Reliability and validity of the Thai version of the Rapid Mood Screener (RMS-T)
C. Som-On, C. Suradom, S. Suttajit, S. Kawilapat, M. SrisurapanontIntroduction
Bipolar disorder (BD) is frequently misdiagnosed as unipolar depressive disorders, resulting in adverse clinical consequences. The Rapid Mood Screener (RMS) was developed to improve BD detection. (McIntyre RS, et al. Curr Med Res Opin. 2021; 37, 135-144).
Objectives
To translate the RMS into Thai (RMS-T) and evaluate its psychometric properties in Thai adults with Bipolar I (BD-I) and Bipolar II (BD-II) disorders.
Methods
The RMS underwent forward translation, back-translation, and cultural adaptation. Outpatients with major depressive disorder (MDD; n = 111) and BD (n = 31; BD-I = 21, BD-II = 10) were enrolled. Psychometric properties included internal consistency, interrater reliability, content validity, and test–retest reliability (2-week interval in a sample of 30), and criterion validity against the Mini International Neuropsychiatric Interview, version 5 (MINI-5). Criterion validity of RMS-T versus the Mood Disorder Questionnaire (MDQ) was compared using receiver operating characteristics (ROC) area under the curve (AUC) analyses, and optimal cut-off were identified using Youden’s index.
Results
Internal consistency was acceptable (Cronbach’s alpha = 0.612, P < 0.001; McDonald’s Omega = 0.633, P < 0.001). Interrater reliability was fair to moderate (Cohen’s Kappa = 0.326, P < 0.001). Content validity indices ranged from 0.36 to 0.75 (P < 0.001). Image 1 shows the ROC curves differentiating BD from non-BD (n = 111), AUC (95% CI; P): RMS-T = 0.798 (0.720 to 0.876; P < 0.001), T-MDQ = 0.823 (0.731 to 0.915; P < 0.001), and AUC difference = -0.02530 (−0.108 to 0.0573; P = 0.549). Image 2 shows ROC curves differentiating BD-I from non-BD-I (n = 121), AUC (95% CI, P): RMS-T = 0.755 (0.657 to 0.854; P < 0.001), T-MDQ = 0.802 (0.700 to 0.904; P < 0.001); and AUC difference = -0.0468 (−0.144 to 0.0507; P = 0.347). For BD-II, AUC (95% CI, P) of the ROC curves of RMS-T was 0.786 (0.669 to 0.902; P < 0.001). Optimal RMS-T cut-offs were ≥4 for BD and BD-I (sensitivity 64.52% and 61.90%; specificity 77.48% and 73.55%) and ≥3 for BD-II (sensitivity 90.0%, specificity 53.79%): Youden’s indices of 42, 35.45, and 43.79, respectively. Positive predictive values were 44.44% (BD), 28.89% (BD-I), and 12.86% (BD-II); negative predictive values were 88.66%, 91.75%, and 98.61%. Test–retest reliability was strong (Spearman’s rho = 0.855, P < 0.001).
Image 1: Long description.