Predicting discharge from long-term forensic treatment: patients characteristics, protective factors, needs and treatment-related factors study in the Czechia

ABSTRACT

One of the challenges in forensic psychiatry is determining when an inpatient is ready to be discharged and return to the community. The comprehension of factors that predict extended treatment or discharge is relatively limited. We assessed the treatment progress of a cohort of forensic inpatients (N=80) divided into two groups: discharged patients and patients who remain detained. We derived socio-demographic and clinical variables from each patient’s medical records and scores on the HoNOS-Secure, GAF, and SAPROF scales. The dataset was subjected to logistic regression and Chi-square analysis to determine the relevant factors. We gained insights into illness as a strong predictor of discharge, which is also associated with the patient’s general compliance with the facility program and participation in occupational therapy. The majority of our sample has moderate or severe functional impairment according to GAF. The instruments used can capture dynamic factors related to discharge or continuing hospitalization, namely the SAPROF total or external factors score, the HoNOS-Secure subscale, and significant items from the HCR-20 clinical and risk subscales.

KEYWORDS:

• Long-stay forensic patient
• prolonged detention
• SAPROF
• HoNOS-Secure
• GAF

Acknowledgments

Deinstitutionalization of Mentally Ill Project, reg. No.: CZ.03.2.63/0.0/0.0/15_039/0006213, Project No. KA 5, Support For the New Mentally Ill Services Development, Reg. No.: CZ.03.2.63/0.0/0.0/15_039/0008217, Grant: Progres Q 06 1LF

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Unless otherwise specified.

2. CI estimates for this model are based on 3141 bootstrapped samples

3. CI estimates for this model are based on 3141 bootstrapped samples

4. CI estimates for this model were based on 4779 bootstrapped samples

5. CI for this model were based on 1000 bootstrapped samples