Jill Dawson,1 Kristina K Harris,2 Helen Doll,3 Ray Fitzpatrick,1 Andrew Carr2 1Nuffield Department of Population Health, 2Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, 3ICON Patient Reported Outcomes, Oxford, UK Background: To explore and compare the factor structure of the 12-item Oxford shoulder score (OSS) and 13-item shoulder pain and disability index (SPADI).Methods: Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) of data from 660 patients attending 46 hospitals in the UK. Complete OSS and SPADI data were available for 648 (98.2%) and 628 (95.2%) participants, respectively.Results: For both instruments, either one or two factors were indicated, depending on the extraction method. On EFA, most OSS items loaded saliently on either of two “Pain” (4 items) and “Function” (8 items) factors, although some items cross-loaded. Cronbach’s alphas were 0.75, 0.90, and 0.91 for “Pain” and “Function” subscales, and all 12 OSS items, respectively. CFA suggested marginally better fit for two factors, with neither one- nor two-factor models rejected. EFA indicated two factors for the SPADI, with three of the eight “Disability” items contributing to an 8-item “Pain factor”, with 2 items within the 5-item “Disability factor” cross-loading. Cronbach’s alpha was 0.87 and 0.93 for the original 5- and 8-item pain and disability scales; 0.94 for all 13 SPADI items, respectively. CFA suggested marginally better fit for the two-factor (original conceptualization) model of the SPADI, with neither one- nor two-factor models rejected. Conclusion: EFA and CFA demonstrated that, in addition to single summary scales usage, separate information on pain and self-reported disability/function can be extracted in a meaningful way, as subscales, from both the OSS and the SPADI. This information can help researchers in choosing primary study endpoints appropriately. Keywords: shoulder, Oxford shoulder score, shoulder pain and disability index, patient-reported outcome measures, factor analysis