OBJECTIVES:  To inform the design of comparative effectiveness studies (CER), a head-to-head psychometric comparison of the PROMIS Global and EQ-5D instruments is needed.  METHODS: In 2013, 2289 US adults completed an online survey including 10 items from the PROMIS Global and 5 items from the 3 EQ-5D versions (EQ-5D-3L, EQ-5D-Y, and EQ-5D-5L) in random order. After testing for unidimensionality between each pair of these 25 items, we conducted 3 separate exploratory factor analyses (EFA) between the PROMIS Global and each EQ-5D version. Next, we performed an item-response-theory (IRT) analysis for factors shared between the 2 instruments. Item levels with insufficient responses (n<40) were collapsed. The relative degree of information obtained by each instrument was assessed using an unconstrained graded response model.  RESULTS: All items were positively correlated (0.32 to 0.98). Regardless of EQ-5D version, EFA analyses identified 3 factors (eigenvalue > 1): Physical Health (PH), Mental Health (MH), and Quality of Life (QL). Each item uniquely loaded to a single factor after rotation. Unlike PH and MH, QL included only PROMIS items. At a threshold of 0.5 in standard error, the IRT analyses showed similar PH information function ranges by instrument (-0.2 to 2.8 for EQ-5D-5L vs. -0.2 to 2.4 for PROMIS Global). However, the MH information range for EQ-5D-5L was substantially narrower than the range for the PROMIS Global (0.3 to 2.4 vs. -0.8 to 2.8). CONCLUSIONS:  The PROMIS Global includes 5 items that extend the measurement of general health beyond the 2 factors shared with the EQ-5D. When comparing the remaining 5 items of the PROMIS Global to the EQ-5D, the instruments appear to share information ranges in PH, but the PROMIS Global has a broader MH range than the EQ-5D. These similarities and differences are important considerations when choosing between the PROMIS Global and EQ-5D for CER.