OBJECTIVES: Simple and complex computer MMs are being used to research infectious disease problems. Analysis of in vitro AB activity is usually limited to simple comparison of National Committee for Clinical Laboratory Standards (NCCLS) tube dilution SUS and MIC90. Similar ABs with similar MIC90s for a bacterium are assumed equal in activity. This mimics the problem of analyzing the survival time of patients, which is solved by statistical methods like Kaplan Meier survival table (KM) analysis. Using AB dilution instead of time in KM analyses may identify differences in SUS that cannot be found by traditional methods. A MM was developed to test this assumption. METHODS: Computer simulations were done to illustrate scenarios where statistical modeling would find differences between antibiotics with identical MIC50, MIC90 and % SUS. The Base2 logarithm of AB dilution was substituted for time in KM analyses. The KM analyses detected differences in AB activity (p<0.00005 to 0.0281) because KM compares across the entire range of dilutions instead of only at single points like MIC50 and MIC90. After developing the application with computer simulation the method was tested by comparing MIC data for piperacillin/tazobactam (PT) versus ticarcillin/clavulanate (TC) against Klebsiella pneumoniae (4,784 isolates) and Morganella morganii (719 isolates) from a previous study (ICAAC abstract A-57, 1996). RESULTS: In spite of identical MIC90 and % SUS using NCCLS criteria the MM of survival using KM analyses showed TC superior against Klebsiella pneumoniae and PT superior against Morganella morganii (both p<0.00005). CONCLUSION: Application of this Mathematical Model and Kaplin Meier survival table analyses can discover differences in antibiotic activity hidden from traditional methods and may improve the clinical utility of in vitro data.