exclusively on the direct statements of an inmate who denies that they are suicidal and/or have a prior history of suicidal behaviour, particularly when their behaviour, actions and/or history suggest otherwise. Often, despite an inmate’s denial of suicidal ideation, their behaviour, actions, and/or history speak louder than their words. Take, for example, the inmate who is on suicide precautions for attempting suicide the previous day. He is now naked in a cell with only a suicide smock, given finger foods, and on lockdown status. The mental health clinician approaches the cell and asks the inmate through the food slot (within hearing distance of others on the cellblock): “How are you feeling today? Still feeling suicidal? Can you contract for safety?” Will this inmate’s response be influenced by their current predicament? How would any of us respond? (Hayes, 2013, p. 6).”

Intervention. It is estimated that 25% of mental health professionals and 50% of psychiatrists will lose a client to suicide during their career (Chemtob, Hamada, Bauer, Torigoe, & Kinney, 1988; McAdams & Foster, 2000; Pope & Tabachnick, 1993). Thus, it is not surprising that several retrospective suicide studies (all non-correctional samples) have revealed that a sizeable portion of suicide victims (20% to 91%) had recently sought help, or were formally under the care of a health care professional at, or near the time of their suicide (Capstick, 1960; Isometsa et al., 1995; Luoma, Martin, & Pearson, 2002; McCarty et al., 2011; Pearson et al., 2009; Robins, 1981). Particularly disconcerting is the close proximity in time between the last health care visit and eventual suicide. Isometsa et al.’s (1995) retrospective review of 500 suicides in Norwegian psychiatric patients revealed that 20% had been to see their therapist the same day of the suicide.

During the last thirty years researchers have attempted to systematically review the effectiveness of numerous suicide prevention programs (Crawford, Thomas, Khan, and Kulinskaia, 2007; Fawcett, 2014; Gunnel & Frankel, 1994; Ward-Ciesielski and Linehan, 2014). In the judgment of these researchers, the suicide prevention studies evidenced serious methodological limitations, nevertheless, some meaningful progress has been made in identifying some evidence-based practices.

First, researchers have considered whether the standard of care model should either a) treat suicidal behaviour as a symptom of a mental disorder (the traditional approach), or b) treat suicidal behaviour as the core disordered behaviour in and of itself, recognizing that suicide may result from any number of environmental and individual factors that may, or may not include mental disorder. Ward-Ciesielski and Linehan (2014) have stated that the standard of care model should favour the latter hypothesis.

Second, there is recent evidence that pharmacological therapies may be effective to some extent. Meta-analytic studies using antidepressants over 8 to 12 weeks were effective in reducing suicide ideation but, unfortunately, did not reduce actual suicide rates (Sher et al., 2012; Khan et al., 2003). When compliance with medication (antidepressants, lithium, and antipsychotic) is enhanced, however, there is evidence that suicide rates will decrease (Cipriani, 2013; Leon et al., 2011). Third, there is evidence that psychological-based interventions – cognitive therapy and DBT in particular – may reduce suicide behaviour but the evidence is less clear for reductions in actual suicide rates (Fawcett, 2014).

Contracting safety. Busch et al. (2003) reported that 28% of the suicide victims had contracted for their safety, thus, as others have noted (see Rudd, 2014), there is mixed evidence that contracting safety in and of itself is an effective intervention strategy. As Jamison opined, if “suicidal patients were able or willing to articulate the severity of their suicidal thoughts and plans, little risk would exist” (Jamison 1999, p. 150).

Best practice interview guidelines. Typically patients will not disclose major life events without careful probing, thus making the interviewing process critical. Fawcett (2014) recommends the use of strong clinical interviews that are not rushed, that include the use of several open-ended probes dispersed throughout the interview. Additionally, he proposes posing general question such as “Everybody worries, “what are you worrying about?”, “how is your love life?”, “any health worries?”. In essence, Fawcett argues that the goal is to illicit responses in a non-threatening manner. Thus, Fawcett questions the implicit assumption that suicidal clients cannot be helped if they do not expressly tell you they are having problems.

Current best practice guidelines. There are numerous clinical practice guidelines available to assist in the assessment and prevention of suicide. Bernert et al.’s (2014) comprehensive review identified 10 clinical practice guidelines and 12 additional resource documents with detailed non-formalized guidelines or tool-kits. In sum, current practice guidelines consistently provide recommendations in the following areas: (1) the inclusion of evidence-based risk factors and protective factors, (2) how to assess suicide intent, and (3) how to manage suicide risk once identified. Although fewer in number and typically less detailed, clinical practice guidelines specific to the correctional environment also exist (American Association of Suicidology, 2014; Federal-Provincial-Territorial Working Group on Mental Health (n.d.); World Health Organization, 2007). Table 7 summarizes corrections-specific best practice guidelines offered from five independent sources. Importantly, corrections-specific best practice guidelines typically overlap in core elements. However, evaluations of these commonly agreed upon best practice guidelines are noticeably absent. Furthermore, corrections-based guidelines like those found in the general literature fail to delineate how risk factors should be combined to arrive at a risk classification level.

Table 7

A synthesis of suicide prevention best practice guidelines in correctional settings

  • Intake and on-going screening at critical points (e.g., transfers)
  • Comprehensive assessment and on-going re-assessment for those at risk
  • Monitoring
  • Meaningful Communication between all parties
  • Meaningful Suicide Prevention Training (provided annually)
  • Treatment without barriers
  • Non-punitive suicide interventions
  • Housing - avoid isolation unless constant observation is possible, house with other inmates in safe environment, increase interaction between inmates and correctional, medical and mental health staff, require regular follow-ups of inmates released from mental health monitoring
  • Suicide-resistant housing (clothing, showers) commensurate with risk level
  • Post suicide procedures (procedures for notification, reporting and reviewing)

Note. Guidelines extracted from: Federal-Provincial-Territorial Working Group on Mental Health, (n.d.); Metzner, & Hayes, (2012); Tripodi & Bender, (2007); Hayes (2013); World Health Organization, (2007)

The previous IRC also recommended best practice for suicide prevention in prison. Weinwrath et al, (2012) summarized the literature as follows:

“First of all, staff culture and cooperation seem to be critical to the successful implementation of prison suicide prevention programmes. Best practices for preventing suicides in jail and prison settings are based on the development and documentation of a comprehensive suicide prevention plan with the following elements:

  • A training programme (including refreshers) for correctional staff and care givers to help them recognize suicidal inmates and appropriately respond to inmates in suicidal crises.
  • Attention needs to be paid to the general prison environment (levels of activity, safety, culture and staff-prisoner relationships).
  • In particular, the quality of the social climate of prisons is critical in minimizing suicidal behaviours. While prisons can never be stress free environments, prison administrators must enact effective strategies for minimizing bullying and other violence in their institutions, and for maximizing supportive relationships among prisoners and staff.
  • The quality of staff-prisoner relationships is critical in reducing prisoners’ stress levels and maximizing the likelihood that prisoners will trust staff sufficiently to disclose to them when their coping resources are becoming overwhelmed, feelings of hopelessness, and suicidal ideation.
  • Procedures to systematically screen inmates upon their arrival at the facility and throughout their stay in order to identify those who may be at high risk.
  • A mechanism to maintain communication between staff members regarding high-risk inmates.
  • Written procedures which outline minimum requirements for housing high-risk inmates; provision of social support; routine visual checks and constant observation for acutely suicidal inmates; and appropriate use of restraints as a last resort for controlling self-injurious inmates.
  • Inmates with mental disorders in need of treatment should receive it (pharmacological or psychosocial interventions) and be kept under strict observation.
  • Development of sufficient internal resources or links to external community-based mental health services to ensure access to mental health personnel when required for further evaluation and treatment.
  • A strategy for debriefing when a suicide occurs towards identifying ways of improving suicide detection, monitoring, and management in correctional setting.” (p.27)

Of the guidelines cited above, the IRC did not note strategies for debriefing when a suicide occurred that could lead to identifying ways of improving detection, monitoring, and management in correctional settings.

In our opinion all of the above recommendations are of value, but from our perspective a, c, d, h, and j merit special attention at the present time.


Analysis of trends

The evidence presented in Tables 1–3 on homicides, suicides and accidental deaths point to the reality that there has been a decline or at least a stabilization (after initial high numbers) in accidental deaths, homicides and suicides since 2002. It is plausible that these trends may in part be due to the fact that CSC has responded to the initial harsh critiques of their lack of appropriate policies (see Gabor, 2007) in protecting the lives of their inmates and generated more effective policies in this regard, which certainly appears to be the conclusion reached by Weinwrath et al., (2012).

This is, however, a qualified conclusion. The time spans evaluated were relatively brief. For example, it could be argued that in the case of homicides and suicides, the first averages recorded (3.4; 16.4) in Tables 1 and 2 were outliers. Data on these outcomes from earlier time periods (e.g., 1970 –1990) would be invaluable to be more confident of the credibility of the trends reported to date. Moreover, it would be informative to compare CSC outcomes and policy data with other jurisdictions.

The results for regional variation on non-natural deaths (see Table 4) indicate two regions have lower rates when the results are weighted by inmate population. The reasons for this could be due to sampling error as only one time period was assessed. The other possibilities might be the composition of inmates, management and custodial practices, the physical structure of settings amongst others. Before speculating, it is necessary to tabulate regional variations data well before 2009.

Review of Sample Cases 23 of 46 non-natural deaths in custody

The current sample (n =23) investigated by the review team varied from data presented by previous evaluators (see Weinwrath et al., 2012) and the institutional factors may have changed somewhat from previous years. However, we have no way of confirming this hypothesis. Suffice it to say that situational factors that this IRC found affecting vulnerable inmates are vitally important to assess on an on-going basis. Some examples are, impending charges, recent transfers, changes in medication, stresses from other inmates and stresses external to the prison situation.

We disagree with the findings of the Boards of Investigation that self-injurious behaviours such as head banging are attention seeking behaviours and not potentially related to suicidal behaviour. The correlation between non-suicidal self-injury and suicide is moderate to high (Muehlenkamp, 2014). The point is that empirically non-suicidal self-injury and suicide attempts/completed are correlated and that non-suicidal self-injury in and of itself is a risk factor for later lethality. A clinical approach, which does not rely predominately on what the person is thinking or explicitly verbalizing regarding suicide ideation, is required. Quick measurable behavioural cues that are research-based and tailored to take into account unique prison environment will likely estimate suicide risk more accurately.

Corrective measures and management of the action plans

A recurring, implicit theme in many of the Boards of Investigation reports was one of “nothing could have been done to prevent this death” because “he contracted for safety but took his life anyway” or “he chose not to seek help, or he evidenced forward thinking, therefore his death was not preventable”. In closing, the findings of the Boards of Investigation conclude that the majority of the suicides could not have been prevented. This conclusion is inconsistent with the recent trends in the suicide literature. It is hoped that the next chapter in the evolution of CSC’s suicide policy development will actively incorporate the literature’s cutting edge findings. In order to promote best practice it is suggested that staff intervention that has produced good results be highlighted.

Communication and information sharing

In reviewing all individual file information, the IRC noted that the Boards of Investigation reports reflect that, for the most part, communication and information sharing among stakeholders during the occurrence of an incident appear to be satisfactory and/or appropriate corrective measures have been recommended. What is unclear from the Boards of Investigation reports is how communication and information sharing occurs between the various stakeholders (e.g., security personnel, mental health staff and who has referred someone potentially suicidal) in regards to the indicators of a possible suicide risk. For example, does a mental health professional performing a comprehensive suicide risk assessment know that an inmate might have lost his support system (protective factors) therefore elevating the risk or that a transfer to another institution is threatening the loss of support system?

Review of suicide risk levels

In view of the results of the comparison of the Boards of Investigation reports and the IRC reviewers’ analysis of suicide risk levels, there is a need to review how suicide risk assessments are determined and the extent to which they include what is reported in the literature on best practice. We understand that the suicide assessment protocol was changed in 2011 (this was noted in an individual file). We did not have sufficient time within this review to find out what changed for the better.

Best practices in suicide prevention and intervention

CSC’s suicide assessment practices, as best as we could determine from the information provided to us, appear to be firmly aligned with the historic “gold” standard of care in suicide prevention. This model, often referred to in the literature as a “clinical” approach (see Bonta, Laws & Hanson, 1998) includes reliance on client self-report (e.g., recent verbalizations of suicide ideation), and loosely structured clinical judgment (e.g., an agency provides the clinician with a list of suicide risk factors; the clinician then decides how to weight and sum these factors on a case-by-case basis but there is no quantitative summary for risk levels). As noted previously, it is clear from the literature review that these practices are increasingly being questioned in favour of using a statistical approach which was used by the reviewers in calculating the results reported in Table 6.

The latest research findings reflect that suicide risk assessment tools with demonstrated predictive utility can be empirically developed, and that these tools need not rely solely on self-report. Thus, the future of suicide risk assessment in correctional settings must become more structured, and more firmly rooted in empiricism.

Suicide prevention in prisons will be further enhanced by (1) actively assessing protective factors (Rudd, 2014), (2) ensuring compliance with pharmacological interventions (Cipriani, 2013), and (3) conceptualizing suicide in and of itself as the core disordered behaviour in need of treatment, rather than a symptom of a mental disorder (Ward-Ciesielski & Linehan, 2014).

As for intervention, the lack of an adequate rigorous evaluation precludes the provision of firm directives on “how to do it”. The aforementioned best practice guidelines (e.g., ensuring compliance with medication, using DBT) are recommended when cognitive interventions are applied to individuals with mental health disorders, short-term cognitive therapies may be effective with mood disorders but requires long-term cognitive therapies for personality disorders.

Given the preceding findings we now summarize the recommendations forthcoming from our work.


Recommendation 1

It is recommended that the comprehensive suicide/self-injury assessment tool be reviewed to reflect the latest best practice in the literature and be quantified to enhance reliability across time and users. The tool should include both acute and chronic stressors within and external to the prison environment.

Recommendation 2

All relevant internal representatives (e.g., mental health, security, research etc…) within CSC should develop a short and long term program of research. Retrospective studies using Offender Management System data, psychological data, and if possible, all existing Boards of Investigation findings should be prioritized in the short term to evaluate the predictive validity and the utility of the existing screening and comprehensive tools. CSC should also explore the feasibility of developing an empirically-based structured suicide risk assessment tool that does not rely on self-reported suicide ideation. There are studies in the literature that could serve as models (see Niculescu et al., 2015).

Recommendation 3

CSC should establish a national advisory committee comprised of external experts as well as internal CSC representatives with expertise along with other relevant stakeholders with the expressed mandate to: (1) review existing screening, comprehensive assessment methods, mental health monitoring practices, training materials, suicide intervention practices and (2) advise on the development of a pilot intervention that does not involve the use of isolation.

Recommendation 4

Mandate and automate information sharing amongst all stakeholders.

Recommendation 5

CSC should ensure that all staff (including employees on contract) are indeed being trained in suicide risk prevention including staff who interact with inmates but not necessarily in a clinical or security capacity (e.g., librarians, chaplains, teachers).

Recommendation 6

CSC should collect data on non-natural deaths in custody as far back as possible to examine trends over time. This examination should focus on whether noticeable changes in deaths in custody rate might be due to changes in policy, operations, socio/cultural/political factors both internal and external to CSC.

Recommendation 7

Account for regional disparities when non-natural deaths vary noticeably across regions.

Recommendation 8

That the Incident Investigations Branch reframe the Terms of Reference within Convening Orders for suicides to ensure that the questions get to the root causes of non-natural deaths in custody including dynamic security, case management practices, information sharing, quality of communications, suicide prevention, intervention, and post-intervention etc.

Recommendation 9

The IRC recommends that future independent reviews be carried out every two years rather than three as there are sufficient incidents and recommendations to warrant an external review. The two year timeframe for a review should be formalised and put forth in policy.


American Association of Suicidology. (2014). Inmate Suicide 2014. Retrieved from

Bernert, R.A., Hom, M.A., Weiss Roberts, L. (2014). A review of multidisciplinary clinical practice guidelines in suicide prevention: Toward an emerging standard in suicide risk assessment and management, training and practice. Academic Psychiatry, 38, 585-592. Doi: 10.1007/s40596-014- 0180-1

Bolton, J. (2015). The cutting edge: Suicide risk assessment. Depression and Anxiety, 32, 73-75. Doi: 10.1002/da.22320

Bonta, J., Law, M., & Hanson, K., (1998). The prediction of criminal and violent recidivism among mentally disordered offenders: a meta-analysis. Psychological Bulletin, 123(2), 123-142.

British Ministry of Justice (2012). Guide to safety in custody statistics. London UK.

Busch, K.A., Fawcett, J., & Jacobs, D.G. (2003). Clinical correlates of inpatient suicide. Journal of Clinical Psychiatry, 64 (1), 14 – 19.

Capstick, A. (1960). Urban and Rural Suicide. Journal of Mental Science, 106, 1327-1336.

Chemtob, C., Hamada, R., Bauer, G., Torigoe, R., & Kinney, B. (1988). Patient suicide: Frequency and impact on psychologists. Professional Psychology, Research and Practice, 19, 416-420.

Cipriani, A. (2013). Review: lithium reduces the risk of suicide compared with placebo in people with depression and bipolar disorder. Evidence Based Mental health, 16(4), 1-13.

Crawford, M.J., Thomas, O., Khan, N., & Kulinskaia, E. (2007). Psychosocial interventions following self-harm: Systematic review of their efficacy in preventing suicide. British Journal of Psychiatry, 190, 11-17.

Fawcett (2014). Is suicide clinically preventable? What is the evidence? In K.E. Cannon and T.J. Hudzik (Eds.), Suicide: Phenomenology and Neurobiology (pp. 141 - 153), Switzerland: Springer International Publishing. Doi: 10.1007/978-3-319-09964-4_8

Federal-Provincial-Territorial Working Group on Mental Health (n.d.). Mental Health Strategy for Corrections in Canada: A federal-provincial-territorial partnership. Retrieved from

Frottier, P., Koenig, F., Seyringer, M., Matschnig, T., & Fuehwald, S. (2009). The distillation of “VISCI”: Towards a better identification of suicidal inmates. Suicide and Life-Threatening Behavior, 39, 376-385.

Gabor, T., (2007). Death in custody. Office of the Correctional Investigator, Ottawa, ON.

Gunnel, D., & Frankel, S. (1994). Prevention of suicide: Aspirations and evidence. British Medical Journal, 308, 1227-1233.

Hastings, R., Foristell, G. & Graceffo, G. (2011). The final report of the independent review committee into deaths in custody: 2009-2010. Correctional Services Canada. Ottawa, ON.

Hayes, L.M. (2013). Suicide prevention in correctional facilities: Reflections and next steps. International Journal of Law and Psychiatry, 36, 188–194.

Horon, R., McManus, T., Schmollinger, J., Barr, T., & Jimenez, M. (2013). A Study of the use and interpretation of standardized suicide risk assessment: Measures within a psychiatrically hospitalized correctional population. Suicide and Life-Threatening Behavior, 43, 17-38.  Doi: 10.1111/j.1943-278X.2012.00124.x

Isometsa, E.T, Heikkinen, M.E., Marttunen, M.J., et al. (1995). The last appointment before suicide: is suicide intent communicated? American Journal of Psychiatry, 152, 919–922.

Jamison, K. (1999). Night Falls Fast: Understanding Suicide. New York: Knopf.

Khan, A., Khan, S., Kolts, R., & Brown, W.A. (2003). Suicide rates in clinical trials of SSRIs, other antidepressants, and placebo: analysis of FDS reports. American Journal of Psychiatry, 160, 790–792.

Leon, A.C., Solomon, D.A., Li, C. et al. (2011). Antidepressants and risk of suicide and suicide attempts: A 27-year observational study. Journal of Clinical Psychiatry, 72, 580–586.

Luoma J.B., Martin, C.E., & Pearson, J.L. (2002). Contact with mental health and primary care providers before suicide: A review of the evidence. American Journal of Psychiatry, 159, 909–916.

McAdams, C.R., & Foster, V.A. (2000). Client suicide: Its frequency and impact on counsellors. Journal of Mental Health Counseling, 22, 107-121.

McCarty, C.A., Russo, J., Grossman, D.C., Katon, W., Rockhill, C., McCauley, E., Richards, J., Richardson, L. (2011). Adolescents with suicidal ideation: Health care use and functioning. Academy of Pediatrics, 11, 422–426.

Metzner, J. & Hayes, L. (2012). Suicide prevention in jails and prisons. In R. Simon & R. Hales (Eds.), Textbook of Suicide Assessment and Management (pp. 389-403). Washington, D.C.: American Psychiatric Association.

Muehlenkamp, J.L. (2014). Distinguishing between suicidal and nonsuicidal self-injury. In M.K. Nock (Ed.), Oxford handbook of suicide and self-injury (pp.23-46). New York: Oxford University Press.

Niculescu, A.B., D F Levey, D.F.,, Phalen, P.L Le-Niculescu, H.,, Dainton, H.D., Jain, N. Belanger,E. James, A. George, S., Weber, H., Graham, D.L., Schweitzer, R., Ladd, T.B., Learman, R., Niculescu, E.M., Vanipenta, N.P. Khan, F.N., Mullen, J., Shankar, G., Cook, S., Humbert, A., Ballew, C., Yard, M. Gelbart,T., Shekhar, A., Schork, N.ZJ., Kurian, S.M., Sandusky, G.E., & Salomon, D.R. (2015). Understanding and predicting suicidality using a combined genomic and clinical risk assessment approach. Molecular Psychiatry, 1-20; doi: 10.1038/mp.2015.112

Nock, M.K., Hwang, J., Sampson, N., Kessler, R.C., Angermeyer, M., Beautrais, A. et al (2009) Cross-national analysis of the association among mental disorders and suicidal behavior: findings for the WHO World Mental Health Surveys. PLoS Medicine 6(8), 1-17.

Office of the Correctional Investigator (2014). Annual report: 2013-2014. The Correctional Investigator. Ottawa, ON.

Office of the Correctional Investigator (2014). A Three Year Review of Federal Inmate Suicides
(2011–2014). Retrieved from

Pearson, A., Saini, P., Da Cruz, D., Miles, C., While, D., Swinson, N., Williams, A., Shaw, J., Appleby, L., Kapur, N. (2009). Primary care contact prior to suicide in individuals with mental illness. British Journal of General Practice, 59, 825-832. Doi: 10.3399/bjgp09X472881.

Pokorny, A.D. (1993). Suicide prediction revisited. Suicide and Life-Threatening Behavior 23(1), 1–10.

Pope, K., & Tabachnick, B. (1993). Therapists’ anger, hate, fear, and sexual feelings: National survey of therapist response, client characteristics, critical events formal complaints, and training. Professional psychology: Research and Practice, 24, 142-152.

Posner, K., Subramany, R., Amira, L., & Mann, J.J. (2014). From uniform definitions to prediction of risk: The Columbia Suicide Severity Rating Scale approach to suicide risk assessment. In K.E. Cannon, & T.J. Hudzik (Eds.), Suicide: Phenomenology and neurobiology (pp. 59-84). Switzerland: Springer International Publishing. Doi: 10.1007/978-3-319-09964-4

Robins, E. (1981). The final months. New York: Oxford University Press.

Rudd, M.D. (2014). Core competencies, warning signs, and a framework for suicide risk assessment in clinical practice. In M.K. Nock (Ed.), Oxford handbook of suicide and self-injury (pp. 323-336).  New York: Oxford University Press.

Sher, L., Stanley B.H., Posner. K., Ardendt, M., Grunebaum, M.F., Neria, Y., Mann, J.J., & Oquendo, M.A. (2012) Decreased suicidal ideation in depressed patients with or without comorbid stress disorder treated with selective serotonin reuptake inhibitors: An open study. Psychiatry Research, 190, (2–3), 261-266.

Tripodi, S.J., & Bender, K. (2007). Inmate Suicide: Prevalence, Assessment, and Protocols. Brief Treatment and Crisis Intervention, 7(1) 40-54. Doi: 10.1093/brief-treatment/mhl016

Wainer, H. (2014). Medical illuminations using evidence, visualization, statistical thinking to improve health care. Oxford, U.K.: Oxford University Press.

Ward-Ciesielski, E.F. & Linehan, M.M. (2014). Psychological treatment of suicidal behaviours. In M.K. Nock (Ed.), Oxford handbook of suicide and self-injury (pp. 367-384). New York: Oxford University Press.

Weinwrath, M., Wayte, T. & Arboleda-Florez, J. (2012). The final report of the independent review committee into federal deaths in custody: 2010-2011. Correctional Services Canada. Ottawa, ON.

World Health Organization. (2007). Preventing suicide in jails and prisons. Geneva SW. Retrieved from

World Health Organization. (2014). Preventing suicide: A global imperative. Retrieved from

Author Note

Yvette L. Thériault, M.A. p.s., is a licensed psychologist and honorary member of the College of Psychologists of New Brunswick as well as a certified psychologist of the College of Psychologists of Ontario. She is a Fellow of the Canadian Psychological Association, a private practitioner and a criminal justice consultant. She is past chairperson for Health Care Advisory Committee – Correctional Service of Canada (HCAC).

Alan Leschied, Ph. D., is a certified psychologist of the College of Psychologists of Ontario, and a full professor with the Faculty of Education at the University of Western Ontario. He is a Fellow of the Canadian Psychological Association, a private consultant for youth and adult courts. He is a past member of the HCAC.

Shelley Brown, Ph. D., is an Associate Professor of forensic psychology within the Department of Psychology, and the Research Ethics Chair of the Carleton University Research Ethics Board. She has 10-year research career with Correctional Service of Canada prior to her current position.

We would like to acknowledge the contribution of Paul Gendreau, O.C., Ph. D. for trend analysis, general input on relevant content and international expertise.

Our appreciation goes to Maria Hill for her assistance with the many requirements of this committee.

Appendix A


FROM APRIL 1st, 2011 TO MARCH 31st, 2014

WHERE AS it is provided by Section 20 of the Corrections and Conditional Release Act (CCRA) , S.C. 1992, c. 20 that the Commissioner of the Correctional Service of Canada (CSC) may appoint a person or persons to investigate and report upon any matter relating to the operations of the Service.


WHERE AS in June 2007, the Office of the Correctional Investigator (OCI) released a report reviewing inmate deaths from 2001 to 2005 that brought to the attention of CSC a number of issues with respect to “Deaths in CSC Custody”.  In response, I, Don HEAD, Commissioner of Corrections, wrote a letter in which I made the commitment to “implement an independent review group to assess, on an annual basis, CSC’s actions and responses to various investigations and other reports on deaths in custody”.  The second Independent Review Committee (2012) recommended that future reviews be convened every three years.  In November 2014, CSC published its response to the second IRC, wherein the Service committed to convening a third IRC in 2015.


WHERE AS, from April 1st, 2011 to March 31st, 2014, there were 46 non-natural deaths in our custody which resulted in an investigation being convened in compliance with CCRA requirements.  A random sampling of 50% of the investigations resulted in the identification of 23 deaths in custody (see Annex B) to be examined by the current IRC.

NOW THEREFORE I, Don HEAD, Commissioner of Corrections, do hereby appoint, by virtue of Section 20 of the CCRA, Yvette THÉRIAULT (Private Psychologist, and Criminal Justice Consultant) as Chairperson of the Independent Review Committee, and Dr. Shelley BROWN (Associate Professor, Department of Psychology, Carleton University), and Dr. Alan LESCHIED (Psychologist, and Professor, Faculty of Education, University of Western Ontario) as members of the Independent Review Committee.

All other aspects of the Convening Order and Terms of Reference dated the 5th day of June 2015 remain in effect.

I DIRECT AND CHARGE the persons so appointed to faithfully execute the duties entrusted to them and provide me with:

  • Analysis of the appropriateness and adequacy of the corrective measures and the management of the action plans supported by CSC’s National Investigations Meetings (NIM) in order to address the identified gaps.
  • Analysis of the quality of the communication and information sharing amongst CSC’s key players (health Services / Security / Correctional Operations and Programs / site management) pre- and post-incidents.
  • The identification and analysis of any trends observed within the following areas such as:
    • the contributing risk factors or precipitating events to the deaths;
    • the physical and mental health care provided to the inmates with mental health needs prior to their deaths;
    • the case management practices pertaining to the inmates prior to their deaths;
    • the security practices such as staff presence and monitoring of the inmate activities; and,
    • the management of and response to the emergencies in regards to the deaths of the inmates.

AND I FURTHER DIRECT the Independent Review Committee to provide me with successful and best practices in other international correctional jurisdictions with respect to deaths in custody.

AND FURTHER, to ensure the success of this review, I authorize the Independent Review Committee to:

  • adopt such procedures and methods as may be deemed necessary for the proper conduct of this review;
  • be provided with adequate and secure working conditions and administrative support as required;
  • have complete access to personnel under the employment of, or under contract with, the Correctional Service of Canada; and,
  • communicate with any outside person, agency, office or organization which may assist in the successful completion of this review.

AND I FURTHER DIRECT the Independent Review Committee to:

  • provide me an interim briefing on your findings including any potential recommendation(s) no later than the 1st of November 2015,

Given under my hand in the City of Ottawa, in the Province of Ontario, this 24th day of June 2015.

Original signed by:

Don Head
Correctional Service of Canada

Annex to convening order

This investigation is convened under Section 20 of the Corrections and Conditional Release Act which reads: “The Commissioner may appoint a person or persons to investigate and report on any matter pertaining to the operations of the Service.”  Section 21 of the Act stipulates that, for investigations convened under Section 20, Sections 7 to 13 of the Inquiries Act apply.

By virtue of Sections 7 to 13 of the Inquiries Act, for the purpose of their investigation, members of national Boards of Investigation have all the powers of “commissioners” under the Inquiries Act, and those special powers and responsibilities are as follows:

INQUIRIES ACT. R.S., c. I-13, s.1


  1. For the purposes of an investigation ..., the commissioners
    • may enter into and remain within any public office or institution, and shall have access to every part thereof;
    • may examine all papers, documents, vouchers, records and books of every kind belonging to the public office or institution;
    • may summon before them any person and require the person to give evidence, orally or in writing, and on oath or, if the person is entitled to affirm in civil matters on solemn affirmation; and
    • may administer the oath or affirmation under paragraph (c). R.S., c. I‑13, s. 7.
  1. (1) The commissioners may, under their hands, issue a subpoena or other request or summons, requiring and commanding any person therein named;
    • to appear at the time and place mentioned therein;
    • to testify to all matters within his knowledge relative to the subject-matter of an investigation; and

Appendix B

ANNEX A – Statement of Work

The Correctional Service Canada (CSC) has a requirement for and independent review of a random sampling of deaths in custody which occurred during the 2011/2012, 2012/2013, and 2013/2014 fiscal years*.

*A fiscal year is defined as a period of twelve (12) months starting the 1st of April of any given year and ending the 31st of March of the following year.

1. Background

In 2008, the Office of the Correctional Investigator released a report on deaths in custody entitled “A Preventable Death” which called for better coordination between correctional and mental health systems across the country. Following the release of this report, CSC identified several priority actions that would be taken by the Service, and initiated additional measures to improve the response to deaths in custody.  One of these actions was the implementation of an Independent Review Committee (IRC) which would conduct an annual external review of CSC’s actions and responses to deaths in custody.

To date, two IRCs have been convened into deaths in custody which occurred during the 2009/2010 and 2010/2011 fiscal years*, from which many recommendations were formulated and implemented. After the second IRC, it has been decided that an IRC into deaths in custody will be convened every three years. Therefore, CSC has committed to convein a third IRC in 2015 with the intent of reviewing deaths in custody which occurred during the 2011/2012, 2012/2013, and 2013/2014 fiscal years.

2. Objectives

As a Chairperson of the IRC, to conduct a review of a random sampling of deaths in custody which occurred during the 2011/2012, 2012/2013, and 2013/2014 fiscal years. 

3. Tasks

The Contractor must perform, without being limited to, the following tasks:

  1. A review of 50% of the investigations convened into non-natural deaths in custody during the review period (2011/2012, 2012/2013, and 2013/2014 fiscal years). This percentage must equitably represent the review period;
  2. An analysis of the appropriateness and adequacy of the corrective measures and the management of the action plans supported by CSC’s National Investigations Meetings (NIM) in order to address the identified gaps;
  3. An pre- and post-incidents analysis of the quality of the communication and information sharing amongst CSC’s key players (Health Services / Security / Correctional Operations and Programs / Site Management);
  4. The identification and analysis of any trends observed within the following areas such as:
    • the contributing risk factors or precipitating events to the deaths;
    • the physical and mental health care provided to the inmates with mental health needs prior to their deaths;
    • the case management practices pertaining to the inmates prior to their deaths;
    • the security practices such as staff presence and monitoring of the inmate activities; and
    • the management of and response to the emergencies in regards to the deaths of the inmates.
  5. An interim briefing (in person or via teleconference) to the Commissioner or his/her designated representative on the IRC’s findings including any potential recommendation(s) no later than the 01-November-2015.

4. Deliverables

The Contractor must submit the following deliverables:

  1. A copy of the interim briefing to the Commissioner on findings including any potential recommendation(s) no later than the 01-November-2015; and
  2. A final report containing analysis pertaining to the tasks identified at Section 3 – Tasks, and findings or recommendations (if any), no later than the 01-December-2015.
  • All Deliverable documents must be done with Microsoft Suite 2007 products;
  • The Contractor must provide one (1) soft copy of all deliverables;
  • All deliverables are to be provided in English, CSC is responsible for the translation.

5. Applicable Documents

The work described in this SOW will be based on the documents provided by CSC that will include:

  • A Briefing Note to the Commissioner;
  • A Convening Order and Terms of Reference; and
  • The final Board of Investigation reports and consultation grids for each of the deaths in custody reviewed by the IRC.

Appendix C

The material provided by CSC at the beginning of the review.

  • Commissioner’s Directive 041, Incident Investigations
  • 2010 IRC Report and CSC Corrective Measures and Management Action Plan
  • 2012 IRC Report and CSC Corrective Measures and Management Action Plan
  • The 23 deaths - Boards of Investigation final Reports, Grids, and Closure Memos, and inmates’ RADAR Custom Reports (when applicable)
  • Incident Investigations Branch Deaths in Custody Discussion Guides (July 2013 to April 2015)
  • Hierarchy of Effectiveness Decks and Consultation Grid template
  • CSC’s Annual Deaths in Custody Report (2013/2014)

Appendix D

Review of cases

Variables used in the review of individual files and collated for analysis

Demographics. This section included: gender, age, ethnicity, employment status prior to custody, highest level of education, and gang affiliation.

Offense Specific. This section included: most serious offense leading to custody, were further charges pending, and was inmate serving a life sentence.

Institution Specific. This section included: the region, specific institution, and level of custody designation.

Custody Specific. This section included: length of sentence, length of custody, whether the level of custody was considered appropriate, whether there was a placement in a mental health unit,and whether there was a placement in segregation.

Custody Specific Stressors. This section included: whether there was/were a recent custody specific stressor(s), whether / what were the conditions of deprivation, transfer to another institution, and was the institution considered to be over capacity.

Death Specific. This section includes: cause of death: if homicide include the means, name the weapon; if the cause of death was suicide what was the means (drug overdose, laceration, hanging); if suicide what was considered the major contributing factor (i.e., a mental health disorder, being in custody, conflict with another person); if suicide, was there a suicide note: if suicide, was there a previous attempt, and if there was a previous attempt had a suicide risk assessment been completed; if there was a previous suicide assessment what was the rating (low, moderate, high or very high); was there evidence of repeated harm, and if yes to repeated self-harm was it considered suicidal with intent or considered non-suicidal self-injurious; if the cause of death was considered accidental, what was the nature of the accident (i.e., overdose), if it was an overdose what was the intoxicant, poisoning.

Location of the Death. This section identifies whether the location of the death was: single cell, multiple cell, dormitory, courtyard, medical facility, mental health facility, or segregation.

Timeline Surrounding the Death. This section identifies when the death occurred relative to: admission of the inmate to the facility (i.e., less than 30 days, 30-60 days, 90 days to one year or over one year; when did the death occur relative to the discharge date of the inmate (i.e., less than 30 days, 30-60 days, 90 days to one year, or over one year).

Access to Contraband. This section identifies: whether there was access to contraband, and if so, what was it.

Mental Health Specific. This section relates to: whether there was a mental health diagnosis, if yes, what was it (i.e., bipolar, major depression, other depressive disorders, schizophrenia, adjustment disorder, impulse disorder, personality disorder, substance abuse disorder; when was the last clinical mental health service provided prior to death (i.e., same day, the previous day, previous three days, previous week or previous month, more than previous month).

Appendix E

Boards of Investigation findings for the fiscal years 2011 – 2014

  1. Systemic
    1. Having a single national coordinator to monitor safe custody
    2. Provide 24 hour nursing coverage at all institutions
    3. Better educate staff on management of inmates at risk of self-injury or suicide
    4. Outdated suicide refresher training of more than two years
    5. That Regional and National Headquarters assess and implement appropriate measures to record cameras that are mounted in front of the units at all times
    6. Improve the identification of gang affiliations at intake using the Security Threat Elimination Program
    7. Improved information sharing between mental health staff and correctional staff within segregation (in one case, segregation unit staff were notified via email, but not the segregation unit log book that the inmate had ‘contracted safety’ with psychologist 
    8. Implement a national review of the inadequacy of institutional resources to manage offenders with moderate mental health needs; review should also consider the feasibility of intermediate mental health care units within each region (last phase of five-part mental health strategy that remains unfunded)
    9. Review the feasibility of acquiring wheeled (emergency) stretchers in all institutions
    10. Review policy regarding the activation and deactivation of alerts in OMS when an inmate is placed and then terminated from Mental Health Monitoring
  1. Security
    1. CSC considers developing standards for the installation, financing and operation of cells used for recovering human excrements
    2. Review the adequacy of how security level is currently determined for Aboriginal offenders
    3. Review existing practices for the identification of weapons-related contraband (e.g., shanks as well as missing knives) home-made weapons as well
    4. Increase security patrols to ½ hour  between 22:00 and 7:00 hours at the institution
    5. Encourage correctional officers to do more frequent security patrols in institutions with older living units during high inmate activity time (e.g., during weekends and evenings)
    6. Afford correctional officers more professional discretion during medical emergencies
    7. Research and potentially purchase a better 911 Rescue Tool
    8. Inmate cell was not searched prior to be placed in it
    9. Counts were not done right, “stop, turn, and look into each cell”.
    10. First correctional officer on scene did not have a radio (cost valuable time as he had to run for help)
  1. Clinical observations
    1. In the majority of suicides no behavioural indicators were found preceding the successful attempts except for historical static indicators; extreme impulsivity implicated in one case.
    2. No evident pre-incident indicators suggestive that someone might attempt suicide or immediate risk factors to an elevated risk of suicide were evident to staff
    3. When there was a comprehensive suicide assessment the inmates did not appear to be agitated or in any distress during the brief period.
    4. No available detection or testing for the illegal lethal substance Fentanyl.
    5. Mix of prescription drugs in combination with illegal drug is potentially lethal.
    6. An inmate diagnosed with Hepatitis C had no recorded visits with a nurse
    7. Releasing institution did not communicate one inmate’s mental health status to the receiving institution, albeit the inmate had been assessed as low suicide risk
    8. In one case, the EpiPen being carried by the inmate had not been approved; there was also no recent allergy testing
    9. Review local procedures regarding the practice of granting external decision makers the authority to terminate a current prescribed medication without internal review
    10. Staff (parole officer (PO)) erroneously dismissed an inmate’s threat to hang himself because in the PO’s opinion it was a manipulation tactic against the inmate’s wife; unclear whether the PO relayed the information to the psychologist, minimally the PO did not follow standard operating procedures by not documenting suicide threat
    11. Inmate did not participate in Suicide Awareness and Prevention Workshop at intake
  1. Policy compliance

In some cases not all policies were followed resulting in less than desirable practices. Generally they were in the area of locations and/or dysfunction of security cameras, lack of preservation of crime scenes, emergency simulation exercises not carried out, next-of-kin information not properly recorded, health checklist not completed at time of death, protective facial mask not used during cardiopulmonary resuscitation, someone leaving the site of an incident to call for help, patrol not carried out as required, correctional officer uses an unauthorized knife to cut down an inmate, delays in starting resuscitation due to delayed communication from security to health care, institutional count started late, staff not wearing radios at the time of death to notify backup, security patrol issues, staff checklist describing medical procedures absent, missing CASA, failure to notify health services when inmate was found unwell, incomplete security classification assessment (i.e., Aboriginal Status not completed), correctional officer entered cell following inmate death contrary to policy, delays in calling ambulance, staff not clearing the emergency scene expediently, hospital escorts not fully briefed regarding required duties. 

One-for-one razor exchange policy was utilized in the Segregation Unit where a suicide by razor blade occurred, however the standing order and post orders did not include the razor protocol.

Institutional psychologists did not document routine contacts with inmates. Staff involved with the inmate were not offered Critical Incident Stress Management. Correctional officer II was not aware that the inmate was on his caseload. Suicide risk forms were not completed in a timely fashion upon admission to segregation.

Only one officer conducted the 2200 hour count, policy dictates two.

Cell not fully secured and provided with crime scene protection

  1. Best practice
    1. To offer Critical Incident Stress Management services to those more indirectly affected by stressful events.
    2. To create an inmate peer support program for suicide prevention, mental health and wellness. Peer helpers in suicide and self-harm education and support to be extended to national level.
    3. That health care staff target medications with potential lethal effect particularly in combination with illegal drugs.
    4. Have clean clothes readily available for staff to change if needed
    5. That National Headquarters (NHQ) health services study the introduction of Fentanyl in the prisons and the possibility of finding a fast detection of this dangerous substance
  2. Corrective measures
    1. In one case that local management review and develop an institutional protocol for bullying behaviour
    2. Review CPR policy
    3. Document police removal of evidence
    4. Implement canine searches of stored inmate effects
    5. Review policies and practices related to self-administered medication
    6. Review practices for ensuring inmate compatibility in the yard
    7. Issue security notification regarding Fentanyl
    8. Review the gathering and analyses of security intelligence information
    9. Review procedures for ensuring cameras are activated at all times in critical areas
    10. Purchase better 911 rescue tools for cutting through thick ligatures

The third Independent Review of Deaths in Custody Report is submitted on October 21, 2015 by the three authors as signed below:

Original signed by

Yvette Thériault, Chairperson
Alan Leschied, Member
Shelley Brown, Member