Rasch Analysis of the Standard Patient Evaluation of Eye Dryness Questionnaire

This study sought to investigate, if the Standard Patient Evaluation of Eye Dryness (SPEED) questionnaire acts as a measure of dry eye severity and whether it is unidimensional using Rasch analysis.

This is a hospital-based cross-sectional study. The study included 127 patients with glaucoma with either a clinical diagnosis of dry eye or an Ocular Surface Disease Index score of ≥13. Participants completed questionnaires (SPEED) delivered directly to them.

A total of 127 participants met the inclusion criteria and were included in the analysis. Category response analysis revealed that the average logit measures for the response category increased with higher response category from -4.93 to +5.41 logit. Item infit mean square statistics of the individual items ranged from 0.57 to 1.33, with a mean of 0.99±0.2. The individual item outfit mean square ranged from 0.58 to 1.36, with a mean of 1.02±0.0. The person separation index (PSI) and separation reliability of the SPEED were 2.23 and 0.83, respectively. Principal component analysis (PCA) of the standardized model residuals revealed evidence of multidimensionality, with first contrast eigenvalue of 2.5 and the second contrast eigenvalue of 1.7. Because fit statistics of all items were within 0.5 to 1.7 and considered adequate for clinical observations, no item was removed. However, the six items of the SPEED that asked about the commonest dry eye symptoms as reported by patients with dry eye were analyzed separately to determine how they behaved. The item fit statistics of these six items were within the range of 0.6 to 1.4. The outfit statistics range from 0.62 to 1.26, with a mean of 1.00±0.1. The infit statistics range from 0.61 to 1.24, with a mean of 0.99±0.2. The PSI and reliability were 2.2 and 0.83, respectively. Finally, PCA of the standardized model residuals revealed no evidence of multidimensionality, with first contrast eigenvalue of 2.0.

In summary, this study showed that the SPEED had an ordered response category and was a measure of dry eye severity. However, PCA of the standardized model residuals revealed evidence of multidimensionality. A proposed six-item SPEED was unidimensional and was also a measure of disease severity. Hence, the SPEED questionnaire score may be used as a measure of dry eye severity in clinical practice and epidemiological studies.