The relationship of demographic, clinical, cognitive and personality variables to the discrepancy between self and clinician rated depression

Abstract

Background

The measurement of depression severity is an important aspect of both clinical and research practices. However, studies examining the self-report Beck Depression Inventory (BDI) and the clinician Hamilton Depression Rating Scale indicate only moderate correlations. The aim of this study was to examine the correlation between two self-report measures, the revised BDI, the Hopkins Symptom Checklist with the clinician rated, Montgomery–Asberg Depression Rating Scale. The secondary aim was to investigate patient factors which contribute to discordant ratings.

Methods

Depression severity and demographic, clinical, personality, cognitive, and personality factors that may contribute to a self-report-clinician rated discrepancy were examined in 177 adult outpatients with a Major Depressive Episode (DSM-IV) participating in a randomised clinical trial comparing CBT and IPT for depression. All assessment was conducted prior to treatment randomisation.

Results

Self-report and clinician rated depression were moderately correlated. Individuals with higher clinician rated depression severity, increased levels of rumination and females were more likely to have higher self-report rated depression (BDI-II and SCL-90) than clinician rated depression. In addition, younger patients and those with melancholic depression had higher BDI-II compared to MADRS scores.

Limitations

Results require replication.

Conclusions

Self-reported and observer rated depression were only moderately correlated. Researchers and clinicians interpreting the level of depression need to be cognizant of the patient factors that may contribute to either underreporting or overreporting self-report scores relative to observer ratings.

Introduction

Depression rating scales are used for a variety of purposes in community and clinical populations (Hamilton, 1976, Snaith, 1981). The most widely used self-report depression severity measure has been the Beck Depression Inventory (BDI) (Beck et al., 1961) which was revised (BDI-II) in 1996 (Beck et al., 1996). The Hamilton Depression Rating Scale (HDRS) (Hamilton, 1960) has been the most widely used clinician rated measure of depression severity, however, during the past decade the Montgomery–Asberg Depression Rating Scale (MADRS; Montgomery and Asberg, 1979) has been used with increasing frequency.

The concordance between self and clinician (observer) ratings of depression severity is variable. In a review, Bech (1992) reports acceptable correlations between self-report and observer total scores ranging from 0.6 to 0.9, however in another review, Moller (1991) reported poorer correlations ranging from 0.2 to 0.8.

The reasons for discordant rating are not clearly understood. It may be due to differences in how depression has been conceptualised, however, even between almost totally corresponding expert rated and self-rated scales concordance is only moderate (Carroll et al., 1981, Rush et al., 1987, Svanborg and Asberg, 2001). Other factors such as patient and observer biases could also contribute to the discrepancy. Severity of illness has been shown to affect the level of correlation, with self-report depression scores tending to be higher than clinician rated scores in individuals with mild to moderate depression (Bailey and Coppen, 1976, Domken et al., 1994, Moller, 1991, Rush et al., 1987). Differences in ratings have also been related to subtype of depression with individuals with nonedogenous depression appearing to rate their illness as much more severe than clinician ratings (Domken et al., 1994, Rush et al., 1987, White et al., 1984). The effect of age has been mixed (Enns et al., 2000, Lyness et al., 1995, Rush et al., 1987, Wallace and Pfohl, 1995). In depressed adolescents, higher correlations between self and observer are evident for girls (Enns et al., 2000). It also appears that self-report measures may be more prone than clinician rated scales to being influenced by maladaptive personality traits (Domken et al., 1994, Enns et al., 2000, Svanborg and Asberg, 2001) and cognitive variables (e.g. dysfunctional attitudes, low self-esteem) (Domken et al., 1994). Further a number of studies demonstrate an association between neurotic and immature defense styles and the extent of overall distress self-reported (Nickel and Egle, 2006).

To date research in this area has primarily focused on the original version of the BDI and on the HDRS. In this study we evaluated the relationship between the MADRS and the BDI-II and the depression subscale of the Hopkins Symptom Checklist (SCL-90, Derogatis et al., 1974). We also examined the relationship of demographic, clinical, cognitive and personality variables to any discrepancy between the self and clinician ratings.