Validação e calibração do Banco de Itens Fadiga do Patient- Reported Outcomes Measurement Information System (PROMIS®) versão para a língua portuguesa

Abstract

Introduction: In the past few years, the evaluation in health has gone through profound changes. The development of the Item Response Theory (IRT) allied to the technological evolution has been significantly contributing to its evolution. The system PROMIS® represents an advancement in this area and allows for practical and dynamical evaluations in the Patient Reported Outcomes (PRO). Objective: The purpose of this study was to validate the Item Bank Fadiga of PROMIS® and perform its calibration, obtaining the parameters of the translated version and culturally adapted for the Brazilian population. Method: The 82 items of the Bank were answered by 1002 people. We used the bifactor model and the optimized implementation of the parallel analysis to verify the assumptions of the item response theory. The gradual response model was used to calibrate the items. Differential item functioning was investigated for sex (male x female) and age group (<X> 60 years). Results: In total, 76 items were calibrated and presented satisfactory properties of validity and reliability. The discrimination parameter ranged from 1.13 to 2.67 and the difficulty parameter ranged from -2.29 to 4.43. Ten items with differential functioning were detected, but the change in the coefficients of determination of these items showed a small impact (R > 0.26) in the evaluations. Conclusion: We conclude that the Item Bank Fadiga of PROMIS® presented 76 items with satisfactory psychometric properties that confer quality and validity in Brazil.