Premium
Development and validation of the McGill body image concerns scale for use in head and neck oncology (MBIS‐HNC): A mixed‐methods approach
Author(s) -
Rodriguez Ana Maria,
Frenkiel Saul,
Desroches Justin,
De Simone Avina,
Chiocchio Francois,
MacDonald Christina,
Black Martin,
Zeitouni Anthony,
Hier Michael,
Kost Karen,
Mlynarek Alex,
BolsterFoucault Clara,
Rosberger Zeev,
Henry Melissa
Publication year - 2019
Publication title -
psycho‐oncology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.41
H-Index - 137
eISSN - 1099-1611
pISSN - 1057-9249
DOI - 10.1002/pon.4918
Subject(s) - rasch model , item response theory , head and neck cancer , reliability (semiconductor) , psychology , patient reported outcome , test (biology) , convergent validity , classical test theory , scale (ratio) , physical therapy , clinical psychology , medical physics , medicine , internal consistency , psychometrics , cancer , psychotherapist , developmental psychology , power (physics) , paleontology , physics , quantum mechanics , quality of life (healthcare) , biology
Objective The aim of this study was to develop and validate a patient‐reported outcome measure to evaluate body image concerns in head and neck cancer (HNC) patients. Methods Items were created using a combination of deductive (eg, US Food and Drug Administration Qualification of Clinical Outcome Assessments, literature review) and inductive approaches (eg, subject matter experts, HNC patients). Items were translated for use in both Canadian English and Canadian French using back‐translation. A two‐step empirical validation process using the Classical Test Theory (CTT) and Rasch Measurement Theory (RMT) was conducted with 224 and 258 HNC patients, respectively, having undergone disfiguring surgery within the past 3 years. Results Analyses suggest two subscales for MBIS‐HNC: social discomfort (10 items) and negative self‐image (11 items). The McGill Body Image Concerns Scale–Head and Neck Cancer (MBIS‐HNC) is reliable with high internal consistency (0.98), high test‐retest reliability over a two‐week period (ICC = 0.88), moderate to high convergent validity (range r = 0.43‐0.81), and divergent validity (range r = 0.12‐0.15). RMT was used in addition to CTT. Disordered thresholds led to the modification of the number of response options, and items were deleted based on differential item functioning and high local dependency. Unidimensionality of both subscales and supporting a total score was confirmed. The measure was however characterized by the presence of an important floor effect, confirmed with poor targeting as demonstrated by the person‐item threshold distribution. Conclusion Evidence gathered from our theory‐driven validation study using CTT and RMT provides practitioners and researchers with a useful and easy to use self‐report measure.