Reliability of computerized perimetric threshold tests as assessed by reliability indices and threshold reproducibility in patients with suspect and manifest glaucoma.

BACKGROUND: High reproducibility of test measurements is often considered an indication of high reliability of test results. The aim of the current study was to analyse the role of the traditional perimetric reliability indices, False Negative and False Positive responses, and Fixation Losses, as indicators of test reliability in comparison with threshold reproducibility in patients with suspect or manifest glaucoma.
METHODS: Perimetry was performed in one eye in each of 76 patients. Each eye was tested twice within approximately one week using the Humphrey II 30-2 SITA STANDARD program. Frequencies of False Positive and False Negative answers, and rates of Fixation Losses were related to threshold reproducibility and to general field status as expressed by Mean Deviation from age-normal threshold values using stepwise multiple linear regression analysis.
RESULTS: Substantial field loss was associated both with low threshold reproducibility (p<0.0001) and with increased frequency of False Negative answers (p=0.047). The traditional reliability indices contributed marginally compared to amount of field loss when predicting threshold reproducibility; the coefficient of determination decreased non-significantly from 0.37 to 0.33 when excluding the three reliability indices as explanatory variables from the regression model. Frequencies of False Positive answers and Fixation Losses showed no association to field status or to threshold reproducibility.
CONCLUSION: Reliability of visual field test results in patients with glaucoma expressed as threshold reproducibility, can be predicted by amount of field loss alone, and traditional patient reliability indices contribute surprisingly little in this regard.