The actual process of rating the global assessment of functioning scale

• Published in: Comprehensive psychiatry 2001-09, Vol.42 (5), p.403-409
• Main Authors: Yamauchi, Keita, Ono, Yutaka, Baba, Kunihiro, Ikegami, Naoki
• Format: Article
• Language: English
• Subjects: Adult; Aged; Biological and medical sciences; Female; Health staff related problems. Vocational training; Humans; Japan; Length of Stay - statistics & numerical data; Male; Medical sciences; Mental Disorders - diagnosis; Mental Disorders - epidemiology; Mental Disorders - psychology; Middle Aged; Personality Assessment - statistics & numerical data; Psychiatric Status Rating Scales - statistics & numerical data; Psychology. Psychoanalysis. Psychiatry; Psychometrics; Psychopathology. Psychiatry; Reproducibility of Results; Social Adjustment; Social psychiatry. Ethnopsychiatry
• ISSN: 0010-440X; 1532-8384
• DOI: 10.1053/comp.2001.26268

The Global Assessment of Functioning Scale (GAF) was developed for the overall assessment of psychological, social, and occupational functioning. While the advantage of the GAF lies in this comprehensiveness, questions have been raised on whether clinicians rate appropriately. To clarify this issue, the actual process of how clinicians assign GAF scores was investigated. A total of 2,462 inpatients of 19 psychiatric hospitals in Japan were assessed by their primary psychiatrists using the following rating scales: GAF, Brief Psychiatric Rating Scale (BPRS), World Health Organization Psychiatric Disability Assessment Schedule (DAS), and physical Activities of Daily Living (ADL) index. A tree-based model analysis (also referred to as Automatic Interaction Detector [AID] or Classification and Regression Tree [CART]) was used to construct a statistical model with the GAF score as the dependent variable. The statistically best-fitted tree to predict the GAF score is as follows. The first split is based on the “conceptual disorganization” score in the BPRS, followed by splits based on DAS item scores, such as “conversation” and “underactivity.” The tree model obtained suggests that Japanese clinicians judge the level of global functioning by integrating the information on both the severity of psychiatric symptoms and the level of impaired behaviour and social functioning. This logic structure was clinically acceptable and agreed well with the concept of the GAF.