Standardized rater training for the Hamilton Depression Rating Scale (HAMD-17) in psychiatric novices.
Müller MJ and Dragicevic A;
Commented by , 21 Nov 2003
Hamilton Depression Rating Scale is one of the most widely used scales in clinical practice when the effect of treatment on a depressive episode is to be evaluated. Accordingly, new doctors and other novices within psychiatry are worldwide taught how to use the scale in a valid fashion. Surprisingly, it has never been investigated what it takes to obtain a sufficient degree of reliability for psychiatric novices.
A total of 21 psychiatric novices with negligible prior experience with the Hamilton Depression Rating Scale (HAMD-17) rated three videotaped interviews of depressed patients with the HAMD. Rating agreement with expert standards were estimated for single items and the total score of HAMD. According to historical suggestions by Hamilton (1967), items within the range of the expert rating +/- 1 are acceptable and the total score of HAMD-17 should be within the range of the expert rating +/- 4 points.
The proportion of ratings of single items within the range of the expert standards +/- 1 increased from 88.8 +/- 10.3 % after the first training session, to 96.1 +/- 5.0 % after the second and 94.7 +/- 5.2 % after the third training session. The proportion of participants with ratings in the range of the expert standards increased from 33 % at the first session to 100 % at the third session.
At the first rating five items showed Cohens weighted coefficient kappa < 0.60 and at the second and third session only one out of the 17 items yielded values below 0.60. The intraclass coefficients were 0.57, 0.65 and 0.73 for the total HAMD score at the three sessions, respectively.
The study shows that three standardised training sessions with videotaped interviews of depressed patients is sufficient to obtain satisfactory agreement among psychiatric novices. The results need confirmation in other settings with a larger number of patients and raters.
Nevertheless, the study confirms that HAMD-17 is an easy and economical tool and the finding implies that HAMD-17 should be used widely in the evaluation of depressive symptoms in clinical practice and in research.