Abstract
Purpose:
To determine if a single valid dry eye severity variable can be estimated from a battery of clinical tests and patient self-reports.
Methods:
This case control study included 150 case and 50 control patients over age 18. Inclusion in the case group required a physician diagnosis of dry eye. Controls were individuals who are not seeking or have not sought eye care services for dry eye. The following tests were performed in the order they are listed here: review of systems, OSDI questionnaire, tear osmolarity, Schirmer’s, automated tear break-up time (T-BUT), corneal staining with fluorescein, and conjunctival staining with lissamine green (SICCA grading system was used for grading). A single dry eye severity variable was estimated for each participant from Rasch analysis using the Masters partial credit model of clinical test results and patient responses to the OSDI. For purpose of analysis, continuous clinical test variables (osmolarity, Schirmer’s, and T-BUT) were binned into quintiles and assigned rank scores. Rank scores were also assigned to ordinal OSDI response categories for each question and ordinal clinician ratings of corneal and of conjunctival staining.
Results:
Measures of dry eye severity were estimated for each case and control (person measures) and measures of sensitivity to dry eye severity were estimated for each sign and symptom (indicator measures). The average person measure for cases (-0.39, SD=0.59) was significantly greater (p<0.001) than the average person measure for controls (-1.2, SD=0.83), but the distributions overlapped heavily. The distribution of indicator measures was well matched to the distribution of person measures for cases. Schirmer’s at 1 minute was too sensitive to discriminate dry eye severity among cases, whereas the other 17 indicators fell within the range of severity values for cases. Schirmer’s at 5 minutes was the most sensitive of these 17 indicators, followed closely by tear osmolarity; OSDI item 9 (watching TV) was the least sensitive. The most informative indicators were corneal and conjunctival staining; OSDI item 7 (driving at night) was the most informative OSDI item. No indicator carried more than 10% of the total information about dry eye severity than that carried by all indicators together.
Conclusions:
Standard dry eye clinical signs and symptoms can work together to define and measure a latent dry eye severity variable. There is no single “best” dry eye severity measure.