Standardized rater training for the Hamilton Depression Rating Scale (HAMD-17) in psychiatric novices

Abstract 

Background: Despite the long and widespread use of the Hamilton Depression Rating Scale (HAMD), standardized reliability studies in inexperienced raters are not available. Methods: Rater training was carried using three videotaped interviews with depressed patients in 21 psychiatric novices who had negligible previous experience with the HAMD. Chance-corrected coefficients of rating agreement with expert standards (weighted κ, ICC) were computed for single items and the total score of the HAMD. Results: The results demonstrate sufficiently high interrater reliability (κ>0.60) for most of the HAMD items and the total score (ICC=0.57–0.73). Three standardized HAMD training sessions seem adequate to establish satisfactory agreement among psychiatric novices. Limitations: The sample of video-taped interviews and, hence, the generalizability of the results, was restricted. Conclusions: High inter-rater reliability of the HAMD justifies the use by clinically inexperienced researchers after standardized training.

Keywords:  Observer variation, Psychometrics, Reproducibility of results, Depression, Psychiatric status rating scales