OBJECTIVE: To evaluate the metrologic properties of composite disease activity indices in rheumatoid arthritis (RA), utilizing information derived from clinical, gray-scale (GS), and power Doppler (PD) ultrasound examinations, and to assess the classification of patients according to disease activity using such indices. METHODS: This ancillary study utilized data from a multicenter, prospective, randomized, parallel-group study conducted in subjects with moderate RA randomized to receive etanercept and methotrexate (ETN + MTX) or usual care (various disease-modifying antirheumatic drugs [DMARDs]). In multimodal indices, the 28 swollen joint count was either supplemented or replaced by clinically nonswollen joints in which the presence of synovitis was detected either by GS and/or PD and was calculated according to the Disease Activity Score in 28 joints (DAS28) or the Simplified Disease Activity Index (SDAI). Reliability, external validity, and discriminative capacity were calculated at baseline/screening by intraclass correlation coefficient, Pearson's correlation, and standardized response mean, respectively. RESULTS: Data from 62 patients (mean ± SD age 53.8 ± 13.2 years, mean ± SD disease duration 8.8 ± 7.7 years, mean ± SD disease activity 4.6 ± 0.5 [DAS28] and 20.9 ± 5.9 [SDAI]) were analyzed, with 32 receiving ETN + MTX and 30 receiving DMARDs. The metrologic properties were at least as good for GS- and/or PD-based indices as for their clinical counterparts. Using GS- and PD-supplemented indices, an additional 67.8% and 32.3% of patients (DAS28-derived and SDAI-derived indices, respectively) could be classified as having high disease activity at the screening visit. CONCLUSION: Multimodal indices incorporating ultrasound and clinical data had similar metrologic properties to their clinical counterparts; certain indices allowed for a significantly larger number of patients to be classified to either high or moderate disease activity at the screening visit.