Fewer study participants needed to demonstrate superior antidepressant efficacy when using the Hamilton melancholia subscale (HAM-D6) as outcome measure

BACKGROUND: In the development of new antidepressant treatments, the failed study has unfortunately become a prevalent problem. The number of failed studies could probably be reduced significantly by applying more informative outcome measures. Previous studies have indicated that the 6-item melancholia subscale (HAM-D6) of the 17-item Hamilton Depression Rating Scale (HAM-D17) may be more informative than other scales, due to its superior psychometric properties. In the present study we investigated whether the HAM-D6 had higher informativeness than the HAM-D17 based on data from a randomized placebo-controlled trial (RCT) testing the effect of erythropoietin (EPO) as augmentation therapy in patients with treatment-resistant depression.

METHODS: We assessed the scalability (Mokken analysis of unidimensionality), responsiveness (item responsiveness analysis) and ability to show drug-placebo separation (estimation of sample size needed to detect statistically significant difference between EPO and placebo) of the HAM-D6 and the HAM-D17.

RESULTS: The HAM-D6 demonstrated higher scalability, higher responsiveness, and better drug-placebo separation compared to the HAM-D17. As a consequence, only 39 participants per group would be required to detect a statistically significant difference between EPO and placebo when using the HAM-D6 as outcome measure, whereas the required group size for HAM-D17 would be 146 participants.

LIMITATIONS: The EPO RCT was not originally designed to investigate the research questions addressed in this study.

CONCLUSIONS: Both for ethical and financial reasons it is of interest to minimize the number of participants in clinical trials. Therefore, we suggest employing the HAM-D6 as outcome measure in clinical trials of depression.