Abstract
Purpose: :
To estimate the specificity of the Guided Progression Analysis (GPA, Carl Zeiss Meditec, CA), in individual patients.
Methods: :
One eye of 30 glaucoma patients (Mean Deviation [MD], mean -3.3 dB, range -10.7 to +1.2 dB) was tested 12 times in 3 months (SITA-Standard 24-2 test). "Likely progression" or "possible progression" was determined with an analysis similar to the GPA where black triangles identify locations with significant change from baseline (mean of tests 1 and 2) in 3 or 2 consecutive follow-up tests, respectively. These analyses were repeated after the tests were randomly rearranged in time (100 permutations per series).
Results: :
The average specificity of the GPA was high - for the entire sample, the false positive (FP) rates for "likely progression" and "possible progression" after 10 follow-up tests were 1.3% and 10%. However, FP rates varied substantially between patients, from less than 1% to approximately 8% with "likely progression" and from less than 1% to approximately 40% with "possible progression". Factors associated with FP events were visual field variability (SD of MD, Spearman rank correlation coefficient r=0.23, p=0.03) but not overall severity of damage (average MD, Pattern Standard Deviation, p>0.1) or any of the reliability indices (average rates of false-positive and false negative responses, fixation losses, p>0.3).
Conclusions: :
On average, progression criteria currently employed in the GPA have high specificity, but some patients are up to 40 times more likely to show a false-positive event than others. These finding are important when the GPA is used in the clinical management of patients and in clinical trials. Because the consistency of visual field results varies greatly between patients, criteria for point-by-point change should be adapted to individual patients.
Keywords: visual fields • perimetry