AUTHOR=Jones Roland M , Patel Kiran , Moscovici Mario , McMaster Robert , Glancy Graham , Simpson Alexander I.F.
TITLE=Adaptation of the Clinical Global Impression for Use in Correctional Settings: The CGI-C
JOURNAL=Frontiers in Psychiatry
VOLUME=10
YEAR=2019
URL=https://www.frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2019.00687
DOI=10.3389/fpsyt.2019.00687
ISSN=1664-0640
ABSTRACT=Background: Provision of mental health care in correctional settings presents unique challenges. There is a need for a simple-to-use tool to measure severity of mental illness in correctional settings that can be used by mental health staff from different disciplines. We adapted the severity scale of the Clinical Global Impression for use in correctional settings, which we have called CGI-C, and carried out a reliability study.
Method: Clinical descriptions of typical inmate presentations were developed to benchmark each of the seven possible ratings of the CGI. Twenty-one case vignettes were then developed for study of inter-rater reliability, which were then rated using the CGI-C by five forensic psychiatrists (on three occasions) and 11 multidisciplinary health care clinicians (twice). The tool was introduced into clinical practice, and the first 57 joint assessments carried out by both a psychiatrist and a clinician in which a CGI-C was rated were compared to measure inter-rater reliability.
Results: We found very good inter-rater and test–retest reliability in all analyses. Gwet’s AC, calculated on initial ratings of the vignettes by the psychiatrists, was 0.85 (95% CI 0.81–0.90, p < 0.001) and 0.87 (95% CI 0.83–0.91, p < 0.001) for clinician ratings. Inter-rater reliability based on 57 joint face-to-face assessments of inmates showed Gwet’s AC coefficient of 0.93 (95% CI 0.88–0.97).
Conclusion: The CGI-C is simple to use, can be used by members of the multidisciplinary team, and shows high reliability. The advantage in correctional settings is that it can be used even with the most severely ill and behaviorally disturbed, based on observation and collateral information.