gms | German Medical Science

Motivation: In Germany, approximately 15,000 blood reserves are given every day. The University Hospital in Essen (UK-Essen) transfuses 45,000 blood reserves each year: 30,000 red blood cell transfusions, 10,000 platelets transfusions, and 5,000 frozen fresh plasma transfusions [1].

Considering the high demand for blood products, the AutoPiLoT project aims to optimize the use of blood products and minimize the management so that fewer products expire or are discarded. This poster presents one part of the project – a recommendation system providing individualized transfusion recommendations for patients derived from the national cross-sectional transfusion guidelines [2], the Guideline Expert System (GES). It focuses on the technical aspects of the prototype as well as the planned study for a first evaluation of the potential benefits.

Related Work: Studies have shown multiple potential effects of knowledge-based clinical decision support systems (CDSS) [3] – from better adherence to a decreased risk of preventable harm by digital order entries or reduced costs [4]. In 2019, a study of a CDSS implementation in packed red blood cell (PRBC) transfusion in cardiothoracic surgery has shown lower pre-transfusion hemoglobin levels, fewer transfusions, decreased infection rates, and decreased health care costs [5]. GES supports PRBC and thrombocyte transfusion, while the project’s focus primarily is set on the thrombocytes for their reduced preservability and high costs. In the case of GES showing a positive effect on the clinical processes within our first study, the prototype will be further developed into a CDSS also covering the needs of the European Union Medical Device Regulations.

Method: Technically GES will be integrated as a NodeJS Express server application into the existing hospital information system environment of the UK-Essen for close workflow integration. When executed, it requests actual data from the focused patient of SHIP (Smart Healthcare Integration Platform), UK-ESSEN’s FHIR-based data integration platform. A form, prefilled with this information, is displayed for the physician to check, modify, and complete. All communication between back- and frontend is handled via REST-calls. After the confirmation of data by the physician, GES analyzes the submitted information and returns a guideline-compliant transfusion recommendation for the patient. The planned study will answer the following questions:

1.

Are there differences in physician’s decision and guideline recommendation and why?

2.

What benefits does the physician see for daily work and patients' safety?

3.

How many patients are identified as “special cases”, meaning that the guidelines are not generally applicable?

4.

Which further patient data might additional be useful to complement and improve the results of the cross-sectional guidelines?

The study will be performed at the UK-Essen with data on non-intraoperative transfusions of red blood cells and platelets.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.