Differentiating moderate and severe depression using the Montgomery–Åsberg depression rating scale (MADRS)

Abstract

Background: MADRS cut-off scores for moderate and severe depression were estimated in relation to the Hamilton Depression Rating Scale (HAMD₁₇) and the Clinical Global Impressions Scale (CGI). Method: HAMD₁₇, MADRS, and CGI ratings from patients with major depression (DSM-IV) were analyzed (N=85). Receiver operating characteristics (ROC) curves were applied. Results: Mean age was 51.4±14.5 years, 69% were female. Mean MADRS scores were 23.4±13.2, HAMD₁₇, MADRS, and CGI scores were highly correlated (r>0.85; P<0.0001). Best separation between moderate and severe depression according to CGI criteria was achieved with a MADRS score of 31 (sensitivity 93.5%, specificity 83.3%). Limitations: Studies to validate severity gradations including DSM-IV or ICD-10 diagnostic severity categories are recommended. Conclusions: Empirically based MADRS cut-off scores to separate moderate from severe depression on the basis of HAMD₁₇ and CGI severity ratings in patients with major depression were yielded.

Introduction

The Hamilton Depression Rating Scale (HAMD₁₇) (Hamilton, 1960) and the Montgomery–Åsberg Depression Rating Scale (MADRS) (Montgomery and Åsberg, 1979) are the world-wide most extensively used observer instruments in clinical and psychopharmacological depression research to assess severity of depression after a categorical diagnosis has been ascertained (Guelfi, 1988; Maier et al., 1988). Depression seems to cover multiple facets (Gibbons et al., 1983; Snaith, 1993), and the multi-dimensional view of severity of depression has several advantages compared to more or less arbitrary categorical subdivisions.

Nevertheless, research practice and clinical reality make it necessary to agree on cut-off values of the HAMD₁₇ and the MADRS to distinguish patients with mild, moderate, severe, and extremely severe depression. That concept is widely used and has several implications for treatment and prognosis. Consequently, a reliable and clinically acceptable severity gradation should be based on data from empirical studies. However, only rare data exist regarding the severity grading of the MADRS (Kearns et al., 1982; Snaith et al., 1986; Mittmann et al., 1997; Müller et al., 2000; Senra Rivera et al., 2000). More specifically, no empirically based MADRS cut-off scores to separate ‘moderate’ from ‘severe’ depression were available so far. In an independent previous study (Müller et al., 2000), HAMD₁₇ and MADRS ratings of 40 clinically at least moderately depressed patients with a diagnosis of major depression (according to DSM-III-R criteria) were investigated by means of linear regression analysis. Categorical MADRS severity categories (‘moderate’ vs. ‘severe’) were established by correlation with conventional gradations of the HAMD₁₇. Accepting an HAMD₁₇ score of ≥28 as cut-off value for severe depression (Angst et al., 1985), a corresponding cut-off value of 35 points on the MADRS was calculated. This study (Müller et al., 2000) had limitations as the sample size was restricted, and no ‘direct’ assessments of depression severity, e.g. diagnostic severity categories as used in the DSM or ICD manual, or direct scorings of depression severity as the Clinical Global Impression scale (CGI) (CIPS, 1998) were utilized.

Thus, the present study was conducted to assess depression severity in patients with major depression (DSM-IV) using the HAMD₁₇, MADRS, and the CGI, to establish MADRS cut-off scores for moderate vs. severe depression in relation to widely used HAMD₁₇ severity definitions (25 or 28 points) as a cross-validation of the previous study (Müller et al., 2000); and to find MADRS and HAMD₁₇ severity grades in relation to the more ‘direct’ severity rating using the CGI severity item.