Abstract
An evaluation of the Automicrobic System (AMS) for Urines (Vitek Systems, Inc.) was carried out under the auspices of the Product Evaluation Committee of the College of American Pathologists from the period June 1977 through October 1978. Data generated during this evaluation indicated that, when comparing the AMS methodology to our clinical microbiology laboratory methodology, a 37% time saving could be realized by utilizing the AMS. Quantitation with the AMS showed a 99% correlation with the clinical microbiology laboratory method except for yeast which correlated only 50% of the time. The average overall identification accuracy was 95%. Negative response accuracy was 99%. Other members of the Enterobacteriaceae which the instrument is not designed to identify may produce erroneous results if they occur in urine specimens. Specimens containing two organisms were identified with a 94% correlation when compared to our conventional methodology. The time when a well becomes positive may be used as a fairly reliable indicator of significance (count greater than 70,000 colony-forming units per ml) for Escherichia coli, Klebsiella-Enterobacter group, and group D Enterococcus, but not for Proteus sp., Pseudomonas aeruginosa, and yeast. Specimen collection must be performed properly since specimens considered as contaminated by conventional plating-out techniques may be reported out by the AMS as only one or two organisms and thus lead to an erroneous assumption as to significance. Cost per specimen was $1.83 more by utilizing the AMS method as compared to our conventional method. This is offset by a saving of 1.74 h daily of personnel time and a final report in 13 h. At least 30 urine specimens would be needed daily to pay for the instrument and specimen costs in 1 year. The AMS can provide significant aid to a clinical microbiology laboratory when all factors are considered.