Rasch analysis of the six items, self-reported Hamilton depression scale in Chinese patients with major depressive disorder.

OBJECTIVE

To analyze the psychometric properties of the adapted 6-item Hamilton Depression Rating Scale (HAMD-6) in Chinese patients with major depressive disorder (MDD) using the Rasch model.

METHODS

The HAMD-6 was administered to 279 patients with MDD. The Rasch analysis was conducted to evaluate the results in terms of unidimensionality, local dependency, item fit, reliability, orders of rating category, differential item functioning (DIF), and item-person targeting. Transformation from raw score to interval data was generated.

RESULTS

Rasch analyses showed the unidimensional construct of HAMD-6. Item2 (Self-esteem and guilt) and item4 (Psychomotor retardation) presented slight dependency (residual correlation = -0.34) and item1 (Depressed mood) presented suspicious misfit (INFIT and OUTFIT ZSTD > 2.00). The reliability was ideal (Cronbach's α = 0.91; person separation index = 2.24). The difficulties of response categories were monotonically increased, with no disorder was found. No DIF was found across gender, age and education. The HAMD-6 showed optimal discriminative ability for individuals with moderate levels of depression. A conversion table was established for raw score to interval data.

CONCLUSION

The HAMD-6 was confirmed acceptable psychometric properties by Rasch model in MDD patients. It provided more optimal discriminative ability for individuals with moderate depression. The conversion table could be used for more precise measurement. However, item1, item2 and item4 needed further validation.