Predictive model for patient blood management in elective cardiac surgery: blood bag saving 2013-2020

BACKGROUND: Blood shortage is a worldwide persistent emergency: it is a relevant topic in scientific research and policy making. There are several research topics about blood management strategies, but there is not much evidence on the quantity of blood bags needed for an elective cardiac surgery. Once they are unfrozen and prepared, they can't be stored anymore even if unused. This study aims to show how many blood bags are saved with the usage of a model, developed by Siena University researchers, that estimates the need for each elective cardiac surgery. METHODS: Data about the number of blood bags predicted by the model for patients that underwent elective cardiac surgery in University Hospital of Siena from 2013 to 2020 were retrospectively collected. Surgeries for aortic dissection, heart transplants and artificial heart implant were excluded. Before developing the model, the general policy of the hospital was to use about 10 blood bags for each patient. The algorithm establishes the number of blood bags based on HCT, therapy, combined and/or a redo operation, Euroscore2 threshold, comorbidities, and type of operation. 4 extra bags were added to every prediction as safety buffer. RESULTS: This study included 2307 surgical interventions. The sum of all the bags calculated by the model and, thus, prepared for the surgeries, with 4 extra bags as safety buffer per operation, was 13044, compared to the 23070 calculated with the previous criteria. Euroscore 2 score had the closest algorithm prediction to real usage and a mean of 5,4 bags per operation for Euroscore2 <6% and 6,6 for Euroscore2 >6%. Roughly 1.814.706 euros were saved, considering only the rough price of the single blood bag (181€ each). CONCLUSIONS: Machine-learning odels like the one used in this study can improve patient blood management by predicting how many blood bags a patient needs for elective cardiac surgery. KEY MESSAGES: • An algorithm can improve patient blood management by predicting how many blood bags a patient needs for elective cardiac surgery. • Euroscore2, among other criteria such as therapy, comorbidities, and type of operation, improves the accuracy of this algorithm.