Psychometric Properties of the Pediatric Patient-Reported Outcomes Measurement Information System Item Banks in a Dutch Clinical Sample of Children With Juvenile Idiopathic Arthritis

Michiel A.J. Luijten, Caroline B. Terwee, Hedy A. van Oers, Mala M.H. Joosten, J. Merlijn van den Berg, Dieneke Schonenberg-Meinema, Koert M. Dolman, Rebecca ten Cate, Leo D. Roorda, Martha A. Grootenhuis, Marion A.J. van Rossum, Lotte Haverman*

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Objective: To assess the psychometric properties of 8 pediatric Patient-Reported Outcomes Measurement Information System (PROMIS) item banks in a clinical sample of children with juvenile idiopathic arthritis (JIA). Methods: A total of 154 Dutch children (mean ± SD age 14.4 ± 3.0 years; range 8–18 years) with JIA completed 8 pediatric version 1.0 PROMIS item banks (anger, anxiety, depressive symptoms, fatigue, pain interference, peer relationships, physical function mobility, physical function upper extremity) twice and the Pediatric Quality of Life Inventory (PedsQL) and the Childhood Health Assessment Questionnaire (C-HAQ) once. Structural validity of the item banks was assessed by fitting a graded response model (GRM) and inspecting GRM fit (comparative fit index [CFI], Tucker-Lewis index [TLI], and root mean square error of approximation [RMSEA]) and item fit (S-X2 statistic). Convergent validity (with PedsQL/C-HAQ subdomains) and discriminative validity (active/inactive disease) were assessed. Reliability of the item banks, short forms, and computerized adaptive testing (CAT) was expressed as the SE of theta (SE[θ]). Test–retest reliability was assessed using intraclass correlation coefficients (ICCs) and smallest detectable change. Results: All item banks had sufficient overall GRM fit (CFI >0.95, TLI >0.95, RMSEA <0.08) and no item misfit (all S-X2 P > 0.001). High correlations (>0.70) were found between most PROMIS T scores and hypothesized PedsQL/C-HAQ (sub)domains. Mobility, pain interference, and upper extremity item banks were able to discriminate between patients with active and inactive disease. Regarding reliability, PROMIS item banks outperformed legacy instruments. Post hoc CAT simulations outperformed short forms. Test–retest reliability was strong (ICC >0.70) for all full-length item banks and short forms, except for the peer relationships item bank. Conclusion: The pediatric PROMIS item banks displayed sufficient psychometric properties for Dutch children with JIA. PROMIS item banks are ready for use in clinical research and practice for children with JIA.

Original languageEnglish
Pages (from-to)1780-1789
Number of pages10
JournalArthritis Care and Research
Volume72
Issue number12
DOIs
Publication statusPublished - Dec 2020

Cite this