Hedlund and Vieweg summarized over 30 studies that reviewed the HRSD as an indicator of change in clinical trials; they commented: “The HRSD is consistently reported to reflect clinically observed treatment changes . . . the HRSD and BDI tend to be about equally sensitive to severity of illness, while the SDS tends to be somewhat less reliable and sensitive” (12, p156). A meta-analysis of 36 studies compared effect sizes for the HRSD, the Zung Self-rating Depression Scale, and the BDI (38). A second review compared effect sizes for BDI and HRSD in 19 studies (39). In both, the HRSD was found to provide the largest consistent index of change. Compared with other clinician ratings, however, the HRSD may not be superior. Change scores on the HRSD were compared with a clinician’s classification of 35 patients under treatment into improved or not improved; the point biserial correlation was 0.59, comparing unfavorably with a figure of 0.70 for the MADRS (19, p385). Maier et al. have compared the discriminatory power of the HRSD and MADRS in two studies; in one the HRSD was superior (15, p8) and in the other the MADRS proved superior (17). Eight of the HRSD items were not sensitive to change. Montgomery and Åsberg found correlations between their MADRS and a clinical judgment of improvement (0.70) to be superior to that of the Hamilton scale (0.59) (19, p385).