Creation and evaluation of an expert system for disk diffusion antimicrobial susceptibility testing;
Detection of antimicrobial resistance is one of the most important responsibilities of the clinical microbiology laboratory. The purpose of antimicrobial susceptibility testing is to detect resistance to antimicrobial agents against the infecting pathogen and to assist the clinician in selection of appropriate therapy. Knowledge of a specific pathogen’s identification and the resistance mechanisms available to the pathogen can help predict resistance. An expert system was created to assist the microbiologist with the complex responsibility of antimicrobial susceptibility testing. This expert system included 87 rules that were subdivided into four groups: (i) unlikely results (Rules U1 through U28), (ii) resistance detection (Rules R1 through R26), (iii) suggestion (Rules S1 through S22), and (iv) quality control (Rules QC1 through QC11). The basis for each rule includes the antimicrobial agent, the test results, and the organism involved. A screen warning displays the reason for the firing of the expert rule and possible actions to ensure an accurate report. Other information given, if applicable, includes methodology recommendations, resistance mechanisms and a printout of results changed by the rule. Printed footnotes provide clinically relevant information for the clinician. All rules have references cited. The expert system was incorporated into the BIOMIC system (Giles Scientific, Inc, New York, NY.) BIOMIC is a susceptibility testing system on disk diffusion methodology. The evaluation of the expert system included 960 organisms: 589 gram negative bacilli and 362 gram positive cocci, representing 11.385 different antimicrobial agent/organism combinations. Using the unlikely result and resistance diction rules, 390/960 (40.6%) of the organisms evaluated fired expert rules. Analysis by the expert rules changed 171/11,385 (1.5%) antimicrobial agent reports. Detection of resistant organisms, including the most common detection problems encountered by the clinical laboratory, was helped by the expert rules. In conclusion, the expert system is a too that assists the microbiologist in education, detection of resistant mechanisms, and flagging of potential errors and problems. An expert system provides clinically relevant information that can have an impact on patient care. I recommend an expert system for susceptibility testing in a clinical microbiology laboratory.
University of Utah
Microbial Sensitivity Tests;
University of Utah;
Relation-Is Version Of
Digital reproduction of “Creation and evaluation of an expert system for disk diffusion antimicrobial susceptibility testing”. Spencer S. Eccles Health Sciences Library.