The Development of Validated Bulbar Redness Grading Scales

Abstract

PURPOSE: To develop a perceptually and physically based bulbar redness grading scale. METHODS: Digital conjunctival hyperemia photographs were taken using a photo-slit lamp at controlled exposures. Nine participants arranged 25 images on a tabletop over a range of 1.5 m, using separation to represent changes in redness. The position of each image was recorded and normalized for a 0 to 100 scale, and compared to chromaticity of each image obtained using a spectrophotometer. The performance of two versions of the scale (5 and 10 images) and a continuous grading scale was evaluated based on repeatability data collected from nineteen observers who used each scale twice to grade 30 randomly presented images of bulbar redness. RESULTS: Psychophysical scaling was highly correlated between single observers (Pearson's r >or= 0.92, p < 0.05). The averaged subjective grades significantly correlated with chromaticity (r = 0.95 and r = 0.99, p < 0.001 for CIE u* and log u*, respectively). Across all observers, test and retest ratings were highly correlated with either scale (r >or= 0.98), and showed high levels of repeatability expressed by intraclass correlation coefficients (ICC >or= 0.98), correlation coefficients of concordance (CCC >or= 0.96), and coefficients of repeatability (COR <or= 5.64). Despite single unit increment options, the majority of grade values assigned using the discrete scales were distributed in multiples of 5. CONCLUSIONS: Combining psychophysical and physical attributes is a promising method for the development of novel anterior segment scales; the newly developed scales performed well in a clinical setting.