Comparison of frequency doubling perimetry with humphrey visual field analysis in a glaucoma practice: The reply

Abstract

Thank you for your interest in our paper “Comparison of frequency doubling perimetry with Humphrey visual field analysis in a glaucoma practice”. Although 29 patients is a relatively small sample size, our study adds to the growing body of literature showing the good sensitivity and specificity of the frequency doubling technology in detecting visual field loss. In particular, our study looked at frequency doubling technology in screening and threshold modes. There may be some misunderstanding as to how we used the frequency doubling technology classification for severity of visual field loss. The categories of mild, moderate, or severe defects, as provided by the frequency doubling technology, were used to determine a useful cutoff to use with the frequency doubling technology; that is, should people with any defect (mild, moderate, or severe) be considered as positive for glaucoma on the screening test, or would it be better to consider only those with moderate or severe defects? We did not evaluate the sensitivity and specificity of the frequency doubling technology by severity of visual field defects, as determined by Humphrey visual field analysis. Table 1 provided the definitions we used in Figure 1. “Mild” defects refer to a broad definition of mild, moderate, or severe loss, as determined by the frequency doubling technology. “Moderate” includes moderate or severe defects, and “severe” is the most stringent category, which includes only those with severe relative loss. As stated in the first paragraph of page 330, “As the criterion for an abnormal result of frequency doubling perimetry in screening mode became more stringent, from requiring any defect to requiring severe defects to define an abnormal case, the area under the receiver operating characteristic curve decreased for all definitions of an abnormal Humphrey visual field.” Our results cannot be compared with the results of Cello and associates. We did not evaluate sensitivity and specificity in terms of severity of defects, which is what Cello and colleagues did in their study.