Class of 2008 Abstract / Objectives: A novel method of predicting vancomycin trough levels at steady state was studied to determine whether it could effectively predict vancomycin trough levels compared to an established predictor method (Matzke). Methods: Adult patients who received at least two consecutive doses of vancomycin and had at least one reported vancomycin trough at steady state were considered. Data extracted and analyzed included patient gender, age, weight, height, and serum creatinine as well as vancomycin dose and interval, number of consecutive doses prior to the trough, time between trough and preceding dose, and measured vancomycin trough level. This data was applied to each of the prediction methods to determine how accurately they predicted actual measured vancomycin trough levels at steady state. Results: Data from 103 patients was analyzed. Vancomycin trough predictions using the Bergstrom method averaged 12.2 mg/dl, with a standard deviation of 3.4. The average actual trough concentration was 10.7 mg/dl with a standard deviation of 3.9, while the Matzke method predicted an average trough concentration of 19.2 mg/dl with a standard deviation of 8.6. Predictions made using the Bergstrom Method were not significantly different than the actual trough concentrations (p = 0.91). The Bergstrom method predicted concentrations within 25% of actual concentrations 42% of the time and within 50% of actual concentrations 78% of the time. Conclusions: The Bergstrom method was a more reliable predictor of vancomycin trough concentrations than the Matzke method in this patient population. Although more research is needed, the Bergstrom method may prove to be a useful tool for pharmacists to predict vancomycin trough concentrations quickly and with relative accuracy for individual patients.