Small areas of damage due to glaucoma are less likely to be identified by SITA-Faster than SITA-Standard

To examine if clusters of colocated damaged locations repeatedly found during prior SITA-Standard (SS) visual field tests are not replicated in a subsequent SITA-Faster (SFR) test.

Visual field series from the Lions Eye Institute, Perth, Australia were extracted if they had at least 4 SS fields followed by a SFR field and all fields had false positive rates and fixation losses less than 15%. Clusters of damaged locations were defined as either an isolated location with no damaged neighbors or a group of connected 24-2 points (not crossing the horizontal) that all had Pattern Deviation (PD) probability values below 1% of normal in all 3 SS tests immediately preceding the SFR test. For each cluster, the number of locations that were damaged in the SFR field that overlapped the established and repeatable SS clusters was computed and reported. We compare this proportion of overlap against the same calculations using the last SS as the final field. We also examine the effect of relaxing the PD criteria for damage in the final field on the proportion of overlap.

2519 eyes were included. For established, repeatable, and clusters of size 1, 2, 3, and 4 locations, the follow-up SFR field completely missed 46%, 26%, 11%, and 14% of the clusters whereas SS missed significantly less 26%, 11%, 4% and 1% (t-test with mixed paired and unpaired observations and p < 0.05). Relaxing the criteria for damage in the final field to PD < 5% reduced the miss rates for both methods, but SFR was still significantly worse than SS (SFR: 24%, 8%, 3%, and 6% vs. SS: 13%, 3%, 1%, and 1%).

SITA-Faster tends to miss established clusters of damaged locations of size 4 or fewer locations in follow up visits at a higher rate than SITA-Standard.