|| Checking for direct PDF access through Ovid
Through a heuristic and probabilistic approach, we evaluated blood culture operating characteristics (sensitivity, specificity, and predictive values) as a function of several pretest parameters, together with their variability. On the basis of a meta-analysis of quantitative data from the literature, a model was developed and an estimation of the operating characteristics through numerical simulations (Monte Carlo method) was performed. The model evaluates the influence of ordering and drawing parameters on the ability of blood culture to distinguish bacteremic from nonbacteremic patients, regardless of the causative species. By considering the total blood volume to be cultured (six 5-10-mL bottles), results were found to confirm the current guidelines. On the basis of this hypothesis, the results, together with an analysis of the literature, failed to show any benefit of a strategy that involves obtaining multiple samples. The best strategy when performing blood culture is to obtain blood for 6 bottles (for a total volume of 35-42 mL), preferably at the same time.