Published March 14, 2019 | Version v1
Journal article Open

Evaluation and improvement of the psychometric properties of the Hamilton Rating Scale for Depression using Rasch analysis for applying in Belorusian Population

Authors/Creators

  • 1. Grodno State Medical University

Description

Introduction:  The  Hamilton  Rating  Scale  for  Depression  (HAM-D)  has  not  been  tested  for  the psychometric  properties  in  Belarusian  population.  In  this  regard  conducting  psychometric assessment of the HAM-D used in Belarusian population has special importance.  The aim of this study  was  to  investigate  and  improve  the  psychometric  properties  of  the  HAM-D  applying  in Belarusian population through Rasch analysis. 

Methods: Data from 551 Belarusian patients with confirmed diagnosis of depression were used in 
this study. The Partial Credit Model was used in the psychometric analysis. 

Results:  The  analysis  of  item-person  map  showed  disagreement  between  distributions  of  person  measures  and  item  categories.  Only  8  items  had  appropriate  values  of  Rasch  fit-indexes.  I  have improved  the  psychometric  properties  of  the  HAM-D.  Five  items  with  bad  values  of  fit-indexes were  recombined  into  two  new  items.  New  items  were  added  to  the  remaining  8  items  with acceptable  fit-indexes  values  and  then  Rasch  analysis  was  repeated.  All  items  showed  adequate 
construct  validity.  The  person  separation  index  (PSI)  for  original  HAM-D  was  2.32.  The  person reliability  was  0.84.  Reliability  analysis  of  the  final  model  included  10  items  showed  higher estimates: the PSI was 3.61, the number of strata rose to 5.14, the person reliability was 0.93. 
Conclusion:  This  study  revealed  that  the  HAM-D  applied  in  Belarusian  population  has  several 
psychometric  problems.  After  removing  invalid  items  and  changing  score  rules  of  bipolar  items 
psychometric properties of the scale were improved.

Files

AssanovichEvaluation2017статьяВАК.pdf

Files (769.4 kB)

Name Size Download all
md5:62c0ef605b0d5ae9199f9093c15255be
769.4 kB Preview Download