Abstract
Purpose :
Previous work has investigated whether a significant regression slope in the first two years for the summary index Mean Deviation (MD) is predictive of rapid (≤−2dB/year) visual field progression in glaucoma. This work assumed six visual fields were obtained in the first two years, as recommended by several guidelines. In clinical practice it is common that only two or three fields are measured in the first two years, however. Here we use simulation methods to investigate how reducing the number of visual fields measured influences the prediction of rapid visual field progression, along with the influence of including criteria based on the magnitudes of the regression slope.
Methods :
We simulated visual field series (N=100,000) spaced annually in the first two years and then biennially, consistent with recent data on test frequency within a UK hospital setting. MD values had a standard deviation of 1dB. The true underlying rates of progression were selected from a modified hyperbolic secant with parameters averaged from fits to large datasets from Canada, Sweden and the USA. We calculated positive and negative predictive values (PPV & NPV) for detecting rapid progression, based on a criterion of a significant negative regression slope of any magnitude. We performed a second simulation using test frequency and disease prevalence parameters designed to match those of a dataset of 255 glaucoma patients from Tokyo, to check the validity of our simulation method.
Results :
PPV at 2 years was 0.09, differing little from when six visual fields in the first two years were performed (PPV=0.10). PPV peaked at 0.11 at 4 years. Adding an additional regression slope criterion of ≤−2dB/year increased the 2-year PPV to 0.35. NPV values were 0.98 at two years, rising to 1.0 by six years, and remained above 0.98 with the addition of a regression slope criterion. Our simulated PPV and NPV values for the Tokyo dataset were close to those determined empirically.
Conclusions :
Infrequent visual field testing does not substantially alter predictive values for detecting rapid visual field progression compared with more frequent testing. Our results do not, however, suggest that the rate of visual field progression can be accurately estimated when a small number of visual fields are performed.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.