A standardized and reliable method to apply the Global Assessment of Functioning (GAF) scale to psychiatric case records

The Global Assessment of Functioning (GAF) scale is widely used both in routine clinical practice and in research. However, its reliability has never been assessed when used to rate information in clinical records. The present study focuses on the development of a standardized method (an ongoing modelling process between raters) for establishing desired levels of inter-rater reliability (IRR) in the application of the GAF to psychiatric case records. Fifty-one patients at first-ever contact with mental health services were included in the study. They were selected from a total sample of 662 first-ever patients by using a systematic sampling. Three raters (resident psychiatrists at their third year of training) took part in a 12-hour training programme, during which they were asked to assess the global psychological functioning of patients, taking into account information recorded in case records. The extent of agreement between raters was estimated by applying the ‘limits of agreement’ method and the ‘concordance correlation coefficient’. The training programme proved to be feasible, easy to administer and acceptable to psychiatrists in training with limited previous experience of using rating scales. Very high levels of concordance (all greater than 0.95) emerged between the three raters. The GAF, completed using information from case records included in the initial assessment form, appeared to be a reliable instrument, even when used by clinical psychiatrists in training. Copyright © 2000 Whurr Publishers Ltd.