J Clin Epidemiol. 2025 May 29:111855. doi: 10.1016/j.jclinepi.2025.111855. Online ahead of print.
ABSTRACT
OBJECTIVE: This study investigated psychometric properties and reference values of the Patient-Reported Outcomes Measurement Information System (PROMIS®) pediatric v2.0 Mobility, Upper Extremity and Pain Interference item banks, short forms (SF) and computerized adaptive tests (CATs) in the Dutch general population, supplemented with a clinical sample to improve low-end item parameter estimates.
STUDY DESIGN AND SETTING: Children (8-18 years) completed PROMIS item banks and legacy instruments (Pediatric Quality of Life Inventory 4.0 subdomain Physical Health, Numeric Pain Rating Scale). Structural validity of item banks was evaluated by fitting a graded response model and inspecting item-fit statistics. Reliability of item banks, SFs, and post-hoc CATs was expressed as standard error of measurement/theta (SE(θ)). To compare measurement efficiency of instruments, relative efficiency was calculated. Construct validity was assessed by correlating item banks with legacy instruments. Differential item functioning (DIF) between Dutch and U.S. samples was evaluated.
RESULTS: 783 children participated: 555 children from the general population, and 228 children receiving physical therapy. Structural validity was sufficient for all banks. PROMIS Pain Interference was reliable at the sample mean (SE(θ)<0.32) and up to 2 SD in the clinically relevant direction (indicating worse health). PROMIS Mobility and Upper Extremity scales were reliable in the clinically relevant direction, but less so within the normal range. CAT outperformed other assessment methods in efficiency. Construct validity was sufficient. No items displayed DIF.
CONCLUSION: The PROMIS v2.0 pediatric Mobility, Upper Extremity, and Pain Interference item banks displayed sufficient validity in the Dutch general population and sufficient reliability in the clinically relevant direction.
PMID:40449761 | DOI:10.1016/j.jclinepi.2025.111855
