Psychometric Properties and Validation of the Standard Patient Evaluation of Eye Dryness Questionnaire

Abstract

To characterize the psychometric properties of the standard patient evaluation of eye dryness (SPEED) questionnaire and to validate and compare its performance with 4 existing dry eye questionnaires. A total of 50 subjects (40 female and 10 male) were enrolled; of these, 30 were symptomatic and 20 asymptomatic, as determined using the ocular surface disease index (OSDI). This study consisted of 2 visits in which all subjects completed 5 different dry eye questionnaires (SPEED, OSDI, dry eye questionnaire, McMonnies dry eye questionnaire, and subjective evaluation of symptom of dryness) in random order at each visit. Clinical measurements were obtained on the first visit. Repeatability was determined using concordance correlation coefficient; dimensionality was determined using principal component, factor, and Rasch analyses; and validity was determined by comparing SPEED scores with dry eye diagnosis based on OSDI (primarily using receiver-operator curve analysis). The SPEED questionnaire data were found to be unidimensional and repeatable. Three principal components (dryness, burning, and soreness/fatigue) were identified and SPEED between visit concordance correlation coefficient was 0.923 (95% confidence interval, 0.868-0.955). The area under the receiver-operator curves was 0.928. The only clinical measures that correlated "well" with SPEED questionnaire scores were corneal staining (P < 0.05), meibomian gland score (P < 0.05), and meibomian glands yielding liquid secretion score (P < 0.05). The SPEED questionnaire was shown to be a repeatable and valid instrument for measurement of dry eye symptoms. The SPEED score also correlated significantly with ocular surface staining and clinical measures of meibomian gland function.