[Evaluation of the GNF computer-coding system for the identification of non-fermentative Gram-negative bacilli]

In order to evaluate the effectiveness of the GNF computer-coding system for the identification of glucose non-fermenting gram-negative bacilli, we employed 406 strains of bacteria including 367 clinical isolates and 39 standard strains for testing. These strains were inoculated into the following eleven conventional biochemical test media: Triple Sugar Iron Agar, Simmon's Citrate Agar, Christensen's Urea Agar, Sulfide-Indole-Motility Medium, Semisolid Voges-Proskauer Test Medium, Moeller's Ornithine Decarboxylase Test Medium, Pyocyanin Test Medium, Oxidation/Fermentation (O/F) Glucose, O/F Fructose, Nitrate Broth, Moeller's Arginine Dihydrolase Test Medium. The results of these tests plus those from the hanging drop motility test and the oxidase test were converted into bacterial code number and then checked with the GNF computer-coding system. It was found that the first preference of agreement was 75.6%, second 15.3%, third 5.9%, and fourth or more 3.2%. In regard to the speed of bacterial identification by using the GNF system and information from hemolysis pattern and flagella stain, it was indicated that 84.7% would be correctly identified within 36-48 hours after isolation. If more confirmational tests were employed, the accurate identification rate would reach to 98.7% after 4 days of isolation. In addition, the use of the GNF computer-coding system can standardize identification procedures, shorten the identification period, and save cost in terms of materials supply, inoculation time, media preparation and media-storing space. Therefore, we conclude that the GNF computer-coding system is an effective tool in the identification of the glucose non-fermenting gram-negative bacilli.