Abstract
Purpose: :
To describe reliability of visual field testing in a developing country setting.
Methods: :
Retrospective review of all threshold visual fields performed at 5 glaucoma clinics affiliated to Pontificia Universidad Catolica de Chile from 2002-2010.Criteria for unreliability was defined as either false positives (FP)>=15%, false negatives (FN) >=25% or fixation losses (FL) >= 25%. For eyes with repeat testing on the same day, only the second test was included in the analysis.
Results: :
63104 visual field examinations were reviewed, 2637 corresponded to repeat examinations, leaving 60467 tests for analysis.The number of exams per patient was 1,2,3,4 and >=5 in 56%, 21%, 9%, 5% and 9% respectively.The overall proportion of unreliable fields was 33.3%. The proportion of patients failing to meet the reliability criterion for FP, FN and FL was 4.8%, 8.8% and 26.5% respectively. FP were more common in the right than in the left eyes (5.19% vs 4.55%, X2 p-value<0.001). FL were more common in left than right eyes (25.4% vs 27.6%, X2 p-value <0.001).The proportion of unreliable tests on first, second, third and fourth visit was 27.6%, 34.2%, 42.3% and 46.4% respectively. After the fourth test the proportion of unreliable tests remained above 46% up to the 13th visit.
Conclusions: :
Unreliability when conducting visual fields is a major problem in our setting. Studies to detect unreliable field takers early during follow-up and train them to perform better are needed.
Keywords: visual fields • perimetry • clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology