Purchase this article with an account.
David A. Leske, Jonathan M. Holmes, Pediatric Eye Disease Investigator Group; Evaluation of the Intermittent Exotropia Questionnaire (IXTQ) using Rasch Analysis. Invest. Ophthalmol. Vis. Sci. 2012;53(14):5450.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The Intermittent Exotropia Questionnaire (IXTQ) is a 3-part (child, proxy, parent) patient-derived intermittent exotropia-specific instrument designed to evaluate health-related quality of life in children with intermittent exotropia (IXT) and their parents. The present study evaluated the IXTQ using Rasch analysis.
575 children 1 to 17 years of age with IXT and one parent were enrolled. All parents completed the proxy and parent IXTQ and 295 children completed the child IXTQ. Rasch analysis was performed to evaluate response ordering, dimensionality, local dependence, fit, differential item functioning (DIF; assessed for gender, age, and previous surgery), and targeting. The original questionnaires had 3 response options for 5- to 7-year olds, and 5 options for 8- to 17-year olds, proxy, and parent.
One question in the 12-item child and proxy questionnaires was removed because it showed a marked ceiling effect. Two response options were not distinguishable in the older child questionnaire; therefore response options were reduced from 5 to 3 and older and younger child questionnaires were combined. The combined child questionnaires showed no strong indication of multidimensionality, no local dependence, no large infit, outfit, or DIF. Targeting was relatively poor, person separation index was 1.19, and reliability was 0.58. One proxy IXTQ response option was not distinguishable from the others, therefore response options were collapsed to 4. There was no strong multidimensionality, no local dependence, no large infit, outfit, or DIF. Targeting was poor, with 35% of the sample scoring at the ceiling. The person separation index was 0.80, and reliability was 0.39 but improved to 1.45 and 0.68 when restricted to 8- to 17-year olds. The 17-item parent questionnaire showed evidence of multidimensionality, with 3 possible subscales. These corresponded to a priori subscales (psychosocial, function, and surgery) and had proper response option utilization with no additional multidimensionality. No local dependence or large infit or outfit were observed. Some DIF was observed, but no items were removed. Targeting was appropriate for each parent subscale and person separation and reliability was 2.69 and 0.88, 2.68 and 0.88, 2.06 and 0.81 for the respective subscales.
Rasch analysis indicated that the 8- to 17-year-old child questionnaire would benefit from reducing response options to 3 and it then may be analyzed with the 5- to 7-year-old child questionnaire. The proxy IXTQ would benefit from collapsing response options to 4 and may be most appropriate for 8- to 17-year olds. The 3 a priori subscales of the parent IXTQ were confirmed and therefore should be analyzed separately.
This PDF is available to Subscribers Only