Descriptive characteristics of the sample
The complete sample comprised 327 forensic patients. The vast majority were male (89.6%). Male forensic patients were older than female patients at the time of our assessment (mean 45.2 years v 42.9 years), but not significantly so (t = 1.079, p = 0.281).
In terms of their placement at the time of assessment, approximately one third were placed in high security, one third were in secure hospital settings or prison, and one third were in the community (see Table
1).
Table 1
Forensic Patient Placement by Gender
Prison | 25 (8.5%) | 0 (0%) | 25 (7.6%) |
High Security | 101 (34.5%) | 8 (23.5%) | 109 (33.3%) |
Medium Security | 45 (15.4%) | 13 (38.2%) | 58 (17.7%) |
Low Security | 8 (2.7%) | 1 (2.9%) | 9 (2.8%) |
Open Security | 22 (7.5%) | 3 (8.8%) | 25 (7.6%) |
Community | 92 (31.4%) | 9 (26.5%) | 101 (30.9%) |
Total | 293 | 34 | 327 |
Thirty-five (10.7%) of the forensic patients had identified as Aboriginal or Torres Strait Islander (ATSI), with significantly more females than males (20.6% v 9.6%, χ2 = 3.879, p = 0.049). Comparing their country of birth, approximately three quarters of the forensic patients were born in Australia, with no significant differences by gender (Male = 73.6%, Female = 73.5%, p = 0.981).
Regarding their forensic patient status, the overwhelming majority (98.2%) were found not guilty by reason of mental illness (NGMI), compared to 6 (1.8%) forensic patients with a limiting term: 4 (1.2%) in the high secure unit, and 2 (0.6%) in the prison setting.
With respect to the index offence, the most serious index offence (according to the Australian and New Zealand Standard Offence Classification, ANZSOC) and the first offence (if there was more than one offence over time) were considered. Male forensic patients were slightly younger, 34.7 years old (SD 11.04), than the female forensic patients, 35.3 years old (SD 8.96), at the time of the first offence but not significantly so (
t = 0.271,
p = 0.787). The majority of the offences involved offences committed against a person (
n = 305), these included: homicide and related offences; acts intended to cause injury; sexual assault and related offences; dangerous or negligent acts endangering persons; abduction, harassment and other offences against the person; and robbery, extortion and related offences [
34]. While globally there were no significant differences in broad offence categories type by gender, a more nuanced analysis of the 16 ANZSOC categories revealed that a significantly higher proportion of females engaged in homicide and related offences compared to males (χ
2 = 3.71,
p = 0.05).
Considering the sample as a whole, the most common primary diagnosis was schizophrenia (n = 234, 71.6%). Comparing gender however, revealed that significantly more male forensic patients were diagnosed with schizophrenia (n = 217, 74.1% v n = 17, 50.0%, χ2 = 8.67, p = 0.003) and significantly more female forensic patients were diagnosed with schizoaffective disorder (n = 12, 35.3% v n = 42, 14.3%, χ2 = 9.70, p = 0.002). The majority of forensic patients had a substance use disorder diagnosis (n = 240, 73.4%); moreover there was a significant association between substance use disorder and gender, with more males than females attracting this diagnosis (n = 222, 75.8% vs. n = 18, 52.9%, χ2 = 8.129, p = 0.004). Seventeen (5.2%) forensic patients were diagnosed as having, or maybe having, an intellectual disability. Proportionally more female forensic patients were identified as having an intellectual disability, but this did not reach statistical significance (Fisher’s Exact Test p = 0.656). One in five (n = 67, 20.5%) of the entire sample was diagnosed with a personality disorder, and there were no significant differences between the genders. The most common personality disorder type for both male and female forensic patients was antisocial personality disorder (Male n = 44, 15.0%, Female n = 6, 17.6%).
There were no differences in the average length of stay across the sites for males and females (Male = 2.18 years, Female = 2.2 years, p = 0.958).
DUNDRUM quartet results
DUNDRUM-1 across the sites
The mean DUNDRUM-1 score of patients in high security was significantly greater than all of the other sites. The mean DUNDRUM-1 score in the prison setting was significantly higher than all sites apart from high security. The mean DUNDRUM-1 score in medium security was significantly greater than the low and community mean scores, but not the open security sites. There was no significant difference between the low and open, and low and community security mean scores, but the mean DUNDRUM-1 score in open security was significantly higher than the mean score for community forensic patients (see Table
2).
Table 2
DUNDRUM Results Across the Sites by Gender
Prison | 2.53 | 0.83 | 2.79 | 0.68 | 2.48 | 0.70 | | | | | | |
High Security | 3.04 | 0.50 | 2.81 | 0.77 | 2.90 | 0.70 | 3.26 | 0.27 | 3.71 | 0.24 | 3.69 | 0.31 |
Medium Security | 2.10 | 0.46 | 1.89 | 0.80 | 1.70 | 0.64 | 2.15 | 0.39 | 1.85 | 0.76 | 1.88 | 0.55 |
Low Security | 1.51 | 0.45 | 1.70 | 1.29 | 1.52 | 1.04 | | | | | | |
Open Security | 1.74 | 0.48 | 1.34 | 0.60 | 1.45 | 0.52 | | | | | | |
Community | 1.29 | 0.40 | 1.50 | 0.94 | 1.25 | 0.71 | 1.31 | 0.47 | 1.68 | 1.08 | 1.30 | 0.79 |
For the female forensic patients, there were significant differences between all three levels of security (high, medium and community) for the mean DUNDRUM-1 scores. The mean scores for the females were higher in high security than medium security, and higher in medium security than in community services.
Cronbach’s alpha for the DUNDRUM-1 ratings was 0.843.
DUNDRUM-3 across the sites
For male forensic patients, there was no significant difference in mean DUNDRUM-3 scores between the prison and high security placed forensic patients, however both cohorts scored significantly higher (i.e., less programmes completed) than all other sites. There was no significant difference in mean DUNDRUM-3 scores between medium security, low security, open security, and community placed forensic patients, although there was a pattern of improving programme completion (as evidenced by decreasing mean scores as security levels decreased).
Considering the female forensic patients, the mean DUNDRUM-3 score for forensic patients was significantly higher in high security compared to medium security and community. There was no significant difference in programme completion between females in medium security and those in the community.
Cronbach’s alpha for the DUNDRUM-3 ratings was 0.897.
DUNDRUM-4 across the sites
There was no significant difference in the mean DUNDRUM-4 recovery scores between the prison and high security placed male forensic patients, however both cohorts were rated as having recovered significantly less (i.e. their mean DUNDRUM-4 score was higher) than all other sites. There was no significant difference in the mean DUNDRUM-4 scores between medium, low and open security, and no significant difference between the low, open and community, though community forensic patients were rates as being significantly more recovered than medium security forensic patients.
Similar to the DUNDRUM-1 scores, the female forensic patients mean DUNDRUM -4 scores varied significantly across all three levels of security (high, medium and community), with the mean score reducing between high security and medium security, and medium security and community.
Cronbach’s alpha for the DUNDRUM-4 ratings was 0.889.
CANFOR results
For the males, there was no significant difference in the mean total number of needs between high security, medium security, low security, open security, and the community. However, the male forensic patients in prison settings were rated as having significantly less needs than those in high security, medium security, and the community. The male forensic patients in high security had significantly more unmet needs than those in prison, medium security, open security, and the community, but there was no significant difference compared to the low security cohort (see Table
3).
Table 3
CANFOR Results Across the Sites by Gender
Prison | 11.28 | 3.11 | 8.84 | 3.25 | 2.44 | 1.73 | | | | | | |
High Security | 15.29 | 2.66 | 10.70 | 3.17 | 4.58 | 2.70 | 18.75 | 2.49 | 15.00 | 3.78 | 3.75 | 1.75 |
Medium Security | 14.56 | 2.87 | 12.56 | 2.77 | 2.00 | 1.38 | 14.69 | 3.40 | 13.08 | 3.15 | 1.62 | 1.56 |
Low Security | 13.50 | 3.63 | 10.75 | 2.61 | 2.75 | 1.91 | | | | | | |
Open Security | 13.23 | 3.01 | 11.32 | 2.55 | 1.91 | 0.92 | | | | | | |
Community | 14.95 | 4.42 | 12.41 | 4.22 | 2.53 | 2.53 | 12.44 | 4.36 | 9.89 | 4.65 | 2.56 | 2.30 |
Female forensic patients in high security had significantly more needs than those in medium security and the community. The female forensic patients in high security had significantly more unmet needs than those in medium security, but there was no significant difference compared to the community based female forensic patients.
Cronbach’s alpha for the CANFOR ratings was 0.504.
HCR-20 V3 results
The male forensic patients in the community had significantly lower H-scores than those in high and medium security, but otherwise there were no significant differences across levels of security. Male forensic patients in high security had significantly higher C-scores than patients in prison, medium security, open security, and the community, but there was no significant difference compared to the low security cohort. There was no significant difference in the mean C-scores between male forensic patients in prison, medium security, low security, open security, and the community. There were no significant differences across any of the sites with respect to R scores. For the mean Total-scores, male forensic patients in high security had significantly higher mean Total-scores than patients in prison, medium security, open security, and the community, but there was no significant difference compared to the low security cohort. There was no significant difference in the mean Total-scores between male forensic patients in prison, medium security, low security, open security, and the community (see Table
4).
Table 4
HCR-20 V3 Results Across the Sites by Gender
Prison | 13.16 | 3.82 | 3.24 | 2.85 | 3.20 | 1.96 | 19.60 | 6.70 | | | | | | | | |
High Security | 14.93 | 3.09 | 5.87 | 2.72 | 3.81 | 2.28 | 24.61 | 6.48 | 18.00 | 3.02 | 8.25 | 1.67 | 8.63 | 1.41 | 34.88 | 4.29 |
Medium Security | 14.49 | 3.38 | 3.49 | 2.42 | 3.22 | 2.12 | 21.20 | 5.81 | 13.92 | 3.86 | 4.23 | 1.74 | 3.92 | 1.93 | 22.08 | 5.51 |
Low Security | 14.38 | 2.07 | 3.50 | 3.55 | 2.38 | 2.33 | 20.25 | 4.74 | | | | | | | | |
Open Security | 13.77 | 3.93 | 2.73 | 2.03 | 2.95 | 1.40 | 19.45 | 5.04 | | | | | | | | |
Community | 12.16 | 3.22 | 2.76 | 1.97 | 3.48 | 2.34 | 18.40 | 5.72 | 11.33 | 3.57 | 2.22 | 1.99 | 3.11 | 2.03 | 16.67 | 6.12 |
Female forensic patients in high security had a significantly higher mean H-score than those in medium security and the community, but there was no significant difference between medium security and the community. There was a significant difference across all sites for the mean C-scores. The mean R-scores showed the same pattern as the mean H-scores, with the female forensic patients in high security having a significantly higher mean R-score than those in medium security and the community, but there was no significant difference between medium security and the community. There was a significant difference across all sites for the mean Total-scores.
Cronbach’s alpha for the HCR scores was calculated: H-scale - 0.656; C-scale - 0.834, R-scale - 0.741.