To the Editor.
—In their prospective cohort study, Bates and Lee1 developed a model that may help in the differentiation of true-positive bacterial cultures from false-positive cultures within 1 day of culture. Their study design avoided most common biases and the use of a logistic function was an appropriate and powerful tool.There are a few areas that need clarification. In Table 6, "Organism category 2" was identified as an independent predictor of bacteremia and assigned "points," yet the 95% confidence interval for the odds ratio crossed one (0.9-17), which indicates nonsignificance. The number of points assigned to "Time until blood culture became positive" are apparently 8, 6,4, and 2 for 1,2,3, and 4 days, respectively. Also, for "Risk category," which is based on clinical-risk score, the points are apparently 0,1, 2, and 3 for groups 1, 2, 3, and 4, respectively. These numbers are important if clinicians are