Page 47 - Contributed Paper Session (CPS) - Volume 7

CPS2027 Olayan A. et al.

Table 2: Developing the risk scoring algorithm for reoffending (overall):
Development Dataset (N= 5,162, 67%) Validation Dataset (N=2,581, 33%)
Characteristics

% Adjusted HR p-value βX10 Score % Adjusted HR p-value βX10 Score
Group
Diverted 26% 1 0.0 0 26% 1 0.0 0
Punitive 74% 1.23 (1.11 – 1.36) <0.001 2.1 2 74% 1.37 (1.18 – 1.59) <0.001 3.1 3
Age at the first offence
<20 years 4% 1.70 (1.38 – 2.10) <0.001 5.3 5 4% 1.62 (1.23 – 2.16) 0.001 4.8 5
20-29 years 31% 1.52 (1.37 – 1.69) <0.001 4.2 4 31% 1.60 (1.38 – 1.86) <0.001 4.7 5
30-39 years 33% 1.42 (1.29 – 1.59) <0.001 3.5 4 33% 1.31(1.13 – 1.52) <0.001 2.7 3
40+ years 32% 1 0.0 0 32% 1 0.0 0
Indigenous status
No 90% 1 0.0 0 91% 1 0.0 0
Yes 10% 2.09 (1.87 – 2.34) <0.001 7.4 7 9% 2.00 (1.70 – 2.36) <0.001 6.9 7
Psychosis type
Affective psychosis 10% 1 0.0 0 11% 1 0.0 0
Schizophrenia and related psychoses 69% 1.16 (1.01 – 1.36) 0.041 1.5 2 69% 1.24 (1.01 – 1.52) 0.040 2.2 2
Substance related psychosis 21% 1.19 (1.01 – 1.40) 0.043 1.7 2 20% 1.29 (1.03 – 1.62) 0.028 2.5 3
Contact with mental health services
No 17% 2.50 (2.27 – 2.74) <0.001 9.1 9 16% 2.40 (2.09 – 2.75) <0.001 8.8 9
Yes 83% 1 0.0 0 84% 1 0.0 0
First offence type
Violent 43% 1 0.0 0 42% 1 0.0 0
Non-violent 57% 1.21 (1.11 – 1.31) <0.001 1.91 2 58% 1.22 (1.09 – 1.38) 0.001 2.0 2

34 | I S I W S C 2 0 1 9