Development of transfusion guidelines for injured children using a Modified Delphi Consensus Process.

Background: There is wide variability of transfusion practices for children with hemorrhagic injuries across trauma centers. We are planning a multicenter, randomized clinical trial evaluating tranexamic acid in children with hemorrhage. Standardization of transfusion practices across sites is important to minimize confounding. Therefore, we sought to generate consensus-based transfusion guidelines for the trial.

Methods: We used a modified Delphi process utilizing a multi-site, multi-disciplinary panel of experts to develop our transfusion guidelines. A survey of 23 clinical categories on various aspects of transfusion practices was developed and distributed via SurveyMonkey®. Statements were graded on a 5-point Likert scale ("Strongly agree" to "This intervention may be harmful"). Statements were accepted if ≥ 80% of the panelists rated the statement as "Strongly agree" or "Agree". After each round, the responses were calculated and the results included on subsequent rounds.

Results: 35 panelists from four pediatric trauma centers participated in the study, including 11 (31%) pediatric EM physicians, 8 (23%) pediatric trauma surgeons, 5 (14%) transfusionists, 5 (14%) pediatric anesthesiologists, and 6 (17%) pediatric critical care physicians (range of 8 to 10 from each clinical site). Four survey iterations were performed. In total 176 statements were rated and 39 were accepted by criteria across all 23 categories. An rational algorithm for transfusion in trauma was then developed.

Conclusions: We successfully developed transfusion guidelines for various aspects of the management of children with hemorrhagic injuries using a modified Delphi process with broad interdisciplinary participation. We anticipate implementation of these guidelines will help minimize heterogeneity of transfusion practices across clinical sites for the upcoming clinical trial evaluating tranexamic acid in children with hemorrhage.