Abstract
BACKGROUND: The modified Fresno test is a questionnaire developed to evaluate Evidence-Based Practice among physical therapists. A previous study has indicated that a shorter version with 9 items would be more appropriate for Brazilian physical therapists. OBJECTIVES: To test the measurement properties of the modified Fresno test - Brazilian-Portuguese short version. METHODS: To analyze the reliability of the modified Fresno test - Brazilian-Portuguese short version by two raters (intra and inter-rater) in a sample of 133 physical therapists. The intraclass correlation coefficient (ICC(2,1)) and standard error of measurement (SEM) was used to assess the test-retest reliability. Internal consistency was tested using the Cronbach's alpha coefficient. An expert committee analyzed content validity using the content validity index (CVI). Confirmatory factor analysis (CFA) using the principal components method was used to assess construct validity. Responsiveness was estimated using effect size, and ceiling and floor effects were also investigated. RESULTS: Inter-rater and intra-rater reliability were, respectively: ICC= 0.93 (95 % CI 0.91, 0.95); rater 1 - ICC= 0.95 (95 % CI 0.94, 0.97); rater 2 - ICC= 0.98 (95 % CI 0.98, 0.99). The agreement was very good (values ≤ 5 %). Internal consistency was good for most instrument items (≥ 0.80). The CVI showed agreement among the expert committee members (0.96). The Cronbach's alpha coefficients calculated for the corrected item total showed values greater than 0.40. In the CFA, the "model 2" showed acceptable indices (≥ 0.90). Responsiveness was classified as very small. No ceiling and floor effects. CONCLUSION: The Fresno Modified Test - Brazilian-Portuguese short version has good to excellent reliability. CFA showed that the fit indices were adequate to be used in the population of interest.