OBJECTIVES: To assess inter-rater reliability and discriminative validity of Brazilian Portuguese HUI questionnaires for a sample of survivors of childhood cancer. METHODS: A sample of consecutive patients attending the long-term follow-up clinic at Centro de Treatmento e Pesquisa Hospital do Cancer in São Paulo was recruited. Self- and proxy-assessment versions of self-complete, one-week health-status recall HUI questionnaires were used. A questionnaire was completed independently by each patient (self-assessment), and nurse and physician (proxy-assessments). HUI single-attribute utility scores (n=14) and health-related quality of life (HRQL) scores (n=2) were derived using standard coding algorithms and published utility functions. Single measure, one-way random effects models were used to calculate intra-class correlations (ICC) of inter-rater agreement. Analysis of variance (ANOVA) assessed the statistical significance (p<0.05) of differences in mean utility scores for patients in various diagnostic groups (n=15). RESULTS: HUI data for 138 cancer survivors (45.7% female) were collected from the three types of assessors. Patient mean age at diagnosis was 6.3 (min=0.2, max=17.4) years and 22.8 (13.4, 40.2) years at survey. There was substantial to almost perfect agreement (ICC >0.85, p<0.001) between all pairs of raters for all types of utility scores. ANOVA detected statistically significant (p<0.013) differences among the diagnostic groups, in means of HUI2 and HUI3 HRQL utility scores, HUI3 vision (p<0.0005), HUI3 ambulation (p=0.002), HUI2 mobility (p=0.001) and HUI2 self-care (p<0.045). CONCLUSIONS: There was substantial or better agreement in all single-attribute and HRQL utility scores between pairs of patient, nurse and physician assessors. Differences in mean HRQL among diagnostic groups is evidence of discriminative validity. The Brazilian Portuguese HUI questionnaires should be considered for future PRO-studies of morbidity and HRQL in survivors of cancer in childhood in Brazil. Future research on the measurement properties of the Brazilian Portuguese HUI questionnaires should focus on concurrent validity and other health problems.