Table 1 Comparison of visual function tests vs ETDRS VA in MVP cohort (n = 130 MS patients, 90 controls) using ROC curve analysis

Visual function test (units)ROC curve area (95% CI)p value vs VA
The ROC curve area represents the probability that each test will correctly classify MVP cohort participants as MS patients vs disease-free controls. Areas range from 0.5 (no ability to distinguish MS vs disease-free controls beyond chance [flipping a fair coin]) to 1.0 (perfect capacity to distinguish).
* Visual function test scores were converted to SDU in the abnormal direction from the control group mean to ensure uniformity of scaling between measures.
† Sloan charts (contrast letter acuity) and Pelli–Robson (contrast sensitivity) distinguished MS patients from disease-free controls to a degree that was significantly better than ETDRS VA (significance level p < 0.01 used because multiple comparisons performed.
ETDRS = Early Treatment Diabetic Retinopathy Study; VA = visual acuity; MVP = Multiple Sclerosis Vision Prospective; MS = multiple sclerosis; ROC = receiver operating characteristic; SDU = standard deviation unit.
ETDRS VA (no. letters identified correctly: SDU)*0.67 (0.60, 0.74)
Sloan charts, 1.25% contrast (no. letters identified correctly: SDU)*0.83 (0.77, 0.88)<0.0001
Pelli–Robson contrast sensitivity (log contrast sensitivity: SDU)*0.79 (0.74, 0.84)0.003
L’Anthony D-15 DS color test (D-15 C index: SDU)*0.71 (0.64, 0.77)0.47
Esterman binocular visual field (Esterman efficiency score: SDU)*0.58 (0.52, 0.64)ROC area < VA