Monitoring transfusionist practices: a strategy for improving transfusion safety

BACKGROUND: Data from New York State indicate that about 1 of every 33,000 red cell units transfused is ABO-incompatible with the recipient. National application of these data suggests that as many as 360 ABO-incompatible whole blood and red cell transfusions might occur annually in the United States. Phlebotomy and blood bank laboratory errors cause some of these ABO-incompatible transfusions, but the greatest number result either partially or solely from the failure of transfusionists to identify properly either a patient or the blood component a patient receives. STUDY DESIGN AND METHODS: A quality assessment/quality improvement (QA/QI), process is described that allowed for the direct oversight (monitoring) of transfusionists' practices and for the assessment of institutional policies for blood administration. RESULTS: At the beginning of the QA/QI process, monitoring of blood administration practices revealed that a variance from institutional blood administration policy occurred during 50 percent of blood and component transfusions. As a result of the QA/QI process, the percentage of transfusions with an associated variance from institutional policy dropped to nearly zero. CONCLUSION: The QA/QI process described in this report, or one similar to it, could improve transfusion safety and serve as a model for increased involvement by transfusion service medical directors in the oversight of transfusionists' practices.     

Failure of bedside ABO testing is still the most common cause of incorrect blood transfusion in the Barcode era.

BACKGROUND AND OBJECTIVES. ABO-incompatible red blood cell (RBC) transfusions are a major risk in transfusion medicine. Identification of factors leading to this hazard is important to improve transfusion safety. MATERIAL AND METHODS. All consecutive erroneous ABO-incompatible transfusions occurring from January 1997 to December 2004 at the Charité University Hospital in Berlin, Germany were analysed. RESULTS. A total of 343,432 RBC units were transfused, and eight patients erroneously received 13 ABO-incompatible RBC concentrates. The most frequent error was incorrect bedside testing (n=7). Intensive care treatment was required in two cases, but there were no fatal mistransfusions. Four patients had no or only mild reactions. CONCLUSION. Mistransfusions are still a considerable risk in transfusion medicine despite quality control systems and electronic data processing. An increase in transfusion safety may require the introduction of further systems, e.g. radio-frequency identification (RFID) tags.     

Evaluation of pediatric transfusion practice using criteria maps

Pediatric blood transfusion practice in a tertiary-care pediatric hospital was evaluated retrospectively by using the technique of criteria mapping. A total of 630 transfusion episodes involving red cell concentrates, frozen plasma (plasma frozen within 24 hours of collection), platelet concentrates, and albumin were reviewed: 243 (86.2%) were reviewed only by a technical assistant, and 87 (13.8%) required additional physician review. Of these, 138 were red cell concentrate transfusions: 79.7 percent of that group were considered appropriate, 11.6 percent of unknown benefit/risk ratio, 5.8 percent inappropriate, and 2.9 percent impossible to evaluate. Some 246 frozen plasma transfusions were reviewed: 42.3 percent were considered appropriate, 32.5 percent of unknown benefit/risk ratio, 17.5 percent inappropriate, and 7.7 percent impossible to evaluate. A total of 139 platelet concentrate transfusions were reviewed: 64.7 percent were considered appropriate, 16.5 percent of unknown benefit/risk ratio, 10.1 percent inappropriate, and 8.6 percent impossible to evaluate. Some 107 albumin transfusions were reviewed: 90.6 percent were considered appropriate, 1.9 percent inappropriate, and 7.5 percent impossible to evaluate. The criteria maps developed for this study were easy for the technical assistant to use, and areas of appropriate and inappropriate pediatric transfusion practice were clearly identified.     

Transfusion management of sickle cell patients during bone marrow transplantation with matched sibling donor

BACKGROUND: Sickle cell disease (SCD) patients have unique transfusion considerations during bone marrow transplantation (BMT), including prophylaxis against stroke and alloimmunization. Characterization of transfusion requirements is important for blood bank and clinician patient management.
STUDY DESIGN AND METHODS: A retrospective analysis of red blood cell (RBC) and platelet (PLT) transfusion of SCD patients during myeloablative matched sibling donor (MSD) BMT at one institution from 1993 to 2007 was performed. Patient characteristics (RBC blood group antibodies, ABO-incompatible donor, BMT-related morbidity) and transfusion practices (RBC phenotype matching, transfusion threshold, and blood age) were assessed for effect on total RBC transfusion volumes.
RESULTS: Twenty-seven patients received MSD BMT with 96% survival and 0% rejection. Six alloimmunized patients received RBCs with extended phenotype matching (C, c, E, e, K, Fy^a, Jk^b), 14 nonalloimmunized received limited matching (C, c, E, e, K), and 7 did not have phenotype matching. Among 26 survivors, a median seven RBC transfusions (range, 3-15) and 13.5 PLT transfusions (range, 4-48) per patient were administered, equivalent to 64 mL/kg RBCs (range, 22-122 mL/kg) and 106 mL/kg PLTs (range, 26-343 mL/kg). BMT-related morbidity predicted increased RBC transfusions (p = 0.006). Venoocclusive disease was associated with greater RBC (p = 0.016) and PLT transfusion volumes (p = 0.016). Greater phenotype matching was associated with decreased RBC transfusions (p = 0.0247).
CONCLUSIONS: SCD patients have high transfusion support during MSD BMT. Communication of BMT complications to the blood bank is essential for transfusion inventory management. Phenotype matching decreased RBC transfusions in this cohort and warrants further investigation in SCD transfusion therapy.     

Enhanced detection of blood bank sample collection errors with a centralized patient database

BACKGROUND: Pittsburgh's Centralized Transfusion Service (CTS) provides transfusion support to 16 hospitals and features an electronic database that contains patient transfusion and serologic histories. This database can be accessed from any hospital in the system. A major cause of ABO-incompatible transfusions is the "wrong blood in tube" (WBIT) phenomenon, that is, the sample is not from the recipient identified on the label. We hypothesized that having access to patient historical ABO types from anywhere in the CTS system can identify WBIT errors and prevent mistransfusions.
STUDY DESIGN AND METHODS: The transfusion committee records of the 16 CTS hospitals from March 2005 to September 2007 were reviewed for major collection errors, that is, the current ABO type differed from the historical type in the database. The patient's historical ABO type, the discrepant type, and the hospital(s) where these samples were collected were recorded.
RESULTS: In 6 of 16 major collection errors for which complete information was available, the current and historical ABO types were obtained from different hospitals within the CTS system. In 3 cases, selection of ABO type–specific blood based on the current sample would have led to an ABO-compatible transfusion (e.g., correct type A, current type O). In the other 3 cases, an ABO-incompatible transfusion would have resulted (e.g., correct type O, current type A).
CONCLUSIONS: Access to a centralized patient database detected 38 percent more ABO typing errors and prevented six mistransfusions, which would not have been prevented at a single institution. Centralization of patient transfusion data should be encouraged.     

Reports of 355 transfusion-associated deaths: 1976 through 1985

From 1976 through 1985, the United States Food and Drug Administration received reports of 355 fatalities associated with transfusion, 99 of which were excluded from further review because they were unrelated to transfusion or involved hepatitis or acquired immune deficiency syndrome. Of the remaining 256 reported deaths, 51 percent resulted from acute hemolysis following the transfusion of ABO-incompatible products. These deaths were due primarily to managerial, not clerical, errors. Other causes of death (in order of frequency of report) included acute pulmonary injury (15%), bacterial contamination of product (10%), delayed hemolysis (10%), damaged product (3%), and graft-versus-host disease (0.4%). Management systems for transfusion facilities should be created or revised to include the specific identification of personnel eligible to administer transfusions to provide written guidance and appropriate training (including recognition and management of errors), and to implement measures that target safe transfusion practices. Continued research into acute pulmonary injury, the immunologic hazards of transfusion, and the prevention of bacterial contamination of blood components is necessary.     

An analysis of errors in blood component transfusion records with regard to quality improvement of data acquisition and to the performance of lookback and traceback procedures

BACKGROUND: Quality assurance of blood transfusion covers institutions, personnel, and procedures involved in preparing, issuing, and using blood components. The accuracy of data related to blood component transfusions is a tool for quality control in the transfusion service. STUDY DESIGN AND METHODS: A study of the accuracy of data records of the transfusion service at the University Hospital of Erlangen, Germany, between June 1994 and May 1996 was carried out. All returned blood component transfusion report forms were examined for discrepancies between primary data records and clinical transfusion reports. RESULTS: Blood components (n = 49,224) from allogeneic and autologous donations, packed red cells, fresh-frozen plasma, and platelet components that had been issued for transfusion were included in this evaluation. For 27.3 percent of all components issued, no transfusion report was returned to the blood bank. For the remaining 35,786 units, errors were found in 3.8 percent of the records. For 1.24 percent of all components, discrepant information related to the recipient's identity or the component's status was found; this affected the feasibility of lookback or traceback searches. CONCLUSION: A remarkably high frequency of discrepancies exists between computerized blood bank records and the information recorded on returned blood transfusion forms. The processes of data acquisition and entry must be included in quality assurance efforts in transfusion medicine.     

Reporting of near-miss events for transfusion medicine: improving transfusion safety

BACKGROUND: Half of the reported serious adverse events from transfusion are a consequence of medical error. A no-fault medical-event reporting system for transfusion medicine (MERS-TM) was developed to capture and analyze both near-miss and actual transfusion-related errors. STUDY DESIGN AND METHODS: A prospective audit of transfusion-related errors was performed to determine the ability of MERS-TM to identify the frequency and patterns of errors. RESULTS: Events and near-miss events (total, 819) were recorded for a period of 19 months (median, 51/month). No serious adverse patient outcome occurred, despite these events, with the transfusion of 17,465 units of RBCs. Sixty-one events (7.4%) were potentially life-threatening or could have led to permanent injury (severity Level 1). Of most concern were 3 samples collected from the wrong patient, 13 mislabeled samples, and 22 requests for blood for the wrong patient. Near-miss events were five times more frequent than actual transfusion errors, and 68 percent of errors were detected before blood was issued. Sixty-one percent of events originated from patient areas, 35 percent from the blood bank, and 4 percent from the blood supplier or other hospitals. Repeat collection was required for 1 of every 94 samples, and 1 in 346 requests for blood components was incorrect. Education of nurses and alterations to blood bank forms were not by themselves effective in reducing severe errors. An artifactual 50-percent reduction in the number of errors reported was noted during a 6-month period when two chief members of the event-reporting team were on temporary leave. CONCLUSION: The MERS-TM allowed the recognition and analysis of errors, determination of patterns of errors, and monitoring for changes in frequency after corrective action was implemented. Although no permanent injury resulted from the 819 events, innovative mechanisms must be designed to prevent these errors, instead of relying on faulty informal checks to capture errors after they occur.     

The release of unsuitable units through misinterpretation of laboratory results transmitted by facsimile

BACKGROUND: Preventable errors in transfusion medicine that have a significant risk of adverse outcome include the erroneous administration of blood of the wrong type or blood with unsuitable laboratory test results. Mandatory reports of errors by facilities providing blood services in New York State offer the opportunity for review and analysis of common factors. The state also collects statistics on the collection, laboratory testing, distribution, and disposition of blood in these facilities. CASE REPORTS: Three serious errors in transfusion medicine occurred within 3 months as a result of the misinterpretation of laboratory test results transmitted by facsimile. Two unsuitable units of blood were erroneously released, and a unit mislabeled as to the ABO group was incorporated into the hospital inventory. One of the unsuitable units was repeatedly reactive for HIV (although negative on confirmatory testing), and the other was confirmed positive for HCV. CONCLUSIONS: The vast majority of blood collected in New York State is tested by reference laboratories. Results are often transmitted by facsimile. Facsimile results may be misinterpreted because of distortion during transmission, misreading, or failure to note a separate report of pending results. Such misinterpretation results in an increased risk of adverse outcome for transfusion recipients. Laboratory results to be transmitted by facsimile could readily-and should-be clarified.     

Nationwide survey of home transfusion practices

BACKGROUND: Limited information exists on home transfusion practices. STUDY DESIGN AND METHODS: In 1995, a survey requesting data for 1994 was sent to 1273 American Association of Blood Banks (AABB) institutional members and 113 non-AABB home health care agencies that provide out-of-hospital transfusions. RESULTS: Of 943 respondents, 102 provide blood to a home transfusion program, 37 provide blood and run a home transfusion program, and 13 run a home transfusion program only, for a total of 152 (16%) with some involvement in home blood transfusions. Most of the 50 respondents with a home transfusion program are licensed by their state and accredited by the Joint Commission on Accreditation of Healthcare Organizations. All respondents have written policies for home transfusion, and 90 percent require a signed informed-consent document before initiating transfusions in the home. Most have policies requiring that there be a second adult and a telephone in the home, that the home be deemed safe for transfusion, that the patient's physician be readily available, and that the patient have had prior transfusions. The most common component issued by the blood providers was red cells, followed by platelets. White cell-reduced components were always provided by 36 percent of respondents. The most common patient diagnosis was cancer. Home transfusions were provided primarily by registered nurses. Only 14 percent of respondents indicated that the medical director of the blood bank is responsible for approving a patient for home transfusion. A posttransfusion visit is performed by 46 percent of respondents. CONCLUSION: Although most facilities have policies for the administration of home transfusions, there remains marked heterogeneity among blood providers and transfusionists regarding home transfusion practices.     

Blood transfusion costs: a multicenter study

The cost of delivering a unit of blood (whole blood or red cells) to a hospitalized patient was examined in 19 United States teaching hospitals. The average hospital acquisition cost was calculated by using the prices charged by regional blood centers for blood products. To this cost was added an estimate of costs incurred by hospitals for handling, testing, and administering blood. Across study sites, the average hospital cost per unit transfused was $155 and the average charge to the patient was $219. Acquisition cost, the price that hospitals pay for blood, was 37 percent of the total cost to the hospital; the other 63 percent of the hospital cost included costs for blood bank handling (13%), laboratory tests (43%), and blood administration (7%). Significant variations in blood transfusion cost were found within our sample. Most of the variability can be attributed to geographic location of the blood supply source, type of red cell product transfused, prices charged by blood transfusion services, and the frequency of laboratory tests. The results of this transfusion cost study may be helpful in determining the costs of health care delivery, especially when blood transfusions are indicated.     

Incidence of red cell antibodies after multiple blood transfusion

A retrospective study was performed to estimate the frequency of alloimmunization against red cell (RBC) antigens in a multiply transfused group. Patients (n = 186) were studied who had received at least six blood transfusions during a period of at least 3 months. Some 6944 units of blood were transfused. One hundred forty patients had hematologic disorders. The patients' sera were investigated every 3 months with indirect antiglobulin tests and enzyme-treated RBCs. Twenty-two patients (11.8%) made 33 antibodies. Seven patients made more than one antibody. Eight of the 22 patients (36.4%) made their first antibody before or at the 10th transfusion. The risk of immunization increased with the number of transfusions. Influence of gender and age was not demonstrable. Nor was a relationship demonstrated between blood transfusion reactions and RBC antibody formation; no delayed hemolytic transfusion reactions occurred. Anti-E was demonstrated in 12 patients and anti-K in 15. When the gene frequencies were taken into account, it appeared that anti-E was made by 11.5 percent of E-negative patients, most of whom were immunized after an estimated three transfusions with E-positive blood. Anti-K was made by 8.7 percent of the K-negative patients, after an estimated 2.1 units of K-positive blood. It might be desirable to match red cell units for the E and K antigens in patients at relatively high risk. These are primarily patients who have already formed an antibody and are going to receive many transfusions and women of childbearing age who are to receive more than 4 units of blood.     

利用两次血型制减少ABO不相容输血风险的研究

目的评估两次血型制在防止标本差错导致ABO不相容输血中的作用。方法对于需要或可能需要输血的患者,实施输血前采集2份单独的标本测定ABO/RhD血型。记录和分析输血前标本中错误标识和错误采集的比率。结果2007年7月—2009年6月,共接收到79115人份输血前相容性测试标本,检出273份错误标识的标本,错误标识标本比为1/290;校正的错误采集标本比为1/1485(基于47507份多次标本中检出23份错误采集标本,校正因子为1.3912)。在23份错误采集的标本中,9份(39%)可导致ABO不相容的红细胞输注。结论两次血型制可减少因标本错误标识和错误采集所致ABO不相容输血的风险。     

Immunochemical characterization of the anti-B response in an ABO-incompatible blood transfusion: presence of antibodies recognizing glucosylceramide

Summary. A male patient of blood group A was accidentally given two units of group B blood when operated upon because of injury. This resulted in an immediate haemolytic transfusion reaction. The patient was treated with exchange transfusions in order to remove the incompatible red cells and thus prevent renal damage. Two weeks after the transfusion, an anti-B humoral immune response appeared. The antibodies were mainly IgG and IgA types, of IgG3 and IgA1 subclasses. The anti-B antibodies seemed specific for the B trisaccharide [Galα1–3(Fucα1–2)Gal] as tested by radioimmunoassay and chromatogram binding assay. No antibodies recognizing part of the core saccharide chain could be detected, i.e. antibodies capable of differentiating between B type 1 and B type 2 chains did not occur. An unsuspected finding was that the patient had IgG antibodies recognizing glucosylceramide (and weakly also galactosylceramide) before and after the transfusion. These antibodies were still present 9 years after the ABO-incompatible blood transfusion.     

Blood component therapy during the neonatal period: a national survey of red cell transfusion practice, 1985

A questionnaire to determine patterns of neonatal red cell transfusion practice during 1985 was mailed to 2200 blood banks of American Association of Blood Banks (AABB) institutional members and children's hospitals. There were 915 responses (41.6%); 785 responses (86%) contained sufficient data for analysis. The majority (70.6%) of 785 responding hospitals were community/urban institutions. However, more highly specialized, pediatric hospitals were also represented by 92 university/tertiary-care hospitals (11.7% of respondents) and 29 children's hospitals (3.7% of respondents). Two-thirds of hospitals performed a major antiglobulin crossmatch (rather than an abbreviated one) before all neonatal red cell transfusions. The red cell preparation most frequently selected for small-volume transfusions was ABO and Rh group-specific red cell concentrates. When performing only large-volume exchange transfusions, 19.2 percent of hospitals used whole blood; all others prepared reconstituted units of red cells plus fresh-frozen plasma, a practice that frequently causes exposure to two donors per unit. Another practice likely leading to multiple donor exposure is the use of fresh-frozen plasma to adjust the hematocrit of red cell preparations to a predetermined value prior to a small-volume transfusion. Over one-half of hospitals adjusting hematocrits used plasma, presumably from one donor, to dilute packed red cells from another donor, a practice that has no apparent medical benefit. Most hospitals (63.4%) provided red cells with a reduced risk of transmitting cytomegalovirus; blood from seronegative donors was selected by 65 percent of hospitals. The majority of hospitals, including most of the community/urban hospitals, did not irradiate blood products before transfusion.(ABSTRACT TRUNCATED AT 250 WORDS)     

ABO血型不合异基因造血干细胞移植后并发纯红系再生障碍

为了研究ABO血型不合异基因造血干细胞移植（allo—HSCT）后并发纯红细胞再生障碍（pure red cell aplasia，PRCA）的发病情况及危险因素，对本医院以往血型不合异基因造血干细胞移植进行回顾性分析。探讨移植后患者PRCA的发病危险因素。研究结果表明，72例ABO血型不合allo—HSCT患者中，4例发生PRCA，其中A供O3例，A供B1例。PRCA的发生不影响急性移植物抗宿主病（GVHD）或巨细胞病毒（CMV）感染的发生。PRCA患者红系恢复的时间显著长于未PRCA发生患者。结论：PRCA是ABO血型不合移植的主要并发症。A供O可能是ABO血型不合allo—HSCT后并发PRCA的危险因素。     

Measures to decrease the risk of acquired immunodeficiency syndrome transmission by blood transfusion. Evidence of volunteer blood donor cooperation

We studied whether volunteers giving blood to the Greater New York Blood Program (GNYBP) cooperated with procedures implementing public health recommendations intended to decrease the risk of acquired immunodeficiency syndrome (AIDS) transmission by blood transfusion. Predonation medical screening was expanded to exclude donors who might be ill with AIDS. To exclude possible asymptomatic carriers of the disease, members of groups at increased risk of AIDS were asked either not to give blood or to give it for laboratory studies. A confidential questionnaire, administered to all donors after medical screening, provided the vehicle for donors to advise the GNYBP whether their donation was for laboratory studies or for patient transfusion. We found that the number of male donors decreased; AIDS-related questions in medical history led to a 2 percent increase in donor rejections; 97 percent of donors said their blood could be used for transfusions; 1.4 percent said their blood could be used for laboratory studies only; and 1.6 percent did not respond. Only units designated for transfusion were released to hospitals. People who indicated that their donation was for laboratory studies had a higher prevalence of markers for hepatitis B virus and of antibodies to cytomegalovirus. White cell counts and helper/suppressor T lymphocyte ratios were not significantly different in the two groups. We conclude that volunteer donors have cooperated with the established procedures. None of the laboratory assays identified blood units donated by individuals who, based on information about AIDS high-risk groups, designated their donation for laboratory studies.     

IMMUNOHEMATOLOGY: Identification and evaluation of false‐negative antibody screens

Background: Non‐ABO alloantibodies are frequently implicated in hemolytic transfusion reactions and are a leading cause of transfusion‐related mortality. Detection of clinically significant non‐ABO alloantibodies is reliant on an antibody screen, which is prone to clerical, technical, and reagent error. Data on the frequency of false‐negative antibody screens due to the occurrence of these errors are scarce, and the true incidence of false‐negative antibody screens in everyday practice is unknown. STUDY DESIGN AND METHODS: Monitoring for false‐negative antibody screens is routinely performed in our institution. All cases of false‐negative antibody screens involving clinically significant antibodies were identified through review of the blood bank quality assurance records from 2004 to 2007. The clinical impact was recorded in each case. RESULTS: Twenty‐one cases of false‐negative antibody screens due to clinically significant antibodies were detected. Sources of error included testing error (12 cases), reagent red blood cell (RBC) failure (one case), and reagent limitations (one case). The cause of error was inconclusive in seven cases. Nine patients were found to have received antigen‐incompatible blood as a consequence of these errors, resulting in a single nonfatal hemolytic transfusion reaction. CONCLUSIONS: The identification and investigation of false‐negative antibody screens is a valuable quality assurance measure which can serve to monitor staff performance, identify cases of reagent RBC failure, and identify patients who have received antigen‐incompatible blood at risk for hemolytic transfusion reactions.