OBJECTIVES: To assess differences in the reliability of HUI Mark 3 (HUI3) health-related quality of life (HRQL) utility scores for patients between self and parent assessments across two Dana-Farber Cancer Institute (DFCI) clinical trials for treatment of ALL during childhood. METHODS: Patients were enrolled in either the DFCI 95-001 or 00-001trial, and were ≥12 years of age at the time of HUI survey. Patients and parents, blind to each other, completed HUI questionnaires at each of 5 trial phases: induction; CNS prophylaxis; intensification; continuation; and post-treatment. Reliability was assessed in terms of inter-rater agreement of individual scores and differences in mean scores. Agreement was quantified using the single-measure two-way mixed model intra-class correlation coefficient (ICC). An ICC of 0.41-0.60 represents moderate reliability, 0.61-0.80 good reliability, and 0.81-1.00 very good reliability. Mean differences of >0.03 are clinically important. Statistical significance was set at p<0.05. RESULTS: The number of patient and parent paired assessments varied by assessment phase for both the 95-001 (minimum=29, maximum=50) and the 00-001 (minimum=28, maximum=54) trials. ICCs in the two trials ranged from 0.49 (p<0.05) to 0.88 (p<0.05). There was substantial overlap of ICC 95% confidence bounds across the two trials at each of the five assessment phases. There was no significant difference (p>0.06) between patient-parent pairs of scores at any assessment phase in either trial.  The difference between trials in mean patient-parent scores was ≤0.03 and insignificant (p>0.08) for each of the 5 assessment phases. CONCLUSIONS: Agreement between patient and parent scores was moderate or better for all assessment phases in both trials. There were no important differences in mean patient and parent scores for any of the assessment phases of the two trials. Inter-rater reliability of scores was similar across the two trials.  Parental assessments provide acceptable and consistent estimates of HRQL for children.