Multi-laboratory evaluation of an automated microbial detection/identification system.

Area of Science:

• Clinical Microbiology
• Medical Diagnostics
• Bacteriology

Background:

• Accurate and efficient detection of bacteria in clinical urine specimens is crucial for timely diagnosis and treatment of urinary tract infections.
• Conventional methods for bacterial detection and identification can be labor-intensive and time-consuming.

Purpose of the Study:

• To evaluate the performance of an automated and computerized system, the Automicrobic System (AMS), for detecting frequently encountered bacteria in clinical urine specimens.
• To determine the sensitivity, specificity, reliability, and reproducibility of the AMS and compare it with conventional methods.

Main Methods:

• A collaborative study involving six laboratories.
• Testing of the AMS using pure cultures and mixtures of pure cultures to simulate clinical urine specimens.
• Comparison of AMS results with conventional bacterial detection and identification systems.

Main Results:

• The AMS achieved an average sensitivity of 92.8% and specificity of 99.4% for identifying common urine organisms using pure cultures.
• Reliability of positive results averaged 92.1%, influenced by false positive Escherichia coli identifications.
• High reproducibility was observed both within and among laboratories; the system detected fast-growing organisms below its enumeration threshold.

Conclusions:

• The Automicrobic System (AMS) demonstrated satisfactory overall performance for detecting bacteria in urine specimens under the tested conditions.
• The system showed high sensitivity and specificity, even in mixed cultures, although false positive E. coli reports occurred, particularly with Serratia species.
• The AMS is a reliable automated tool for bacterial detection in clinical urine samples, offering high reproducibility.