The Sankey diagram: An exploratory application of a data visualization tool

Abstract

Hospitals utilize a massive transfusion protocol (MTP) for the management of massive hemorrhage. MTPs not only are associated with improved patient outcomes1 but also cause blood product wastage and a burden on the laboratory.2 To audit recent MTP activations that did not meet our criteria per policy, we utilized data visualization methods to examine our MTP data. The Sankey diagram is a visualization technique, originally used for depicting energy flow, that can be used to understand complex processes. The diagram consists of a series of “Nodes” and “Arcs”, that are read from left to right, with the thickness of each line proportional to the value it is representing.3 This tool has been applied in other clinical settings,3, 4 yet it is underutilized in Transfusion Medicine. The diagram can be created using the free Plotly, pySankey, or Matplotlib Python libraries (available at http://www.python.org) or as we used here, a free, open-source, browser-based application, SankeyMatic (available at http://www.sankeymatic.com), which does not require any programming experience. Development of the diagram involves the labeling of each node, as well as the value assigned to each starting node, branching arc, and destination node. In this diagram, the node on the far left indicates 66 MTP activations, followed by MTP with blood issued, then MTP location. The final nodes address whether the MTP met our criteria. We visualize that 26 of 66 MTP activations did not meet our criteria for massive hemorrhage. Only 77% of MTPs activated in the OR met criteria, compared with 31% activated in the ICU. The Sankey diagram was instrumental in the discussion on the appropriate activation of MTP at our hospital and can be used to visualize data at other centers as well (see Figure 1). The authors have disclosed no conflicts of interest.

Topics & Concepts