Research Reports

The Prison Careers of Offenders with Mental Disorders 1992, N°. R-33
Frank J. Porporino
Correctional Service Canada
Laurence L. Motiuk
Correctional Service Canada
December, 1992
Published in April, 1994   Author's Note: The points of view expressed in this research are those of the authors and do not necessarily reflect the views or policies of the Correctional Service of Canada. Address all correspondence to the Research and Statistics Branch, Correctional Service Canada, 340 Laurier Ave. W., Ottawa, Ontario, Canada K1A 0P9. In press. International Journal of Law and Psychiatry

TABLE OF CONTENTS

THE PRISON CAREERS OF MENTALLY DISORDERED OFFENDERS

METHOD

Subjects

Instrument

Procedure

Matching Procedure

Data Gathering Process

RESULTS

Criminal Background

Custody Level

Prison Incidents

Temporary Absences

Prison Releases

Post-release Outcome

DISCUSSION

REFERENCES

TABLE 1 Case Characteristics of Non-disordered and Disordered Offenders

TABLE 2 Criminal Conviction History of Non-disordered and Disordered Offenders

TABLE 3 Court Disposition History of Non-disordered and Disordered Offenders

TABLE 4 Prison Custody/Transfer History of Non-disordered and Disordered Offenders

TABLE 5 Prison Incident History of Non-disordered and Disordered Offenders

TABLE 6 Temporary Absence History of Non-disordered and Disordered Offenders

TABLE 7 Prison Release History of Non-disordered and Disordered Offenders

TABLE 8 Post-release Outcome of Non-disordered and Disordered Offenders

TABLE 9 Other Correlates and Re-admission

TABLE 10 Co-morbidity, Release and Re-admission Rates

THE PRISON CAREERS OF OFFENDERS WITH MENTAL DISORDERS

     There is now widespread agreement, both within the research literature and among practitioners, that individuals with mental disorders are increasingly, and often unintentionally, caught in the ambit of the criminal justice system (Roesch & Golding, 1985; Teplin, 1983). Evidence suggests, for example, that individuals with mental disorders are more vulnerable to detection and arrest for nuisance offences (Robertson, 1988; Teplin, 1984), that they are more likely to be remanded in custody for these minor offences, and that they spend disproportionately more time on remand and awaiting a sentencing disposition (Gingell, 1990). Furthermore, it has been shown that many of these individuals are entangled in a cyclical pattern of recurrent and brief encounters with both the mental health and criminal justice systems (Menzies & Webster, 1989). One consequence of this phenomenon of criminalization of the individuals with mental illness is that local jail facilities, holding offenders on remand or serving short terms, have become crowded repositories for people with mental disorders (Adler, 1986; Teplin, 1990).

     Although research in recent years has attended to the consequences of criminal justice processing of individuals with mental disorders, one group of offenders has been relatively ignored -- those offenders who are diagnosed as having a mental illness but are found legally sane, who commit relatively serious crimes and receive lengthy terms of imprisonment.

     Adopting established diagnostic criteria (American Psychiatric Association, 1980), a few studies have now documented the prevalence of mental health problems in random samples of longer-term prison inmates. Consistently, whether examining American state prisoners (Collins & Schlesinger, 1983; Neighbours et al., 1987), Canadian federal inmates (Hodgins & Côté, 1990; Motiuk & Porporino, 1991), or British offenders (Gunn, Maden, & Swinton, 1991), these studies have found rates of psychotic disorders, major depression and other less serious disorders that are substantially higher than in the general population.

     What remains unanswered regarding offenders with disorders serving lengthy prison terms is whether their criminal and prison careers unfold in a fashion that is distinct from the careers of their counterparts without disorders.

     A first question relates to the correctional system's ability to manage these individuals. This subdivides into two related issues -- are there differences in how individuals with mental disorders react to imprisonment (e.g., in terms of rule infractions and other disruptive behaviour) and, as a corollary, are there differences in how the correctional system responds (e.g., in terms of discretionary decisionmaking such as security placement or granting of temporary passes).

     A number of studies have noted higher rates of mental disturbance among inmates who are persistent management problems (Hodgins & Côté, 1991; Toch, 1982; Uhlig, 1976). For example, Adams (1983) found significantly higher rates of rule infractions among inmates who had been hospitalized for mental health difficulties during their prison term. Toch and Adams (1989) arrived at a similar conclusion in a thorough analysis of the patterns of disturbed and disruptive behaviour among prison inmates. Classifying prisoners' behaviour into "episodes", that can include multiple incidents that are either disturbed or disruptive in nature, Toch and Adams (1989) observed that longer episodes, with more incidents, were characteristic of inmates who had been hospitalized or had received outpatient treatment.

     Although suggestive, a major methodological flaw cuts across these studies. Typically, what is used to define the presence of mental disorder is whether there has been involvement with mental health services during the period of imprisonment. However, offenders who have been diagnosed as having a mental disorders and who are also disruptive may be a subgroup who are particularly likely to have frequent mental health contacts. Looking only at this subgroup results in a spuriously inflated rate of prison misbehaviour associated with mental disorder. Rather than relying on mental health contacts to define the presence of mental disorder, a fairer test would follow a sample of offenders, diagnosed as being disordered at one point in time, and assess the frequency of their involvement in rule infractions as compared with a matched sample of offenders who have not been diagnosed with mental disorders.

     A similar approach is needed to assess other aspects of the correctional process in so far as it may differentially affect offenders with mental disorders. It has been shown that media depictions of people suffering from mental illness are characteristically unfavourable, often ascribing unpredictability and dangerousness as co-occurring with mental disorder (Gerbner, Gross, Morgan, & Signorielli, 1981). At least one study has observed that the perceptions of correctional officers are similarly biased (Kropp, Cox, Roesch, & Eaves, 1989). If these sorts of perceptions are pervasive within the correctional environment, then one would expect that they would be reflected in more negative outcomes for offenders with mental disorders flowing from the decisions of correctional authorities.

     Perhaps most fundamental in examining the correctional system's response to offenders with mental disorders is the issue of risk for re-offending, and whether mental disorder in and of itself is a reliable and prominent risk factor. The usefulness of a mental disorder diagnosis in predicting risk for criminal offending continues to be debated. The question emerges as to whether offenders with mental illness are given equitable opportunities for release under community supervision, commensurate with their risk for re-offending.

     As Monahan (1992) has noted recently, the extensive body of research on forensic populations is unequal to the task of illustrating a clear link between mental disorder and propensity for violence. For example, research on samples of patients hospitalized for mental disorders, whether focusing on behaviour before, during or after hospitalization, are confounded by selection biases and the lack of a comparative reference. Some recent epidemiological evidence, from methodologically sound studies of community samples, has indeed suggested that individuals with mental disorders may be more prone to acting out behaviour (Swanson, Holzer, Ganju, & Jono, 1990), especially during times when psychotic symptomology is active (Link, Andrews, & Cullen, in press). However, the nature of the behaviour examined in these studies was self-reported and mostly in the realm of impulsive interpersonal assaultiveness (e.g., hitting and throwing things at your partner; fighting while drinking). It has yet to be determined clearly whether the people with mental disorders are at greater risk for serious offending that demands a criminal justice response. The suggestion has been made that among offenders who are mentally disordered, the same factors are predictive of risk as with other offenders; namely, early onset of criminality and antisocial personality disorder (Hodgins & Côté, in press).

     The present study follows a sample of federally incarcerated offenders who were diagnosed as having experienced a serious mental disorder. On a continuum of psychopathology, we examined only the offenders with extreme disturbance (i.e., using stringent criteria for diagnosis of psychosis). The diagnoses were arrived at as part of a unique national study on the prevalence of mental health problems among offenders within the Canadian federal correctional system (Motiuk & Porporino, 1991). In Canada, offenders who are given prison sentences of two years or more fall under federal jurisdiction. Therefore, the sample of offenders studied were all serving fairly lengthy prison terms. The study examined aspects of the prison careers of these offenders with mental disorders, as well as their pre-imprisonment criminal histories and their post-release outcomes. A matched sample of similarly situated offenders without diagnosed mental disorders was used as a comparative reference to examine possible differences in the correctional system's response to offenders with mental illness.

METHOD

Subjects

     The study focused on a sample of 36 federally sentenced adult male offenders who were identified as meeting stringent diagnostic criteria for manic episode (n = 21), schizophrenia (n = 13) or schizophreniform disorder (n = 2). These disordered cases were identified within one region of the country (i.e., Ontario) as part of a National Mental Health Survey that was conducted by the Correctional Service of Canada (CSC) in 1988 (Motiuk & Porporino, 1991). A matched group of 36 similarly situated offenders who did not meet any of the criteria for having a major psychotic disorder were used to conduct comparative analyses. Instrument

     The study relied on a structured interviewing instrument called the Diagnostic Interview Schedule (DIS) Version III-A. This instrument, developed by the National Institute of Mental Health (Robins & Helzer, 1985), adopts objective diagnostic criteria described in the Diagnostic and Statistical Manual (DSM III) of the American Psychiatric Association (APA, 1980). While the DIS allows for diagnoses on a historical (lifetime) basis, it also allows for alternative diagnostic approaches for some mental disorders, such as with or without premptions by other diagnoses, with or without severity criteria, as well as milder forms of particular disorders, with and without current symptoms. Procedure

     The design of the Mental Health Survey involved a modification of simple random sampling of all male inmates in federal institutions, with the exception of Regional Treatment Centres where a complete census was attempted. In order to improve the national representativeness of the survey, each of the five CSC regions ( i.e., Atlantic, Quebec, Ontario, Prairies, and Pacific ) was stratified into regional subgroups. Sample sizes were calculated to achieve prevalence estimates of mental health problems for the regional populations that would be correct within a 5% margin of error with 99% confidence.

     For each regional portion of the survey, the eligible respondent population was listed by institution and ordered by age. This procedure was followed to ensure that there would be proportional representation by institution, and by age within institutions. Of the 2,772 federal inmates targeted for sampling in the Ontario region, 629 (22.7%) were selected. Of these, 428 (68%) were actually interviewed, and further analyses were based on 420 cases for which there was complete case-file information.

     The organization and administration of fieldwork entailed the careful recruitment and selection of interviewers, formal DIS training with an experienced instructor, and ongoing field supervision and quality control during approximately four months of interviewing (Motiuk & Porporino, 1991). Offenders were asked to volunteer their participation with confidentiality ensured.

     The overall response rate (i.e., complete DIS interviews) in the Ontario region was 68.0% (n = 428), and no institutions had response rates below 60%. In comparing the characteristics of offenders who agreed or declined to participate in the survey (i.e., refusal, no show or not available), Motiuk and Porporino (1991) reported that there was little variation in response rates across security levels and that the distribution of respondent characteristics (i.e., age, marital status, major offence, sentence length and time served) compared favourably with that of non-respondents.

     Prevalence rates of mental disorder for the Mental Health Survey were measured with respect to two defining parameters: temporal reference period and breadth of diagnostic criteria employed. For the purpose of the present investigation, the reference period was lifetime and the breadth of diagnostic criteria was severe and exclusive. Severity criteria included only those cases where all identifying behaviours were present; exclusive criteria removed from the prevalence figures those diagnoses which may have been due to other mental health problems. From the Ontario region survey results, a total of 36 (8.6%) offenders in the region's general population were found to have met the DIS lifetime criteria for a manic episode, schizophrenia or schizophreniform disorder. The prevalence rates of other mental disorders using stringent diagnostic criteria were found to be the following: major depression (11.9%), generalized anxiety (27.9%), psychosexual dysfunction (19.8%), antisocial personality (59.0%), alcohol use/dependence (69.1%) and drug use/dependence (36.7%). It should be noted that an additional 48 cases were interviewed; these were patients at the Regional Treatment/Psychiatric Centre, but they were excluded in order to situate the sample for the present study within the general inmate population for comparative analysis. Matching Procedure

     To construct a matched group for our sample of 36 offenders with disorders, matching criteria were applied as follows. First, a regional sampling population was created of inmates who had been interviewed using the DIS during the Mental Health Survey but who were not found to have met the criteria for lifetime prevalence of either a manic episode, schizophrenia or schizophreniform disorder. For each offender with a disorder, a similarly situated offender without a disorder was identified based on age at time of survey, type of major admitting offence (i.e., offence for which the inmate received the longest sentence) and length of sentence. Sentence length was difficult to match precisely and an attempt was made to arrive at the closest sentence length match within a particular offence category.

     Table 1 presents comparisons between "without" and "with" offenders on a number of characteristics as well as the three matching variables. As Table 1 shows, the "with" offenders did not differ significantly from the "without" offenders on any of the three matching variables. However, there was a tendency for offenders with disorders to have somewhat longer sentence lengths overall. Not listed in the table is the distribution of major admitting offences which was chosen to be equivalent for the two groups. Offence types included the following: homicide (22.2%), sex offences (19.4%), robbery (22.2%), assault(11.1%), drug offences (8.3%) and other Criminal Code offences (16.7%).

TABLE 1

Case Characteristics of Offenders "With" and "Without" Mental Disorders

	"Without" Disorders (n = 36)		"With" Disorders (n = 36)
Variable	M	SD	M	SD	F
Age (years):
at admission	31.8	8.70	32.2	10.7	0.03
at survey	33.6	8.98	34.6	10.9	0.18
	n	%	n	%	X2
Admission Type:
Warrant of Committal	28	77.8	30	83.3	0.36
Other	8	22.2	6	16.7
Admitting Offence:
Violent	28	77.8	28	77.8	0.00
Non-violent	8	22.2	8	22.8
Federal Term:
First	24	66.7	23	63.9	0.06
Second +	12	33.3	13	36.1
	M	SD	M	SD	F
Sentence Length* (days):	2,000	1,138	2,628	1,628	3.00
Convictions (#):
Prior	13.4	13.8	11.4	9.2	0.51
Current	3.9	6.6	3.8	2.9	0.01
Total	17.2	18.9	15.1	10.7	0.33
Violent Convictions (#):
Prior	0.05	0.2	0.11	0.40	0.52
Current	0.08	0.3	0.08	0.28	0.00
Total	0.14	0.4	0.19	0.47	0.33

Note: * Life sentences excluded (n = 6 from each group).

Data Gathering Process

     Data were gathered on the entire study sample of 72 federal inmates from a variety of sources. The CSC's automated Offender Information System provided general demographics (e.g., age), current offence characteristics (e.g., type) as well as correctional process variables (e.g., admission date, transfers, release date, release type). CSC's automated Security Incident System data base yielded information on all recorded incidents (e.g., possession of contraband, inmate assault) for each offender's period(s) of federal supervision. CSC's automated Parole Supervision System data base gave information on suspension warrants for those offenders who were released under community supervision. The Canadian Police Information Centre was also utilized and provided the complete offence history (e.g., criminal convictions, dispositions) for each offender.

     Information on all the correctional process variables was gathered over the period of approximately four years following the completion of the Mental Health Survey (i.e., to November 15, 1992).

RESULTS

Criminal Background

     When we examine the criminal conviction histories of the two matched groups of offenders "with" or "without" disorders, the similarities are striking. Table 1 shows that about the same proportions in the two groups were first-time federal offenders (67% of the "without" disorders and 64% of the group "with" disorders). There were no significant differences in the mean number of prior convictions, either in total or specifically for prior violent convictions. Similarly, there were no differences in the number of current convictions. Table 2 shows the pattern of prior convictions for each of ten different categories of offences. Again we see very little difference in the number and type of criminal convictions on record.

TABLE 2

Criminal Conviction History of Offenders "With" and "Without" Disorders

	"Without"		"With"
Type of Conviction	M	SD	M	SD	F
Homicide	0.28	0.51	0.25	0.60	0.04
Attempted Murder	0.28	0.17	0.28	0.17	0.00
Sex Offence	0.50	1.95	0.75	1.44	0.38
Robbery	1.17	2.62	1.14	2.71	0.00
Assault	1.36	1.55	0.94	1.54	1.30
Property	5.67	6.61	5.06	4.72	0.20
Fraud	3.03	14.11	2.11	6.59	0.12
Public Morals	2.08	2.69	1.67	1.82	0.59
Drug	0.64	0.99	1.11	2.76	0.93
Alcohol	0.19	0.52	0.03	0.17	3.30

     Another perspective in assessing severity of criminal conviction history is to examine court dispositions. Table 3 presents the total volume of court dispositions (e.g., absolute discharges, fines, probation terms, etc.) recorded for both the groups "without" and "with" disorders. Again, statistical analyses revealed no significant differences between the two groups in the number and types of court dispositions.

TABLE 3

Court Disposition History of Offenders "Without" and "With" Disorders

	"Without"		"With"
Type of Conviction	M	SD	M	SD	F
Absolute Discharges	0.03	0.17	0.03	0.17	0.00
Suspended Sentences	0.56	0.88	1.06	1.30	3.63
Fines	1.81	2.64	1.19	1.88	1.28
Restitution Orders	0.08	0.28	0.19	0.47	1.50
Probation Terms	1.50	1.56	2.11	1.70	2.52
Incarcerations	4.47	2.92	4.50	1.54	0.00

     Custody level assignments can be viewed as the correctional system's generalized response to risk; the greater the perceived risk posed by the offender, the higher the level of custody. In particular, transfers to minimum security, are a preparatory response to consideration for release. Table 4 provides a breakdown of the custody level assignments and transfer histories for the two groups of offenders "without" or "with" disorders. At the time that the Mental Health Survey was conducted, no overall difference was observed in the distribution of cases across levels of security, X² (2, N = 72) = 3.26, ns. However, a clear trend was evident with a smaller proportion of offenders "with" disorders seen as suitable for minimum security placement (5.6% for those with disorders versus 19.4% for those without disorders). Subsequent to the survey, a similar proportion of cases in both groups was transferred (72%), and the total mean number of transfers was equivalent for the two groups. When the first transfer post-survey was examined, no overall difference was observed between the two groups with reference to changes in security designation (i.e., either no change or an upward or downward change). However, again, a trend is evident with fewer offenders "with" disorders having moved downward in custody level and more having moved upward.

TABLE 4

Prison Custody/Transfer History of Offenders "With" and "Without" Disorders

	"Without"		"With"
Variable	n	%	n	%	X2
Custody Level at Survey
Minimum	7	19.4	2	5.6	3.26
Medium	16	44.4	20	55.6
Maximum	13	36.1	14	38.9
Transferred Post-survey
Yes	26	72.2	26	72.2	0.00
No	10	27.8	10	27.8
First Transfer Post-survey
No Change	10	38.5	10	38.5	2.13
Upward	4	15.4	8	30.8
Downward	12	46.2	8	30.8
	M	SD	M	SD	F
# of Transfers	2.42	2.87	3.17	4.10	0.81

     Not surprisingly, during their current federal prison term, a significantly larger proportion of the "with" group had been in the Regional Treatment Centre on a previous occasion relative to the "without" group (36.1% and 13.9%, respectively X² (1, N = 72) = 4.74, p<.05). Similarly, a larger proportion of the "with" group than of the "without" group was transferred to the Regional Treatment Centre subsequent to the Mental Health Survey (33.3% and 13.9%, respectively); however, the difference was not statistically significant. With respect to those inmates subsequently transferred for some form of mental health treatment (including treatment for sex offending), 73.3% were from the "with" group whereas 26.7% were from the "without" group. Prison Incidents

     Involvement in significant prison incidents was assessed by classifying all recorded instances post-survey as violent (e.g., assaults ), escape, possession of contraband, general behaviourial disruption, and other. Table 5 presents the mean number of these types of security incidents for both the "without" and "with" groups. The results indicated that the level of involvement in prison incidents was quite similar for the two groups. Moreover, offenders "with" disorders showed no particular pattern in the types of incidents in which they were involved while in prison.

TABLE 5

Prison Incident History of Offenders "Without" and "With" Disorders

	Non-disordered		Disordered
Variable	n	%	n	%	X2
Incident
Yes	18	50.0	17	47.2	0.06
No	18	50.0	19	52.8
Violent
Yes	6	16.7	7	19.4	0.09
No	30	83.3	29	80.6
Escape
Yes	3	8.3	0	0.0	3.13
No	33	91.7	36	100.0
Contraband
Yes	11	30.6	7	19.4	1.19
No	25	69.4	29	80.6
Behavioral
Yes	7	19.4	4	11.1	0.97
No	29	80.6	32	88.9
Other
Yes	12	33.3	9	25.0	0.61
No	24	66.7	27	75.0
	M	SD	M	SD	F
# of Incidents	1.75	2.60	1.67	2.45	0.02

Temporary Absences

     Temporary absences are granted to offenders either for essential medical or administrative (e.g., court appearance) purposes, usually under escort, or as a precursor to release to assist in gradual reintegration. The post-survey history of experience with temporary absences for the offenders "without" disorders and offenders "with" disorders is shown in Table 6. There were no significant differences in the proportions of offenders in the two groups who were granted various forms of temporary absences. There was some evidence that the offenders "without" disorders were granted a greater number of both unescorted and escorted passes, however, the difference in the means for the two groups was not statistically significant.

TABLE 6

Temporary Absence History of Offenders "Without" and "With" Disorders

	"Without"		"With"
Variable	n	%	n	%	X2
Temporary Absence(s) Granted
Yes	26	72.2	28	77.8	0.30
No	10	27.8	8	22.2
Escorted
Yes	26	72.2	27	75.0	0.07
No	10	27.8	9	25.0
Unescorted
Yes	11	30.6	8	22.2	0.64
No	25	69.4	28	77.8
Medical
Yes	20	55.6	24	66.7	0.94
No	16	44.4	12	33.3
Administrative
Yes	6	16.7	4	11.1	0.47
No	30	83.3	32	88.9
Community Service
Yes	11	33.6	6	16.7	1.93
No	25	69.4	30	83.3
Family Contact
Yes	13	36.1	8	22.2	1.68
No	23	63.9	28	77.8
Compassionate
Yes	2	5.6	3	8.3	0.22
No	34	94.4	33	91.7
Rehabilitative
Yes	6	16.7	10	27.8	1.29
No	30	83.3	26	72.2
	M	SD	M	SD	F
# of Temporary Absences	7.7	33.2	9.9	16.7	1.58
# Escorted	16.5	32.8	9.2	16.6	1.43
# Unescorted	1.2	2.3	0.7	1.5	1.27

Prison Releases

     Conditional release from federal custody in Canada occurs either in the form of parole, which is dependent on eligibility requirements and granted at the discretion of the National Parole Board, or in the form of mandatory supervision, which typically occurs after two thirds of the sentence is served. Generally, 40% of federal offenders are released on full parole and 60% on mandatory supervision.

     Table 7 shows the patterns of release from federal custody for both the offenders "without" disorders and those "with" disorders. The findings indicate that, during the four-year follow-up, approximately as many offenders in the "with" group were released as those in the without group (67% and 75% respectively). Of those who were not released, 5 of the 9 in the "without" and 6 of the 12 in the with group were serving life sentences. The remainder in both groups were all past their parole eligibility dates.

     Although offenders "with" disorders were equally likely to be released, when we look at type of release, a clear and significant difference emerges. Offenders with disorders as a group, were much more likely to be released on mandatory supervision, while the offenders without disorders were released more often on parole, at about the same frequency as the base-rate in the general federal inmate population X² (1, N = 72) = 4.55, p<.05. Table 7 also shows the average time served prior to release, in number of days, and the average proportion of the sentence served before release for the two groups. Although the differences do not reach significance, there is a tendency for offenders with disorders to serve more time prior to release, in absolute terms, and to serve a greater portion of their sentence, in relative terms.

TABLE 7

Prison Release History of Offenders "Without" and "With" Disorders

	"Without"		"With"
Variable	n	%	n	%	X2
Released
Yes	27	75.0	24	66.7	0.61
No	9	25.0	12	33.3
Type
Parole	12	44.4	4	16.7	4.55*
Mandatory	15	55.6	20	83.3
	M	SD	M	SD	F
Time served (days)	1,058	592.2	1,371	836.3	0.00
Proportion of Time Served	.59	.21	.65	.39	0.93

Note: *p<.05.

Post-release Outcome

     To conduct a post-release follow-up of the two groups, four outcome measures were used (suspension warrant executed, re-admission to federal custody, readmission with a new offence and re-admission with a violent new offence). Two follow-up periods were taken in order to examine both early and more sustained post-release performance (6 and 24 months). The analyses presented in Table 8 reveal no significant differences between the "without" and "with" groups in relation to early outcome while on conditional release (i.e., at a 6 month follow-up). There is a trend, however, for more of the offenders without disorders to be returned for a new offence or a new violent offence.

     With an extended follow-up period of 24 months, the offenders with disorders were now significantly more likely to be suspended during their period of community supervision X² (1, N = 72) = 4.46, p<.05. On the other hand, the "without" group was more likely to return to federal custody with a new offence X² (1, N = 72) = 4.9, p <.05. Upon closer examination, it was found that the largest proportion of the "with" disorders group (85.7%) was re-admitted for revocations without a new offence, whereas the largest proportion of the "without" disorders group (54.5%) was re- admitted for new offences.

TABLE 8

Post-release Outcome of Offenders "Without" and "With" Disorders

	"Without"		"With"
Follow-up Period/Measure	n	%	n	%	X2
6 Months/ (n = 47)
Suspended
Yes	10	38.5	11	47.6	0.91
No	16	61.5	10	52.4
Re-admission
Yes	11	42.3	7 3	3.3	0.40
No	15	57.7	14	66.7
New offence
Yes	6	23.1	1	4.8	3.07
No	20	76.9	21	95.2
Violent New Offence
Yes	3	11.5	1	4.8	0.69
No	23	88.5	20	95.2
24 Months/ (n = 37)
Suspended
Yes	6	30.0	11	64.7	.46*
No	14	70.0	6	35.3
Re-admission
Yes	8	40.0	6	35.3	0.09
No	12	60.0	11	64.7
New Offence
Yes	5	25.0	0	0.0	4.91*
No	15	75.0	17	100.0
Violent New Offence
Yes	2	10.0	0	0.0	1.80
No	18	90.0	17	100.0

Note: * p<.05.

     We were interested in factors that might be generally predictive of post-release outcome (i.e., age, conviction history experience of, other mental disorders) and, more specifically, whether the same factors would correlate with outcome for the "without" disorder and the "with" disorder groups. These correlates are shown in Table 9, with reference to the outcome measure of any re- admission during the entire follow-up period. Interestingly, the total number of prior convictions was the only factor significantly related to re-admission for the "with" disorders group. No factor was significantly related to re- admission for the "without" disorders group. Recalling that offenders with disorders were re-admitted disproportionately for revocations without a new offence, this suggests that when a lengthy criminal history combines with evidence of mental disorder, conditional release may be more readily revoked by supervision authorities.

TABLE 9

Other Correlates and Re-admission

	"Without" Disorders (n = 27)	"With" Disorders (n = 24)
Measure	r	r
Age at admission	-.15	-.12
Age at survey	-.19	-.11
Convictions # - Prior Convictions # - Total	.34 .35	.56** .53**
Violent # - Prior Violent # - Total	-.16 -.13	.31 .27
Major Depression	-.16	.06
Anxiety	.37	-.26
Psychosexual Dysfunction	.08	.13
Antisocial Personality	.27	-.16
Alcohol Use/dependence	.27	-.16
Drug Use/dependence	.02	.00

Note: ** p<.01.

     It is noteworthy that the presence of any of the other disorders assessed with the Mental Health Survey, including antisocial personality disorder (APD), was generally not found to be related to post-release outcomes. This was examined further with a set of analyses focusing on co- morbidity, with APD and alcohol and/or drug dependence, separately within both the "without" disorders and "with" disorders groups. The results are shown in Table 10.

TABLE 10

Co-morbidity, Release and Re-admission Rates

			Released		Re-admission		Readmission with Offence
Diagnoses	N	%	n	%	n	%	n	%
"Without" Disorders +
Non-APD + Non-alcohol/drug	11	30.6	9 81.8		2 22.2		1 1.1
Non-APD + Alcohol/drug	1	2.8	0 0.0		0 0.0		0 0.0
APD + Non-alcohol/drug	11	30.6	7 63.6		4 57.1		2 28.6
APD + Alcohol/drug	13	36.1	11 84.6		5 45.5		3 27.3
Subtotal	------------- 36	------ 100.0	------------- 27	------ 75.0	----------- 11	----- 40.7	-------- 7	------ 25.9
"With" Disorders +
Non-APD + Non-alcohol/drug	6	16.7	4	66.7	2	50.0	1	25.0
Non-APD + Alcohol/drug	1	2.8	0	0.0	0	0.0	0	0.0
APD + Non-alcohol/drug	8	22.2	6	75.0	2	33.3	0	0.0
APD + Alcohol/drug	21	58.3	4	66.7	4	28.6	0	0.0
Subtotal	------------- 36	------ 100.0	------------- 24	------ 66.7	----------- 8	------ 33.3	-------- 1	------- 4.2

Note: APD = Antisocial Personality Disorder.

     In both the "without" and "with" disordered groups, the greatest proportion of offenders had both a co-occurring diagnosis of APD, as well as alcohol and/or drug dependence (36% and 58%, respectively). When the percentage of cases released is examined by category of co- morbidity, the pattern is quite similar within both the "without" and "with" disorders group. Offenders showing evidence of APD with alcohol and/or drug use/dependence are as likely to be released into the community as those exhibiting neither APD nor an alcohol/drug disorder. When the percentage of cases readmitted is considered, however, a different pattern emerges in the "without" disorders and "with" disorders groups. APD, whether with or without alcohol and/or drug dependence, clearly distinguishes the "without" offenders as higher risk for re-admission. However, the pattern is unclear for offenders with disorders. Part of the explanation for these results is found in the fact that re-admission for offenders with disorders is occurring not for new offences, but for violations of supervision conditions. This likelihood of revocation for offenders "with" disorders seems to be more highly correlated with criminal history than with evidence of co-occurring disorders that are generally predictive of risk.

DISCUSSION

     Consistent with findings at other stages in the criminal justice process, the results of the present study suggest that there is a distinguishable and differential correctional response in the management of offenders "with" mental disorders who are convicted of serious offences. Our approach contrasted a sample of Canadian federal offenders who had met the criteria for a psychotic diagnosis with a matched group of similarly situated offenders "without" disorders. Beyond the factors that were controlled in the matching process (i.e., age, nature of current offence, and sentence length), the two groups were also found to be equivalent in criminal history and in the nature of their institutional comportment (i.e., involvement in institutional incidents). Despite these equivalencies, the offenders with disorders were given fewer opportunities for early release on full parole (released instead under mandatory supervision), and, when released, they were significantly more likely to receive suspension warrants or have their supervision revoked without commission of a now offence (issued as a result of failure to abide by supervision conditions). This exercise of caution in managing offenders with mental disorders becomes particulary noteworthy when release outcomes are examined; it was the "without" disorders group, and not the "with" disorders group, who were more likely to commit a new offence while under supervision.

     There was evidence that caution in dealing with offenders with disorders also occurred earlier in the correctional process. Custody level assignments showed some disadvantage for offenders "with" disorders, with fewer of these offenders being considered appropriate for lower levels of security. Similarly, offenders with disorders seemed to have less access to escorted or unescorted temporary absence passes.

     One possible explanation for these findings is that there may be a biased presumption of greater risk for offenders with mental disorders, particularly for those who also have a lengthy record of criminal convictions. Some support for this is found in the significant relationship between number of prior convictions and likelihood of re- admission for "with" disorders group not for offenders "without" disorders. It would seem that the combination of mental disorder and a lengthy criminal record is interpreted by parole supervisors as warranting special attention, and revocation at the earliest sign of breakdown in the case. A biased presumption of greater risk would also explain the National Parole Board's hesitancy to release offenders with mental disorders on full parole.

     Another explanation is that the types of specialized mental health services, which may give correctional staff more confidence in managing offenders with disorders, are simply unavailable. With a scarcity of alternatives for support and for ongoing monitoring of symptomology, correctional staff resort to exercising caution. Similarly, in assessing the quality of release plans, the National Parole Board would appropriately demand evidence of availability of mental health services. Offenders with disorders may be deemed unsafe to release without such services.

     We can conclude, as have others in commenting on how the criminal justice system may become better equipped to deal with mental illness, that a coordinated network of specialized services is needed to target the needs of offenders with disorders (Corrado, Doherty, & Glackman, 1989; Rogers & Bagby, 1992). Use of incarceration as a cautious but more restrictive option might thereby be minimized.

REFERENCES

American Psychiatric Association (1980). DSM III: Diagnostic and statistical manual of mental disorders. New York, Masson.

Adams, K. (1983). Former mental patients in a prison and parole system: A study of socially disruptive behaviour. Criminal Justice and Behaviour, 10, 358-384.

Adler, F. (1986). Jails as a repository for former mental patients. International Journal of Offender Therapy and Comparative Criminology, 30, 225-236.

Collins, J.J., & Schlenger, W. E. (1983). The prevalence of psychiatric disorder among admissions to prison. Paper presented at the American Society of Criminology, Denver, Co.

Corrado, R.R., Doherty, D., & Glackman, W. (1989). A demonstration program for chronic recidivists of criminal justice, health, and social service agencies. International Journal of Law and Psychiatry, 12, 211-229.

Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1981). Health and medicine on television. The New England Journal of Medicine, 305, 901-904.

Gingell, C. (1990). Criminalization of the mentally ill: Some data to examine the hypothesis. Doctoral Dissertation, Simon Fraser University.

Gunn, J., Maden, T., & Swinton, M. (1991). Mentally disordered prisoners. A report commissioned and published by the Home Office, United Kingdom.

Hodgins, S., & Côté, G. (1990). The prevalence of mental disorder among penitentiary inmates. Canada's Mental Health, 38, 1-5.

Hodgins, S., & Côté, G. (in press). The criminality of mentally disordered offenders. Criminal Justice & Behaviour.

Kropp, F.R., Cox, D. N., Roesch, R., & Eaves, D. (1992). The perceptions of correctional officers toward mentally disordered offenders. International Journal of Law and Psychiatry, 12, 181-188.

Link, B.G., Andrews, H., & Cullen, F. T. (in press). Reconsidering the dangerousness of mental patients: Violent and illegal behaviour of current and former patients compared to controls. American Sociological Review.

Menzies, R.J., & Webster, C. D. (1989). Mental disorder and violent crime. In Weiner, N., & Wolfgang, M. (Eds.), Pathways to criminal violence (pp. 109-136). London: Sage.

Motiuk, L.L., & Porporino, F. J. (1991). The prevalence, nature and severity of mental health problems among federal male inmates in Canadian Penitentiaries. Report No. 24, Research and Statistics Branch, Correctional Service Canada.

Neighbours, H.W., Williams, D. H., Gunnings, T. S., Lipscomb, W. D., Broman, C., & Lepkowski, J. (1987). The prevalence of mental disorder in Michigan prisons. Final report submitted to the Michigan Department of Corrections, MI.

Robins, L.N., & Helzer, J. E. (1985). Diagnostic Interview Schedule (DIS) Version 111-A. Department of Psychiatry, Washington University School of Medicine.

Robins, L.N., Helzer, J. E., Croughan, J., & Ratcliff, K. S. (1981). National Institute of Mental Health Diagnostic Interview Schedule: Its history, characteristics and validity. Archives of General Psychiatry, 38, 381-389.

Robertson, G. (1988). Arrest patterns among mentally disordered offenders. British Journal of Psychiatry, 153, 313-316.

Rogers, R., & Bagby, M. (1992). Diversion of mentally disordered offenders: A legitimate role for clinicians. Behavioral Sciences and the Law, 10, 407-418.

Roesch, R., & Golding, S. L. (1985). The impact of deinstutionalization. In D. P. Farrington & J. Gunn (Eds.), Aggression and Dangerousness (pp. 209 - 239). New York: J. Wiley & Sons.

Swanson, J.W., Holzer, C. E., Ganju, V. K., & Jono, R. T. (1990). Violence and psychiatric disorder in the community: Evidence from the epiderniologic catchment area surveys. Hospital and Community Psychiatry, 41, 761-770.

Teplin, L. (1983). The criminalization of the mentally ill: Speculation in search of data. Psychological Bulletin, 94, 54-67.

Teplin, L. (1984). Criminalizing mental disorder: The comparative arrest rate of the mentally ill. American Psychologist, 39, 794-803.

Toch, H. (1982). The disturbed disruptive inmate: Where does the bus stop? Journal of Psychiatry and the Law, 10, 327-349.

Toch, H., & Adams, K. (1989). The disturbed violent offender. London: Yale University Press.

Uhlig, H. (1976). Hospitalization experience of mentally disturbed and disruptive incarcerated offenders. The Journal of Psychiatry and Law, 4, 49-59.