Reply to: The Implementation of Diagnostic Stewardship for Blood Cultures in Critically Ill Children

The multi-center Bright STAR Collaborative demonstrated an association between diagnostic stewardship and reduction in blood cultures and antibiotics in the pediatric intensive care unit.¹ Dr. Xing and colleagues raise several points that warrant discussion, and we appreciate the opportunity to reply here.

First, we agree with Dr. Xing that sepsis is a clinical diagnosis requiring bedside judgment, and a universally applicable standardized criteria for sepsis does not exist for critically ill children. The Bright STAR Collaborative did not intend to determine whether individual patients had sepsis or to guide treatment of children with suspected sepsis. Bright STAR sites aimed to reduce avoidable blood cultures in children with non-specific symptoms when there was not a suspicion for sepsis. Local teams reviewed patients with positive blood cultures to monitor safety, specifically whether there was a delay in blood culture collection.¹ Reviewing clinical information related to patients with confirmed bacteremia is quite distinct from “adjudication of patients with suspected sepsis.” The Bright STAR sites used tools to guide decision making related to blood cultures, but excluded patients with suspected sepsis. Collaborative teams spent much time discussing how to distinguish non-specific symptoms associated with sepsis and frequent in intensive care unit (ICU) patients, but that are often due to non-sepsis related causes. As expected, the clinical assessment of sepsis risk in individual patients likely varied site by site and provider to provider. This variation existed before and after implementation of this program and thus was unlikely to have systematically influenced the findings.

Second, we also agree with Dr. Xing that the generalizability of our findings to settings such as non-ICUs or ICUs in low- or middle-income countries may be limited. A recent study demonstrated similar impact in a non-ICU setting, but further studies are needed to extend the Bright STAR findings to other environments.² In our ecological study, we did not evaluate patient-level data, limiting our ability to examine the impact of diagnostic stewardship on an individual patient, such as their risk of mortality, length of stay or readmission. We also did not collect detailed microbiologic data from each site to compare before and after implementation, though this could be included in future studies. As we emphasized in the closing section of our manuscript, we strongly believe that continued monitoring of balancing metrics is critically important as diagnostic stewardship efforts spread to additional settings in the future.

Finally, we acknowledge that the importance of an initiative to reduce blood cultures may vary significantly depending on the setting. Overuse of tests may not be present in all settings, owing to resource limitations or differences in local practices. Practice change in pediatric ICUs should be guided by local context, prioritize the needs of local patients and key stakeholders, and be designed with safety and sustainability as the ultimate goal.³ We are in the process of writing a separate manuscript focused on these aspects of implementation and look forward to sharing those results with the pediatric critical care community soon.