Abstract
Purpose::
To examine the sensitivity, specificity and repeatability of categorical (e.g., Likert Scale) and continuous (e.g., Visual Analogue Scale) questionnaires used to identify subjects with dry-eye syndrome (DES) in studies of contact lens wear.
Methods::
18 experienced contact lens wearers were assigned a DES grade at an initial screening by a trained observer (1=no symptoms, 2=mild symptoms, 3=symptoms with significant discomfort, 4=symptoms with severe discomfort requiring discontinuation of lens wear). Subjects then completed 2 questionnaires detailing the severity and frequency of their DES symptoms: (1) a categorical scale questionnaire with severity ranging in 4 steps from no dryness to extreme dryness and with frequency ranging in 4 steps from never to all the time, and (2) a Visual Analogue Scale (VAS) questionnaire in which continuous lines with endpoints defined by these same extrema were marked with a vertical slash to indicate severity and frequency of dryness. The order of the questionnaires presented was randomized and subjects did not have access to the first questionnaire when completing the second. Subjects repeated the protocol at a second visit 2 weeks later.
Results::
The sensitivity and specificity (95% CI) of the categorical scale for the diagnosis of DES were 0.75 (0.51, 0.99) and 0.17 (0.00, 0.47), respectively. At optimum threshold values, the sensitivity and specificity of the VAS questionnaire were 0.58 (0.30, 0.86) and 0.67 (0.29, 1.00), respectively. . Repeatability was better with the categorical scale. Among subjects who reported no change in dryness after 2 weeks (n=4), 75% also reported no change in categorical scale response, compared to 50% with no change in VAS response. Among subjects who did report an increase or decrease in dryness after 2 weeks (n=14), 64% showed a concordant change in categorical scale response, compared to 50% with VAS.
Conclusions::
Identification of DES subjects for contact lens studies may be better achieved with a continuous VAS questionnaire, but may reject a greater number of potential subjects than the categorical questionnaire. In longitudinal trials the VAS questionnaire may be less reliable due to poor repeatability. A rigorous classification of DES subjects must depend, at least in part, on objective measurement in addition to questionnaire responses.
Keywords: cornea: tears/tear film/dry eye