A controlled trial of the cost benefit of computerized bayesian aminoglycoside administration.

We studied the effect of a bayesian pharmacokinetic dosing program on the outcome of aminoglycoside therapy in patients with clinical infections. Patients were randomized to a control (dosing based on physician choice; n = 75) or experimental group (dosing based on the bayesian program; n = 72). Both groups used serum aminoglycoside concentration data when making dosing decisions. Improved response rates were seen in the experimental (60%; 42/68) compared with the control group (48%; 36/68). A higher, but not statistically significant, incidence of toxicity was found in the control (7/75; 9.7%) versus the experimental group (4/72; 5.1%). Mean length of total hospital stay was significantly longer for patients in the control group (20.3 days) compared with the experimental group (16.0 days) (p = 0.028). The variables from multivariate analysis with a significant impact on length of stay were patient group and length of aminoglycoside therapy. On the basis of a reduced length of stay, a potential cost savings of $1311 per patient can be achieved.