(P-QU-13) Reducing Group and Crossmatch Errors in a Tertiary Academic Hospital in Singapore

Group and Crossmatch (GXM) testing is common and necessary prior to blood transfusions. However, it is prone to errors which may have fatal consequences including haemolytic transfusion reactions. Mistakes during collection of GXM samples contribute significantly to such errors, many being due to human factors.

The National University Hospital is a tertiary academic hospital in Singapore. Our blood bank has noticed increasing GXM error rates in recent years, especially after the switch to a new Electronic Medical Records (EMR) system in 2022. The mean error rate was 1.18% between 2016 to 2021. After switching, it increased to 1.63% between January 2022 to September 2023. This project was designed to identify the root causes, and implement solutions to reduce the GXM error rate.

Study

Design/Methods:

Errors were classified into major errors (e.g. wrong blood in tube), minor errors of significant impact (e.g. unlabelled sample), and minor errors of less significant impact (e.g. no collector signature). Through a multi-disciplinary team, multiple quality improvement methodologies were applied to identify the causes and implement solutions.

Root causes were identified through Process Mapping, an Affinity Diagram, and a Cause and Effect Diagram. A Pareto Chart was used to prioritise various causes. Interventions were designed and implemented using the Plan-Do-Study-Act model. Six interventions were explored, of which five were implemented. These targeted a broad range of areas, including behavioural change through new reminders, simplification of existing reminders, improving education of the GXM sample collection process, and punitive measures. Technology was leveraged through EMR system changes, incorporation of reminders in the EMR, and enhancements of the pre-existing electronic GXM sample labels. Interventions were implemented from October 2023 onwards.

Sampling error data was prospectively collected from October 2023 and compared to data collected prior to implementation of the measures.

Results/Findings: There was a total of 23,027 GXM samples tested between April to September 2023, with an error rate of 1.35%. Rate of major errors was 0.01%, rate of minor errors of significant impact was 0.07%, and rate of minor errors of less significant impact was 1.26%. Following the implementation of this project’s new measures, with a total of 23,614 GXM samples between October 2023 to March 2024, the total error rate decreased to 0.73% (p< 0.01). During this period, the rate of major errors remained at 0.01% (p=0.68), rate of minor errors of significant impact remained at 0.07% (p=0.92), and rate of minor errors of less significant impact decreased to 0.65% (p< 0.01).

Conclusions: Our project demonstrates that quality improvement methodologies and leveraging of technology can result in significant reduction in GXM error rates with meaningful improvement in the quality and safety of blood transfusion programmes.