Purchase this article with an account.
Ling Bei, Carolina Gracitelli, Linda M Zangwill, Robert N Weinreb, Felipe A Medeiros; Use of Mixture Item Response Theory to Identify Classes of Functionally Impaired Glaucoma Patients from Analyses of Patient-Reported Outcomes. Invest. Ophthalmol. Vis. Sci. 2015;56(7 ):1041.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To use mixture item response theory (IRT) to identify subclasses of functionally impaired glaucoma patients by analysis of results from the National Eye Institute Visual Function Questionnaire (NEI VFQ-25).
This is a cross-sectional study involving 586 patients with glaucomatous optic neuropathy. All subjects had testing with standard automated perimetry (SITA Standard 24-2) and responded the NEI VFQ-25 questionnaire for assessment of quality of life. The 25 items of the NEI VFQ-25 were subdivided into visual function and socio-emotional subscales. Mixture item response theory was used to identify subclasses of patients according to their questionnaire responses. The Vuong-Lo-Mendell-Rubin likelihood ratio test was used to investigate number of classes.
Four classes of subjects were identified by the mixture IRT model. The model had entropy of 0.836. The first class contained the majority of subjects (411 patients) and showed normal results for both the visual function and socio-emotional subscales. The second class (51 patients) had abnormal results on both scales. The third class had abnormal results only on the visual function scale (57 patients), whereas a fourth class (67 patients) showed abnormal results only on the socio-emotional scale. Compared to the reference (normal) class, subjects in classes 2 and 3 had worse results on SAP binocular mean sensitivity. However, subjects in class 4 (socio-emotional) did not show abnormal visual field results compared to the reference class.
Different subclasses of functionally impaired glaucoma patients were identified from analysis of NEI VFQ-25 results by mixture modeling. These findings suggest that the NEI VFQ-25 lacks unidimensionality and that this needs to be taken into account when evaluating the relationship between results of this questionnaire and tests of visual function in glaucoma.
This PDF is available to Subscribers Only