The repeatability of tests for dry eye signs and symptoms in the dry eye assessment and management (DREAM) study

Purpose: To assess the repeatability of measures for dry eye disease (DED) symptoms and signs in the DREAM study. Methods: At screening and baseline visits approximately 2 weeks apart, participants were assessed for symptoms by Ocular Surface Disease Index (OSDI) and Brief Ocular Discomfort Index (BODI), and signs by the same physician in the same order: tear break-up time (TBUT), corneal staining, conjunctival staining, Meibomian gland evaluation, and the Schirmer test. The repeatability of DED symptoms and signs was assessed by interclass correlation coefficient (ICC), 95 % limits of agreement, and the percent of eyes with inter-visit difference above the clinically significant threshold. Results: Among 1046 eyes (523 participants), ICC for signs ranged from 0.53 (TBUT) to 0.73 (corneal staining). A substantial percentage of eyes showed clinically significant inter-visit differences: ≥2 points in 17.8 % of eyes for conjunctival staining; ≥3 points in 18.8 % for corneal staining; >2 s in 14.1 % for TBUT; ≥5 mm/5 min in 29.9 % for the Schirmer test, and ≥ 2 points in 27.5 % for Meibomian gland plugging and lid secretion. The OSDI and BODI had ICC of 0.64 and 0.63 respectively, and nearly 40 % of participants had inter-visit score differences ≥ 10 points. Conclusion: In DREAM participants with moderate-to-severe DED, DED signs and symptoms had moderate repeatability, with ocular surface staining scores being the most repeatable and TBUT the least repeatable. A notable percentage of participants had inter-visit differences above the clinically meaningful threshold. These test–retest variabilities in DED signs and symptoms should be considered for designing clinical trials and monitoring disease progression.