Description

Wolf et al reported a clinical prediction rule for a patient committing a violent crime after discharge from a secure psychiatric hospital. This can help to identify a patient who may benefit from careful follow-up. The authors are from the University of Oxford, Karolinska Institute, Orebro University and the Oxford Health NHS Foundation Trust.


FoVOx stands for Forensic Psychiatry and Violence tool Oxford.

 

Patient selection: inpatient of a secure psychiatric hospital

 

Parameters:

(1) sex

(2) age at discharge in years

(3) history of previous violent crime

(4) primary diagnosis at discharge

(5) current drug use disorder

(6) current alcohol use disorder

(7) personality disorder

(8) employment prior to admission

(9) number of previous inpatient episodes

(10) lifetime drug use disorder

(11) length of stay in forensic hospital in months

 

Parameter

Finding

Points

sex

male

0

 

female

-0.8407

age at discharge

 

-0.0299 * (age)

previous violent crime

no

0

 

yes

1.1682

primary diagnosis

schizophrenia spectrum

0

 

bipolar

0.5994

 

unipolar depression

0.2867

 

anxiety disorder

0.1142

 

other

0.304

drug use disorder

no

0

 

yes

-0.1188

alcohol use disorder

no

0

 

yes

0.2288

personality disorder

no

0

 

yes

0.3052

employment

no

0

 

yes

-0.578

previous inpatient episodes

< 5

0

 

>= 5

-0.4676

lifetime drug use

no

0

 

yes

0.7964

length of stay

< 12 months

0

 

>= 12 months

-0.4576

 

where:

• In Appendix Table 6 the entries for sex and age appear to be flipped, as compared to the hazard ratios shown in Table 2 of the article.

 

score =

= SUM(points for all of the parameters)

 

X = EXP(score)

 

probability of violent offense within 12 months =

= 1 - (0.9280^X)

 

probability of violent offense within 24 months =

= 1 - (0.8762^X)

 

Performance:

• The area under the ROC curve is 0.77.


To read more or access our algorithms and calculators, please log in or register.