Comparing transfusion practice at multiple hospitals using electronically collected and analysed data

Comparisons of transfusion practice between organisations are time-consuming using manual methods for data collection. We performed a feasibility study to determine whether large-scale transfusion data from three English hospitals could be combined to allow comparisons of transfusion practice.

Clinical, laboratory and transfusion data from patients discharged between 1 April 2016 and 31 March 2017 were extracted from Patient Administration Systems (PAS), Laboratory Information Management Systems (LIMS), and electronic transfusion systems at three NHS hospitals, which are academic medical centres based in large cities outside London. A centralised database and business intelligence software were used to compare the data.

The dataset contained 748 982 episodes of patient care with 91 410 blood components transfused. The study confirms the results of previous studies finding peaks in the ages of transfusion in the 0-4 years age range, in women of childbearing ages, and in males over 60 years. The number of components transfused per 1000 bed days was used as a standardised comparator. Red cell utilisation was 42.4, 40.4 and 49.5 units/1000 bed days and platelet utilisation 11.69, 7.76, and 11.66 units/1000 bed days. 60.5% (6848/11 310) of Group O D negative red cell units were transfused to non-group O D negative recipients. An analysis of component usage highlighted variations in practice, for example platelet usage for cardiac surgery varied from 2.4% to 7.3% across the three hospitals.

This feasibility study demonstrates that large electronic datasets from hospitals can be combined to identify areas for targeted interventions to improve transfusion practice.