Computerized expert system for evaluation of automated visual fields from the Ischemic Optic Neuropathy Decompression Trial: methods, baseline fields, and six-month longitudinal follow-up.

PURPOSE To validate a computerized expert system evaluating visual fields in a prospective clinical trial, the Ischemic Optic Neuropathy Decompression Trial (IONDT). To identify the pattern and within-pattern severity of field defects for study eyes at baseline and 6-month follow-up. DESIGN Humphrey visual field (HVF) change was used as the outcome measure for a prospective, randomized, multicenter trial to test the null hypothesis that optic nerve sheath decompression was ineffective in treating nonarteritic anterior ischemic optic neuropathy and to ascertain the natural history of the disease. METHODS An expert panel established criteria for the type and severity of visual field defects. Using these criteria, a rule-based computerized expert system interpreted HVF from baseline and 6-month visits for patients randomized to surgery or careful follow-up and for patients who were not randomized. RESULTS A computerized expert system was devised and validated. The system was then used to analyze HVFs. The pattern of defects found at baseline for patients randomized to surgery did not differ from that of patients randomized to careful follow-up. The most common pattern of defect was a superior and inferior arcuate with central scotoma for randomized eyes (19.2%) and a superior and inferior arcuate for nonrandomized eyes (30.6%). Field patterns at 6 months and baseline were not different. For randomized study eyes, the superior altitudinal defects improved (P = .03), as did the inferior altitudinal defects (P = .01). For nonrandomized study eyes, only the inferior altitudinal defects improved (P = .02). No treatment effect was noted. CONCLUSIONS A novel rule-based expert system successfully interpreted visual field defects at baseline of eyes enrolled in the IONDT.