Psychometric Analysis of the Patient-Reported Outcomes Measurement Information System (PROMIS) Parent Proxy Physical Function-Upper Extremity Item Bank for Children With Duchenne Muscular Dystrophy

Linda Pax Lowes ^1* Corinne M. Le Reun ² Teofil Ciobanu ³ Lindsay N. Alfano ¹ Natalie F. Reash ¹ Megan A. Iammarino ¹ Shivangi Patel ⁴ Ivana F. Audhya ⁴

• ¹ Center for Gene Therapy, Abigail Wexner Research Institute, Nationwide Children's Hospital, Columbus, Georgia, United States
• ² Biostatistics, Sainte-Anne, Guadeloupe, France
• ³ F. Hoffmann-La Roche Ltd., Basel, Switzerland
• ⁴ Patient-Centered Outcomes, Sarepta Therapeutics, Inc., Cambridge, United States

Background: The Patient-Reported Outcomes Measurement Information System (PROMIS) is a collection of generic patient-reported outcome instruments used to quantify disease impact on a variety of functional subdomains, including physical, cognitive, emotional, and other domains. The reliability and validity of the PROMIS Parent Proxy (PP) Physical Function–Upper Extremity (UE) item bank is not established in children with Duchenne muscular dystrophy (DMD). This study evaluated the psychometric properties and measurement quality of the PROMIS PP UE item bank v1.0 (29 items) in DMD using a Rasch psychometric analysis.Methods: The PROMIS PP UE item bank was completed by caregivers of children with DMD aged at least 8 years, under care at Nationwide Children’s Hospital (Columbus, OH, USA). Rasch analysis was used to evaluate the psychometric performance of the measure and its items in DMD, based on several criteria, including item–trait interaction, individual items fit, Person Separation Index (PSI), individual persons fit, and response dependency.Results: Rasch analysis was conducted on 206 observations. Several items had weak clinical utility in measuring upper extremity functioning in DMD. Additionally, the analysis identified specific response options that could be restructured to improve the reliability and precision of the items in evaluating upper extremity function in DMD. A new customized 21-item measure demonstrated overall good fit to Rasch model expectations (p=0.095; nonsignificant) and the ability to discriminate among respondents with different levels of upper extremity function (0.95 PSI; excellent reliability). Upper extremity function was generally well targeted across the severity spectrum, except for the least severe patients.Conclusion: The customized PROMIS PP UE measure conformed to Rasch assumptions, indicating that it can serve as a reliable option for caregiver-reported upper extremity assessment in DMD.