Retest Variability as a Function of Sensitivity and Zero Thresholds in Humphrey Size–III SITA and Full–Threshold Size V in Glaucoma Patients

Abstract

 

Purpose:

 

to compare how retest variability differs across the threshold range using size III SITA and size V full threshold methods of the Humphrey Field Analyzer (HFA).

 

Methods:

 

Once a week for five consecutive weeks glaucoma patients (N=17) were given two HFA tests: (1) SITA standard Size III and (2) Full–threshold Size V. Point–wise standard deviations (sds) were calculated across the five tests for: (1) all test locations and (2) for locations without a zero threshold on all of the five tests. These sds were aggregated and averaged across thresholds and were then banded into five similarly sized bins for inferential statistics. A 2x2x5 (test type, sd type, threshold bin) independent groups factorial ANOVA was performed on the averaged sds.

 

Results:

 

Size III yielded slightly more retest variability (.324 sd; p=.027). Removal of the floor effect (zero values) reduced retest variability at the two lowest thresholds bins (1.43 and 1.38 sd difference); that is, the interaction between sd type and dB group was significant (p<.001; Figure). However, this pattern did not vary between size III SITA and size V full threshold (p=.723).

 

Conclusions:

 

Retest variability is slightly higher for size III than size V perimetry and decreases in magnitude at lower sensitivities when zero thresholds are excluded.  

 

Keywords: perimetry • visual fields