Clinical analysis of screening and threshold visual field data for glaucoma detection

The value of clinical interpretation in differentiating between glaucomatous and normal fields from threshold (Humphrey Field Analyser) and screening (Henson CFS3000) measures was determined using a masked prospective experimental design. The visual field plots of 20 primary open-angle glaucoma (POAG) patients, 19 ocular hypertensive and 19 normotensive glaucoma suspects, and 21 age- matched normals measured with the Humphrey Field Analyser (Humphrey) and Henson CFS3000 (Henson) were categorised by two experienced clinicians. Significant differences in interpretation of the field plots were demonstrated between the two clinicians (χ^{2 McNemars}= 19–36; p<0.001). The sensitivity of clinical interpretation was shown to lie between 65 per cent and 90 per cent (dependent upon the individual clinician) for the Humphrey plots, but was as low as 40 per cent with the Henson plots. Specificity was, however, higher for the Henson overall, regardless of the clinician (between 90 per cent and 95 per cent) compared to the Humphrey (between 75 per cent and 100 per cent). These levels of sensitivity and specificity do not reach the levels reported when interpretation is based on the visual field indices alone. It was concluded that a screening instrument, such as the Henson, should only be employed for testing large unselected populations, in which the prevalence of glaucoma is low. Visual fields should not be judged in isolation, but in conjunction with measures of optic nerve and nerve fibre layer integrity, intra-ocular pressure and family history.