Determination of the Minimal Clinically Important Difference of the University of North Carolina Dry Eye Management Scale.

PURPOSE:To establish an initial estimation of the MCID of the University of North Carolina Dry Eye Management Scale (UNC DEMS) and assess its association with patient perceptions of symptom change. METHODS:Thirty-three patients (33.3% men, 67.7% women, mean age 60.5 yrs) with previous DEMS scores were recruited from a UNC ophthalmology clinic in spring 2014. We used anchor-based methods, categorizing important symptom change, to compare the change in the DEMS scores across visits to patient assessments of change; linear regression coefficients estimated the MCID. We correlated clinical assessments, patient perceptions, and DEMS scores. RESULTS:DEMS score changes correlated with global anchors [-0.4229 (P = 0.014)]. Unadjusted linear regression yielded a beta coefficient of -0.54 (confidence interval, -0.97 to -0.12, R = 0.18, P = 0.014), which estimated the DEMS MCID. Adjusting the regression model for days since the last visit and DEMS score improved the association (beta = -0.56; confidence interval, -0.99 to -0.13; R = 0.43; P = 0.013). Descriptive statistics produced an MCID of 1 point. Patients said that 2 points would represent a significant change. The DEMS modestly correlated with the Schirmer test (-0.4045, P = 0.0266), Oxford Grading Scheme (+0.3713, P = 0.0364), and tear breakup time (-0.3559, P = 0.0456). CONCLUSIONS:The UNC DEMS is a valid, responsive patient-reported outcome measure instrument, which is easy to use in the clinic and capable of showing an MCID of 1 point.