The predictive ability of a four-point sampling method versus a two-point sampling method for vancomycin was assessed in 11 patients with various staphylococcal infections. All steady-state predictions were based on first-dose pharmacokinetic parameters. The mean vancomycin serum concentrations achieved at 4 and 8 h postinfusion were not significantly different from the predicted concentrations derived from either the four- or two-point method. Also, there was no significant difference between the two methods in predictive ability or accuracy. Both methods underpredicted the steady-state concentration to the same degree, 2.9 micrograms/ml at 4 h and 3.1 micrograms/ml at 8 h, which would appear to be clinically acceptable. A one-compartment pharmacokinetic model, which uses two serum concentrations, appears to be adequate for adjusting vancomycin regimens.