Revisiting Depression Rating Scales: Analysis of a Randomized Trial

Background And Objectives: The Hamilton Depression Rating Scale (HDRS) and Montgomery-Åsberg Depression Rating Scale (MADRS) are both frequently employed to gauge the symptoms of major depressive disorder (MDD). Limited studies have attempted to compare these two scales. The purpose of our study was to compare the HDRS and MADRS scores of the study population as well as to calculate their bias and precision.

Methods: An open-label, randomized, three-arm trial was executed on 96 MDD patients. For 16 weeks, participants were put into three groups (in a 1:1:1 ratio) and administered vortioxetine (5-20 mg/day), escitalopram (10-20 mg/day), or vilazodone (20-40 mg/day). Vortioxetine and vilazodone were experimental drugs, while escitalopram acted as a control. Follow-up visits were scheduled at four-week intervals following the first visit. The per-protocol (PP) population had their HDRS and MADRS scores recorded at each visit. We contrasted the MADRS and HDRS scores through the Bland-Altman analysis and Taffé method. The former method generated the limit of agreement (LoA) plot. Moreover, the latter provided the bias, precision, and comparison plots. We prospectively registered our trial in the Clinical Trial Registry of India (CTRI) (2022/07/043808).

Results: We screened 134 individuals, and 109 (81.34%) were qualified. Ninety-six (88.07%) accomplished the 16-week study. The average age was 46.3 ± 6.2 years. The study population had baseline and final HDRS scores of 30.05 ± 1.52 and 14.29 ± 1.51, respectively. The mean MADRS scores at the first and last visits were 35.73 ± 1.47 and 18.08 ± 1.92, respectively. According to the LoA plot, the mean difference between HDRS and MADRS scores was 4.78 (95% CI: 2.61-6.95). As the bias plot indicates, MADRS scores had an estimated differential bias of 3.478 (95% CI: -2.119 to 9.074) and a proportional bias of 0.816 (95% CI: 0.649 to 0.982). The precision plot demonstrated that HDRS scores had lower standard deviation values and a narrower confidence interval than MADRS scores. This meant that the HDRS was more precise. The comparison plot showed that the regression line of the recalibrated MADRS coincided with that of HDRS. The MADRS was effectively recalibrated to remove bias.

Conclusion: After 16 weeks of therapy, we noticed substantial drops in HDRS and MADRS scores. The HDRS was proved to be more precise than the analogous MADRS. The MADRS was recalibrated to remove differential and proportional biases.