The EQ-5D-Y-3L and the EQ-5D-Y-5L are friendly-child versions of the EQ-5D instruments that measure health-related quality of life in children and adolescents (kids) aged 8-15 years old. However, both instruments' measurement properties have not yet been tested in Brazilian kids yet.

This study aimed to test the EQ-5D-Y-3L and EQ-5D-Y-5L measurement properties in Brazilian kids with disabling musculoskeletal pain.

This is a measurement proprieties study with two periods of measures was conducted in 181 Brazilian kids with disabling musculoskeletal pain (i.e., who reported pain in the back, neck, arm, or legs that lead to school absenteeism and/or interference with normal and/or recreational activities) from public and private schools in Sao Paulo state. Kids answered the self-reported versions of the EQ-5D-Y-3L and the EQ-5D-Y-5L. We tested test-retest reliability using the Kappa coefficient for the descriptive system and intraclass correlation coefficients (ICC) for EQ-VAS. We tested construct validity (classified as sufficient if at least 75% of the results were in accordance with our pre-specified hypothesis) using the Pediatric Quality of Life Inventory questionnaire version 4.0 (PedsQL) and the Child Health Utility 9D (CHU9D). We also tested the ceiling and floor effects of the instruments using the dimensions’ descriptive system and health profile and the feasibility by the missing responses.

Most kids with musculoskeletal pain were female (61%) with a mean age of 12 years old (standard deviation: 3). In the descriptive system, reliability ranged from 0.32 to 0.47 for the EQ-5D-Y-3L and 0.20 to 0.49 for the EQ-5D-Y-5L. There was substantial reliability for the EQ-VAS (ICC: 0.80; 95% CI: 0.71, 0.86). Construct validity was sufficient for the EQ-5D-Y-3L and the EQ-5D-Y-5L compared to the PedsQL, sufficient for the EQ-5D-Y-5L and insufficient for the EQ-5D-Y-3L compared to the CHU9D (89%, 100%, 81%, and 47% in accordance with the hypothesis, respectively). There was as lower ceiling effect of the EQ-5D-Y-5L compared to the EQ-5D-Y-3L for all the dimensions of the descriptive system, except for the ‘having pain or discomfort’, while the health profile (11111) was 18.2% for the EQ-5D-Y-3L and 16% for the EQ-5D-Y-5L. The missing response rate ranged from 1.3% for the EQ-5D-Y-3L and 4% for the EQ-5D-Y-5L.

The descriptive system of the EQ-5D-Y-3L and the EQ-5D-Y-5L presented inadequate reliability and the EQ-VAS presented substantial reliability, but both instruments presented sufficient construct validity, except the EQ-5D-Y-3L compared to the CHU9D. Furthermore, the EQ-5D-Y-5L had lower ceiling effects compared to the EQ-5D-Y-3L and both instruments had good feasibility.

This study tested the measurement properties of the EQ-5D-Y-3L and the EQ-5D-Y-5L in Brazilian kids with disabling musculoskeletal pain. The results of this study could help clinicians to measure health-related quality of life in the youth population. Furthermore, the EQ-5D-Y-3L may facilitate the calculation of the quality-adjusted life of years in economic evaluations conducted in Brazil in the future.