RESUMO
PURPOSE: To carry out the translation, cultural adaptation, and validation in the Brazilian Portuguese version of Knee Osteoarthritis Outcome Score (KOOS-BR) in individuals with knee osteoarthritis (KOA). METHODS: Patients with KOA (n = 136) answered KOOS-BR, Short-Form Health Survey (SF-36) and Numerical Pain Scale (NPS) to access construct validity. KOOS-BR was answered again with an interval of five to eight days. The following were performed: Cronbach's alpha to assess internal consistency, intraclass correction coefficient (ICC2,1) to assess reproducibility, standard error of measurement (SEM) and minimal detectable change (MDC) as error measurements. Dimensionality was tested through confirmatory factor analysis (CFA). Responsiveness was investigated by the correlation between KOOS-BR subscales with global perceived effect (GPE) before and after physical therapy treatment for 6 weeks. Floor and ceiling effects (< 15%) were also assessed. RESULTS: KOOS-BR version showed good test-retest reliability in all KOOS-BR subscales (ICC = 0.77-0.84), with proper internal consistency (α = 0.71-0.94). KOOS-BR showed a moderate direct correlation with physical health domains of SF-36 (r = 0.39-0.68; p < 0.001) and a moderate inverse correlation with pain intensity (r = - 0.51 to - 0.57; p < 0.001). KOOS-BR proved to be responsive, the correlation between the KOOS-BR subscales and GPE before and after treatment ranged from 0.42 to 0.60 (p < 0.001). There was no ceiling and floor effect. CONCLUSION: KOOS-BR is reliable, valid, and responsive in patients with KOA. LEVEL OF EVIDENCE: II.