The development and evaluation of the new Ocular Surface Disease Index-6

Abstract

Aim: To evaluate if the OSDI can be shortened and the score calculation simplified without significantly impacting the outcome of the questionnaire. Methodology: Study#1: 264 participants completed the OSDI questionnaire (174 females; mean age: 34.4 ± 12.3yrs) and the results were analyzed to detect those questions of each subscale that are the most discriminative by multiple regression and RASCH analyses, resulting in the OSDI-6. Study#2: OSDI-6 was compared to the OSDI and the 5-item Dry Eye Questionnaire (DEQ-5) in 120 patients (73 females; mean age: 51 ± 20.2yrs) to evaluate predictive ability of the OSDI-6. Repeatability was analyzed in 50 participants. Results: Study#1: The mean OSDI score was 13.1 ± 11.5. The most discriminant questions were questions 1, 4, 7, 9, 10 and 11. Infits and outfits of the OSDI-6 were between 1.26 and 0.78 (STRATA = 3). The OSDI-6 was significantly correlated to the OSDI (r = 0.898, p < 0.001). Area under the curve (AUC) of the OSDI-6 to predict OSDI was 0.967 p < 0.001. Study#2: The mean OSDI score was 20.3 ± 16.6, the mean DEQ-5 score 7.9 ± 4.6 and the mean OSDI-6 score 10.3 ± 8.6. AUC of the OSDI-6 to predict OSDI was 0.901 (p < 0.001) and 0.803 (p < 0.001) to predict DEQ-5. The OSDI-6-V2 was significantly correlated to the OSDI and its sub-scales (r > 0.842, p < 0.001). Repeatability of the OSDI was 0.72 ± 0.11 (Kappa; p < 0.001), the DEQ-5 was 0.75 ± 0.06 (p < 0.001) and the OSDI-6 was 0.80 ± 0.05 (p < 0.001). Conclusions: The OSDI-6 seems to be a repeatable questionnaire and a good alternative to the full original OSDI for use in clinical practice as well as research.