Objective: To determine whether the sensitivity of the Hamilton Depression Rating Scale (HDRS) and Montgomery-Asberg Depression Rating Scale (MADRS) to treatment effects are comparable in geriatric antidepressant randomized controlled trials by developing and validating an equation that links between the two instruments.
Methods: Literature search for this meta-analysis was based on three sources: MEDLINE, a recent related meta-analysis, and experts in geriatric antidepressant trials. The search resulted in 11 relevant geriatric antidepressant trial studies that administered both instruments for symptom ratings. The authors used baseline ratings as a model-building sample and postrandomization ratings as a validation sample. HDRS scores were prorated into HDRS17, a 17-item HDRS, for analysis. The development and validation was based on a total number of 1,874 subjects.
Results: The correlations were high between baseline mean HDRS17 and MADRS ratings (r = 0.80; Fisher's z = 1.09, N = 25, p <0.0001) and between postrandomization ratings (r = 0.88, Fisher's z = 1.39, N = 65, p <0.0001). The following equation was derived: HDRS17 = -1.58 + 0.86 x MADRS. The difference between observed and estimated HDRS17 in a validation sample consisting of postrandomization follow-up means did not depend on magnitudes of HDRS17.
Conclusion: Although generalizability of findings into a broader population could be limited, and the authors could not assess concordance of changes of particular item constructs between HDRS and MADRS ratings, both ratings are comparable in assessing changes in overall depressive symptom severity in response to antidepressants in depressed elderly at aggregated group mean levels.