Examination of the PROMIS upper extremity item bank.

STUDY DESIGN: Clinical measurement.

INTRODUCTION: The psychometric properties of the PROMIS v1.2 UE item bank were tested on various samples prior to its release, but have not been fully evaluated among the orthopaedic population.

PURPOSE OF THE STUDY: This study assesses the performance of the UE item bank within the UE orthopaedic patient population.

METHODS: The UE item bank was administered to 1197 adult patients presenting to a tertiary orthopaedic clinic specializing in hand and UE conditions and was examined using traditional statistics and Rasch analysis.

RESULTS: The UE item bank fits a unidimensional model (outfit MNSQ range from 0.64 to 1.70) and has adequate reliabilities (person = 0.84; item = 0.82) and local independence (item residual correlations range from -0.37 to 0.34). Only one item exhibits gender differential item functioning. Most items target low levels of function.

DISCUSSION: The UE item bank is a useful clinical assessment tool. Additional items covering higher functions are needed to enhance validity.

CONCLUSIONS: Supplemental testing is recommended for patients at higher levels of function until more high function UE items are developed.

LEVEL OF EVIDENCE: 2c.