Practical Transfusion Medicine 4th Ed.

18. The role of haemovigilance in transfusion safety

James P. AuBuchon1 & Katharine A. Downes2

1Puget Sound Blood Center and University of Washington Seattle, Seattle, Washington, USA

2University Hospitals Case Medical Center and Case Western Reserve University, Cleveland, Ohio, USA

Introduction

In the last decade the quality improvement movement in healthcare has engendered many adages while increasing our ability to improve the delivery of services and patient outcomes. One of these is especially pertinent to haemovigilance: ‘If you can't measure it, you can't improve it.’ This concept sums up the rationale behind the creation of a haemovigilance system in most developed nations over the last decade and the excitement associated with the data they are generating. By defining the frequency of problems encountered in the transfusion system from vein to vein, i.e. from the time of collection through the consequences of transfusion, and documenting any failures to achieve the desired goal of safe and efficacious transfusion every time, we can begin to identify where our attention and resources should be directed in order to allow transfusion medicine to participate fully in modern medicine's attempt to provide care through robust systems that yield dependable outcomes.

Although different entities have applied slightly different definitions to haemovigilance, that of the International Haemovigilance Network (IHN) is one of the most encompassing:

Haemovigilance is a set of surveillance procedures covering the entire transfusion chain (from the donation of blood and its components to the follow-up of recipients of transfusions), intended to collect and assess information on unexpected or undesirable effects resulting from the therapeutic use of labile blood products, and to prevent the occurrence or recurrence of such incidents.

Recognition that transfusion hazards may accrue from any of the multiple steps along the complex pathway from donor selection to recipient transfusion requires that a haemovigilance system maintain a broad, all-encompassing scope in order to define these risks. Inclusion of the expectation that steps will be taken to reduce these risks is critical, since merely collecting data will not prompt improvements in the transfusion system. In essence, then, haemovigilance systems provide the engine through which transfusion systems can improve their services and patient outcomes.

Origin and impetus

Haemovigilance systems arose out of a confluence of events that questioned the safety of the healthcare system to deliver treatment without causing unnecessary harm and the ability of the blood supply system to deliver components with minimal risk. Increasing recognition that the reliability of medical care systems is suboptimal led to broad efforts to reduce the substantial risks associated with delivery of all aspects of healthcare. These efforts alone might have stirred blood bankers into action, but the earlier public debates, commissions of inquiry and prosecutions (and convictions) stemming from how the nascent HIV risk of the 1980s had been handled provided additional impetus for the field to assess the safety of its services through ongoing risk assessment measures. As the infectious disease risks of HIV and HCV that captivated our attention through the 1980s and 1990s were documented to have been greatly diminished through concerted multifaceted interventions, the field of transfusion medicine felt increasingly able to redirect its attention to problems that had been known to exist for many years but that had never been definitively or effectively addressed.

The legal framework and the organizational structure of these systems vary from country to country. An early system was reported to have been established in Japan in 1992 [1]. In Europe the first system was established about the same time in France as a mandatory system in which the reporting of all untoward outcomes from transfusions was required [2]. The second European system implemented arose through the efforts of transfusion medicine professionals in the UK and was organized through the Royal College of Pathologists as a voluntary reporting system focused on Serious Hazards of Transfusion (SHOT) [3]. Subsequently, systems have been created and implemented in most developed nations as a hybrid of these approaches. Some of these reside within and derive reporting mandates from a national ministry of health while others are primarily organized through professional societies or the country's blood collection system with sharing of data among all concerned parties. The European Community currently requires implementation of a haemovigilance system in each member state with reporting to a central office [4]. Development of a voluntary hemovigilance system in the USA evolved through efforts of numerous stakeholder organizations from the private and public sectors.

Although each country's system has characteristics unique for its own healthcare and transfusion systems, these systems bear multiple similarities and have yielded similar results [5], as will be discussed in the next section.

Coincident with these occurrences, more and more healthcare organizations, including blood collection organizations and transfusion services, have recognized the value of applying lean process improvement methodology. By focusing attention on where processes add value that is important to the end customer (the patient, in the case of transfusion) or where they fail to do so, improvements in the system that could add useful benefits can be targeted. Haemovigilance systems identify circumstances where value – in terms of safety and, in some cases, also efficiency – is lost from the transfusion process. Therefore, the experience garnered from a haemovigilance system adds useful information to the attempts by transfusion professionals to provide increased safety to the entire process.

Key elements and residual questions

The structure and content of haemovigilance systems vary by country, but the successful show several similarities in important facets of their philosophy and function (Table 18.1).

Table 18.1 Important features of a haemovigilance system.

Confidentiality of submitted data

Broad participation, supported by education

Use of standardized definitions and terminology

Nonpunitive evaluation of data

Reporting of rates of occurrences

Sufficient detail to make effective recommendations for improved practices

Focus on improved safety and outcomes

Simple and efficient operations

Sustainable organization

Definitions and terminology

Whether dealing with clinical events, such as transfusion reactions or near-miss incidents, a standardized lexicon must be adopted. Clear and precise definitions must be utilized and terms resistant to misinterpretation must be employed to decrease the chance of a misapplication of a classification scheme that would degrade the value of the system's data.

The importance of data conventions and data standardization cannot be overstated [6]. Standardization of data elements and terminology, as well as uniformity in data capture and reporting methods, are required to ensure data integrity for comparative analyses and also enable (1) consistent tracking of internal performance over time, (2) benchmarking across institutions and (3) assessment of outcomes of process improvements. The importance of this standardization for comparison between systems was recognized early on by the IHN, and a working party of the International Society of Blood Transfusion (ISBT) has been developing definitions of transfusion reactions that could be used to achieve commonality and facilitate meaningful comparisons of data between countries. Some systems have taken the additional step of interposing a review of the details of a reported event to ensure that it meets the definitions used in that reporting system. While this rigour adds robustness to the reports of the system, larger systems or those with relatively fewer resources need to depend on participants' accurate application of the definitions embedded in the classification scheme.

Unexpected value from clearly defining the elements of a transfusion reaction have been reported from early implementation of the US Biovigilance Network. In addition to providing a standardized means through which reaction rates could truly be compared between institutions, the standardized definitions that were applied provided a framework through which reactions previously regarded as ‘uncategorizable’ could be tallied appropriately as well as uniformly. The standardized definitions thus provided benefit to the system and to the local transfusion specialist as well as to the practitioner who previously had to deal with the patient without clear guidance. The field of anatomic pathology has long recognized the importance of a clearly enunciated set of diagnostic criteria for diagnosis and grading of pathologic lesions; the transfusion segment of clinical pathology now is applying the same methodology to achieve better uniformity in describing untoward outcomes, thus informing the individual clinician as well as the system as a whole.

Detection and reporting of adverse events

Common to all haemovigilance systems is development and/or application of a system to capture events of interest directly from the site of the transfusion recipient or the hospital transfusion service. Many of the countries that first implemented haemovigilance systems had most of their transfusion expertise headquartered in blood centres and few transfusion medicine specialists located in hospitals. They generally did not have a history of capturing transfusion reaction events in real time for investigation, accumulation and (local) analysis. To provide information for the new haemovigilance network, France identified a ‘rapporteur’ or ‘haemovigilance officer’ in each hospital (usually a physician, such as a haematologist or anaesthesiologist), the UK trained a cadre of nurses as ‘specialized practitioners of transfusion’ and Québec established a network of ‘transfusion safety officers’ (TSOs, usually nurses) in the larger hospitals who also oversaw reporting of events from nearby smaller facilities. In countries with a more established transfusion medicine professional presence at the hospital level, these new assignments or positions would not be required. For example, the haemovigilance reporting system in the USA is based on the pre-existing transfusion reaction reporting system extant in hospitals of all sizes that reports to the hospital's transfusion service and to the facility's transfusion committee. Regulatory and accreditation requirements strongly influence many of the components and requirements of such transfusion reaction reporting systems in US healthcare organizations.

For any reporting system to be reliable, those charged with capturing the event must be cognisant of the commonly recognized signs and symptoms of transfusion reactions as well as remaining alert to events that ‘just don't seem right’ during or after a transfusion. Simply dismissing an unusual sign or an unexpected symptom as causally unrelated to the transfusion will deprive not only the patient of steps that might prevent a recurrence in a future transfusion but also the transfusion system of the knowledge of an event that might disclose a significant new phenomenon. Ensuring that those likely to be the first to identify problems with a transfusion will report the occurrence is a critical first step in creating and maintaining a useful and credible haemovigilance system. Tying haemovigilance reporting to a bedside system that captures standardized and complete data from every transfusion undoubtedly improves the penetration of the haemovigilance system and the believability of the data it generates as well.

Despite the diligence of such efforts, however, haemovigilance systems will always be challenged to tally untoward outcomes that occur long after a transfusion because of the loss of an obvious temporal relationship (Table 18.2). For example, identifying transfusion transmission of an infectious disease would require recognition of the lack of other means of transmission to the recipient and relative rarity of the disease entity in order for the transfusion connection to be recognized, and then the physician caring for the patient would need to contact the transfusion service in order for an investigation and possibly a report to the haemovigilance system to be generated. Haemovigilance-like investigations led to the detection of transmission of West Nile virus (WNV) through transfusion relatively early in the US outbreak of this disease, but making the connection between poor patient outcome and transfusion exposure of an organ donor would have been much more difficult had the incubation period been longer. Additional utility of a haemovigilance system is seen, however, through real-time assessment of regional risk through donor testing and revision of donor testing protocols in response to this information, as is being done in the USA with West Nile virus testing.

Table 18.2 Limitations of haemovigilance systems.

Incomplete reporting

Detection of transfusion relationship of late events, including infections

Limited details

Variation in terminology and definitions

Influence of healthcare system's or institution's ‘culture’ regarding compliance, process improvement and reporting

Acceptance of nonstandard definitions and terminology

Inability to track cases back to their source to ensure that correct and complete reporting has occurred

Scope of reporting

The scope of haemovigilance systems, however, is varied (Table 18.3). The SHOT system clearly specifies its interest only in the serious hazards and precludes reporting febrile and urticarial reactions, for example. This restriction provides natural focus on the hazards that have the largest potential impact on a particular recipient but at some risk of missing events of lesser morbidity that may affect a larger number of patients or that may be harbingers of more serious sequelae after transfusion. The more common approach of seeking to capture all transfusion reactions, on the other hand, risks ‘system fatigue’ from overwhelmed personnel or reporting and analysis tools through which events of major clinical significance are obscured from view by the more numerous but less-informative reports of ‘minor’ events.

Table 18.3 Variations among haemovigilance systems.

Scope

Serious events or all events?

 

Events causing harm or also ‘near-miss’ events?

Breadth

Labile transfusible components

 

Plasma derivative products

 

Tissues and/or organs for transplantation (biovigilance)

Analysis

System level

 

Institutional level

 

Healthcare system level, comparison with all hospitals – locally, regionally, nationally and/or internationally?

 

Comparison with (anonymous) peer institution subset 

Appropriateness of analysis of incidents and events 

Access to full details of incidents and events

The extension of haemovigilance systems to incidents that are not directly associated with a reaction or an untoward outcome for a recipient is an important means of detecting problems in the transfusion system and preventing these from harming patients. Whether called an incident, a deviation or an error, these failures to adhere to standard procedures may represent human frailty or limitations, inadequate training, unique features of a patient's situation or a combination of factors that aligned with weak points in the transfusion system. The most notable among these has been patient and sample identification errors in pretransfusion testing. Inclusion of ‘near-miss’ events where the error is detected and remedied and/or where it does not cause harm to the recipient quickly causes such occurrences to become the most commonly reported events in a haemovigilance system. Although these occurrences may individually appear to be of minor importance, they represent a critical view into the workings of a transfusion system and allow preventive actions to be taken to bolster system safeguards.

Analysis of incident reports

The ability of an individual hospital or some other relevant unit of the healthcare system to assess the outcomes of its transfusions is important in addition to the national review of problems in the transfusion system. At the smallest division in which policies are common and practice is (presumed to be) universal, an analysis of reports of transfusion reactions and transfusion practice should be undertaken in order to understand how this microsystem compares to the larger whole, such as national comparators. In most situations, this analysis would be at the hospital level since enforcement of uniformity of practice outside of one's immediate reach is often difficult. Identification of where one's system is not being applied faithfully is critical knowledge that can be used to strengthen the system and remove ambiguity or opportunities for imprecise or incorrect actions to occur (or at least to go unnoticed). The individual hospital that participates in a national haemovigilance system that captures these kinds of data can benefit by comparing their experience to that of others working in the same type of system. Understanding where in the transfusion chain differences in practice are occurring may identify those steps in the process that are truly critical to be performed with a high degree of fidelity and also identify transfusion practices associated with superior outcomes.

There are several features of haemovigilance systems and hospital ‘cultures’ that are presumed to be associated with higher rates of reporting compliance. The degree to which an institution focuses on outcome improvement and adherence to policies is probably an important determinant of the acceptance of a reporting system such as represented by a haemovigilance system. Related to this may be the extent to which an ‘open learning culture’ is supported in the institution. Similarly, the more that ‘incidents’ are recognized as failures of the (imperfect) system rather than of those working within it, the more likely that staff will feel comfortable reporting the occurrences. When these reports lead to improvements in the process (and particularly when those affected by the incident have been able to participate in the improvement of the transfusion system), the satisfaction that is generated also empowers further reporting towards the end of improving operations and outcomes.

Encouragement of active participation in haemovigilance

This leads naturally to a discussion of whether reporting to a haemovigilance system should be ‘voluntary’ or ‘mandatory’. In all likelihood, a mixture of incentives is most salutary. Error reporting will suffer in a system where reporting may be ‘required’ but more frequently leads to punishment of those caught in a cumbersome, error-prone system rather than to change in that system. The lack of nonpunitive reporting would clearly limit the ability of the haemovigilance system to effect improvements. A nonpunitive approach to reporting – for the individual reporting the case as well as for those involved in it and the institution itself – is essential to compliance. On the other hand, voluntary reporting, even when coupled with confidentiality safeguards, is unlikely to attract respondents unless the importance of their actions is well understood and they see the fruits of their efforts through system reports and process changes. All systems have noted a continual rise in the number of reports submitted over their first few years, as more staff ‘on the front lines’ become aware of the system, its importance and the logistics of reporting; those systems with more extensive educational infrastructure appear to have the most rapid attainment of extensive penetration. Endorsement of participation by a professional association or the ministry of health may boost participation, particularly if the system is easy for participants to use and has already established its credibility. In a legal system where the involved person or institution is subject to liability damages from an ‘error’, assurance of confidentiality for the details of the report and freedom from compelled disclosure are absolutely essential for enabling reporting. Even if no harm befell a patient from an incident, an institution would understandably require assurance that its ‘dirty laundry’ could not be used against it by a plaintiff's attorney alleging a pattern of inappropriate practices.

Whether participation in haemovigilance is voluntary or mandatory, thorough preparation within the participating organization is a key success factor. Effective preparation includes detailed review of current processes and procedures for adverse reaction reporting and interpretation and gap analysis compared to what will be expected for the haemovigilance programme. Such a review optimally includes representation from all stakeholder groups including physicians, nurses, laboratory staff and information system support staff. Sufficient time must be set aside for creating the necessary, locally appropriate definitions, e.g. transfusion location codes, in the computer system(s) and adequate testing. Training requirements, particularly for busy clinician groups, can be easily underestimated or overlooked. Pilot implementation at a hospital may be a useful way to uncover issues for correction prior to data entry into a haemovigilance system.

Data management

Simplicity of reporting is also the key to high participation rates. Most haemovigilance systems began with paper reports but are converting or have converted to electronic submission systems. This approach not only facilitates data management by the coordinating office but also simplifies the reporting mechanism for the participant; an intelligent, web-based data capture system could display only those items that were pertinent to the case's report as it unfolded rather than frightening a respondent with multiple pages of data elements, many of which would not be pertinent in any one particular case. Such a context-sensitive system could also check for completeness and prevent logging of an incomplete case or check for internal congruity in a case and prevent clearly erroneous entry errors. The extent to which a system wishes to go to ensure that its definitions are being followed may be dependent on assumptions made about the support available for data entry. For example, does the system require entry of pre- and posttransfusion temperatures and a check to see if the definition of a febrile transfusion reaction has been met, or does the system expect that the respondent is applying the definition correctly? The former approach would provide better assurance of data integrity but would require more time for data entry. The quality of analysis possible from computerized databases depends upon the appropriateness of the definitions of the data elements that are established in the database and the consistency with which data entry is accomplished.

Optimally, the national haemovigilance system would be interfaced to a facility's internal error management software (such as MERS-TM [7]) and be able to accept the report of the necessary elements of a case automatically. Most haemovigilance systems have not reached that level of sophistication, but centrally coordinated healthcare systems may be able to integrate laboratory information, error management and haemovigilance systems to accomplish this. Since the capabilities of many transfusion services to analyse and report their experience with incidents and events are limited in their laboratory information system, the ability of the national haemovigilance system to construct the tables and graphs illustrating the experience of the reporter's institution and overlay the system's experience (in aggregate form or from institutions with similar characteristics) may reward (and thus encourage) participation.

The extent of data collection ultimately affects the richness of the system's data. The more detail of an incident or a reaction that is captured by a system, the greater the impact on policy development the reports may offer. Capturing detail about where an incident occurred may help identify ‘high-risk clinical areas’, and identifying the steps in the transfusion process that were vulnerable to error will similarly help direct attention to the part of the transfusion process where improvements will have the greatest impact. The performance of a root-cause analysis is beyond the purview of a national haemovigilance system, but capturing the results of such an analysis and estimating the probability of a recurrence and its impact can also help focus attention on ‘big payoff’ targets.

A critical category of data elements in a haemovigilance system is the ‘denominator data’. Although tallying the number of units of each component type transfused or the number of pretransfusion specimens tested does not inform us about transfusion safety, this information is essential in turning reported occurrences into rates. Only through a comparison of the rates of events can we compare meaningfully across institutions and countries of different sizes and different transfusion activity levels. As different countries – or, indeed, different regions within one country – account for outdated and discarded units differently, depending solely on reports from blood centres for component volumes may lead to inaccuracy in calculation of transfusion rates.

Breadth of the system

An analogous consideration is the breadth of the system. What products should be included in the system's tracking? The IHN's definition of haemovigilance focuses on labile components, but there remains much to be learned about the use of and reactions to plasma derivative products, and in some jurisidictions these are handled through the same transfusion service system. In such a case, as in Canada, valuable information can be gleaned through extending the system to these additional components. Should haemovigilance systems be further expanded to incorporate tissues and organs to become biovigilance systems? The principles of haemovigilance – including product traceability, learning from one's practice and a commitment to continual improvement – are applicable to transplantation as well as to transfusion, and transfusion services handle these products in some countries as well. Efforts to create systems applicable to tissues have already begun in earnest in the USA, again spurred by safety concerns, and marrow transplant organizations have for many years been tracking the outcomes of their efforts in order to find keys to improving patient outcomes. The extent to which these efforts will be interdigitated with haemovigilance systems rather than standing alone as complimentary systems will be determined by the extent to which the participating institutions and physicians contribute directly to both fields and whether their systems are mutually supportive. At the least, transfusion medicine specialists should be aware of these parallel efforts in related fields so that, whenever possible, their systems can be made compatible and congruent, even if not communicating directly.

However, inclusion of donor operations should be regarded as integral to a haemovigilance system. Not only are there important donor and patient safety elements to be gleaned from taking a holistic view of the transfusion process, but a system that spans the same vein-to-vein reach of the transfusion process will be better positioned to ‘connect the dots’ to unravel dilemmas and inform policy changes. Already enlightening have been data on the frequency of postdonation reaction rates with different blood volumes collected and delineation of the rate of postdonation death among donors assumed to be healthy. As with the recipient-focused end of the system, standardized terminology and definitions are needed to ensure comparability of recorded observations.

Learning from experience

Haemovigilance systems are sufficiently mature in multiple countries that the medical literature is providing an increasing number of reports of their observations. By their nature, these are observational, but they still provide interesting insights into the problems faced by different transfusion systems and the risks borne by their donors and recipients. It is beyond the scope of this chapter to attempt to replicate all the available data, but the reader is referred to several recent, thorough reports to understand the scope of information available (see Further reading).

Several important, common themes are evident from reviewing these data and some of these observations have led to recommendations that have improved transfusion safety, the intention of haemovigilance systems in the first place.

1. Transfusion of the incorrect unit or component is the most frequent system problem encountered. The problem may manifest as the patient not receiving precisely the component (sub)type that was ordered or may manifest as a fatal haemolytic transfusion reaction. Clearly, inadequate and/or inaccurate identification of the patient or sample/unit at the time of pretransfusion sampling and at transfusion are frequent problems that defy simple solutions but that merit more attention and more capable technology. As shown in the Canadian experience, almost half of all high severity incidents were related to pretransfusion sample collection, and a third of all high severity events where harm occurred were associated with transfusion of the incorrect unit.

2. The greatest mortality risk in most systems currently appears to be transfusion-related acute lung injury (TRALI), although assessment of the frequency of this complication is complicated due to subtle differences in the definitions applied and understanding that significant ‘underrecognition’ is undoubtedly occurring. This highlights again a potentially important role for TSOs in clinician education for the benefit of the transfusion recipient as well as for improved reporting and understanding of the reaction.

3. Data from haemovigilance systems have been helpful not only in defining the frequency of bacterial contamination of platelets but also in investigating the frequency of certain types of contamination after identification of a cluster of incidents and in documenting the effects of implementing interventions to address the problem.

Several applications of haemovigilance data from the SHOT system have been particularly noteworthy in improving the safety of that transfusion system and show the power of haemovigilance systems when their data are applied thoughtfully through evidence-based recommendations (see Further reading). Reports of mistransfusion due to identification problems continued unabated until the recognition that delineation of the problem alone would not effect an improvement. Subsequent implementation of augmented approaches to patient, sample and unit identification was associated with, for the first time, a decline in the number of reported deaths due to mistransfusions. Similar results of interventions have been reported from other systems, such as in France, where ABO-incompatible transfusions were reduced by three-quarters.

Similarly, recognition of the magnitude of the problem posed by TRALI and the high frequency of association of TRALI cases with the plasma of female donors prompted the UK's National Blood Service to reduce the proportion of plasma for transfusion coming from female donors with a subsequent marked reduction in the number of deaths attributable to TRALI. Recognition in one reporting period that transfusion-associated circulatory overload (TACO) was the most common cause of posttransfusion mortality in Québec [8] prompted an increase in education of clinicians on this problem and additional clinical attention to the complication such that a subsequent decline in its frequency was seen.

Haemovigilance data have also led to some unexpected, intriguing observations. Thirteen fatal cases of graft-versus-host disease (GVHD) were reported in a 10-year time span through SHOT, all of them prior to the implementation of universal leucocyte reduction and all but two of them in patients who did not meet usual indications for use of irradiated components [9]. Also, universal leucocyte reduction was associated with a marked reduction in the number of posttransfusion purpura (PTP) cases reported and an apparent shift of these from predominantly red cell recipients (57% from 3%) to include more platelet recipients. This information may be useful in exploring the pathophysiology of these reactions and provide information that may be relevant to blood supply systems not currently employing this approach to component production. Without a haemovigilance system operating ‘in the background’ to amass this experience, these findings might have been missed due to the relative infrequency of GVHD and PTP.

Haemovigilance systems could also be applied to address important yet unanswered questions through the large number of events that they track. For example, the utility of routine premedication with antipyretics and antihistamines has been challenged, but many clinicians continue to believe in their importance. A haemovigilance system could seek to generate a database of sufficient size to address this issue within one jurisdiction or healthcare system so that clinicians could not simply dismiss the data as not being applicable to them. Conversely, the large experience of multiple nations might be able to determine whether the incidence of TRALI is affected by all the plasma in a platelet unit originating from a single donor versus multiple but smaller exposures. There is also considerable interest in applying the power of a haemovigilance system's large purview for ‘Phase IV’ or ‘postmarketing surveillance’ studies to help assess the safety of new interventions, such as pathogen inactivation. This could be accomplished by ensuring that appropriate questions were included on the posttransfusion report form completed by transfusionists or through targeted studies utilizing special reporting tools.

In addition to the evidence-based benefits outlined above, haemovigilance programmes have resulted in number of ‘intangible’ benefits. Participation in a coordinated, standardized programme aimed at improving patient care around transfusions engenders confidence in laboratory and clinical staff, and, particularly for laboratory staff, greater recognition of the role they play in patient care. The benchmarking inherent in haemovigilance is of interest to clinical departments and hospital administrators looking for such data. Computerizing formerly paper-based processes reduces inefficiencies of manual processes. Overall, participating in such national (and international) efforts offers satisfaction from being involved in an even greater good.

Although the stated intention of haemovigilance systems is to improve the safety of transfusion recipients, such systems are also excellently positioned to improve the practice of haemotherapy through the assessment of patient outcomes. There remains wide variation in the application of indications for transfusions and the lack of adoption of published guidelines is sometimes blamed on lack of evidence of their clinical applicability. The data necessary to address these questions are not readily available to most transfusion services at present, but further expansion and integration of laboratory information systems with electronic medical records and linkage with recipient registries may allow future versions of haemovigilance systems wider access to data that could address these and other important transfusion-related questions.

Key points

1. ‘If you can't measure it, you can't improve it’, but measurement alone will not improve systems. Concerted action to improve transfusion systems and reduce transfusion risks is necessary.

2. Haemovigilance systems are most effective when they are based on data reported by clinical practitioners trained to be observant for transfusion-related problems and reported consistently using standardized nomenclature and definitions.

3. Including ‘near-miss’ incidents in haemovigilance reporting provides insights into weak points of the transfusion process and the opportunity to improve the system by reducing the potential for human error to cause harm.

4. The most common problem reported across multiple haemovigilance systems is the transfusion of an ‘incorrect blood component’ and the most dangerous problems encountered frequently relate to sample and patient identification errors.

5. Recognition of serious problems followed by action directed at their cause can improve transfusion safety, as has been seen in steps taken to reduce the frequency of TRALI and ABO-related acute haemolytic events.

6. Haemovigilance systems may be extended to provide important information about haemotherapy decisions and follow-up of new transfusion approaches.

References

1. Juji T, Nishimura M, Watanabe Y, Uchida S, Okazaki H & Tadokoro K. Transfusion-associated graft-versus-host disease. ISBT Sci Ser 2009; 4: 236–240.

2. Rebibo D, Hauser L, Slimani A, Hervé P & Andreu G. The French haemovigilance system: organization and results for 2003. Transfus Apher Sci 2004; 31: 145–153.

3. Stainsby D, Jones H, Asher D et al. (on behalf of the SHOT Steering Group). Serious hazards of transfusion: a decade of hemovigilance in the UK. Transfus Med Rev 2006; 20: 273–282.

4. Faber JC. The European blood directive: a new era of blood regulation has begun. Transfus Med 2004; 14: 257–273.

5. Faber J-C. Work of the European haemovigilance network (EHN). Transfus Clin Biol 2004; 11: 2–10.

6. Robillard P, Chan P & Kleinman S. Hemovigilance for improvement of blood safety. Transfus Apher Sci 2004; 31: 95–98.

7. Callum JL, Merkley LL, Coovadia AS, Lima AP & Kaplan HS. Experience with the medical event reporting system for transfusion medicine (MERS-TM) at three hospitals. Transfus Apher Sci 2004; 31: 133–143.

8. Engelfriet CP & Reesink HW. Haemovigilance. Vox Sanguinis 2006; 90: 207–241.

9. Williamson LM, Stainsby D, Jones H et al. (on behalf of the Serious Hazards of Transfusion Steering Group). The impact of universal leukodepletion of the blood supply in hemovigilance reports of posttransfusion purpura and transfusion-associated graft-versus-host disease. Transfusion 2007; 47: 1455–1467.

Further reading

AuBuchon JP & Whitaker BI. America finds hemovigilance! Transfusion 2007; 47: 1937–1942.

Callum JL, Kaplan HS, Merkley LL et al. Reporting of near-miss events for transfusion medicine: improving transfusion safety. Transfusion 2001; 41: 1204–1211.

de Vries RR, Faber JC, Strengers PF, Board of the International Haemovigilance Network. Haemovigilance: an effective tool for improving transfusion practice. Vox Sanguinis 2011; 100(91): 60–67.

Williamson LM. Transfusion hazard reporting: powerful data, but do we know how best to use it? Transfusion 2002; 42: 1249–1252.