Validation of a Novel, Multidomain Head and Neck Cancer Appearance‐ and Function‐Distress Patient‐Reported Outcome Measure
Evan M. Graboyes, Brittany N. Hand, Mark A. Ellis, Andrew T. Huang, Marci L. Nilsen, Patrik Pipkorn, Courtney H. Marsh, Stacey Maurer, Terry A. Day, Katherine R. Sterba
Abstract
OBJECTIVES: Distress with self-perceived changes in appearance and function can result in body image disturbance (BID), which is common in head and neck cancer (HNC) survivors and a major source of psychosocial morbidity. To address the lack of psychometrically sound patient-reported outcome measures (PROMs) of HNC-related BID, we aim to create and validate the Inventory to Measure and Assess imaGe disturbancE-Head & Neck (IMAGE-HN). STUDY DESIGN: Survey study. SETTING: Multiple academic centers. SUBJECTS AND METHODS: Following item development, HNC survivors from 4 academic centers completed the IMAGE-HN. Item responses were psychometrically analyzed using confirmatory factor analysis (CFA) and Rasch analysis. RESULTS: Item development resulted in a 31-item PROM consisting of 5 individual domains and a global domain. In total, 305 HNC survivors of diverse ages, HNC subsites, and reconstructive paradigms completed the initial items. After removal of 3 items for local dependence, CFA confirmed the unidimensionality and local independence (item residual correlations <|0.20|) for each domain. Rasch analysis indicated acceptable fit (infit and outfit mean squares <2.0), monotonicity of all rating scale categories, and low person misfit (<4%). Person separation indices and person reliability were adequate for each domain except appearance concealment, which was removed (4 items). This resulted in the IMAGE-HN, a psychometrically acceptable 24-item PROM of HNC-related BID consisting of a global scale and 4 subscales measuring unique constructs and comprised independent items. CONCLUSIONS: IMAGE-HN is a novel, psychometrically sound, multidomain PROM of HNC-related BID for use in clinical and research settings.