Testing for Vision Loss in Glaucoma Suspects

The problem of decreasing the amount of testing for vision loss in persons susceptible to glaucoma is formulated as an integer optimization problem. Nonlinear and linear versions of the problem formulation are given, and both heuristic and exact algorithms are proposed. Testing with retrospective clinical data indicates that the suggested procedures can yield substantial improvement in testing time. The relation of the glaucoma testing problem to classical problems in search theory is briefly discussed.