Development of a reliability criterion for the binocular Esterman visual field test.

Abstract

Purpose: To evaluate the repeatability of the binocular Esterman visual field (bEVF) test and develop a corresponding reliability criterion based on repeatability-associated factors.

Study design: Prospective cross-sectional study.

Methods: Ninety-five glaucoma patients underwent two bEVF tests on the same day using a Humphrey field analyzer. The repeatability of the bEVF was assessed by the agreement rate for seen and unseen points over all test points between the first and second tests. Factors associated with the repeatability were analyzed with multivariate linear regression analysis. Receiver operating characteristic (ROC) curve analysis was performed for those associated factors to identify the reliability criterion.

Results: The mean agreement rate was 88.9% (95% confidence interval [CI]: 73.9%, 99.2%). A total of 44 patients (46.3%), 75 patients (78.9%), and 88 patients (92.6%) had agreement rates of more than 90%, 85%, and 80%, respectively. The false positive (FP) response rate of the bEVF test was negatively associated with the agreement rate (p<0.01). The best cutoff value according to the maximum Youden index for the FP response rate of the bEVF test was >7.1% (sensitivity = 71.0% and specificity = 88.0%), while the cutoff value based on a fixed 95% specificity was >12.1% (sensitivity = 28.6%).

Conclusion: An approximately 90% agreement rate existed for the bEVF test when applied to glaucoma patients. The FP response rate was associated with the repeatability of the bEVF test, and 8% or at least 13% may serve as a useful reference to ensure an agreement rate of 80% or higher.