Practical Transfusion Medicine 4th Ed.

23. Blood transfusion in hospitals

Erica M. Wood1, Mark H. Yazer2 & Michael F. Murphy3

1 Monash Medical Centre, Departments of Clinical Haematology and Epidemiology and Preventive Medicine, Monash University, Melbourne, Victoria, Australia

2 The Institute for Transfusion Medicine, Pittsburgh and Department of Pathology, University of Pittsburgh, Pittsburgh, Pennsylvania, USA

3 University of Oxford and NHS Blood and Transplant and Department of Haematology, John Radcliffe Hospital, Oxford, UK

The aim of transfusion practice is to provide ‘the right blood to the right patient at the right time for the right reason’. It focuses on ensuring that, when it is clinically indicated, patients receive the correct transfusion support, in a safe, timely and cost-efficient manner, with incidents and adverse reactions recognized and managed effectively. Specialists in all branches of medicine and surgery are involved in prescribing blood, and engagement, cooperation and coordination is required by staff, and with patients, to manage the complex, interacting sequences of the process.

In many countries, blood for transfusion is neither safe, sufficient, nor reliably available. In these settings, haemorrhage is a major direct cause of mortality. However, even in modern healthcare settings with adequate blood supplies, patients die from transfusion complications [1,2] or from lack of adequate transfusion support: for example, massive haemorrhage is still one of the most common direct causes of maternal death worldwide. Some instances of undertransfusion are due to patient refusal to accept transfusion support or failure by physicians to recognize and respond to clinical manifestations of bleeding. However, others can be attributed to either lack of knowledge of transfusion protocols or failures of communication within, and between, clinical teams. At the other extreme, patients are frequently overtransfused and transfusion-associated circulatory overload (TACO) is increasingly recognized as a common serious adverse event [3].

Mistransfusion, or ‘wrong blood’ events, i.e. administering an incorrect unit of blood, which either does not meet the patient's needs or is intended for another recipient, can also have serious consequences, including severe haemolysis due to ABO incompatibility, and is another well-recognized cause of mortality and morbidity. Human errors leading to mistransfusion can occur at any step in the process and usually result from failures to comply with clerical or technical procedures, or systems that are either poorly constructed or not understood. Multiple errors are frequently involved in these cases. Some can be detected during the bedside check at the time of administering blood, and this remains a final opportunity to prevent mistransfusion. It has been observed that as many as 1 in 19 000 red cells are given erroneously and 1 in 33 000 will involve ABO-incompatible units [4]. Estimates of mortality due to mistransfusion range from 1 in 600 000 units to 1 in 1.8 million.

Enormous investments have been made to reduce the risks of transfusion-transmitted infections, but to date generally there has been much less investment in improving hospital systems required for clinical practice. Consequently, evidence of progress in reducing procedural risks and improving the safety of hospital transfusion practice is slower to accumulate. Some interventions, such as the practice of a bedside ABO group check before transfusion or the use of physical barriers to transfusion, such as a code to link the patient's wristband, pretransfusion sample and unlock the designated unit of blood from secure storage, are intrinsically attractive. For a variety of reasons they have been difficult to implement fully and are not yet widely used [5]. Data to support the effectiveness of many procedural interventions are still limited and many serious (and often preventable) adverse events continue to be reported. However, where haemovigilance programmes have been able to highlight these issues and their causes, and action has been taken to address them, progress has been demonstrated in at least some of these areas [1].

Effective quality frameworks are required to minimize transfusion risks and to ensure that the supply of donated blood is managed effectively. These in turn require a patient-centred approach to transfusion, committed leadership and adequate resources.

Key features of hospital transfusion governance

Many countries have established requirements that blood centres or services (termed ‘blood establishments’ in Europe) and hospital transfusion laboratories maintain robust quality systems to ensure good practice – which may include meeting national or regional standards for good manufacturing, laboratory and/or clinical practice (also see Chapter 17). These requirements are typically overseen by national regulatory authorities and/or professional authorities to ensure compliance. For example, to meet EU Directives, the UK Blood Safety and Quality Regulations [6] outline requirements for quality management in transfusion laboratories, including staff training, process validation, documentation, storage and handling, traceability and reporting of adverse events.

However, this regulation has typically not yet extended to the practice in clinical areas and other measures are needed to ensure that processes and systems that influence the quality and governance of clinical transfusion practice at the hospital level are optimized and working as expected. Recommendations for the practice are derived from clinical experience and the peer-reviewed evidence base, along with lessons from haemovigilance and external quality assessment (EQA) schemes. These are translated into policies, standards and guidelines by government agencies and professional groups, who in turn assess implementation and compliance and promote best practice through training, education and communication.

In England, the Care Quality Commission regulates healthcare providers. Many of the national standards for governance and risk assessment are applicable to blood transfusion, including those relating to patient engagement, informed consent, staff training and competency assessment, participation in audit and other quality improvement activities, and reporting of incidents. Specifically, the National Patient Safety Agency in England in 2006 issued a safety notice requiring competency-based training and assessment for all staff involved in blood transfusion, a bedside identity check for administering blood that matches the blood pack with the patient wristband (excluding compatibility form or case notes) and a formal risk assessment of the alternative means of confirming patient identity. The National Health Service Litigation Authority also inspects acute care English hospitals against risk management standards that include blood transfusion. Hospitals are expected to have transfusion policies and provide evidence of implementation and monitoring for effectiveness.

Similar expectations apply in other countries. For example, the Australian Commission on Safety and Quality in Healthcare recently released a standard on clinical transfusion practice as part of new national safety and quality standards [7] that outline requirements against which hospitals will be assessed for accreditation. In the USA, authorities from state health departments through to national regulators like the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) and the Food and Drug Authority are involved in assessing and regulating transfusion practice.

At an institutional level, executive management is responsible for implementation of standards and policies, including through initial and ongoing staff training and appraisal, and for monitoring through clinical audit and other quality system activities.

Hospital transfusion committees

Hospital transfusion committees (HTC) are focal points for overseeing transfusion practice at the institutional level. Their roles are outlined in documents such as the UK NHS Better Blood Transfusion health circulars and statements from regulatory bodies like the US JCAHO.

To be effective and to deliver on their objectives, HTCs require support from dedicated hospital transfusion teams, at a minimum consisting of a medical specialist, transfusion practitioner(s) and blood bank scientist/manager. Other necessary resources include IT and clerical support to facilitate regular meetings, data retrieval and audit. HTCs are essential components of institutional clinical governance, so they must be incorporated into hospital frameworks for clinical governance, performance and risk management, and report findings and activities in a timely and meaningful way, accompanied by recommendations for action. HTCs have the remit to promote best practice, review clinical transfusion practice, monitor performance of the hospital transfusion service, participate in regional or national initiatives and communicate with local patient representative groups as appropriate (Table 23.1).

Table 23.1 Activities of the hospital transfusion committee.

Area or activity


Policies and procedures

Develop and promulgate policies and procedures, including for:

·        Clinical indication and decision to transfuse

·        Establishing and enforcing transfusion thresholds

·        Informed consent process

·        Collection of samples for compatibility testing, including patient identification and specimen labelling requirements

·        Transfusion administration and monitoring

·        Indications for specialized components (e.g. irradiated, CMV seronegative, phenotype-matched)

·        Maximum surgical blood ordering schedule (MSBOS)

·        Blood conservation strategies including use of cell salvage and pharmacological agents

·        Management of patients who decline transfusion

·        Management of adverse reactions

Education, training and assessment

Develop strategies for education, training and assessment of all staff involved in transfusion 

Monitor implementation and results of education and training activities 

Develop and/or promulgate information/materials for patients

Audit, monitoring and review

Develop annual audit plans and monitor performance 

Review adverse event reports 

Conduct incident, ‘near miss’ and sentinel event reviews 

Oversee traceability and record-keeping obligations

System performance


·        Blood component availability, utilization and wastage rates

·        Activation of massive transfusion protocol, use of uncrossmatched emergency red cell stocks

·        Performance of institutional transfusion laboratory and blood service (‘blood establishment’)

·        Participation in regional and national audit, transfusion practice improvement and haemovigilance programme activities

·        Participation in EQA activities

Oversee hospital and laboratory accreditation activities relating to transfusion Contingency and disaster planning 

Function and performance of HTC


A chairperson with understanding and experience of transfusion practice should be appointed by hospital senior management. Ideally, the chairperson should not be the medical specialist responsible for the hospital transfusion service, who could be perceived to have a vested interest.

The following membership is suggested:

·        representatives of all major clinical blood users, including junior medical staff;

·        specialist haematologist/pathologist with responsibility for transfusion;

·        hospital blood bank senior scientist/manager;

·        specialist practitioner(s) of transfusion;

·        senior nursing representative;

·        representatives from hospital management and clinical risk management;

·        local blood centre medical specialist (ex officio); and

·        other co-opted representatives as required, e.g. from medical records, portering staff, clinical audit, training or pharmacy.

Administration of blood and blood components and management of the transfused patient

This process involves multiple steps:

·        counselling patients regarding the need for blood transfusion, when alternative approaches (salvaged blood, iron and/or erythropoietic stimulating agents) are predicted to be insufficient or inappropriate for their circumstances;

·        prescription;

·        requests for blood components;

·        sampling for pretransfusion compatibility testing;

·        collection and delivery from approved storage facility (e.g. monitored blood refrigerator) to clinical area;

·        pretransfusion checking process;

·        administration;

·        patient monitoring; and

·        documentation of all steps.

Errors occurring at blood sampling, collection and administration can lead to patient misidentification and mistransfusion. Prescription errors, however, lead either to failure to provide special components to meet recipient special needs or to transfusions that are unnecessary or inappropriate and carry the potential for complications. For example, TACO has occurred when transfusions have been given on the basis of a spuriously low haemoglobin value resulting from samples taken from IV (‘drip’) arms or measured by gas analysers, or as a result of a clerical error where the decision was based incorrectly on another patient's results. Fatal errors have also occurred in prescribing the volume or rate of transfusion. Failure to monitor transfused patients, particularly in the first 15 minutes of receiving each unit, can lead to life-threatening reactions being overlooked and delays in resuscitation.

Hospitals should have written procedures to cover all these steps, against which relevant staff are trained and assessed, and which are readily available for quick reference at the bedside. Clinical responsibilities, actions, documents, potential errors and some of their consequences are outlined in Tables 23.2 to 23.8. Prescription charts, donation numbers of components and batch/lot numbers of fractionated plasma products issued and transfused, nursing observations and recipient vital signs related to the transfusion should be kept in the medical case notes as permanent records. Regulatory and accreditation authorities require a complete audit trail of blood to the patient's bedside. Many hospitals comply with this requirement by returning signed and dated compatibility forms or compatibility labels to the transfusion laboratory.

Table 23.2 Examples of some errors and other problems in the transfusion process, and their potential outcomes.


Potential outcome

Unnecessary prescription

Patient subjected to unnecessary risks, including transfusion-associated circulatory overload


Blood component wastage

Prescribed components do not meet patient special requirements

Transfusion complications, e.g. transfusion-associated graft-versus-host disease

Blood not stored in controlled environment

Blood component wastage 

Transfusion complications (e.g. risk of bacterial growth)

Pretransfusion samples taken from incorrect patient 

Sample transposition or other laboratory errors 

Incorrect unit of blood collected and/or administered

Mistransfusion and potential for ABO- or RhD-incompatible transfusion

Insensitive techniques in pretransfusion testing

Potential for acute and delayed haemolytic transfusion reactions

Poor laboratory stock control

Blood component wastage


Inappropriate overuse of group O red cells and potential for consequent shortages of that group

Delay in emergency provision of blood components

Patient morbidity/mortality due to hypoxia or coagulopathy

Table 23.3 Prescription of blood components.


Table 23.4 Requests for blood and blood components.


Table 23.5 Sampling for pretransfusion compatibility testing.


Table 23.6 Collection and delivery of blood components from transfusion storage facility to clinical area.


Table 23.7 Administration of blood components.


Table 23.8 Monitoring of transfused patients.


Technologies to reduce patient misidentification errors in administering blood

Additional manual systems of patient identification

Sets of distinctive (e.g. coloured) labels with the same unique number can be allocated to each pretransfusion blood sample. A transfusion label can be incorporated into an additional patient wristband at the time of phlebotomy, affixed to the request form, sample tube and into the current medical notes, and the unique number can also be printed on to the compatibility labels and compatibility report form. At the time of administration, the additional unique number provides a supplementary means of cross-checking.

Electronic bedside processes for safe transfusion practice

The use of bedside hand-held computers, barcoded staff identity badges, barcoded printed wristbands for patients and portable printers for sample tube labels provide the means for improving patient identity and safety [8]. For example, at sample collection, phlebotomist and patient identity can be scanned and barcoded labels generated at the bedside to attach to the sample tube at the time and place where it is collected. In the laboratory, allocated units are labelled to incorporate the patient's unique identification barcode and the unit number. At administration, staff are prompted by a hand-held computer to scan their own identification barcode, the barcoded patient wristband, the compatibility label and the unit number on the blood component. The computer prompts the staff to check the identity of the (conscious) patient verbally and the barcode scans confirm that the unit is correct for the patient. The user and transfusion laboratory are alerted if there is a mismatch. It also provides prompts to check for special requirements, pretransfusion observations and the unit expiry date. Documentation of each step is transmitted to the laboratory information system to confirm traceability of the unit and the competency of staff in safe transfusion practice.

Electronic bedside systems can be linked to similar systems controlling release of blood from remote blood refrigerators to provide full electronic process control and to facilitate electronically controlled remote issue (see later in this chapter and Reference [9]).

Electronic systems for blood transfusion are increasingly being implemented, although further studies are needed to confirm cost-effectiveness. The rationale for their use would be even greater if they were integrated with other processes requiring patient identification, such as medication administration.

Influencing clinical practice

Potential factors influencing transfusion practice and decision making include:

·        physician knowledge and perception based on clinical experience;

·        peer pressure and feedback;

·        effectiveness of hospital governance frameworks;

·        educational prompts at the time of decision making;

·        patient knowledge and preferences [10];

·        financial pressures or incentives;

·        public and political perceptions; and

·        fear of litigation.

Improving transfusion practice within a hospital community requires a planned, consistent approach, endorsed and implemented through clinical governance frameworks, supported over time and monitored for effect.

Guidelines, algorithms and protocols

Guidelines are systematically developed statements to assist practitioner and patient decision making about appropriate healthcare for specific clinical circumstances. A list of websites with some examples of guidelines is included at the end of this chapter. Data from randomized controlled trials are generally not available to assess the impact of professional guidelines, but even national guidelines rarely lead to change without local implementation and dissemination strategies, and these require time and resources.

Developing an institutional strategy to implement guidelines is a useful opportunity to gain ownership and participation. For example, educational opportunities arise from examining the evidence basis for the guidelines, and dissent and other local barriers to implementation, such as limited staff or IT resources, or effects on laboratory turnaround times, can be identified.

Institutions should adopt recommendations from authoritative professional guidelines and carefully review their content to consider whether any customization is required for local use. This may involve separating guidelines into sections and/or incorporating some recommendations into other local protocols for specific conditions, e.g. a fresh frozen plasma guideline incorporated into protocols for management of disseminated intravascular coagulation and massive haemorrhage. These local documents should be incorporated into transfusion policies and disseminated, with training, for all involved staff.

Experience in other medical fields has demonstrated that embedding guideline recommendations into materials used during transfusion decision making and administration processes can significantly improve compliance. Examples include:

·        listing indications for special blood components on transfusion request forms or electronic request screens;

·        using electronic warning systems to alert prescribers when, based on laboratory values, planned transfusions do not meet guidelines (see below);

·        listing, on specific transfusion observation charts, actions to be taken in the event of reactions; and

·        detailing checks to be made on the compatibility form prior to administering blood.

Intraoperative algorithms for the use of platelets and plasma to correct microvascular bleeding during and after cardiac bypass surgery have also proved to be successful in reducing inappropriate use of these components, especially when combined with near-patient testing and rapid availability of results.

Clinical audit

Clinical audit is a quality improvement process that seeks to improve patient care and clinical outcomes, through systematic review of care against explicit criteria or standards, followed by the implementation of change. Analysis of audit findings can lead to recommendations for improvements when deficiencies or non-guideline-based practices are identified, in turn generating cycles where feedback and clarification of hospital policies lead to improved practice.

Audits can be conducted retrospectively or concurrently. Retrospective transfusion audits are often performed under the auspices of the HTC. Some regulatory agencies require a certain percentage of all transfusions to be reviewed by the HTC and those felt to have been administered without reasonable justification brought to the committee's attention. If, from available data, a transfusion is felt to be egregious, further information should be requested from the responsible physician. If the explanation is inadequate, or if the physician fails to reply, other steps, such as letters to department chairs, can also be taken. Advantages of this type of review are that communications from the HTC carry additional weight and they can be educational tools to inform physicians of institutional protocols. The main disadvantages are the limited number of transfusion episodes that can be audited and, because audits are performed after the event, educational opportunities are lost if the staff who ordered the transfusion cannot be located or cannot recall the event. Retrospective audits also cannot influence clinical practice for the episodes being audited.

Audits performed concurrently with blood component ordering, but before product issue, can take several forms. A simple example involves transfusion laboratory staff comparing component orders with hospital guidelines; if criteria are not met, the ordering physician is contacted, the reasons for ordering the transfusion discussed and plans established. Intervention by transfusion medicine physicians has been demonstrated to be effective in reducing unnecessary transfusions [11]. Audits of this type have been criticized for potentially causing delay in providing necessary products (although they would also prevent unnecessary transfusions before they were administered). Significant time, effort and good communication are required to make these audits effective.

Another approach to concurrent audit involves automation to warn clinicians at the time of ordering that the transfusion might not be necessary. Where a hospital uses computerized order entry and institutional guidelines are in place, warnings can appear on the screen when physicians try to order blood for patients whose laboratory values suggest that transfusion is not indicated. Figure 23.1 demonstrates the response where a physician attempts to order red cells for a patient whose latest haemoglobin value is above the threshold set by the HTC. The warning appears, giving the physician the option of either cancelling the order or proceeding, depending on the patient's current clinical situation (which might not be accurately reflected in a historical laboratory value). In the first few months after these warnings were instituted at hospitals of the University of Pittsburgh Medical Center (Pittsburgh, USA), cancellation was observed for 12% and over 25% of non-guideline-based red cell and plasma orders, respectively. The ability to track individual physicians and hospital locations that generate the greatest number of warnings also supports provision of focused education.

Fig 23.1 Warning message displayed when a physician at a University of Pittsburgh Medical Center (UPMC, Pittsburgh, PA) hospital attempts to order red cells using the computerized order entry system on a patient whose most recent haemoglobin value is in excess of the institutional guidelines.


Many countries have regional or national clinical audit programmes, with participation being either voluntary or, increasingly, mandated by accreditation or governmental agencies. Participation provides opportunities to compare performance between similar institutions for the purposes of benchmarking and to promote engagement in practice improvement activities more broadly. The UK national audit programme and several of the practice improvement collaboratives in Australia have made their audit tools available to invite collaboration, comparisons of practice and sharing of resources.


Many activities that fall under an ‘audit’ banner are not comparing practice with a standard, but are monitoring or surveying practice. These activities, many of which can be quantified, often provide information and baseline data that can lead to the development of quality or performance indicators. Trend analysis, or comparison of organizations or blood users with each other, is a powerful means of exerting peer pressure and influencing practice (benchmarking, as above).

Performance indicators can be applied to:

·        clinical and laboratory practice issues: e.g. percentage of primary arthroplasties requiring allogeneic transfusion; proportion of patients receiving platelets after coronary artery bypass grafting; red cell use by surgical procedure (by surgeon or unit); or percentage of anaemic patients being investigated, correctly diagnosed and managed appropriately to minimize unnecessary transfusions; and

·        process issues: e.g. percentage of mislabelled samples received in the laboratory, patient wristband errors; numbers of units crossmatched to units transfused (C : T) ratio; hospital blood wastage; percentage of group O red cells used.

National schemes

Many countries now have national schemes to monitor transfusion practice and promote practice improvement. These may be voluntary or mandatory, and institutions may be anonymous or identified. The programmes can be used to influence policy at national and local level, and to educate clinicians. Examples include:

·        Haemovigilance programmes. The UK SHOT scheme is a voluntary system for collecting data on serious transfusion adverse events and near misses. It produces annual reports with recommendations. Many other regional and national examples exist and experiences presented in these reports have been very valuable in identifying areas for improvement. A national haemovigilance programme in the USA with voluntary hospital participation was launched in 2010 as a joint effort between the AABB and the Centers for Disease Control (Chapter 18).

·        EQA schemes. These programmes periodically provide clinical material to be tested by transfusion laboratories. Results are returned for analysis and collated reports disseminated to participants.

·        Utilization and wastage schemes, such as the UK Blood Stocks Management Scheme, collate and publish details of blood stock inventory and wastage, and allow participants to benchmark against comparable hospitals.

Patient knowledge and preferences, public and political perceptions and fear of litigation

Many patients, and indeed the general public, have a very limited understanding of the true benefits and risks of transfusion and may consequently have considerable anxieties about transfusion. Patients who have received a transfusion often do not recall the consent process, either because they were not given full information or because they rapidly forgot it. Communication with both transfusion recipients and the general public needs to be improved. Involvement of patients in decision making about transfusion and the safety of transfusion procedures such as blood sample collection and the administration of blood are potential important interventions to improve the quality and safety of blood transfusion [10]. However, it is as yet unclear how willing patients and healthcare staff would be for patients to undertake a more active role. Furthermore, the use of advance directives and the clinical team's recognition that some patients have a religious or moral objection to the receipt of some or all blood components will allow patients to receive care that is consistent with their wishes and beliefs.

Transfusion-transmitted human immunodeficiency virus (HIV) led to substantial reductions in allogeneic red cell use in many countries after 1982. These declines are even more significant considering population growth and ageing during this period. Over the same interval, autologous donations increased greatly. Some physicians were sued when transfused patients contracted HIV and the transfusions had not been clinically indicated (also see Chapter 22).

The potential for transfusion-transmitted variant Creutzfeldt–Jakob disease (vCJD) was one of the concerns that led the UK Department of Health in 1998 to require that all hospitals should have HTCs, implement good transfusion practice and explore the feasibility of cell salvage. Universal leucocyte reduction of blood was introduced in the UK in 1999 as a further preventive measure for vCJD. This resulted in a significant increase in the price of blood, which was an additional encouragement for hospitals to implement more judicious approaches to transfusion and use of alternatives to transfusion. As a consequence, red cell use in the UK has decreased by about 20% over the last 10 years, despite an increase in the volume and complexity of clinical care over this period.

Local investigation and feedback following ‘near misses’ and serious adverse events

SHOT defines a ‘near miss’ as any error that, if undetected, could result in the determination of a wrong blood group or the issue, collection or administration of an incorrect, inappropriate or unsuitable component but which was recognized before transfusion took place [1]. In Europe, ‘serious adverse events’ must also be reported to the competent authority. These events are defined as any untoward occurrences associated with the collection, testing, processing, storage and distribution of blood or blood components that might lead to death or life-threatening, disabling or incapacitating conditions for patients, or that result in, or prolong, hospitalization or morbidity. Systematic root cause analyses of these incidents provide opportunities to detect and understand system and process weaknesses and take corrective action to minimize recurrence. Typical weaknesses identified through root cause analyses include: inadequate training; human factors such as fatigue, misconceptions, ignorance of relevant policies; environmental factors such as distractions or interruptions, time pressures or access to equipment and IT support; and defective or risky processes.

Sample errors, most importantly those where the tube is labelled with the intended patient's details but contains blood from another patient, ‘wrong blood in tube’ (WBIT) events, are some of the most common detectable errors reported to haemovigilance programmes. These inevitably arise as a result of failures to identify systematically and positively the patient at the bedside. However, investigations almost always uncover other contributing factors, which need to be understood and addressed, for example:

·        Failure to positively identify the patient. Healthcare workers have often not been trained in and are unfamiliar with hospital policies and procedures, or perceive this activity as unimportant or suggesting an inadequate knowledge of patients under their care.

·        Reduced junior doctors' hours and shift patterns of those involved in direct patient management, and inadequate communication and documentation, leading to unfamiliarity with patients.

·        Admission and discharge practices, which frequently lead to patients having samples taken for pretransfusion testing before case notes are available or wristbands applied, leading to the potential for misidentification.

Exposure to avoidable patient morbidity or fatality often triggers clinical awareness of transfusion hazards and can instigate procedural changes. Corrective action should involve counselling and educating individuals who failed to comply with procedures, but focusing on addressing important, underlying system issues identified above and supporting staff in often traumatic situations [12].

Education and continuing professional development

Education of all individuals in the transfusion process has traditionally been difficult, but UK experience shows it to be achievable when made an integral part of mandatory hospital training programmes and subjected to external inspection. However, it requires considerable dedicated resources, a flexible and pragmatic approach to accommodate shift patterns and availability of staff, including temporary staff. Observational competency assessment is more readily achieved with the help of clinical ‘champions’. Training and knowledge-based assessments can be facilitated by web-based programmes (such as the e-learning modules of the Australian BloodSafe program, which have been completed by over 80 000 staff nationally and internationally), which also permit management oversight of participation.

Education is an essential component of strategies to gain clinician compliance with procedures and guidelines and to modify practice. Educational interventions are more successful when they are interactive, focused on a specific objective and directed at groups of individuals with reflections on their own practice. Continuing professional development schemes for the various craft groups encourage knowledge acquisition with documentation (typically via participant portfolios) of accredited activity in educational, professional and vocational areas.

Centralization of transfusion services

The medical and patient safety benefits of a centralized transfusion service (CTS) vary depending on its organisation [13]. The CTS in Pittsburgh, USA (city population 306 000; catchment population 2.1 million), operates as follows. The main blood supplier, the Central Blood Bank (CBB) delivers blood components to a central laboratory. This stand-alone central facility also houses the red cell reference laboratory and performs most of the automated, batched pretransfusion testing. The laboratory then distributes products to 19 CTS-networked hospitals in a ‘hub and spoke’ manner. Each hospital has an on-site transfusion laboratory, staffed and stocked with products in accordance with the acuity of patients treated and volume of transfusions performed. Each hospital laboratory performs routine pretransfusion testing and basic immunohaematology, thawing of plasma and cryoprecipitate, and some platelet pooling and leucocyte reduction (most is performed centrally).

Perhaps the most important patient safety benefit of a CTS is the ability to access patient records at different hospital sites. Since patients can visit different hospitals within the network, recipient immunohaematology and component modification requirements are available electronically at each hospital's blood bank, reducing the need for re-investigation and ensuring that any special component modifications are fulfilled for each patient whenever and wherever transfusion is required. Having records of recipient historical ABO groups provides additional opportunities to detect WBIT errors. In 16 cases where recipient historical ABO groups on file at the Pittsburgh CTS did not match the ABO group of specimens submitted for pretransfusion testing, 6/16 were detected based on an historical ABO group collected previously at a different hospital [14]. Requiring a second ABO group to be performed on a separate specimen before ABO-specific RBCs are issued on recipients without an historical ABO group on file would achieve the same end.

Other advantages of a CTS include availability of transfusion medicine expertise for community hospitals without experts on the staff. CTS transfusion physicians participate on HTCs of all networked hospitals, supporting rapid implementation of evidence-based practice and benchmarking. Consolidating technical expertise into one reference immunohaematology laboratory permits rapid and expert service provision. There are also numerous opportunities for cost savings, through greater efficiency from technical and nontechnical employees, economies of scale and use of automation. Blood supplier logistics are greatly simplified by delivery to one central location and lower inventory levels can be supported due to the ability to circulate blood components between hospitals to reduce wastage.

Pretransfusion compatibility testing

This typically comprises:

·        determination of the recipient ABO and RhD group;

·        a screen for red cell alloantibodies reactive at 37°C in recipient plasma;

·        a check for previous records or duplicate records, and comparison of current with historical findings (these three elements comprise a ‘group and screen’);

·        identification of the specificity of any alloantibody detected in the antibody screen;

·        selection of appropriate component(s), bearing in mind blood group compatibility (and extended red cell phenotype, where relevant), and any modifications, such as irradiation or washing, to meet individual requirements;

·        a serological or electronic crossmatch; and

·        labelling of the blood with recipient identifying information.

Detection of red cell antigen–antibody reactions

Detection and identification of blood group antigens and auto- and alloantibodies depends on interpretations of serological reactions. Various test systems are available to demonstrate these interactions and methods must be optimized in order to obtain the necessary sensitivity and specificity for their intended clinical use. Failure to follow the instructions provided by reagent manufacturers can lead to incorrect conclusions.

Test methods have been developed to allow detection of antibodies of different isotypes. Antibodies with specificities for red cell antigens are usually IgG or IgM. Pentameric IgM antibodies can cross-link antigens on adjacent cells, causing direct agglutination of red cells. Conversely, IgG antibodies are monomeric and, although divalent, the distance between the Fab regions on a single IgG molecule is generally insufficient to allow direct agglutination. Methods such as the indirect antiglobulin test (IAT) (which uses a secondary anti-isotype antibody; see Figure 23.2) or the enzyme method (which uses proteolytic enzymes such as papain or ficin to cleave negatively charged, hydrophilic residues from red cell membranes) must therefore be used to detect most IgG red cell antibodies.

Fig 23.2 Indirect antiglobulin test.


Test systems for detection of serological reactions can be classified into three broad categories.

Liquid-phase (‘tube’) systems

Liquid-phase systems rely on visualization of haemagglutination reactions in individual glass/plastic tubes or 96-well microplates. The presence or absence of agglutinated red cells distinguishes positive and negative reactions, allowing grading of reaction avidity according to the strength of haemagglutination. While not the most sensitive methods available today, IAT methods using red cells suspended in low-ionic-strength solution remain the gold standard for detection of clinically significant red cell alloantibodies. These methods require meticulous procedural attention, in particular during washing to remove unbound IgG.

Column-agglutination systems

Introduction of column-agglutination systems has resulted in very significant changes to routine laboratory practice. Synthetic gel mixtures or glass microbeads configured into vertical columns on small cards form density barriers, retaining agglutinates and allowing the passage of unagglutinated cells. Positive reactions (antibody/antigen interactions) are distinguished by agglutinates at or near the top of the gel column and negative reactions appear as buttons of red cells at the bottom (Figure 23.3).

Fig 23.3 Column-agglutination technology for blood grouping and antibody screening. Samples may consist of patient cells and reagent antisera or reagent red cells and patient serum/plasma. Positive results are seen in the first and last columns, the other columns show negative reactions.


Reagent (IgM) antibody can be incorporated into the columns, allowing phenotyping simply by addition of test cells to the top of the column. Similarly, the IAT can be performed in columns containing antiglobulin reagent to which plasma and reagent red cells are added. Because plasma proteins are less dense than the gel, washing is not needed. This property, and the relative stability of reaction endpoints, gives column agglutination methods a simplicity and reliability not achieved by other methods. Manual and automated methods for performing and interpreting these tests are now widely available.

Solid-phase systems

These techniques are performed in microplates and provide another alternative to tube or column IAT. Positive reaction endpoints are characterized by red cell monolayers in the wells while discrete buttons of red cells at the bottom of the well indicate negative reactions (Figure 23.4). Solid-phase systems require carefully standardized centrifugation and washing steps; however, unlike liquid-phase test systems, fully automated equipment allows these steps to be performed safely and consistently without operator intervention.

Fig 23.4 Solid-phase blood grouping technology. See text for an explanation of results.


Reduction of error in pretransfusion compatibility testing

Analysis of mistransfusion cases from SHOT in 2011 showed that laboratory errors were implicated in up to 55% of incidents [1]. Many of these were due to human error in sample transposition or in test setup or interpretation. Provided that correct laboratory identifiers (such as barcodes) are placed on patient samples, these events can be avoided by using fully automated systems interfaced to transfusion laboratory computer systems.

Basic features of fully automated (‘walk away’) systems should include:

·        trays or carousels to stack samples;

·        automated liquid handling and other robotic operations;

·        devices to ensure that positive sample identification is maintained;

·        clot sensor and liquid level alarms;

·        an optical device to record reaction patterns; and

·        comprehensive system management software that interprets reaction patterns and flags discrepant results.

When automation cannot be used for the whole process (i.e. antibody identification), steps must be taken to minimize the occurrence of error and its impact. High standards of training, participation in internal and external quality assessment schemes, and strict adherence to validated documented procedures are among the measures that reduce errors.

The most important procedure in the transfusion laboratory is the recipient ABO group determination. Different practices are in place in different countries, but this critical procedure should be performed independently by two people (except in urgent situations where uncrossmatched group O red cells will be issued) if there is no record of a grouping result from a historical sample. Obtaining a second sample from the patient drawn by a different phlebotomist separate from the first sample collection (often called an ‘ABO check type sample’) serves as the preferred means of verifying the results of the first ABO grouping. Similarly, determination of an RhD group should be performed in duplicate, in the absence of full automation.

ABO and RhD grouping

Patient red cells should be tested against monoclonal anti-A and anti-B grouping reagents and patient serum/plasma should be tested against A1 and B reagent red cells, except in neonates. The expected reaction patterns in ABO grouping are illustrated in Table 23.9. Patient red cells should be tested with an IgM monoclonal anti-D reagent, which does not detect the common DVI variant (because individuals with this phenotype can become alloimmunized if transfused with RhD positive RBCs; thus as recipients they should be typed as RhD negative and transfused with RhD negative RBCs). ABO and RhD groups must be repeated when discrepancies are found and these should be performed using a fresh suspension of washed cells, ideally from a new sample.

Table 23.9 ABO grouping patterns.


Antibody screening

The IAT performed at 37°C is the best method available for detection of clinically significant red cell antibodies. It is simple (especially when using a column-agglutination system), sensitive and has a high degree of specificity.

Patient serum/plasma should be tested against two or more ‘screening cells’ using the IAT. The reagent red cells used for screening should between them express antigens reactive with all clinically significant antibodies; ideally the phenotypes R1R1, rr and R2R2 should be represented in the screening cell set. Different national standards and guidelines are in place, but in many countries it is recommended that the screening cells express the Jka, Jkb, S, s, Fya, Fyb antigens and incorporate the following phenotypes: Jk(a+b–), Jk(a–b+), S+s–, S–s+, Fy(a+b–), Fy(a–b+), since stronger reactions may be obtained with cells having double-dose antigen expression.

Antibody screening performed in advance of the requirement for transfusion also provides the laboratory with time to identify the specificity of any antibody detected and, when clinically significant, to select antigen-negative units for crossmatching.

Antibody identification

When an antibody has been detected in screening tests, the specificity should be determined by testing patient serum/plasma against a panel of reagent red cells of known phenotypes. In addition to the IAT, other methods (e.g. using enzyme-treated red cells) may be helpful, particularly when mixtures of antibodies are present. Antibody specificity can be determined when the serum/plasma is reactive with at least two examples of red cells bearing the antigen and nonreactive with at least two examples of red cells lacking the antigen. When a single antibody specificity has been determined, it is essential that additional clinically significant antibodies are also detected, if present. Multiple antibodies can be confirmed only by testing against red cells that are antigen negative for the recognized specificity, but that express other antigens to which clinically significant antibodies may arise.


These may be suspected when patient serum/plasma reacts with all cells used in the reverse ABO group (in the case of cold autoantibodies) or with all cells in the antibody identification panel, including the patient's own red cells. Autoantibodies are common, but not all autoantibodies give rise to clinically significant haemolysis. Serological investigations should focus on obtaining the correct ABO and RhD group of the patient and excluding the presence of underlying alloantibodies.

Cold-type autoimmune haemolytic anaemia

These autoantibodies tend not to cause problems in alloantibody identification unless they react at 37°C. Red cells should be washed at 37°C before performing the direct antiglobulin test (DAT), which will usually be strongly positive due to coating with C3d.

Warm-type autoimmune haemolytic anaemia

Red cells will usually have a positive DAT due to coating with IgG with or without complement and an eluate prepared from these cells typically reacts with all panel cells. Rarely, red cells may be coated with IgA or IgM and IgG. Underlying alloantibodies may be detected following removal of autoantibodies from patient serum. This may be achieved either by absorbing the patient's serum with the patient's own red cells (treated with a combination of papain and dithiothreitol, or ‘ZZAP’) or, if the patient has been recently transfused, performing absorption and elution studies with reagent red cells of specific phenotypes.

These processes can be very time-consuming and the treating clinical unit should be advised that investigations may take some time before any underlying alloantibodies can be identified and antigen-negative RBC units can be available; hence, the urgency of transfusion support requirements must be determined. Chapter 29 has more information on immune-mediated haemolysis.

Selection of red cells for transfusion

Red cell units that are compatible with the recipient's ABO and RhD groups are routinely issued. As group O red cells are the ‘universal donor’ type they can be safely administered to recipients of any blood group; otherwise group A and B (and O) recipients can only receive ABO-identical red cells due to the presence of naturally occurring antibodies to the A and B antigen(s) lacking on their RBCs. Group AB recipients lack naturally occurring anti-A and anti-B antibodies and can thus receive RBCs of any blood group. Group O RBCs may be issued in life-threatening situations where blood is required before the patient has been grouped. In these emergency situations, if the patient is a premenopausal female, group O, RhD-negative uncrossmatched RBCs should be used (see below). Group-specific units can be provided as soon as the patient's group is known. Care should be taken to establish the recipient's ABO group before a large number of group O red cells are transfused, as a massive transfusion with group O units could obscure the recipient's actual ABO group and complicate the further selection of both RBCs and other components.

Premenopausal females should ideally receive RhD- and K-matched red cells and in some countries this is a requirement to prevent alloimmunization, which could lead to severe haemolytic disease of the fetus and newborn (HDFN). RhD positive patients can always safely receive RhD negative units. Patients with anticipated long-term red cell transfusion requirements (see Chapters 28 and 29) should also ideally receive red cells matched at least for Rh and K antigens. Red cells for fetal or neonatal exchange transfusions should also be selected to be compatible with the maternal serum/plasma and any known maternal antibodies.

A recipient of an ABO-incompatible allogeneic haemopoietic stem cell graft will need to be transfused with red cells of the donor's group in the case of a minor ABO mismatch or group O in the case of a combined (bidirectional) ABO mismatch. RhD-negative red cells should also be selected for an RhD-positive recipient of an RhD-negative stem cell donation (see Chapter 28).

Selection of blood for patients with red cell alloantibodies is summarized in Table 23.10. However, in life-threatening situations, the immediate need for red cell transfusion may necessitate the use of potentially incompatible units. Discussions between the treating clinical unit and transfusion service about risks of haemolysis when using uncrossmatched red cells for a patient with a positive antibody screen are essential.

Table 23.10 Recommendations for selection of blood for patients with red cell alloantibodies.


Typical examples


Antibodies considered clinically significant

Anti-RhD, -C, -c, -E, -e

Anti-K, -k

Anti-Jka, -JkbAnti-S, -s, -U

Select ABO-compatible, antigen-negative blood for serological crossmatching


Anti-Fya, -Fyb


Antibodies directed against antigens with an incidence of <5% and where the antibody is often not clinically significant




Anti-Wra (anti-Di3)

Select ABO-compatible blood for serological crossmatching

Antibodies primarily reactive below 37°C and never or only very rarely clinically significant


Select ABO-compatible blood for serological crossmatching, performed strictly at 37°C






Anti-Lea, -Leb, -Lea+b


Anti-HI (in A1 and A1B patients)


Antibodies sometimes reactive at 37°C and clinically significant


If reactive at 37°C, select ABO-compatible, antigen-negative blood for serological crossmatching


If unreactive at 37°C, select ABO-compatible blood for serological crossmatching performed strictly at 37°C

Other antibodies active by IAT at 37°C

Many specificities

Seek advice from blood centre

The laboratory also has responsibility for ensuring that requests for irradiated, CMV seronegative or other requirements are fulfilled in accordance with institutional policies and that patient records are ‘flagged’ for these needs.


Red cell crossmatching techniques have been simplified in recent years and only the immediate spin (IS) and IAT crossmatches remain in common use.

The IAT crossmatch can be abolished in favour of an electronic, or IS only, crossmatch when antibody screening is performed with screening cells that express the most common antigens capable of stimulating clinically significant antibodies and the patient's serum/plasma has never been found to contain clinically significant antibodies. Studies have shown that there is negligible risk in omitting the IAT crossmatch. Although up to 0.2% IAT crossmatches may reveal an unpredicted incompatibility, few of these transfusions result in haemolysis. Antibodies directed against low frequency antigens may be missed, but the majority of these are clinically insignificant.

If the IAT crossmatch is omitted, there must be some check included to detect ABO incompatibility. The IS crossmatch (i.e. agglutination in saline following centrifugation) is a serological check that can be used. However, this technique is fallible when the patient has low levels of anti-A or anti-B antibodies and, unless ethylenediamine tetra-acetic acid (EDTA) saline is used, false-negative results may also arise as a result of steric hindrance of agglutination by complement. False-positive IS results arising from rouleaux or cold agglutinins can also cause ABO discrepancies. Limitations of the IS crossmatch have heralded the acceptance of electronic issue as an alternative method of preventing the release of ABO-incompatible units of blood.

Electronic crossmatch and issue

Electronic crossmatch can only be used for detecting ABO incompatibility between the donor unit and recipient. There are several essential requirements for adopting this approach, which are common to the various professional standards.

·        The computer contains logic to prevent assignment and release of ABO-incompatible blood.

·        No clinically significant antibodies are detected in the current patient plasma/serum sample and there is no record of previous detection of such antibodies.

·        There are concordant results of at least two determinations of patient ABO and RhD groups on file, at least one of which is from a current sample.

·        Critical system elements (application software, readers and interfaces) have been validated on-site and there are mechanisms to verify the correct entry of data prior to release of blood, such as barcode identifiers to enter information when it cannot be automatically transferred. Fully automated blood grouping and antibody screening, although not a requirement in some national guidelines, is strongly recommended.

Electronic issue has been widely used for over a decade and is now routine practice in many countries. It has several potential advantages over serological crossmatching:

·        reduced technical workload;

·        rapid availability of blood;

·        improved blood stock management through reduced numbers of crossmatched red cells and reduced wastage;

·        less handling of biohazardous material;

·        elimination of unwanted false-positive results in the IS; and

·        ability to issue blood electronically at remote sites, using trained nonlaboratory staff.

This last characteristic has allowed development of systems for electronic remote blood issue. When patient details are entered, the system checks that criteria for electronic issue are fulfilled and either allows access to ABO and RhD compatible units in the remote blood refrigerator or dispenses compatible units. A compatibility label is printed and attached to the unit and rescanned to ensure it is the correct one for the unit. Such systems reduce the time taken for the issue of blood, particularly in small hospitals without transfusion laboratories.

Maximum surgical blood ordering schedule (MSBOS)

This consists of a table of elective surgical procedures that lists the extent of pretransfusion testing that is routinely required before the case begins (see Table 23.11). The MSBOS is prepared taking into account the likelihood of transfusion and the response time for having blood available, following an immediate spin crossmatch or electronic issue. An MSBOS reduces the workload of unnecessary crossmatching and issuing of blood and can improve stock management and reduce wastage.

Table 23.11 Example of maximum surgical blood order schedule (MSBOS, general surgery).


Red cells crossmatched or group & screen (G&S)

Lumbar spine disc replacement Caesarean section

No pretransfusion testing required No pretransfusion testing required

Colonoscopy with polypectomy Tonsillectomy

Spinal laminectomy and decompression

Carotid endarterectomy

Lung biopsy

Mammoplasty reduction


Coronary artery bypass graft

No pretransfusion testing required

No pretransfusion testing required







Total hip arthroplasty

Resection/repair ascending aortic aneurysm



Heart transplant


Liver transplant


The successful implementation of an MSBOS depends on all parties agreeing to the schedule, education of blood prescribers, confidence of senior staff that there is a robust system for accessing blood promptly when there is unexpected blood loss and ability to override the schedule when there are reasons to suggest that indicate that greater blood loss will occur. The schedule is constructed by:

·        analysing each surgical procedure in terms of the crossmatch : transfusion (C : T) ratio;

·        routinely managing procedures with a C : T ratio greater than 2 (i.e. a low probability of transfusion) with a group and screen, and issuing blood only when there is a need for transfusion; and

·        allocating an agreed number of units for procedures with a C : T ratio of less than 2.

An overall C : T ratio of 1.5 for elective surgery is achievable when the laboratory is centrally issuing blood in accordance with the MSBOS. However, lower ratios are possible with use of electronic crossmatch (see above) and/or remote electronic issue from blood refrigerators in theatre suites. In addition to reducing the number of allocated, crossmatched red cells for specific surgical patients, which increases the number of available units in the general inventory, another benefit of adhering to recommendations of the MSBOS is that fewer patients with unexpected antibodies will be taken to surgery without appropriate transfusion support. As the MSBOS indicates the extent of pretransfusion testing that should be performed before surgery, adherence to its recommendations will lead to antibody screening being performed on patients with a reasonable chance of requiring intraoperative transfusion, thereby allowing the transfusion service to locate and crossmatch compatible units before the case begins, should unexpected antibodies be detected.

In recipients with red cell alloantibodies who require transfusion, consideration should be given to the time taken to acquire and crossmatch antigen-negative units, and the treating clinical team should be informed.

Selection of platelets and plasma components

Platelets are collected by apheresis or separated from whole blood donations using either the ‘platelet-rich plasma’ or buffy coat methods. ABO and RhD compatible platelets are preferable, but when these are not available, ABO incompatible platelets may be used. Many countries provide platelets suspended in additive solution, to reduce recipient exposure to plasma and to support platelet viability and function.

Plasma for transfusion is collected by apheresis or separated from whole blood donations. Plasma should be ABO compatible with the recipient, but where the recipient's blood type is not known, such as in an emergency, AB plasma can be used. Plasma depleted of cryoprecipitate (cryodepleted or cryopoor plasma) may be used in the treatment of thrombotic thrombocytopenic purpura. Other preparations, such as pooled, solvent-detergent plasma, are also available in some regions.

Cryoprecipitate is prepared from whole blood donations or apheresis plasma collections. Cryoprecipitate should ideally be of the same ABO group as the recipient, but this is not essential. Cryoprecipitate is mainly used as a source of fibrinogen, where a virally inactivated fibrinogen concentrate is not available.

Frozen products must always be thawed using an approved device and method.

Key points

1. The transfusion process is unique as it links blood donors with patients in an altruistic, potentially life-saving activity. For many patients, there is still no substitute for donated blood components.

2. Prescribers of blood components have a duty of care to their patients to ensure that the benefits of the transfusion outweigh the risks, and a moral obligation to donors to ensure that their donations are used appropriately.

3. The transfusion process is multistep and complex, involving many different staff across the broad spectrum of clinical practice and settings, often working under challenging conditions. In this context there are many opportunities for human error to occur.

4. Investments in quality infrastructure, computerization and automation and training in the clinical and laboratory aspects of transfusion practice are essential to minimize or, ultimately, prevent errors in the transfusion process.


This chapter updates the material in the previous edition by Sue Knowles and Geoff Poole.


1. PHB Bolton-Maggs (ed.) & H Cohen on behalf of the Serious Hazards of Transfusion (SHOT) Steering Group. The 2011 Annual SHOT Report; 2012.

2. US Food and Drug Administration. Fatalities reported to FDA following blood collection and transfusion: Annual Summary for Fiscal Year 2010. Available at:

3. Narick C, Triulzi D & Yazer MH. Transfusion-associated circulatory overload after plasma transfusion. Transfusion 2011, 18 July. Epub ahead of print.

4. Linden JV, Wagner K, Voytovich AE & Sheehan J. Transfusion errors in New York State: an analysis of 10 years' experience. Transfusion 2000; 40: 1207–1213.

5. Murphy MF, Stanworth SJ & Yazer M. Transfusion practice and safety: current status and possibilities for improvement. Vox Sanguinis 2011; 100: 46–59.

6. UK Blood Safety and Quality Regulations. Available at:

7. Australian Commission on Safety and Quality in Health Care. National Safety and Quality Health Service Standards. Sydney: ACSQHC; 2011.

8. Turner CL, Casbard AC & Murphy MF. Barcode technology: its role in increasing the safety of blood transfusion. Transfusion 2003; 43: 1200–1209.

9. Staves, J, Davies A, Kay J et al. Electronic remote blood issue: a combination of remote blood issue with a system for end-to-end electronic control of transfusion to provide a ‘total solution’ for a safe and timely hospital blood transfusion service. Transfusion 2008; 48, 415–424.

10. Davis RE, Vincent CA & Murphy MF. Blood transfusion safety: the potential role of the patient. Transfus Med Rev 2011; 25(1): 12–23.

11. Tavares M, DiQuattro P, Nolette N et al. Reduction in plasma transfusion after enforcement of transfusion guidelines. Transfusion 2011; 51: 754–761.

12. Stainsby D, Russell J, Cohen H & Lilleyman J. Reducing adverse events in blood transfusion. Br J Haematol 2005; 131(1): 8–12.

13. Simpson M. Strategies for Centralized Blood Services. Bethesda, MD: AABB Press; 2006.

14. MacIvor D, Triulzi DJ & Yazer MH. Enhanced detection of blood bank sample collection errors with a centralized patient database. Transfusion 2009; 49: 40–43.

Further reading

Department of Health. Better Blood Transfusion, HSC 2007/001. London: HMSO; 2007.

Dzik WH, Corwin H, Goodnough LT et al. Patient safety and blood transfusion: new solutions. Transfus Med Rev 2003; 17: 169–180.

Eisenstaedt RS. Modifying physicians' transfusion practice. Transfus Med Rev 1997; 11: 27–37.

Judd WJ. Requirements for the electronic crossmatch. Vox Sanguinis 1998; 74 (Suppl. 2): 409–417.

Klein HG & Anstee D (eds). Mollison's Blood Transfusion in Clinical Medicine, 11th edn. Oxford: Blackwell Publishing; 2005.

McClelland DBL (ed.). Handbook of Transfusion Medicine, 4th edn. Norwich: TSO; 2007. Available at:

Saxena S & Shulman I (eds). The Transfusion Committee: Putting Patient Safety First. Bethesda, MD: AABB Press; 2006.

Guidelines and other resources

For a range of guidelines and other resources on laboratory and clinical hospital transfusion practice:


Australian and New Zealand Society of Blood Transfusion:

British Committee for Standards in Haematology:

Canadian resources: and

International Society of Blood Transfusion:

Joint Commission on Accreditation of Healthcare Organizations:

Network for Advancement of Transfusion Alternatives:

World Health Organization: