Monitoring the appropriateness of blood transfusion orders is mandated by regulatory agencies as over-transfusion places patients at unnecessary risks, some of them potentially fatal. Transfusion practice oversight poses several challenges, especially those related to meaningful interpretation of large amount of data. Proprietary tools are expensive and typically have a long learning curve. This report describes a novel blood utilization tool developed at our medical center using only in-house resources.

Describe the new knowledge and additional skills the participant will gain after attending your presentation.: Objective
This study aims to assess the feasibility of using a novel blood utilization tool for monitoring transfusion practice at a large academic center.

Assessment of blood utilization at many large hospitals, including our center, frequently involves the review of lengthy reports generated by Laboratory Information Systems (LIS) with limited capabilities to run transfusion related reports displaying granular data in formats that are user friendly for analysis.

We developed TMPy, a programming tool, to extract accurate information from LIS and used it for assessing baseline transfusion practice at UCSF. TMPy is based on Python programming and uses 550 command lines to extract information from pre-existing LIS (Sunquest) monthly blood utilization reports. Information is extracted from two separate LIS reports, one with transfusion data and pre-transfusion specific triggers, and the other with location and provider data. Along with a Data Dictionary, these two reports are integrated into one report, readable in Excel format and amenable to further processing. Deidentified data can be visualized in an interactive histogram via a separate application (transfusionherokuapp.com) that can also extract and display data sub-sets based on specific pre-transfusion trigger range or sort data based on provider, specialty or location.

TMPy was used to generate blood utilization reports for a 3-month period (February-April 2018) to determine baseline transfusion practice. Transfusion orders were extracted from the Excel document and analyzed. Mean pre-transfusion Hb for all RBC orders was calculated. Mean pre-transfusion Hb was also calculated after excluding RBC orders likely associated with surgical procedures or temporally associated with more than three blood product transfusions, scenarios likely representing significant active bleeding.

A total of 5634 orders for 6322 RBC units were placed during the study period. Pre-transfusion Hb was identified in 5223 orders (~83%). Fig 1 shows the distribution of pre-transfusion Hb levels. Average pre-transfusion Hb value for all RBC orders is: 7.97 ± 1.94 g/dL; and slightly lower, 7.72 ± 1.64 g/dL, when RBC orders associated with surgical procedures and active bleeding are excluded from analysis (n= 3209). An example of interactive monthly display in transfusionherokuapp.com application is illustrated in Fig. 2.

TMPy is a user-friendly and flexible tool for the display and analysis of blood product orders with the added ability to correlate metrics with pre-transfusion laboratory values that typically serve as triggers for transfusion. Generation of hospital unit-, specialty- or provider-specific blood utilization data, provides meaningful benchmarks for clinical services and presents opportunities for implementing and tracking service-specific or interdisciplinary patient blood management and quality improvement initiatives.


Jacob Spector (Presenter)
UCSF Medical Center

Ashok Nambiar, UCSF Medical Center
Sara Bakhtary, UCSF Medical Center
Morvarid Moayeri, UCSF Medical Center
Alexandra Tabacu, Stanford University OHS
Russell Thorsen, UCSF Medical Center
Elena Nedelcu, UCSF Medical Center

Presentation Materials: