The Rasch Analysis Shows Poor Construct Validity and Low Reliability of the Quebec User Evaluation of Satisfaction with Assistive Technology 2.0 (QUEST 2.0) Questionnaire.

Abstract

This study aims to test the construct validity and reliability of the Quebec User Evaluation of Satisfaction with assistive Technology 2.0 (QUEST)-device, an eight-item questionnaire for measuring satisfaction with assistive devices. We collected 250 questionnaires from 79 patients and 32 caregivers. One QUEST was completed for each assistive device. Five assistive device types were included. QUEST was tested with the Rasch analysis (Many-Facet Rating Scale Model: persons, items, and device type). Most patients were affected by neurological disabilities, and most questionnaires were about mobility devices. All items fitted the Rasch model (InfitMS range: 0.88-1.1; OutfitMS: 0.84-1.28). However, the ceiling effect of the questionnaire was large (15/111 participants totalled the maximum score), its targeting poor (respondents mean measure: 1.90 logits), and its reliability was 0.71. The device classes had different calibrations (range: -1.18 to 1.26 logits), and item 3 functioned differently in patients and caregivers. QUEST satisfaction measures have low reliability and weak construct validity. Lacking invariance, the QUEST total score is unsuitable for comparing the satisfaction levels of users of different device types. The differential item functioning suggests that the QUEST could also be problematic for comparing satisfaction in patients and caregivers.