Validation de l’Échelle d’Humeur Dépressive (EHD), version autoquestionnaire, dans la sclérose en plaques

Abstract

Depressive Mood Scale (EHD) aims at assessing the various depressive mood dimensions as "blunted affect" and "lack of emotional control". It is an 18 items hetero-evaluation scale. The aim of this study was the validation of an EHD self questionnaire version. Self questionnaire items were generated from genuine scale items. As in the former version, response format was a Lickert 5 point scale. This validation study was carried out on 77 Multiple Sclerosis (MS) patients. Mood disorders are frequent during the course of MS and might be triggered or worsened by immuno-modulation therapies. Principal Component Analysis (ACP) with Varimax rotation revealed a two factors structure. The first one, corresponding to a "blunted affect" dimension, explained 33.5% of the scale variance and was composed of 7 items. The second one, corresponding to a "lack of emotional control" dimension, explained 20% of total scale variance and was composed of 4 items. The questionnaire internal coherence coefficient (Cronbach alpha) was excellent (=0.87) and the two sub-scales ones were satisfactory [0.89 for "blunted affect" dimension and 0.71 for "lack of emotional control" dimension. The questionnaire's external validity was confirmed by a positive correlation between "lack of control" sub-score and state sub-score of the Stait-Trait Anger eXpression Inventory (STAXI)] (r=0.55, p<0.01). Moreover we found a positive correlation between the total EHD autoquestionnaire score and both sub-scores on the one hand, and the Beck Depression Inventory score on the second hand (EHD/BDI: r=0.76, p<0.01; "lack of emotional control"/BDI: r=0.68, p<0.01; "blunted affect"/BDI: r=0.63, p<0.01). Test-retest reliability was good with a positive correlation between all the initial scores and their retests, a week later. Secondarily, a structural equation modeling analysis confirmed the two-factors structure model suggested by ACP. Various indicators showed a good fit between theoretical variance-covariance matrix and the observed one (chi(2)=41.55, p=0.49, ddl=42, Goodness Fit Index GFI=0.91, Root Mean Square Residual RMSEA=0.00). Thus, we proposed a well validated self questionnaire that allows the assessment of "blunted affect" and "lack of emotional control". It should be challenging to correlate those dimensions with neuro-psycho-logical testing and neuro-imagery, in patients affected by CNS diseases. Moreover, the assessment of those dimensions during interferon treatment in MS could allow a more precise evaluation of the emotional changes potentially induced by immuno-modulatory treatments.