The reliability of three depression rating scales in a general population of Dutch older persons

Abstract

Objective: To compare the reliability of three rating scales for assessing depressive symptoms in a community-based, non-clinically depressed older population.

Methods: The study sample comprised of 302 independently living subjects aged 65 years or older. Depressive symptoms were assessed using the centre for epidemiologic studies depression scale (CES-D), the geriatric depression scale (GDS-15) and the Montgomery and Asberg depression rating scale (MADRS) at three time points: at baseline, after 13 weeks (except the GDS-15) and after 26 weeks. Three dimensions of reliability were compared: (i) internal consistency (Cronbach's alpha), (ii) reproducibility (Spearman correlations) and (iii) the intra- and inter-rater reliability (Spearman correlations to compare the differences between correlations of subjects tested by the same vs. different raters at three time points).

Results: Cronbach's alpha was high for the CES-D (0.84), good for the MADRS (0.72) and relatively low for the GDS-15 (0.55). Reproducibility was also higher for the CES-D (0.71) than for the MADRS (0.61) and the GDS-15 (0.52). The rater had little influence on CES-D scores (intra/inter-rater ratio = 0.99). The GDS-15 and the MADRS, however, performed better when administered by the same rater.

Conclusions: The CES-D was the most reliable scale for measuring depressive symptoms in a non-clinically depressed older population.