Section 1: Suicide-Related Research in Canada: A descriptive overview – Biomedical research

1. Biomedical Research

Research investigating the biological and genetic risk factors for suicide has proliferated in recent years and Canadian researchers have been at the forefront in advancing this knowledge base. Approximately thirty Canadian studies were published in this area over the period 1985-2003.

Hrdina has made a significant contribution to our understanding of the possible biological bases of suicidality based on a series of post-mortem studies (Hrdina, 1996; Hrdina and Du, 2001). For example, Hrdina and colleagues (1993) reported an increase in 5-HT2 receptors in the post-mortem brains of suicide victims and depressed patients who died of natural causes. Their findings provided support for the view that an abnormality in the brain serotonergic system is associated with depression and suicidal behaviour. More recent studies (Alda and Hrdina, 2000) have examined the frequency distribution of platelet 5-HT2A receptor densities. Results from this investigation support the notion that high 5-HT2A receptor density is a marker of suicidality, and is possibly determined genetically.

In an effort to elucidate the genetic component of the serotonergic abnormalities found in suicide victims, Turecki and his colleagues at the McGill Group for Suicide Studies (1999) investigated the variance observed in brain serotonin receptor 2A (5-HTR2A) binding in patients who died by suicide. By comparing brain tissue samples of subjects who died by suicide with those who had not, these researchers were able to confirm previous findings, i.e., greater 5-HTR2A binding in subjects who died by suicide. More importantly, this study also provided preliminary support for the hypothesis that the number of 5-HTR2A receptors is genetically mediated. In a study on patients with major depression, Du et al (2000) investigated variance at the 102T/C polymorphism on this locus and found a significant association between this variant and higher level of suicidal ideation as measured by the HAM-D item on suicidal ideation.

Other serotonergic receptors have also been the target of genetic studies carried out by Canadian researchers. For instance, the study of a polymorphism located in the promoter region of the 5-HT1A autoreceptor revealed that the minor allele in this locus has a negative feedback on the repressor activity of a transcriptional factor acting on this autoreceptor in raphe cells. These findings suggested that genetic variation on this locus leads to a reduction of serotonergic neurotransmission and, consequently, to a predisposition to depression and suicide (Lemonde, et al., 2003). The possible role of genetic variation at genes coding for other serotonergic receptors has also been investigated. Turecki and colleagues (2003) investigated variation at seven serotonin receptor genes (5-HTR1B, 5-HTR1D?, 5-HTR1E, 5-HTR1F, 5-HTR2C, 5-HTR5A, and 5-HTR6) in suicide completers. They were unable to find evidence supporting a major role for these loci in the predisposition to suicide.

Canadian researchers (Du, et al., 1999; Du, Faludi, Palkovits, Bakish, and Hrding, 2000; Turecki, 2001; Fitch, et al. 2001, Anguelova, Benkelfat, and Turecki, 2003) have also examined the role of the serotonin transporter gene. Current findings suggest that variation at this gene, particularly at a 44bp insertion/deletion locus on the promoter region of this gene, a variant thought to be functional, may play an important role in the predisposition to suicide and suicidal behaviours. Using a different design and methodology, Filteau and colleagues (1993) reported data that indirectly supports the role of the serotonin transporter in suicidality. They found a significant decrease of suicidal ideation among depressed patients treated with specific serotonin reuptake inhibitors (SSRIs) compared to norepinephrine reuptake inhibitors, mixed norepinephrine serotonin uptake inhibitors and serotonin-2 antagonists.

Other genes coding for components of the serotonergic pathways have been investigated by Canadian biomedical researchers. In particular, the genes that code for tryptophan hydroxylase (TPH) was thought to play a prominent role in suicidality (Du et al., 2000, Turecki et al., 2001; Lalovic and Turecki, 2002); however, more recent data suggest that this particular gene does not code for an isoform of TPH that is expressed in brain tissue. An homologous gene, referred to as TPH2, codes for such a variant.. This is consistent with results of a meta-analysis of TPH1 studies carried out by Lalovic and Turecki (2002).

Du and colleagues (2002) studied a possible contribution of the monoamine oxidase (MAO-A) gene in depressed suicides. Interestingly, they found an association between a high activity-related allele and depressed suicide in males. This finding indicated that the MAO-A gene may be also involved in confers susceptibility to suicide in depressed males.

Turecki (2001) suggests that part of the vulnerability to suicide may be explained through the presence of genetic tendencies toward impulsive and impulsive-aggressive behaviours. The same group of researchers (Sequeira et al., 2003) reported an association of a genetic variant of the Wolfram Syndrome gene (WFS1) with suicide and higher measures of impulsivity. Their preliminary findings indicated a role for this gene in the pathophysiology of impulsive suicide. Investigating the relationship between suicide and impulsive-aggressive behaviours using a different approach, Arato and colleagues (1991) suggested that the hemispheric asymmetry of serotonergic mechanisms found in suicides could be associated to violence and aggression.

Another line of biological investigation has examined the link between serum cholesterol levels and suicide risk. By linking data from the Nutrition Canada Survey with mortality records from the Canadian National Mortality Database, Canadian researchers (Ellison and Morrison, 2001) have found that low serum cholesterol levels are associated with an increased risk of suicide. Furthermore, the association persisted even after controlling for unemployment and receipt of treatment for depression

Canadian researchers have also investigated alterations in signal transduction. In 1999, Reiach and colleagues (1999) provided preliminary evidence consisting in a reduction of adenylyl cyclase (AC) type 4 immunolabeling and its activity in post-mortem temporal cortex of depressed suicide victims. These results suggested that this alteration accounts for the disturbances in the postreceptor cAMP signaling cascade in depression. Furthermore, Young and colleagues (2003) have contributed to the growing evidence supporting the fact that signal transduction abnormalities occur in patients with a mood disorder who died by suicide. Their data indicated that the transcription factor CREB might play an important role in the neurobiology of suicide and the antisuicidal effect of lithium. As part of this effort undertaken by Canadian researchers to investigate the molecular mechanisms involved in suicide, Honer and colleagues (2002) examined molecular components of neural connectivity in severe mental disorders and suicide.

Researchers in Canada have also approached several other aspects that might be involved in suicide. For instance, high catecholamine levels have been reported in depressed patients prior to suicide and in parasuicide (Dent, Ghadirian, Kusalic, and Young, 1986; Mancini and Brown, 1992). In addition, increased levels of ER stress proteins have been found in the temporal cortex of depressed suicides (Bown, Wang, MacQueen, and Young, 2000). The research work carried out by Merali and colleagues (2004) in the suicide brain led to the possibility that the observed changes of the corticotropin-releasing hormone (CRH) and GABA A receptor subunit, or the dysregulation between these GABA A receptor subunits confer a risk to depression and/or suicide or are secondary to the psychopathology related with it.

Finally, more comprehensive biological studies using microarrays are starting to emerge. The McGill Group for Suicide Studies (Turecki, Sequeira, Gwadry, Canetti, Gingras, and French-Mullen, 2003) has recently begun conducting extensive brain expression studies investigating expression patterns in several cortical brain areas. These studies are promising and indicate a series of new candidate systems for future studies, representing new avenues for biomedical research in suicide.

In summary, there is a growing body of evidence confirming that neurobiological and genetic factors play a significant role in the etiology of suicide, and Canadian researchers have played an instrumental role in illuminating some of the molecular pathways and genetic processes that appear to contribute to an increased risk for suicide. Many more neurobiological and genetic studies are currently underway in Canada and build on previous studies by attempting to elucidate more precisely those genes implicated in suicide.

2. Clinical Research

Over 90 research articles examining the topic of suicide from a clinical or treatment perspective have been published in Canada since 1985. These include studies that have (a) examined specific personality characteristics (e.g., dependency, perfectionism) or other psychological dimensions of suicide risk in individuals, and (b) investigated risk for suicide in specific vulnerable or clinical populations (e.g., psychiatric in-patients, youth).

For ease of discussion, the category of clinical research has been divided into five sub-categories:

  1. treatment/intervention approaches
  2. family factors
  3. vulnerable clinical populations
  4. psychological dimensions, and
  5. survivors/bereavement.

Clinical and theoretical discussions which present models for clinical decision making (e.g., Kral and Sakinofsky, 1994; Truscott, Evans, and Knish, 1999) and reviews of the evidence (e.g., Lesage, 2002) are included here.

2a. Treatment/Intervention Approaches

Canadian studies investigating treatment approaches to suicide and suicidal behaviour have been few, with most studies being published in the last ten years. Investigations range from large, international, multi-site trials (Ahrens, Grof, Möller, Müller-Oerlinghausen, and Wolf, 1995) to smaller, in-depth qualitative studies (Hoover and Paulson, 1999) to single case-study designs (Malcolm and Janisse, 1994; Lum, Smith, and Ferris, 2002).

Suicide and treatment of mood disorders

While a large amount of clinical research has been devoted to determining the efficacy of specific treatments for patients suffering from mood disorders, only a few Canadian studies have specifically examined these interventions for their effect on suicidal behaviours. For example, Sharma (2001) investigated the effect of electroconvulsive therapy (ECT) on suicide risk among patients with mood disorders. He found that ECT has an acute but not long-term beneficial effect on suicidal behaviour. According to the author, the findings need to be interpreted with caution, given some of the study's specific limitations. In a large multi-centre trial, Ahrens and colleagues (1995) found that among those patients who received prophylactic lithium treatment for two years or longer, mortality due to suicide and cardiovascular disease was the same as, or only slighter higher than mortality in the general population.

Viewed from yet another perspective, the relationship between suicidality, adverse treatment-induced events (e.g., sedation, mania, decreased libido, psychosis) and pharmacological treatment for depression was explored (Tollefson, Rampey, Beasley, Enas, and Potvin, 1994). Results suggested no relationship between a treatment-emergent adverse event pattern and suicidality in this population.

Facilitation of healing

Hoover and Paulson (1999) used a phenomenological approach to identify themes of healing among those who were previously suicidal. Based on their in-depth analysis they were able to identify a series of processes that could have important implications for guiding future treatment decisions with suicidal individuals.

Strategies that facilitate healing among Aboriginal individuals have been described by McCormick (1996, 1997b) and Paproski (1997). They include: establishing a social connection and obtaining help/support from others; anchoring oneself in tradition; exercising and practicing self-care; involving oneself in challenging activities and setting goals; expressing oneself; establishing a spiritual connection and participating in ceremony; helping others; gaining an understanding of the problem; learning from a role model; and establishing a connection with nature (McCormick, 1997b).

Determination of risk

deMan and his colleagues in Québec have validated the psychometric properties of the Scale for Suicide Ideation (de Man and Leduc, 1995) and have adapted it for use with a French Canadian youth population (de Man, Leduc, and Labreche-Gauthier, 1993). Reynolds (1991) has also contributed to an understanding of the role of screening instruments and the assessment of suicide risk, particularly with children and adolescents. He has developed a series of useful screening tools, including the Suicide Ideation Questionnaire (SIQ) (Reynolds, 1988).

Meanwhile, in an effort to identify potential predictors of suicide, other Canadian researchers (Enns, Inayatulla, Cox, and Cheyne, 1997) examined the relationship among depressive symptoms, anxiety, hopelessness, and suicidal intent in a group of Aboriginal and non-Aboriginal adolescents who had been hospitalized after attempting suicide. Elsewhere, Wright and Adam (1986) found that even though many suicide attempters claimed that they wished to die at the time of hospital admission, all of the surviving patients expressed a desire to live at the time of discharge from hospital.

Holden and colleagues (1985, 1989) examined the relationship between suicidal intent and social desirability and found that "negative desirability responding" represented a distress set in the context of suicidal behaviour. Finally, Truant, O'Reilly and Donaldson (1991) have provided us with a glimpse into how psychiatrists weigh risk factors for suicide when making overall determinations about risk. They found that hopelessness was ranked as the most important risk factor, followed by suicidal ideation, previous attempts, level of mood and affect, quality of relationships, signs and symptoms of depression, and social integration.

Clinical case studies

Other suicide-related investigations with clinical implications have been undertaken in Canada in recent years. For example, Malcolm and Janisse (1994) provided a case-study investigation of a series of imitative suicides in a small, closed organization in British Columbia and found that scores of depression, anxiety and hostility, in combination with clinical interviews, were useful in identifying high-risk groups following the suicide deaths. Leenaars and Wenckstern (1998) presented a protocol analysis of Sylvia Plath's last poems, showing specific differentiating risk factors in these narratives. More recently, Lum and colleagues (2002) relied on a case-study approach to demonstrate how to apply Satir's therapeutic model to the treatment of a young male suicidal client. Leenaars (1997) presented a case, highlighting that even if one attempts to assess risk comprehensively, some patients will dissemble (or mask) their suicidal risk.

2b. Family Factors

Studies that explore the relationships among family dynamics, parenting relationships, attachment patterns, and risk for suicide constitute a small but important part of the Canadian treatment literature. One of the well-known researchers in this area is Ken Adam, who has investigated the role of early attachment patterns in the etiology of suicidal behaviour (Adam, 1985; 1986; Adam, Keller, West, Larose, and Goszer, 1994; Adam, Sheldon-Keller, and West, 1996). Many of his published works pre-date the scope of this review, but some recent studies, which build on his earlier lines of investigation, are highlighted briefly below.

From a different perspective, extensive family studies investigating familial aggregation of suicide and behavioural factors mediating suicide risk are currently being carried out at the McGill Group for Suicide Studies lead by Gustavo Turecki.

Attachment studies

In one study, adolescents who were referred to outpatient and/or residential services in three Canadian cities were assessed for lifetime suicidal ideation and attempts, and these findings were compared with their scores on the Parental Bonding Instrument (PBI) (Adam et al., 1994). Suicidal youth reported lower care and higher over-protection in relation to their mothers than their non-suicidal peers. In a later study (Adam et al., 1996), attachment patterns and history of suicidal behaviour among adolescents in psychiatric treatment were examined. Along these same lines, West and colleagues (1999) and Lessard and Morretti (1998) explored the relationship between perceived levels of attachment to caregivers (attachment-felt security) and suicidal ideation among clinical samples of adolescents.

Parent-child relationships

Approaching the exploration of family-based risk factors from a different perspective, de Man, Labrèche-Gauthier, and Leduc (1993) examined the relationship between parent-child relationships and suicide ideation in French Canadian youth. They found that suicide ideation in adolescent males and females were associated with a parenting style that was characterized by high control, in combination with a lack of sufficient maternal and paternal social support.

Tousignant and colleagues have also examined the role of the family in the context of suicide and suicidal behaviour (1986, 1993) by investigating the respective contributions of a father's and mother's care to determine its potential association with suicidal behaviour. Results showed poor care by the father to be highly associated with suicidal behaviour.

2c. Vulnerable Populations

Canadian researchers have undertaken several studies that highlight the clinical implications of working with particular high-risk populations. Because space limitations do not permit a review of each individual study, many of these works are discussed very briefly.

Individuals with mental disorders

Given the strong association between mental disorders and suicide (Tanney, 2000) it is not surprising that many studies have focused on trying to explicate this relationship with greater precision and across a number of different high-risk groups. For example, in an important case-control study of young men in Québec, which was the first psychological autopsy study to be carried out in Canada, Lesage and his colleagues (1994) found that among young men, suicide is linked to the following mental disorders: major depression, borderline personality disorder, and substance abuse. Further studies in an extended sample indicated the important role of comorbidity in suicide risk (Kim et al., 2003).

In a series of studies that examined risks for suicide among patients diagnosed with borderline personality disorder, Canadian researchers found that the most significant predictors among this clinical population were previous attempts and higher education (Paris, 1987; 1990; Paris, Nowlis, and Brown, 1989).

Several Canadian studies have investigated the specific risks for suicide among psychiatric in-patients. In a Montréal-based study, findings indicated that in-patients suffering from an affective disorder or schizophrenia comprised the majority of the suicide death sample (Proulx, Lesage and Grunberg, 1997), while an Ontario study found that in-patients at greatest risk for suicide fit the following profile: previous suicidal behaviour, suffering from schizophrenia, admitted involuntarily, and living alone (Roy and Draper, 1995). A more recent study from Ontario found that in-patient suicide attempters were more likely to have had a family history of psychiatric problems, a history of previous suicide attempts, with the most common diagnosis being a mood disorder and not schizophrenia (Sharma, Persad, and Kueneman, 1998). Holley and her colleagues (1998) conducted a 13-year mortality review of a regional cohort of 876 suicide attempters who were admitted as inpatients between 1979 and 1981. Compared to the general population, study subjects were four times more likely to die of any cause, but 25 times more likely to die by suicide. Finally, in another Canadian study, Chandrasena and colleagues (1991) found that many of the foreign-born patients who had died by suicide were unemployed with poor social integration.

Parasuicidal patients

Another group of studies with important clinical implications are those that investigate patients who repeatedly engage in self-harming and suicidal behaviour. Reynolds and Eaton (1986) compared multiple suicide attempters with single attempters and found that repeaters had higher levels of depression, hopelessness, and substance abuse as well as higher lethality ratings. Sakinofsky and Roberts (1990) examined why parasuicidal patients continued to engage in self-harming behaviour despite the apparent resolution of their problems. In an earlier study, Goldberg and Sakinofsky (1988) explored levels of "intropunitiveness" among parasuicidal patients across various interview modes. They found that highly intropunitive individuals who were in the cognitive interview group showed the most improvements on self-report measures of depressive symptoms. Leenaars, Lester, Wenckstern and their colleagues (1992), examining attempters and those who died by suicide, concluded that there may well be more similarities than differences in these groups. They may well be two overlapping groups.

Youth

Concern over increased rates of youth suicide in the past 30-40 years has prompted several Canadian researchers to try to better understand the specific risk factors for youth suicide and to highlight some possible implications for treatment and supportive interventions. For example, qualitative studies have been conducted which seek to illuminate some of the processes that led young people to contemplate suicide and the factors that helped them recover (Everall, 2000; McCormick, 1997a; Paulson and Everall, 2001). In Québec, de Man and colleagues surveyed a group of high school students (de Man and Leduc, 1995; de Man, 1999) in order to identify the correlates of suicide ideation. They found that depression was one of the most important factors, a finding which has clear clinical implications. Other groups of high-risk youth have also been studied, including youth in foster care (Charles and Matheson, 1991). Leenaars, De Wilde, Wenckstern and Kral (2001) showed that the suicide of teens may be highly related to cognitive constriction and lack of understanding; suicidal teens may well be blind to many aspects of their own deaths.

Elderly

Continued high rates of elderly suicide have also fuelled a number of recent investigations by Canadian researchers. For example, writing from a nursing research paradigm, Delisle (1992) examined the questions: "What makes new pensioners vulnerable to suicide? Are their suicides preventable? What are the best nursing interventions?" Fortin and colleagues (2001) investigated the relationship between suicide ideation and self-determination among the institutionalized elderly and found that the suicidal elderly did not differ from nonsuicidal individuals on the dimension of self-determination.

Other, very specific, at-risk populations have also been investigated, including Chinese Canadian women patients (Lalinec-Michaud, 1988), patients who are using isotretinoin for the treatment of acne (Jick, Kremers, and Vasilakis-Scaramozza, 2000), and patients who present at emergency departments with chest pain, panic disorder and suicidal ideation (Fleet, Dupuis, Marchand, Burelle, Arsenault, and Beitman, 1996; Fleet, Dupuis, Kaczorowski, Marchand, and Beitman, 1997). Leenaars (1992) found that our understanding of suicide in the elderly may well be the poorest across the life span, requiring much greater attention.

2d. Psychological Dimensions

Antoon Leenaars has contributed substantially to the national and international knowledge base in suicide prevention through his empirical investigation of suicide notes as well as through other studies conducted in Canada and internationally.

Studies of suicide notes

Through the development of specific protocols for analyzing suicide notes, which have been applied across a number of different suicidal subjects (i.e., men, women, young, old, lethal methods, passive methods, attempters, those who died by suicide), Leenaars proposed a multi-dimensional model of suicide, examining both intrapsychic and interpersonal aspects that may assist us to better understand the suicidal individual (see Leenaars, 1996). He then went on to examine some very specific psychological characteristics related to suicide, including unbearable pain, cognitive constriction, indirect expressions, inability to adjust, ego, interpersonal relations, rejection-aggression, and identification-egression.

A recent study (Leenaars, De Wilde, Wenckstern, and Kral, 2001) provides a concise summary of some of the key findings emerging from the analysis of suicide notes. Suicide notes representing four developmental ages (adolescents, young adults, middle adults, old adults) were analyzed for specific protocols along the eight psychological multidimensional dimensions. Despite many commonalities across age groups, the results suggest that suicides of adolescents may be more highly related to cognitive constriction, indirect expressions, rejection-aggression, and identification-egression, than those of other age groups. Young adults too have their unique markers; they show the highest incidence of psychopathology (or inability to adjust).

It is important for progress in health that research findings in one nation be replicated in other nations so that their cross-cultural reliability can be ascertained. Leenaars and a number of international colleagues have applied Leenaars' multidimensional model to suicide notes not only from Canada, but the United States, Northern Ireland, Hungary, Russia and Australia (see Leenaars, Lester, Lopatin, Schustov & Wenckstern, 2002).

Personality characteristics

In addition to Leenaars, other Canadian researchers have examined specific personality traits or psychological characteristics related to suicide. For example, Bettridge and colleagues (1995) examined dependency needs and the perceived availability and adequacy of relationships among female adolescent attempters and non-attempters in Canada. Hewitt and others have studied the dimension of perfectionism and its relationship to suicidal behaviours across a number of different clinical populations including adolescent psychiatric patients and alcoholics (Hewitt, Flett, and Weber, 1994; Hewitt, Newton, Flett, and Callander, 1997; Hewitt, Norton, Flett, Callander, and Cowan, 1998). Finally, the McGill Group for Suicide Studies has been carrying out several studies investigating personality traits in those who died by suicide (see for instance Kim et al., 2003).

2e. Survivors/Bereavement

Very few published Canadian studies examine the issue of suicide bereavement. A brief review of this literature follows.

Grief after suicide

In comparison to parents who had lost a child to a motor-vehicle death, Séguin and colleagues (1995a) found that parents grieving a suicide death were more depressed (although this difference disappeared after nine months), experienced more shame and life events, and had a greater history of loss. In another level of investigation (Séguin, et al., 1995b), mother-survivors of a suicide death were interviewed about their experience of loss, and these findings were augmented with psychological autopsy findings. Subsequent analysis revealed the presence of significant transgenerational loss, separation, and inadequate child-rearing.

In a more recent study, Bailley, Kral, and Dunham (1999) compared grief experiences among university students across a range of different types of deaths. They found that suicide survivors, compared to the other groups, experienced more frequent feelings of rejection, responsibility, "unique" reactions, total grief reactions, increased shame and perceived stigmatization.

Bereavement groups

Hopmeyer and Werk (1994) looked at the structure and membership of various types of bereavement support groups offered in Montréal: those that served widows, those that served family survivors of suicide, and those that served family survivors of cancer deaths. All attendees reported strong satisfaction with their experiences, but the reasons for joining and the most valuable aspects of the group experience differed across settings. Finally, Rubey and McIntosh (1996) conducted a survey of suicide survivor groups in the United States and Canada in order to better understand their composition and character.

3. Health Services and Systems

Fewer Canadian studies have been conducted in this category as compared with the previous one on Clinical Research. Among the works that have been published - approximately 45 in total - a range of interests is apparent, including: reviews of crisis lines and suicide prevention centres; program evaluation studies; retrospective analyses of services following deaths by suicides; and professional development and training activities in suicide prevention and intervention. As mentioned earlier, several studies are included here which have relevance for suicide prevention, as well as broader applications for the health and well-being of the population as a whole, i.e., mental health promotion and prevention programs. Theoretical discussions and commentaries (e.g., Boyer and Loyer, 1996), presentations of models for organizing services (e.g., Boldt, 1985) and reviews of the relevant literature (e.g., Frankish, 1994; Rhodes and Links, 1998) are not included in this paper.

For ease of discussion, studies will be organized under the following sub-headings: suicide prevention centres, organization and evaluation of services, training and professional development, and mental health promotion/prevention programs.

3a. Suicide Prevention Centres

Canadian contributions to the international knowledge base regarding the effectiveness of crisis-based telephone hotlines and other suicide prevention centres have been significant. In one of the first of a series of studies that was done to explore the nature and effectiveness of suicide prevention centres, Mishara and Daigle (1992) investigated volunteer-based telephone crisis services at two different centres in Québec. By listening unobtrusively to a series of incoming calls and coding all responses, researchers were able to assess program effectiveness along the following three dimensions: changes in depression ratings, changes in urgency, and use of no-harm contracts or safety agreements. Later studies (Daigle and Mishara, 1995; Mishara and Daigle, 1997) examined the specific types of interventions provided by crisis-line volunteers and found that overall, a greater proportion of "Rogerian" or nondirective responses were related to decreases in depression ratings. Mishara and Giroux (1993) studied the role of stress in crisis-line volunteers at three different points in time: before the shift, during a call, and at the conclusion of a shift.

Other Canadian researchers have also examined the nature and impact of suicide prevention and crisis intervention centres in cities throughout Canada (Adamek and Kaplan, 1996; Leenaars and Lester, 1995). Adamek and Kaplan (1996) surveyed a number of different crisis lines in the United States and Canada to determine their readiness and capacity to respond to suicidal older adults. On the whole, they found that there was a lack of specific training in this area, poor awareness of recent suicide trends, and limited outreach to older adults.

On balance, the empirical evidence regarding the effectiveness of telephone-based hotlines at reducing suicide rates is scanty at best. Leenaars and Lester (1995) showed that suicide prevention centres had a positive impact on reducing suicide rates in Canada, but the impact did not reach statistical significance. Canadian researchers have been leaders in summarizing the findings to date as well as in articulating the methodological and ethical challenges inherent in conducting these types of investigation (Mishara and Daigle, 2000). Further evaluation studies of crisis centre services are currently being conducted by Canadian researchers and are described later in this paper.

3b. Organization and Evaluation of Services

Suicide prevention programs other than crisis hotlines have also been evaluated by Canadian researchers. For example, de Man and Labreche-Gauthier (1991) evaluated two different community-based suicide prevention programs in Québec by examining levels of self-esteem, stress, and suicide ideation among individuals who attended the programs. Meanwhile, in a study that has important implications for the way services to potentially suicidal youth are organized, Cappelli and others (1995) found that among youth attending an adolescent health clinic, depression and suicidal thoughts represented a significant portion of mental health problems. More recently, Breton and his colleagues (2002) conducted an in-depth review of suicide prevention programs being offered throughout Canada to describe their overall character and identify their underlying theoretical bases. These authors concluded that most of the suicide prevention programs being offered in Canada are not being formally evaluated and most program descriptions remain short on details, particularly in terms of their theoretical foundations.

Reviews identifying "best practices" in youth suicide prevention (Gardiner, 2002; White and Jodoin, 1998) and other evidence-based reviews of the efficacy of youth suicide prevention programs (Guo and Harstall, 2002; Ploeg, et al., 1996) are not included here.

Interventions with suicidal individuals

Allard and others (1992) investigated whether specific follow-up interventions provided after a suicide attempt could decrease the risk of repeat attempts at two-year follow-up. Following their analysis, these authors concluded that intensive follow-up interventions did not reduce risks of repeated suicide attempts. Greenfield and his colleagues (1995) conducted a study to determine the impact of an outpatient psychiatric team on the hospitalization rates of youth in crisis, most of whom were suicidal adolescents. In a more recent investigation, Links (2002) determined that ongoing contact for two years following discharge was associated with greater survival among those patients who were at risk of post-crisis suicide.

Retrospective analyses and audits

Grunberg and his colleagues (1994) examined health service utilization issues based on a retrospective analysis of suicide deaths among young men in Québec. These authors found that almost half of the subjects who had died by suicide had consulted a mental health professional in the year before the suicide, compared to 5% in the control group. More recently, in a retrospective analysis of all suicides at one large psychiatric facility in Ontario, Martin (2002) described the most common patient characteristics and also identified potential deficiencies in care among those who died by suicide. According to the author, this study represents the first Canadian report of a cumulative series of suicides, including documented deficiencies in care, at a single psychiatric facility.

Cost analysis studies

At least three studies have assessed costs related to suicide and its prevention. These studies have relevance for the way services are conceptualized, justified, and delivered. In one of the first reports to calculate the costs of suicide mortality in a Canadian province, Clayton and Barcel (1999) discovered that in New Brunswick, the mean total cost estimate per suicide death in 1996 was (including indirect costs) $849,877.80. Following a different line of investigation, Barnett and colleagues from the University of Alberta (1999) examined the specific costs associated with using flumazenil in cases of drug overdose. Based on a randomized, placebo-controlled study to assess cost-effectiveness, they found that the use of flumazenil in intentional drug overdose of unknown etiology is not cost effective. Finally, in calculating overall costs related to gunshot wounds in Canada in 1991, Miller (1995) found that the estimated cost was $6.6 billion. Suicides and attempted suicides accounted for the bulk of these costs at $4.7 billion, including indirect costs.

3c. Training and Professional Development

Less than a handful of Canadian studies have been undertaken which specifically examine the effectiveness of training in suicide prevention and interventionFootnote 2 despite the fact that training and education activities constitute a major part of most local suicide prevention activities and efforts.

Davis (1991) describes the implementation of a curriculum that was used to train volunteers who were providing follow-up counselling to suicidal clients. Volunteers who evaluated the curriculum found it to be effective, although they were concerned about limited practice time. Both community professionals and trainees felt the project met its objectives.

Tierney (1994) used simulated role plays to evaluate the degree to which learners who had participated in a two-day suicide intervention gatekeeper training program had mastered various skills, knowledge and attitudes. Based on an assessment of immediate training effects, Tierney concluded that significant increases in suicide intervention skills had taken place.

Ross and colleagues (1998) were interested in determining the extent to which Canadian Schools of Nursing included violence-related content (including suicide prevention) in their curricula. Based on a survey with an 88% response rate, they learned that content regarding violence against children and women, and suicide as a response to abuse, formed part of the curriculum for all Schools of Nursing.

3d. Prevention/Mental Health Promotion

Primary prevention and mental health promotion programs typically target healthy populations who have not yet shown evidence of disease or disorder. Examples include the promotion of social competencies among youth, community renewal strategies, and family support programs. Several prevention and mental health promotion programs have been studied and evaluated by Canadian researchers in recent years, with most studies being published in the late 1990s. Not included in this document are studies which synthesize findings from the resiliency literature (e.g., Steinhauer, 2001), discussions of optimal prevention and clinical service delivery components in children's mental health (e.g., Offord et al., 1989), and guidelines for setting up effective Aboriginal mental health promotion programs (e.g., Kirmayer and Boothroyd, 1999).

Children, youth, and their environments

Bélanger and colleagues (1999) evaluated the impact of a program designed to promote social competence among kindergarten children in Montréal. Analysis revealed that there were significant gains in self-esteem and conflict resolution skills among children in the experimental group when compared with a control group.

Recognizing the impact of the social environment on child and youth well-being, another Canadian study examined system-level changes made by schools as a result of their participation in an innovative, comprehensive mental health promotion program (Bond, Glover, Godfrey, Butler, and Patton, 2001). The Gatehouse project in Victoria, BC was designed as a "whole school" intervention which included the following elements: establishment and support of a school-based adolescent health team, identification of risk and protective factors in each school's social and leaning environment from student surveys, and identification and implementation of effective strategies to address these issues.

In an earlier study, Peirson and Prilletensky (1994) also examined how school-level changes could contribute to a prevention climate at the secondary level. A grounded theory of successful school change was generated which included community ownership, attention to human factors, and proper implementation. Lastly, Collins and Angen (1997) highlighted the importance of youth participation in the development of health promotion and suicide prevention programs.

Conceptual models and approaches

In a study from Québec, researchers reviewed current practice in the area of child, youth and family interventions with a view to determining the conceptual underpinnings of various prevention models. Chamberland and colleagues (2000) analyzed a number of different prevention programs being implemented in Québec which were designed to serve individuals 18 and under and their families. Results indicated that intervention strategies were not aimed solely at modifying characteristics of children, youth and their families. Some projects also tried to change living environments, revealing the influence of ecological and social models in program development. Pancer and Cameron (1994) used the primary prevention initiative, "Better Beginnings, Better Futures" to assess the impact of citizen involvement in community-based prevention programs. Through a qualitative research methodology, these authors were able to assess the positive and negative outcomes that residents from seven Ontario communities derived from their involvement in this project.

Although many mental health promotion and prevention programs target children, youth and their families, two recently published Canadian studies have respectively targeted older adults and farming families/communities. Bouffard and others (1996) studied the development, implementation, and evaluation of a mental health promotion program for older adults that focused on assisting elderly women to establish and work towards personal goals. Using a case-study approach, Gerrard (2000) describes how a community psychology model was used to design and implement a farm stress program in Saskatchewan. Among other considerations, Gerrard emphasized the importance of conceptualizing mental health issues, like farm stress, from the perspective of "individuals-in-communities."

4. Health of Populations and Sociocultural Determinants

A large number (approximately 85) of suicide-related research studies in Canada fall within this category. To make the review task manageable, studies have been further sub-divided under the following headings: social/cultural variables, specific at-risk populations, and determinants of health, i.e., healthy child development, income, employment, education, working conditions, social support, and environment.

4a. Social/Cultural Variables

This group of studies includes those that have examined rates of suicide at the population level based on a range of different social and cultural variables including age, sex, marital status, ethnicity, attitudes, and regional issues. Many of the studies included here reflect a distinctly Durkheimian orientation and arise out of the sociological tradition, with Frank Trovato from the University of Alberta furnishing the bulk of these studies. Space does not permit a review of each individual study conducted by Trovato, so they will be discussed briefly.

Social factors

Trovato's studies include an examination of the relationship between Canadian suicide rates and broad social factors such as ethnicity and immigration status (Trovato, 1986a; 1986b; 1992), sex and marital status (Trovato, 1986c; 1987; 1991), labour force participation (Trovato and Vos, 1992), regional and ecological variations (Trovato, 1992) and interprovincial migration (1986d). He has also examined suicide rates at the provincial level, specifically Québec (Krull and Trovato, 1994; Trovato, 1998). In a Durkheimian analysis of youth suicide in Canada, Trovato (1992) investigated the effects of three measures of social integration on youth suicide rates: family integration, religious integration, and unemployment. He found that religious detachment among the young was associated with an increased propensity towards suicide.

Other Canadian researchers have also examined rates of suicide by relying on Durkheim's concepts of social integration and anomie to better understand the social and economic correlates of suicide (Leenaars, Yang, and Lester, 1993; Leenaars and Lester, 1995; Leenaars and Lester, 1998; 1999). Theoretical frameworks describing alienation and specific risks for suicide are not included here.

Additional studies investigating broad-level social and demographic influences which are not necessarily located in the sociological tradition include: an examination of the "relative age effect" and its influence on youth suicide (Thompson, Barnsley, and Dyck, 1999); socioeconomic risk factors for elderly suicide (Agbayewa, Marion, and Wiggins, 1998); a demographic review of child and adolescent suicide (Thompson, 1987); an investigation of unemployment and labour force participation as risk factors for suicide (Cormier and Klerman, 1985); regional and ecological variations in suicide rates in provinces across Canada (Agbayewa, 1993; Sakinofsky and Roberts, 1987); investigations of immigrant suicide rates (Kliewer and Ward, 1988; Singh, 2002); and the psychosocial correlates of suicide among specific populations (Bagley and Ramsay, 1985; de Man, Labreche-Gauthier and Leduc, 1993; Hurteau and Bergeron, 1992).

Cultural factors

Canadian studies investigating the broad role of cultural influences (including gender, attitudes, and cross-cultural factors) on suicide are diverse. In a recent study, Pinhas and colleagues (2002) examined the role of gender in understanding suicidal behaviour among adolescents. These authors found that gender-role conflict may be a potential contributing factor in the etiology of suicidal behaviour among female adolescents. Though illuminating, theoretical discussions about gender and suicide (e.g., Canetto and Sakinofsky, 2000) are not included here.

Bagley and Ramsay (1989) explored attitudes towards suicide, religious values, and suicidal behaviour through a community survey. Domino and Leenaars (1989) utilized the Suicide Opinion Questionnaire to compare Canadian and American college students' attitudes towards suicide. Canadian college students were significantly more likely than their American counterparts to view suicide as "a part of everyday life." Leenaars and Domino (1993) utilized the same questionnaire on a general population in Canada and the United States and noted differences in attitudes here too, but not as extreme as in the youth.

Studies that compare Canada with other nations as a way to further understand potential culture-bound risks for suicide have also been undertaken in recent years. In a series of cross-cultural investigations that compare Canada with the United States, Leenaars and others (Leenaars, 1992; Leenaars and Lester, 1994; Leenaars, 1995; Leenaars and Lester, 1995; Sakinofsky and Leenaars, 1997) have identified some clear differences in these two neighbouring countries, some of which may help to explain differences in suicide patterns. Other cross-cultural studies have compared college student suicide ideation in Canada and Japan (Heisel and Fuse, 1999). Cantor, Leenaars, Lester and their colleagues (1996) compared rates and patterns of suicide in Canada to that in seven other nations.

Lester and Leenaars (1998) examined the ecological relationship between suicides, homicides and accidental deaths from firearms in the Canadian provinces. They found a positive correlation for all these causes of death, interpreting this result as evidence for a regional subculture of firearm violence in Canada.

4b. Specific At-Risk Populations

In addition to the research examining specific clinical populations (summarized in the previous section), a number of other groups have been identified in the empirical literature as being at elevated risk for suicide including youth, elderly, Aboriginal Peoples, gay/lesbian populations, and those who are incarcerated. Each of these populations has been studied from a number of different perspectives in the hope of being able to identify specific risk or protective factors for suicide. Canadian studies investigating specific high-risk populations are summarized next. Reviews of the relevant literature (e.g., Kirmayer 1994; Clarke, Frankish, and Green, 1997) are not included here.

Youth

deMan and others (1992, 1993) examined risks for suicide among French-Canadian adolescents based on a range of variables including stress and social support. Bagley has made a sizable contribution to the suicide research literature in Canada. Of particular relevance here, Bagley has examined a range of factors associated with youth suicide (1989, 1992), including the role of sexual assault as a risk factor for suicidal behaviour (1995, 1997). McBride and Siegel (1997) found preliminary evidence to suggest an association between adolescent suicide and learning disabilities. Barber (2001) was guided by an "absolute misery hypothesis" as a way to account for suicidality among young people, and found that suicide risk among young males is increased when those around them are perceived to be more advantaged. Researchers in Ontario (Kidd, 2001; Kidd and Kral, 2002) investigated the meaning that suicide held for street youth and young people engaged in prostitution. Key themes identified in participant narratives included feelings of worthlessness, loneliness, and hopelessness.

Elderly

At the other end of the age continuum, elderly suicide has also been studied fairly extensively by Canadian researchers. Duckworth and McBride (1996) examined coroners' records of elderly suicides in Ontario and determined, among other things, that elderly people who had died by suicide had rarely received psychiatric treatment prior to their deaths. Researchers in British Columbia examined the impact of weather and season on elderly suicide rates (Marion, Agbayewa, and Wiggins, 1999), while Alberta-based researchers (Quan and Arboleda-Florez, 1999) used coroners' records to describe the characteristics of suicide deaths among those over the age of 55, paying particular attention to gender differences. In a more recent study (Quan, Arboleda-Florez, Fick, Stuart, and Love, 2002), researchers examined the relationship between suicide and physical illness among the elderly. Examining the relationship between suicide ideation and a range of personal and social variables, researchers in Québec found that suicidal ideation among the elderly was associated with infrequent alcohol consumption, gender, depression, social isolation, and dissatisfaction with health and social support (Mireault and de Man, 1996).

Aboriginal

High rates of suicide among Canada's indigenous peoples have sparked a series of investigations in recent years. Many of these studies have attempted to understand Aboriginal suicide within a specific regional and cultural context. Starting in the West, several researchers have investigated Aboriginal suicides in British Columbia (Lester, 1996; Chandler and Lalonde, 1998; Cooper, Corrado, Karlberg, and Adams, 1992). Cooper and others (1992) examined the circumstances surrounding Aboriginal suicides in BC and found that rates of suicide among Aboriginal people living off-reserve more closely approximated the rates in the general population. In a later study of Aboriginal youth in BC, Chandler and Lalonde (1998) found that while the overall rate of youth suicide was higher among Aboriginal communities when compared with the general population, there were a number of First Nations that had low or non-existent rates of suicide. A strong association was found to exist between those communities with low rates of youth suicide and certain markers of "cultural continuity" including: making advances towards self-government and settling land claims, having control over community social services (i.e., police, education, and child welfare), and engaging in traditional cultural healing practices.

In Alberta, correlates of adolescent suicidality among Aboriginal youth were identified and compared with non-aboriginal youth (Gartrell, Jarvis and Derckson, 1985). After controlling for age, risk factors for Aboriginal suicide attempts included heavy alcohol use, absent fathers, sleeping difficulties, and poor psychological well-being. Bagley (1991) found that poverty and suicide were shown to have strong positive correlation in a study of young Aboriginal males living on reserves in Alberta.

In Manitoba, characteristics of suicides among Aboriginal individuals living on-reserve were compared with suicides among Aboriginal people living off-reserve (Malchy, Enns, Young, and Cox, 1997). Findings indicated that there were no significant differences in mean age, sex, blood alcohol level and previous psychiatric care among Aboriginal people who died by suicide living on and off reserve. A five-year review of all youth suicides was also conducted in Manitoba (Sigurdson, Staley, Matas, Hildahl, and Squair, 1994), as was a study of suicide and parasuicide in a remote northern Aboriginal community (Ross and Davis, 1986).

Suicides and psychological distress among Aboriginal communities in Ontario and Québec have also been studied. For example, Spaulding (1985) examined suicide rates among ten Ojibwa bands in northwestern Ontario. In Québec, Barss (1988) summarized the circumstances surrounding suicide deaths and suicide attempts among the Cree of James Bay, while Kirmayer and others (2000) have identified specific risk and protective factors associated with psychological distress among this population.

Suicide and suicidal behaviour among the Inuit have also been studied in recent years (Kirmayer, Malus, and Boothroyd, 1996; Kirmayer, Boothroyd, and Hodgins, 1998; Kral et al., 2000; Leenaars, 1995; Leenaars, Anawak, Brown, Hill-Keddie & Taparti, 1999). Factors associated with attempted suicide among Inuit youth include: substance use (solvents, cannabis, cocaine), recent alcohol abuse, psychiatric problems, and a greater number of stressful life events in the last year. Regular church attendance was negatively associated with attempted suicide (Kirmayer et al., 1998). Inuit elders' first-hand accounts of suicide, distress and healing in the eastern Arctic have also been captured through a series of narratives, adding a richness and dimension to our current knowledge base (Kral et al., 2000; Leenaars, 1995). Leenaars et al. (1999) have extended such study by capturing not only the narratives of some Inuit, but also from some Aborigine in Australia. These international comparisons show rich similarities to the malaise in both peoples, such as social and cultural turmoil created by the policies of colonialism in Canada and Australia.

Looking more broadly at the issue of Aboriginal suicide in Canada, Lester (1995a) found that there was no association between the proportion of Aboriginal people living in a particular region and the overall suicide rate.

Gays/lesbians

Canadian research exploring sexual orientation and suicide risk has primarily focused on gay men. For example, Bagley and Tremblay (1996) found that attempted suicide rates among young gay men in their sample ranged from 20% to 50%. Based on a survey conducted in Calgary, these authors found significantly higher rates of previous suicidal ideation and actions among homosexual males than among heterosexual males (Bagley and Tremblay, 1997). In a more recent Canadian study of gay and bisexual men, just under half of the subjects reported that they had considered suicide, while approximately 20% reported that they had attempted suicide at least once (Botnick et al., 2002).

Incarcerated populations

Several recent Canadian studies have examined suicide risk among prisoners. Bland, Newman, Dyck, and Orn (1990) looked at the prevalence of psychiatric disorders and suicide attempts among prisoners in Edmonton, while Lester (1995b) studied overall rates of suicide among Canadian prisoners. Green and colleagues (1993) summarized the most common characteristics of suicides among federally incarcerated prisoners and noted that suicides were not associated with age, offence, previous convictions or length of sentence. In a later study, risk factors for inmate suicide were identified following a review of suicide deaths in federal institutions: lengthy involvement in the criminal justice system, a greater likelihood of being incarcerated for robbery or murder, and involvement in institutional incidents of a serious nature (Laishes, 1997). More recently, Fruehwald and colleagues (2001) investigated the relevance of previous suicidal behaviour in understanding suicide among Canadian prisoners.

4c. Determinants of Health

Studies in this category are discussed below under the following headings: child development, social support, occupations and working conditions. Reviews of the literature concerning the determinants of health and suicide (i.e., Dyck, Mishara, and White, 1998) and evidence-based reviews of the association between unemployment and suicide (Jin, Shah, and Svoboda, 1996) are not included here.

Child development

In studies examining children's understanding of death and suicide, researchers in Québec found that among children in 1st, 3rd, and 5th grades who knew what suicide meant, the concept was related to age, the concept of death and personal experiences with death (Normand and Mishara, 1992). In a later study, Mishara (1999) found that by the 3rd grade, children have a fairly sophisticated understanding of suicide, learned by many from television and discussions with other children.

Social support

Hanigan and colleagues (1986) studied the relationship between young adults' suicidal behaviours and the level and quality of support received from their social environment following a critical life event. Later, Tousignant and Hanigan (1993) examined the role of social support among suicidal college students who had recently suffered the loss of a love relationship or the loss of a close friend. They found that suicidal students named fewer important persons in the kinship network and had more conflicts with this network than did nonsuicidal peers.

Meanwhile, in a more recent study, Stravynski and Boyer (2001) examined the construct of loneliness and its relationship to suicide. They found strong associations among suicide ideation, parasuicide and subjective and objective experiences of being lonely and alone.

Occupations and working conditions

A series of studies have been conducted in Canada in recent years to determine if certain occupational groups are associated with elevated risks for suicide. For example, Loo (1986) examined risks for suicide in the RCMP and found that the average annual suicide rate was approximately half that of the comparable general population. Among those officers who did kill themselves, the most common method was service revolver.

Sakinofsky (1987) looked at suicide rates among physicians and compared their age-standardized mortality ratios with those of other occupational groups.

In a study that investigated the characteristics of suicides among Canadian farm operators, including the risks posed by exposure to specific hazards like pesticides, Pickett and colleagues (1993, 1998, 1999) found that provincial suicide rates among farm operators were generally lower than or equivalent to those of Canadian males generally. Furthermore, analyses revealed that there was no association between suicide and exposure to herbicides or insecticides.

In a study of suicide risk among electrical utility workers, Baris, Armstrong, Deadman, and Thériault (1996) did not find strong evidence for causal association between exposure to electric fields and suicide among this particular occupational group.

Finally, in a recent study that investigated risks for suicide among U.N. peacekeepers, Wong and colleagues (2001) found no overall increased risk of suicide in Canadian peacekeepers. The male rate of suicide in this military group was half that of a Canadian civilians comparison group. An apparent increased risk among a subgroup of air force personnel was not related to military hazards but to personal problems such as relationship and financial issues, and also possibly to loneliness and isolation. Airforce personnel did not go through the careful pre-screening and training procedures that their army colleagues were subjected to, which also resulted in mutual bonding and support. These authors suggest that while peacekeeping per se does not increase overall suicide risk, military lifestyles may tax interpersonal relationships, promote alcohol abuse, and contribute to psychiatric illness in a minority of vulnerable individuals.

4b. Specific At-Risk Populations

In addition to the research examining specific clinical populations (summarized in the previous section), a number of other groups have been identified in the empirical literature as being at elevated risk for suicide including youth, elderly, Aboriginal Peoples, gay/lesbian populations, and those who are incarcerated. Each of these populations has been studied from a number of different perspectives in the hope of being able to identify specific risk or protective factors for suicide. Canadian studies investigating specific high-risk populations are summarized next. Reviews of the relevant literature (e.g., Kirmayer 1994; Clarke, Frankish, and Green, 1997) are not included here.

Youth

deMan and others (1992, 1993) examined risks for suicide among French-Canadian adolescents based on a range of variables including stress and social support. Bagley has made a sizable contribution to the suicide research literature in Canada. Of particular relevance here, Bagley has examined a range of factors associated with youth suicide (1989, 1992), including the role of sexual assault as a risk factor for suicidal behaviour (1995, 1997). McBride and Siegel (1997) found preliminary evidence to suggest an association between adolescent suicide and learning disabilities. Barber (2001) was guided by an "absolute misery hypothesis" as a way to account for suicidality among young people, and found that suicide risk among young males is increased when those around them are perceived to be more advantaged. Researchers in Ontario (Kidd, 2001; Kidd and Kral, 2002) investigated the meaning that suicide held for street youth and young people engaged in prostitution. Key themes identified in participant narratives included feelings of worthlessness, loneliness, and hopelessness.

Elderly

At the other end of the age continuum, elderly suicide has also been studied fairly extensively by Canadian researchers. Duckworth and McBride (1996) examined coroners' records of elderly suicides in Ontario and determined, among other things, that elderly people who had died by suicide had rarely received psychiatric treatment prior to their deaths. Researchers in British Columbia examined the impact of weather and season on elderly suicide rates (Marion, Agbayewa, and Wiggins, 1999), while Alberta-based researchers (Quan and Arboleda-Florez, 1999) used coroners' records to describe the characteristics of suicide deaths among those over the age of 55, paying particular attention to gender differences. In a more recent study (Quan, Arboleda-Florez, Fick, Stuart, and Love, 2002), researchers examined the relationship between suicide and physical illness among the elderly. Examining the relationship between suicide ideation and a range of personal and social variables, researchers in Québec found that suicidal ideation among the elderly was associated with infrequent alcohol consumption, gender, depression, social isolation, and dissatisfaction with health and social support (Mireault and de Man, 1996).

Aboriginal

High rates of suicide among Canada's indigenous peoples have sparked a series of investigations in recent years. Many of these studies have attempted to understand Aboriginal suicide within a specific regional and cultural context. Starting in the West, several researchers have investigated Aboriginal suicides in British Columbia (Lester, 1996; Chandler and Lalonde, 1998; Cooper, Corrado, Karlberg, and Adams, 1992). Cooper and others (1992) examined the circumstances surrounding Aboriginal suicides in BC and found that rates of suicide among Aboriginal people living off-reserve more closely approximated the rates in the general population. In a later study of Aboriginal youth in BC, Chandler and Lalonde (1998) found that while the overall rate of youth suicide was higher among Aboriginal communities when compared with the general population, there were a number of First Nations that had low or non-existent rates of suicide. A strong association was found to exist between those communities with low rates of youth suicide and certain markers of "cultural continuity" including: making advances towards self-government and settling land claims, having control over community social services (i.e., police, education, and child welfare), and engaging in traditional cultural healing practices.

In Alberta, correlates of adolescent suicidality among Aboriginal youth were identified and compared with non-aboriginal youth (Gartrell, Jarvis and Derckson, 1985). After controlling for age, risk factors for Aboriginal suicide attempts included heavy alcohol use, absent fathers, sleeping difficulties, and poor psychological well-being. Bagley (1991) found that poverty and suicide were shown to have strong positive correlation in a study of young Aboriginal males living on reserves in Alberta.

In Manitoba, characteristics of suicides among Aboriginal individuals living on-reserve were compared with suicides among Aboriginal people living off-reserve (Malchy, Enns, Young, and Cox, 1997). Findings indicated that there were no significant differences in mean age, sex, blood alcohol level and previous psychiatric care among Aboriginal people who died by suicide living on and off reserve. A five-year review of all youth suicides was also conducted in Manitoba (Sigurdson, Staley, Matas, Hildahl, and Squair, 1994), as was a study of suicide and parasuicide in a remote northern Aboriginal community (Ross and Davis, 1986).

Suicides and psychological distress among Aboriginal communities in Ontario and Québec have also been studied. For example, Spaulding (1985) examined suicide rates among ten Ojibwa bands in northwestern Ontario. In Québec, Barss (1988) summarized the circumstances surrounding suicide deaths and suicide attempts among the Cree of James Bay, while Kirmayer and others (2000) have identified specific risk and protective factors associated with psychological distress among this population.

Suicide and suicidal behaviour among the Inuit have also been studied in recent years (Kirmayer, Malus, and Boothroyd, 1996; Kirmayer, Boothroyd, and Hodgins, 1998; Kral et al., 2000; Leenaars, 1995; Leenaars, Anawak, Brown, Hill-Keddie & Taparti, 1999). Factors associated with attempted suicide among Inuit youth include: substance use (solvents, cannabis, cocaine), recent alcohol abuse, psychiatric problems, and a greater number of stressful life events in the last year. Regular church attendance was negatively associated with attempted suicide (Kirmayer et al., 1998). Inuit elders' first-hand accounts of suicide, distress and healing in the eastern Arctic have also been captured through a series of narratives, adding a richness and dimension to our current knowledge base (Kral et al., 2000; Leenaars, 1995). Leenaars et al. (1999) have extended such study by capturing not only the narratives of some Inuit, but also from some Aborigine in Australia. These international comparisons show rich similarities to the malaise in both peoples, such as social and cultural turmoil created by the policies of colonialism in Canada and Australia.

Looking more broadly at the issue of Aboriginal suicide in Canada, Lester (1995a) found that there was no association between the proportion of Aboriginal people living in a particular region and the overall suicide rate.

Gays/lesbians

Canadian research exploring sexual orientation and suicide risk has primarily focused on gay men. For example, Bagley and Tremblay (1996) found that attempted suicide rates among young gay men in their sample ranged from 20% to 50%. Based on a survey conducted in Calgary, these authors found significantly higher rates of previous suicidal ideation and actions among homosexual males than among heterosexual males (Bagley and Tremblay, 1997). In a more recent Canadian study of gay and bisexual men, just under half of the subjects reported that they had considered suicide, while approximately 20% reported that they had attempted suicide at least once (Botnick et al., 2002).

Incarcerated populations

Several recent Canadian studies have examined suicide risk among prisoners. Bland, Newman, Dyck, and Orn (1990) looked at the prevalence of psychiatric disorders and suicide attempts among prisoners in Edmonton, while Lester (1995b) studied overall rates of suicide among Canadian prisoners. Green and colleagues (1993) summarized the most common characteristics of suicides among federally incarcerated prisoners and noted that suicides were not associated with age, offence, previous convictions or length of sentence. In a later study, risk factors for inmate suicide were identified following a review of suicide deaths in federal institutions: lengthy involvement in the criminal justice system, a greater likelihood of being incarcerated for robbery or murder, and involvement in institutional incidents of a serious nature (Laishes, 1997). More recently, Fruehwald and colleagues (2001) investigated the relevance of previous suicidal behaviour in understanding suicide among Canadian prisoners.

4c. Determinants of Health

Studies in this category are discussed below under the following headings: child development, social support, occupations and working conditions. Reviews of the literature concerning the determinants of health and suicide (i.e., Dyck, Mishara, and White, 1998) and evidence-based reviews of the association between unemployment and suicide (Jin, Shah, and Svoboda, 1996) are not included here.

Child development

In studies examining children's understanding of death and suicide, researchers in Québec found that among children in 1st, 3rd, and 5th grades who knew what suicide meant, the concept was related to age, the concept of death and personal experiences with death (Normand and Mishara, 1992). In a later study, Mishara (1999) found that by the 3rd grade, children have a fairly sophisticated understanding of suicide, learned by many from television and discussions with other children.

Social support

Hanigan and colleagues (1986) studied the relationship between young adults' suicidal behaviours and the level and quality of support received from their social environment following a critical life event. Later, Tousignant and Hanigan (1993) examined the role of social support among suicidal college students who had recently suffered the loss of a love relationship or the loss of a close friend. They found that suicidal students named fewer important persons in the kinship network and had more conflicts with this network than did nonsuicidal peers.

Meanwhile, in a more recent study, Stravynski and Boyer (2001) examined the construct of loneliness and its relationship to suicide. They found strong associations among suicide ideation, parasuicide and subjective and objective experiences of being lonely and alone.

Occupations and working conditions

A series of studies have been conducted in Canada in recent years to determine if certain occupational groups are associated with elevated risks for suicide. For example, Loo (1986) examined risks for suicide in the RCMP and found that the average annual suicide rate was approximately half that of the comparable general population. Among those officers who did kill themselves, the most common method was service revolver.

Sakinofsky (1987) looked at suicide rates among physicians and compared their age-standardized mortality ratios with those of other occupational groups.

In a study that investigated the characteristics of suicides among Canadian farm operators, including the risks posed by exposure to specific hazards like pesticides, Pickett and colleagues (1993, 1998, 1999) found that provincial suicide rates among farm operators were generally lower than or equivalent to those of Canadian males generally. Furthermore, analyses revealed that there was no association between suicide and exposure to herbicides or insecticides.

In a study of suicide risk among electrical utility workers, Baris, Armstrong, Deadman, and Thériault (1996) did not find strong evidence for causal association between exposure to electric fields and suicide among this particular occupational group.

Finally, in a recent study that investigated risks for suicide among U.N. peacekeepers, Wong and colleagues (2001) found no overall increased risk of suicide in Canadian peacekeepers. The male rate of suicide in this military group was half that of a Canadian civilians comparison group. An apparent increased risk among a subgroup of air force personnel was not related to military hazards but to personal problems such as relationship and financial issues, and also possibly to loneliness and isolation. Airforce personnel did not go through the careful pre-screening and training procedures that their army colleagues were subjected to, which also resulted in mutual bonding and support. These authors suggest that while peacekeeping per se does not increase overall suicide risk, military lifestyles may tax interpersonal relationships, promote alcohol abuse, and contribute to psychiatric illness in a minority of vulnerable individuals.

5. Health Information and Epidemiology

Many of the studies included in this category have considerable overlap with the studies presented in the previous section on the health of populations. For the most part however, studies included here arise out of the epidemiological tradition, i.e., studies of prevalence rates, distribution patterns of morbidity and mortality at the population level, and analyses of suicide rates and trends over time. There have been approximately 50 studies published in this category during the period under review. For discussion purposes, studies will be presented under the following sub-headings: national and regional analyses of mortality statistics, prevalence of suicidal behaviours, and studies of methods.

5a. National and Regional Analyses of Mortality Statistics

Starting with the study of suicide rates based on an analysis of birth cohorts (Lester, 1988; Reed, Camus, and Last, 1985), there has been a longstanding and consistent interest in studying suicide rates at the national level in Canada. Some researchers have examined overall trends in Canadian suicide rates (Beneteau,1988; Chipeur and Von Eye, 1990; Lester, 2000) while others have looked at more specific contributions: age, gender, and geographical region (Dyck, Newman, and Thompson, 1988); country-of-origin (Strachan, Johansen, Nair, and Nargundkar, 1990); birth rates (Lester, 2000); sociodemographic factors (Hasselback, Lee, Mao, Nichol, and Wigle, 1991); age-period-cohort effects (Newman et al., 1988; Trovato, 1986); sex-specific trends (Hutchcroft and Tanney, 1988, 1990); and cohort size (Leenaars and Lester, 1996). Other national studies have relied on time-series designs (Lester, 1997) or broad epidemiological perspectives (Mao, Hasselback, Davies, Nichol, and Wigle, 1990) to study suicides in Canada, while others have turned their attention to the phenomenon of murder-suicides (Gillespie, Hearn, and Silverman, 1998).

Examining rates of suicide at the provincial and regional levels, researchers in Canada have been able to track important trends and highlight interprovincial differences. For example, researchers in Ontario examined suicide rates in that province from 1877 to 1976 (Barnes, Ennis, and Schrober, 1986), while another study investigated suicide rates in Québec from 1951-1986 (Lester, 1995b). Other researchers in Québec have looked at suicide rates in a particular region of the province from 1986 to 1991 (Caron, Grenier, and Béguin, 1995) or among specific age groups (Chiefetz, Posener, LaHaye, Zajdman, and Benierakis, 1987). Isaacs and others (1998) conducted a descriptive review of all suicides in the Northwest Territories and Aldridge and St. John (1991) looked at rates of child and adolescent suicide in Newfoundland and Labrador. Researchers in Québec have looked more closely at homicides which are followed by suicides (Buteau, Lesage, and Kiely,1993; Bourget, Gagne, and Moamai, 2000), while the phenomenon of cluster suicides has been a subject of investigation among researchers in Alberta (Davies and Wilkes, 1993) and Manitoba (Wilkie, Macdonald, and Hildahl, 1998).

5b. Prevalence of Suicidal Behaviours

Several Canadian researchers have conducted large-scale epidemiological studies that seek to estimate the prevalence of suicidal behaviours in specific populations. Bland and his colleagues examined hospital attendance records to estimate the prevalence of parasuicidal behaviours in Edmonton, Alberta and found a rate of 466 per 100,000 per year among those 15 years and older (Bland, Newman, and Dyck, 1994). Newman and Bland (1988) also found that suicide risk varied across the spectrum of affective disorders. Other Alberta researchers (Ramsay and Bagley, 1985) relied on a community survey to better understand the association between suicidal behaviours, attitudes, and experiences among an adult population in Calgary.

Using the data collected through the Ontario Child Health Study, Joffe and his colleagues (1988) found that between five and ten percent of young males, aged 12 to 16, reported engaging in suicidal behaviour within a six-month period. For young females of the same age, the percentage increased to 10 to 20 percent. In a study of adolescent suicide attempters, Grossi and Violato (1992) noted that the lack of an emotionally significant other differentiated suicidal youth from those who had never made an attempt. A number of researchers in Québec have also examined prevalence rates of suicide ideation and behaviours among young people in their province (Bouchard and Morval, 1988; Coté, Provonost, and Ross, 1990; Coté, Provonost, and Larochelle, 1993; Pronovost, Coté, and Ross, 1990). Finally, a national profile of intentional and unintentional injuries among Aboriginal people in Canada has recently been published for the period 1990-1999 (Health Canada, 2001).

5c. Studies of Methods

In an effort to illuminate potential opportunities for prevention, several Canadian studies have analyzed the use of specific methods among those who died by suicide in Canada. For example, Avis (1993, 1994) has summarized the most common characteristics of suicide deaths by firearms and by drowning. In Québec, researchers retrospectively analyzed the circumstances surrounding suicide deaths at a particular bridge site (Prevost, Julien, and Brown, 1996), while Mishara (1999) reviewed a series of suicides on the Montréal subway system.

Examining more rare methods of suicide, Bullock and Diniz (2000) investigated self-suffocation by plastic bags in Ontario, Ross and Lester (1991) looked at suicide deaths at Niagara Falls and Shkrum and Johnston (1992) looked at self-immolation deaths over a three-year period.

Finally, Killias (1993) looked at the association between gun ownership and rates of suicide and homicide. Several other Canadian investigations have looked more specifically at the impact of gun control legislation on suicide and these studies will be reviewed in the next section.

6. Knowledge Development and Policy Research

This small but important category of research concerns itself with the practical matters of research methods and dissemination. Epistemological studies that interrogate both, "how we know," as well as studies that raise questions about "what counts as knowledge," are also included here. The few studies that have been published in this area have appeared in the last five years. Policy-related research, including an examination of specific legislative effects on rates of suicide in Canada, is also in this category. Two sub-headings will be used to review the 16 studies in this category: research issues and the development of knowledge, and policy research.

6a. Research Issues and the Development of Knowledge

Studies reviewed in this category are quite diverse and include those that address specific research issues as well as those that explore the philosophical and ethical underpinnings of certain research traditions and approaches to prevention.

At the most practical levels are studies that specifically address methodological considerations in studies of suicide. For example, Speechley and Stavraky (1991) have raised questions about the adequacy of suicide statistics for epidemiological research purposes. Despite some evidence of underreporting, these authors concluded that underreporting is not large enough to threaten the overall validity of officially reported suicide rates.

More recently, van Reekum and Links (1997) have highlighted some important research considerations in the study of impulsivity and suicide, while Breton and his colleagues (2002) have identified the importance of using "informant-specific correlates of suicidal behaviour" in the development of research and interventions targeting youth suicidal behaviour.

Another level of suicide-related research includes studies that investigate the knowledge development process itself. Included here is the study by Ramsay and colleagues (1990) that retrospectively examined the suicide prevention gatekeeper training model used in Alberta and developed by LivingWorks Education Inc., using Rothman's developmental research model. More recently, Breton (1999) reviewed a series of prevention and mental health promotion programs being delivered across Canada, and noted a distinct lack of conceptual coherence as well as a great deal of confusion about appropriate program aims. To correct some of these difficulties, Breton (1999) recommends that a new understanding of the emergence of mental health problems among children and youth may be required.

Building on this line of inquiry, Schrecker and colleagues (2001) recently published an article that raises some important questions about how current prevention research models and methods may direct our attention towards some specific dimensions of risk while potentially concealing other contributions to the emergence of overall vulnerability. More specifically, these authors suggest that biological and biomedical approaches tend to dominate the research agenda and have the effect of distracting attention away from public health. Furthermore, they believe that discussions of public health and mental illness should refer not only to prevention, but also to social risk reduction.

6b. Policy Research

Several Canadian studies have specifically examined the impact of Bill C-51 on firearm-related suicide deaths in Canada. This legislation was introduced in 1977 to restrict the use of firearms in Canada and many researchers hypothesized that it would have a favourable effect on reducing firearm suicides in Canada.

In one of the first studies, Rich and colleagues (1991) examined suicide rates and methods in Toronto and Ontario for five years before and five years after the enactment of gun control legislation. They found that the legislation led to an overall decrease in firearm-related suicide deaths among men, but the difference was apparently offset by an increase in suicide by jumping from a height, providing preliminary support for a substitution-of-method hypothesis.

Lester and Leenaars' (1993) analysis of Canadian data suggested that there was a decreasing trend in the overall firearm suicide rate and the percentage of suicides by firearms, following the introduction of Bill C-51. Carrington and Moyer (1994) reviewed the suicide mortality data for Ontario before and after the introduction of the gun control legislation and found an overall decreasing trend in firearm suicides with no concomitant evidence of substitution. Leenaars and Lester (1998) conducted a thorough review of the gun control legislation studies in Canada and concluded that while the legislation may have had some impact on suicide rates by firearm, more studies were warranted. Meanwhile, Carrington (1999) re-analyzed the earlier data summarized by Lester and Leenaars and challenged some of their original findings, but subsequent studies support the original Leenaars and Lester finding. Furthermore, more extensive international comparisons show great utility of controlling the environment to reduce suicide (Leenaars, et al., 2000). Public health approaches appear to be most effective in reducing suicide, not only in Canada, but also world wide.

In addition to examining the issue of gun availability in Canada and its effect on overall rates of suicide and homicide (Lester, 1994, 2000), Lester has also looked more broadly at the legislative effect of decriminalizing the act of suicide in Canada (1992).

Finally, there has been one other study in Canada that has examined the effects of provincial policy on the development of programs designed to prevent mental health problems (Nelson, Prilleltensky, Laurendeau, and Powell, 1996). These authors concluded that while the rhetoric of prevention permeated many of the provincial policy documents, there had not been a reallocation of funding in the health field from treatment and rehabilitation services to prevention programs, and funding for prevention remained minimal.


Report a problem or mistake on this page
Please select all that apply:

Thank you for your help!

You will not receive a reply. For enquiries, contact us.

Date modified: