FORUM on Corrections Research

An examination of suicide attempts among inmates

Ralph C. Serin, and Larry Motiuk¹
Research Branch, Correctional Service of Canada
Cherami Wichmann²
Family, Children and Youth Services, Department of Justice

To date research suggests that prison settings have a higher incidence of suicidal behaviour than other settings.³ An important and consistent observation has been that many offenders who complete suicide have a history of attempts, suggesting that an increased understanding about these individuals may also reduce the prevalence of completed suicide. Presently, much of the research in this area has been descriptive, such that the relative contribution of risk factors is unclear. Also, the research has often been retrospective rather than postdictive. Quantitative analyses and the use of comparison groups are also infrequent in this area, thereby impacting on being able to draw conclusions.

This article presents an overview of an ongoing research initiative regarding the assessment and prediction of suicidal behaviour among offenders to address the aforementioned limitations. First, research regarding the prediction of suicide attempts among male offenders is presented. Second, research regarding self-harm among women offenders is described. Third, a comparison between attempters and completers for a sample of male offenders is provided. An aspect of this research is the emphasis on suicidal behaviour as a process, with static, vulnerability and protective, and precipitating factors being considered important to improve our understanding.⁴

Suicide attempts in male offenders

Using work by Polvi⁵ as a guide, static factors (previous self-harm, psychiatric history, adjustment problems) were assessed in addition to general demographics (age at admission, race), criminality, and family history. Further, indicators from the Offender Intake Assessment⁶ were used to assess a range of criminogenic needs and psychological constructs. From a population of 14,500 offenders, a sample of 731 attempters was identified plus a random comparison sample of 731 offenders who had never attempted suicide while in federal custody. The attempter group were younger, more likely to be single, but there were no differences regarding ethnicity.⁷ The attempter group also had different criminal histories, being more likely to have committed homicides, break and enters or robberies, and less likely to have committed sex offences.

Attempters were over-represented as being in maximum security.

From this sample of attempters, a subsample of 76 was identified for whom comprehensive information was available. A comparison sample of 76 offenders was identified, matched according to age at admission (above and below 30 years), sentence length (4 years or less and 5 years and greater), and type of offence (violent crimes and non violent). Multivariate and univariate analysis of variance revealed that attempters had more serious mental health difficulties (externalizing and social cognitive problems, internalizing problems, dysfunctional family relationships). Criminal risk also significantly distinguished between groups, even after the matching strategy. The attempters also had more problematic institutional adjustment. They had significantly, more violent incidents, contraband incidents, and more requests for protective custody, relative to the matched comparison group.

At intake offenders are assessed for suicide risk potential using a standardized rating scale that has 9 indicators reflecting a standard of care for such practice. The indicators represent prior history, recent loss, depressive symptoms, current suicidal ideation, and presence of a suicide plan. One aspect of this research was to determine the efficacy of this assessment procedure. The internal consistency of the scale was 0.77. Also, the attempter group had a significantly higher mean total score than the comparison group (r2 = .17). Comparisons indicated significant differences between attempters and the comparison sample on 4 scale items - a rating the offender may be suicidal; a prior suicide attempt; recent psychological/psychiatric intervention; and, signs of depression. Finally, including proximal (institutional disciplinary problems, contraband incidents) and distal factors (previous adult convictions) in an exploratory logistical regression with these suicide scale items, served to enhance the prediction of a suicide attempt during incarceration.

Summary

These findings are consistent with prior research. Attempters were predominantly young unmarried Caucasian men who had committed violent (person-related) crimes. Attempters were classified as higher criminal risk at intake and placed in higher security. Importantly, older offenders were not more likely to attempt suicide. When age was controlled, there was no direct association between marital status and suicide attempts; suggesting marital status is more related to age than risk of suicide. Psychological functioning assessed at intake differentiated offenders who later attempted suicide from those who did not. Attempters displayed more externalizing and internalizing problems, had more extensive psychiatric histories, and more dysfunctional families. Augmenting the suicide risk scale with institutional adjustment and criminal history information improved the prediction of subsequent suicide attempts.

Self harm among women offenders

Estimates indicate that almost half of women offenders have attempted suicide8 and that other forms of self-harm are prevalent in this high-risk population.9 Such findings underscore the merits of conceptualizing self-harm as a mental health concern for women offenders.

The sample included 155 federally sentenced women, a target group of 78 who had engaged in self-harm while in federal custody and a comparison group of 77 who had not.¹⁰ The groups were matched on admission year (before or after 1994), age at admission (under 30, above 30), sentence length (3 years or less, 4 to 9 years, 10 years or more), and offence type (violent, non violent). The self-harm group was more likely to be unmarried, Caucasian or Aboriginal. In terms of criminal history, the attempter group was significantly more likely to have prior convictions, escapes, disciplinary infractions, less time crime free, a violent history, and prior failures on conditional release. Based on intake ratings, the attempter group had significantly greater problems in terms of employment, marital/family, substance abuse, community functioning, and personal/emotional stability. Women offenders who engaged in self-harm were also significantly more likely to have difficulties in terms of externalizing and social cognitive problems, internalizing and psychiatric problems, and lacked education and cognitive functioning skills.

Investigation of institutional adjustment indices revealed that the self-harm group was more likely to have been segregated and to have been involved in a range of disturbances (contraband, disciplinary infractions, been victimized, violent incidents, and substance abuse). While preliminary and descriptive, these findings are comparable to those for male offenders. Although there may be gender differences regarding the etiology of self-harm, their expression appears similar for male and women offenders. For those women who exhibited suicidal behaviour, there were a host of adjustment difficulties that were also prevalent. These difficulties were not apparent in a matched group of women offenders. Also, while coping deficits may be related to the expression of suicidal behaviour in these women offenders, notably their violent behaviour was not exclusively self-directed.

Completers versus attempters in male offenders

This group was a random sample of 48 offenders who had attempted suicide but were unsuccessful and 48 offenders who succeeded in committing suicide. A comparison group of 48 offenders who had not attempted suicide were matched in the same manner as described previously. Inter-rater reliability for coding variables was calculated for 10% of the sample and was found to be acceptable (92% of variables had agreement > 75%). Consistent with the literature and not surprisingly, method of self-injury varied by group. Attempters were significantly more likely to overdose and slash, whereas completers were more likely to hang/suffocate themselves. Suicide attempters, however, were heterogeneous with respect to intent, with greater than 20% reporting very serious intent and 22% using methods with high lethality.

In terms of demographics, relative to the attempters, the completers had a slightly higher mean age, were more somewhat likely to be Aboriginal, and were more likely to be in maximum security. Regarding criminal history, the comparison group was significantly less likely to have violated parole/probation and was less likely to have had escapes than completers and attempters. While both the completers and attempters were somewhat more likely to have a suicide alert on file relative to the comparison group, fewer than 20% of offenders who actually did attempt or commit suicide were flagged as high suicide risk at intake. Significantly more offenders who attempted or completed suicide during their present federal sentence were more likely to have made previous attempts and to have engaged in self-injury than the comparison group. More importantly, the attempters were significantly more likely to have made previous attempts and to have engaged in self-injury than the completers. Also, while attempters and completers had more serious psychiatric histories than the comparison group, there were no between group differences. The type of psychiatric diagnoses, however, was instructive: attempters were more likely to have been diagnosed as having antisocial personality disorder, whereas the completers were more likely to have been diagnosed as schizophrenic, paranoid, or depressed; the comparison group had fewer institutional adjustment difficulties than those offenders who engaged in suicidal behaviour; those who attempted suicide had more adjustment difficulties in terms of acting-out behaviour than completers. Proximal to the target date of suicide, very few attempters were in minimum security, but attempters and completers were over-represented in maximum security. Finally, significantly fewer completers participated in correctional programs, relative to the attempters, whose rate of participation was significantly lower than the comparison group.

Assessment of psychological adjustment at intake also differentiated among groups. Completers were rated as less impulsive and with poorer coping skills than the other groups. Their social problem-solving skills were comparable to the comparison group, but notably poorer than the attempter group. At intake, there were no significant between group differences for depression, anger/hostility, suicidal ideation, or insight.

Proximal to the target date, the attempters had significantly poorer family relationships than the completers, whose family relationships were poorer than the comparison group. Also, the completers had markedly poorer relationships with other offenders than the attempters and both groups were significantly poorer than the comparison group. Finally, proximal to the target date, attempters and completers had more adjustment difficulties in terms of depression, impulsivity, anger/hostility, coping, psychiatric symptoms, isolation, insight, and social problem-solving.

Additional investigation of proximal events was informative. There appeared to be deterioration in adjustment prior to the target date for 19% of completers, 10% of attempters, and 2% of the comparison group. Further, for attempters and completers, precipitants were identified proximal to the target date that could affect level of suicidal ideation (e.g., negative personal events/decisions, transfer, placement in segregation, being under the influence, inmate pressure). Importantly, relative to the comparison group, the attempters and completers had significantly more precipitants.

Conclusions

The research described suggests there are identifiable factors, both static and proximal, that distinguish attempters from non-attempters and that predict suicidal attempts and self-harm in male and women offenders. The validity of the suicide risk scale completed at intake was supported, and the inclusion of institutional adjustment and criminal history variables augmented its utility. As well, proximal precipitants distinguished offenders who engaged in suicidal behaviour from those who did not. Finally, comparisons between completers and attempters highlighted important differences in terms of social adjustment, coping skills, and program involvement. Together, these findings support the importance of assessment at intake and the monitoring of offenders’ behaviour over time in order to enhance clinical practice.

---

1.  340 Laurier Avenue West, Ottawa, Ontario K1A 0P9

2.  284 Wellington Street, Ottawa, Ontario K1A 0H8

3.  We define suicidal behaviour as encompassing attempts and completions. Burtch, B., and Ericson, R. (1979). The silent system: An inquiry into prisoners who suicide. Toronto, ON: Centre of Criminology, University of Toronto. See also Dooley, E. (1990). Prison suicide in England and Wales 1972-1987. British Journal of Psychiatry, 156, 40-45.

4.  Heikkinen, M., Aro, H., & Lönnqvist, J. (1993). Life events and social support in suicide. Suicide and Life-Threatening Behaviour,
23 , 343-358.

5.  Polvi, N. H. (1997). Assessing risk of suicide in correctional settings. In C. D. Webster & M. A. Jackson (Eds.), Impulsivity, Assessment, and Treatment (pp. 278-301). New York, NY: Guilford Press.

6.  Motiuk, L. L. (1997). Classification for correctional programming: The offender intake assessment process. Forum on Corrections Research 9(1), p. 14-18.

7.  Wichmann, C. G., Serin, R., and Motiuk, L. (2000). Suicide and parasuicide in federal institutions: Predicting suicide attempts by male offenders. Research Report R-91. Ottawa, ON: Correctional Service of Canada.

8.  Blanchette, K. (1997). Maximum-security female and male federal offenders: A comparison. Research Report R-53. Ottawa, ON: Correctional Service of Canada. See also Loucks, A. D., and Zamble, E. (1994). Some comparisons of female and male serious offenders. Forum on Corrections Research, 6(1), 22-24.

9.  Heney, J. (1990). Report on self-injurious behaviour in Kingston Prison for Women. Ottawa, ON: Correctional Service of Canada.

10.  Wichmann, C., Serin, R., and Abracen, J. (2002). Women offenders who engage in self-harm: A comparative investigation. Research Report R-123. Ottawa, ON: Correctional Service of Canada.