Structural validity of the Finnish Manchester-Oxford Foot Questionnaire (MOXFQ) using the Rasch model. Issue 1 (January 2021)

• Record Type: Journal Article
• Authors: Ponkilainen, Ville T.; Miettinen, Mikko; Sandelin, Henrik; Lindahl, Jan; Häkkinen, Arja H.; Toom, Alar; Tillgren, Tomi; Ilves, Outi; Latvala, Antti O.; Ahonen, Katri; Sirola, Timo; Sampo, Mika; Väistö, Olli; Repo, Jussi P.
• Is Part Of: Foot and ankle surgery. Volume 27:Issue 1(2021)
• Page Start: 93
• Page End: 100
• Publication Date: 2021-01
• Subjects: Foot -- Ankle -- Patient-reported -- MOXFQ -- PROM -- Psychometrics -- Validation -- Surgery -- Rasch; Foot -- Surgery -- Periodicals; Ankle -- Surgery -- Periodicals; Ankle -- surgery -- Periodicals; Foot -- surgery -- Periodicals; Ankle -- Surgery; Foot -- Surgery; Periodicals; Electronic journals; 617.58
• DOI: 10.1016/j.fas.2020.02.012
• Languages: English
• ISSNs: 1268-7731

Abstract

Highlights: The Manchester-Oxford Foot Questionnaire has been claimed to be the most suitable patient-reported outcome measure for foot and ankle patients. The response categories were ordered in all but three items (in the subscale of pain). The pain subscale did not support unidimensionality. Removing one item led to unidimensional structure. The subscales had good coverage and targeting for both first metatarsal joint pathologies and other foot and ankle pathologies. The Finnish version of MOXFQ has acceptable structural validity for assessing outcomes in orthopaedic foot and ankle population. Abstract: Background: The 16-item patient-reported Manchester-Oxford Foot Questionnaire (MOXFQ) with subscales of pain, social interactions, and walking/standing has been claimed for strongest scientific evidence in measuring foot and ankle complaints. This study tests the validity of the Finnish MOXFQ for orthopaedic foot and ankle population using the Rasch analysis. Methods: We translated the MOXFQ into Finnish and used that translation in our study. MOXFQ scores were obtained from 183 patients. Response category distribution, item fit, coverage, targeting, item dependency, ability to measure latent trait (unidimensionality), internal consistency (Cronbach's alpha), and person separation index (PSI) were analyzed. Results: Fifteen of the items had ordered response categories and/or sufficient fit statistics. The subscales provided coverage and targeting. Some residual correlation was noted. Removing one item in the pain subscale led to a unidimensional structure. Alphas and PSIs ranged between 0.68–0.90 and 0.67–0.92, respectively. Conclusions: Despite some infractions of the Rasch model, the instrument functioned well. The subscales of the MOXFQ are meaningful for assessing patient-reported complaints and outcomes in orthopaedic foot and ankle population.