The objectives of this study were to estimate the prevalence of low or excessive vancomycin dosing after initiation of treatment in pediatric patients and to determine the factors that are most predictive of optimized vancomycin dosage in this group. Among 74 patients who underwent vancomycin concentration monitoring, low trough (< 4.0 micrograms/ml) and/or peak (< 15.0 micrograms/ml) concentrations were noted in 28 (38%) patients after the initiation of therapy but in only four of the 28 (14%) patients (p = 0.29) after optimization of the initial dosage. There were not toxic peak concentrations (> 60 micrograms/ml) reported during the study. In patients older than 1 month old, 11 low peaks were associated with troughs less than 7.5 micrograms/ml, whereas no low peaks were associated with troughs more than 7.5 micrograms/ml. The significant predictive variables of optimized vancomycin dosage in the reduced regression model (p < 0.00001; adjusted r2 =0.85; n = 36) were (log) initial dose (p < 0.0001), initial trough (p < 0.0001), and age (p = 0.009). Initial peak concentrations were not associated with the optimized dosage (p = 0.50). The results of this study indicate that approximately 40% of all pediatric patients will be at risk of significant underdosing if standard vancomycin dosing guidelines are followed and that patients older than 1 month old with initial trough concentrations less than 7.5 micrograms/ml are at a greater risk of low peak concentrations than individuals with trough concentrations more than 7.5 micrograms/ml. Monitoring vancomycin concentrations appears to be essential to prevent the underdosing of many pediatric patients; however, only initial trough vancomycin concentrations may be needed to optimize dosages.