Errors in blood transfusion in Britain: survey of hospital haematology departments

BMJ 1994; 308 doi: (Published 07 May 1994) Cite this as: BMJ 1994;308:1205
  1. D B L McClelland,
  2. P Phillips
  1. Edinburgh and South East Scotland, Blood Transfusion Service, Royal Infirmary of Edinburgh, Edinburgh EH3 9HB
  1. Correspondence to: Dr McClelland.

    Blood transfusion laboratories aim to provide a high quality service with minimum risk to patients. British guidelines for good practice in transfusion medicine exist,1 and most hospitals have local protocols. If these procedures fail incompatible blood may be transfused, which could lead to potentially fatal haemolytic reactions. As no system of collecting data centrally exists in Britain failures of the transfusion process are not documented. In contrast, in the United States the Food and Drug Administration requires all establishments that are registered to process blood to report all errors and deaths associated with transfusion. We aimed to investigate the incidence of recognised transfusion errors in Britain in 1990 and 1991 and the cause and clinical outcome of these errors.


    and results

    A short questionnaire about errors in blood transfusion procedures and the outcome of these errors was sent to the 400 hospital haematology laboratories in Britain in August 1992. In all, 245 (61%) laboratories responded: these supplied 3.3 million red cell units for transfusion (about three quarters of all the red cell and whole blood units collected annually in Britain). A third of responding laboratories reported incidents in which patients received the wrong blood. The table shows the results of the survey.


    The error rates that we found are similar to those reported in studies from the United States.2,3 Our data do not allow the calculation of error rates per patient transfused, which must be substantially higher than the rates in the table since most patients receive several units of blood. Several respondents indicated that multiple errors had contributed to the wrong blood being transfused; similar findings have been reported elsewhere.4

    Twenty respondents reported (without having been asked in the questionnaire) 100 incidents in which the wrong blood sample was submitted in the compatibility tube and the error was detected in the laboratory because of a previous blood sample on the same patient. On the basis of this information and comments volunteered by other respondents, we estimate that the incidence of wrong blood being submitted in tubes is about 1/6000 red cell units issued.

    Only a third of unmatched transfusions are incompatible with ABO blood groups; of these, only about a tenth are associated with a fatal outcome.4 We should not, however, be complacent as these figures emphasise that data on mortality and morbidity, even if complete, can give only a substantial underestimate of the incidence of important failures in the transfusion process.

    The data available are inadequate to determine the true incidence of errors in transfusion. All the errors found in this survey were reported by only a third of the responding laboratories; it would be surprising if the remaining laboratories had experienced no errors over two years.

    We propose several ways of improving the quality and safety of the blood transfusion process in Britain. Firstly, a national system should exist for reporting critical transfusion incidents, especially those in which the wrong blood is transfused and “near misses.” Regular reports to transfusion laboratories and hospital transfusion committees could be incorporated in the national external quality assurance scheme. Secondly, all hospitals should establish clear and coordinated managerial responsibility for the transfusion process. Thirdly, transfusion laboratories should be required to have a procedure for recording all transfusion errors and the corrective action taken and to report regularly to the proposed national scheme. Fourthly, statistics on the number of blood units transfused and the number of patients who receive transfusions should be collected and reported nationally to provide denominators for any reporting scheme. Finally, pilot projects should be set up to identify and report cost effective ways of improving the safety of the clinical transfusion process.

    Results of questionnaire about errors in blood transfusion sent to 400 hospital haematology departments, of which 245 responded

    View this table:


    We thank the organisers of the national external quality assurance scheme in haematology, which distributed the questionnaire and the final report to all hospital transfusion laboratories in Britain.


    View Abstract