[Computer-assisted diagnosis and therapy for glaucoma].

Abstract

Boolean algebra, or combinatory analysis and their related computer routines, can provide invaluable help in resolving classic diagnostic problems. However complex each case may be, the diagnosis is always made from a finite set of data, and the fundamental problem is thus how to exploit this data. Invention no longer has a place in ascertaining a diagnosis. Traditional ways of reasoning are numerous, personal, and fragile, but fortunately redundant. They may give rise to four types of error: omission or mistake (an error of judgment), either during the semiotic or the dialectic stages. Whereas the physiological capacity of the human brain and memory only enables it to make a limited number of hypotheses concerning certain aspects of glaucoma, computer programs can take the total number of hypotheses into account, i.e., 3000. For every input the program explores each of the 3,000 items, thus eliminating the four types of error. The probabilistic nature of data, which compromises the confidence one can have in conclusions resulting from such complex reasoning, is treated by the adjusted probabilities. The use of such diagnostic aids, whose thesaurus is updated regularly, is reserved for ophthalmologists, the only authority capable of assessing the pertinence of the computer responses. Consequently, the specialist can rest assured that the patient has benefited from the most comprehensive and updated knowledge in medical science.