Validation of the Standardized Universal Pain Evaluations for Rheumatology Providers for Children and Youth (SUPER-KIDZ)
Abstract• STUDY DESIGN:
Longitudinal observational clinimetric study with repeated measures.• BACKGROUND:
No validated multidimensional pain measure for children and youth with juvenile idiopathic arthritis exists.• OBJECTIVE:
To determine the test-retest reliability, construct validity, and responsiveness of English and French versions of the Standardized Universal Pain Evaluations for Rheumatology Providers for Children and Youth (SUPER-KIDZ).• METHODS:
Measurement properties of the SUPER-KIDZ (older child, younger child, and parent versions) were prospectively evaluated in patients (aged 4 to 18 years) with juvenile idiopathic arthritis at 2 centers. Internal consistency of the 3 subscales was measured using ordinal reliability alpha. Test-retest reliability for each subscale was evaluated with intraclass correlation coefficients (ICCs) from participants assumed to have stable pain (over a 1-week period with no change in treatment). Correlations of SUPER-KIDZ scores with validated measures determined construct validity. Responsiveness of SUPER-KIDZ subscales was evaluated in patients with improvement in pain, using standardized response mean and linear mixed-model regression.• RESULTS:
Seventy-one children aged 8 to 18 years and 29 parent-child dyads aged 4 to 7 years were included. Seventy-four percent of participants were female, with a median of 3 active joints (interquartile range, 1–5). Internal consistency was strong (α = .78-.96) for pain characteristics, interference, and emotional functioning SUPER-KIDZ subscales. Good test-retest reliability (ICC≥0.80) was found for the pain characteristics subscale in older- and younger-child versions. Most other subscales had satisfactory reliability coefficients (ICC≥0.70). Correlations of 0.50 or greater were found between the older-child SUPER-KIDZ scores and the Childhood Health Assessment Questionnaire and Patient-Reported Outcomes Measurement Information System depressive symptoms items, as well as the younger-child pain-intensity item and the Faces Pain Scale-Revised. Strong responsiveness was found for all subscales (standardized response mean, 0.63–1.54; significant linear mixed-model regression), except for the older-child emotional functioning subscale.• CONCLUSION:
The SUPER-KIDZ has shown good internal consistency and responsiveness, and satisfactory test-retest reliability. Construct validity was moderate for the younger- and older-child versions, but weak for the parent version.