Are GAF scores reliable in routine clinical use?

Acta Psychiatr Scand. 2007 Apr;115(4):326-30. doi: 10.1111/j.1600-0447.2006.00925.x.


Objective: The Global Assessment of Functioning (GAF) scale has been considered as a reliable tool. However, most studies of GAF reliability have been based on special conditions, such as prior training and test awareness. This study takes a different approach investigating inter-rater reliability of GAF scores in a routine clinical context.

Method: Eighty-two consecutively admitted acute psychiatric patients were rated according to routine department procedures and by two researchers at admission and discharge. Intraclass correlations (ICC) were computed using two-way mixed models.

Results: Inter-rater reliability between routine scores and research scores were low (ICC coefficients between r = 0.39 and 0.59). Inter-rater reliability between the two researchers were excellent (ICC coefficients r = 0.81 and 0.85).

Conclusion: Inter-rater reliability of GAF scores in a routine clinical context seems to be insufficient.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Hospitals, Psychiatric
  • Humans
  • Medical Staff, Hospital
  • Observer Variation*
  • Personality
  • Psychiatric Nursing
  • Psychiatric Status Rating Scales / standards*
  • Psychiatry
  • Reproducibility of Results
  • Students, Medical