An audit of error rates in a UK district hospital transfusion laboratory

Quantitative assessment of prevalence of pre-analytical variables and their effect on coagulation assay. Can intervention improve patient safety?

BACKGROUND

Very few Indian studies exist on evaluation of pre-analytical variables affecting "Prothrombin Time" the commonest coagulation assay performed. The study was performed in an Indian tertiary care setting with an aim to assess quantitatively the prevalence of pre-analytical variables and their effects on the results (patient safety), for Prothrombin time test. The study also evaluated their effects on the result and whether intervention, did correct the results.

METHODS

The firstly evaluated the prevalence for various pre-analytical variables detected in samples sent for Prothrombin Time testing. These samples with the detected variables wherever possible were tested and result noted. The samples from the same patients were repeated and retested ensuring that no pre-analytical variable is present. The results were again noted to check for difference the intervention produced.

RESULTS

The study evaluated 9989 samples received for PT/INR over a period of 18 months. The prevalence of different pre-analytical variables was found to be 862 (8.63%). The proportion of various pre-analytical variables detected were haemolysed samples 515 (5.16%), over filled vacutainers 62 (0.62%), under filled vacutainers 39 (0.39%), low values 205 (2.05%), clotted samples 11 (0.11%), wrong labeling 4 (0.04%), wrong vacutainer use 2 (0.02%), chylous samples 7 (0.07%) and samples with more than one variable 17 (0.17%). The comparison of percentage of samples showing errors were noted for the first variables since they could be tested with and without the variable in place. The reduction in error percentage was 91.5%, 69.2%, 81.5% and 95.4% post intervention for haemolysed, overfilled, under filled and samples collected with excess pressure at phlebotomy respectively.

CONCLUSION

Correcting the variables did reduce the error percentage to a great extent in these four variables and hence the variables are found to affect "Prothrombin Time" testing and can hamper patient safety.

Blood sample collection and patient identification demand improvement: a questionnaire study of preanalytical practices in hospital wards and laboratories

Scand J Caring Sci; 2010; 24; 581-591 
 Blood sample collection and patient identification demand improvement: a questionnaire study of preanalytical practices in hospital wards and laboratories

BACKGROUND

  Most errors in venous blood testing result from human mistakes occurring before the sample reach the laboratory.

AIMS

  To survey venous blood sampling (VBS) practices in hospital wards and to compare practices with hospital laboratories.

METHODS

  Staff in two hospitals (all wards) and two hospital laboratories (314 respondents, response rate 94%), completed a questionnaire addressing issues relevant to the collection of venous blood samples for clinical chemistry testing.

RESULTS

  The findings suggest that instructions for patient identification and the collection of venous blood samples were not always followed. For example, 79% of the respondents reported the undesirable practice (UDP) of not always using wristbands for patient identification. Similarly, 87% of the respondents noted the UDP of removing venous stasis after the sampling is finished. Compared with the ward staff, a significantly higher proportion of the laboratory staff reported desirable practices regarding the collection of venous blood samples. Neither education nor the existence of established sampling routines was clearly associated with VBS practices among the ward staff.

CONCLUSIONS

  The results of this study, the first of its kind, suggest that a clinically important risk of error is associated with VBS in the surveyed wards. Most important is the risk of misidentification of patients. Quality improvement of blood sample collection is clearly needed, particularly in hospital wards.

Preanalytical mistakes in samples from primary care patients

BACKGROUND

Preanalytical mistakes (PAMs) in samples usually led to rejection upon arrival to the clinical laboratory. However, PAMs might not always be detected and result in clinical problems. Thus, PAMs should be minimized. We detected PAMs in samples from Primary Health Care Centres (PHCC) served by our central laboratory. Thus, the goal of this study was to describe the number and types of PAMs, and to suggest some strategies for improvement.

METHODS

The presence of PAMs, as sample rejection criteria, in samples submitted from PHCC to our laboratory during October and November 2007 was retrospectively analysed.

RESULTS

Overall, 3885 PAMs (7.4%) were detected from 52,669 samples for blood analyses. This included missed samples (n=1763; 45.4% of all PAMs, 3.3% of all samples), haemolysed samples (n=1408; 36.2% and 2.7%, respectively), coagulated samples (n=391; 10% and 0.7%, respectively), incorrect sample volume (n=110; 2.8% and 0.2%, respectively), and others (n=213; 5.5% and 0.4%, respectively). For urine samples (n=18,852), 1567 of the samples were missing (8.3%).

CONCLUSIONS

We found the proportion of PAMs in blood and urine samples to be 3-fold higher than that reported in the literature. Therefore, strategies for improvement directed towards the staff involved, as well as an exhaustive audit of preanalytical process are needed. To attain this goal, we first implemented a continued education programme, financed by our Regional Health Service and focused in Primary Care Nurses.

The suitability of hemolyzed specimens for compatibility testing using automated technology

BACKGROUND

Blood banks have historically rejected hemolyzed specimens for ABO type and antibody screen based on concerns of artifactual interference with test performance or the detection of incompatibility. Samples from emergency departments (EDs) are commonly discarded due to hemolysis or mislabeling.

STUDY DESIGN AND METHODS

This study investigated whether hemolysis produced via experimental mechanical stress has a threshold for introducing discrepancy in ABO-Rh typing and antibody screening using an automated gel testing system (ProVue, Ortho-Clinical Diagnostics, Inc.).

RESULTS

Twenty-three samples from healthy subjects were shown to have threshold supernatant hemoglobin (Hb) levels producing discrepancy in the results for both ABO reverse type and antibody screen at 75 and 125 mg per dL, respectively. Above these levels of Hb, both tests became uninterpretable and reported as "no result determined." Twenty of 31 positive antibody screens became uninterpretable after experimental hemolysis without any threshold supernatant Hb being evident, likely explained by the duration of specimen storage before experimental hemolysis. No false-positive or false-negative samples were observed in the antibody screen of the hemolyzed specimens.

CONCLUSIONS

Properly collected hemolyzed specimens present an opportunity for ABO-Rh forward typing, and a significant proportion of such specimens give valid (concordant) results for reverse type and antibody screen. Our data, although device-specific, are adequate to suggest that sample discard and recollection due to hemolysis in blood bank specimens should be reconsidered with potential benefits for patient safety and efficiency.

Tubes for pretransfusion testing should be collected by blood bank staff and hand labelled until the implementation of new technology for improved sample labelling. Results of a prospective study

BACKGROUND AND OBJECTIVES

The greatest risk in transfusion medicine is actually human error, resulting in the use of the incorrect blood component. The aim of our study was to identify and evaluate the risk factors involved in the collection and labelling of pretransfusion blood samples.

MATERIAL AND METHODS

We prospectively evaluated 6446 samples submitted to the blood bank for pretransfusion testing. Inappropriate samples were classified as 'mislabelled' or 'miscollected'. After 4 months of study, an educational approach was taken.

RESULTS

The frequency of inappropriately labelled samples was 6.45%. Such samples were associated with the use of addressograph labels (vs. hand-written labels) [23.4% vs. 1.4%, P < 0.0001], collection by clinical staff (vs. blood bank staff) [8.8% vs. 2.1%, P = 0.001] and emergency situations (vs. routine sampling) [10.1% vs. 6.1%, P = 0.005]. Following educational intervention, the percentage of inappropriately labelled samples decreased from 7.3% (pre-educational) to 5.8% (post-educational), P = 0.005.

CONCLUSION

Ongoing monitoring and analysis of labelling and collection should be mandatory in order to improve the safety of transfusion.

Single nucleotide polymorphism profiling assay to exclude serum sample mix-up

BACKGROUND AND OBJECTIVES

Sample mix-ups are a threat to the validity of clinical laboratory test results. To detect serum sample mix-ups we developed a single nucleotide polymorphism (SNP) profiling test. SNPs are frequent sequence variations in the human genome. Each individual has a unique combination of these nucleotide variations.

MATERIALS AND METHODS

Predeveloped SNP amplification assays are commercially available. We recently discovered that these SNP assays could be applied to serological samples, which is not self-evident because a key step in serum preparation is removal of white blood cells, the major source of DNA, from blood. DNA was extracted from serum samples. Real-time polymerase chain reaction (PCR) analysis of the purified DNA using a selection of 10 SNP assays provided SNP profiles.

RESULTS

The applicability of the SNP profiling test was demonstrated by means of a case where hepatitis E virus serological determinations of four serum samples of one patient seemed inconsistent. SNP profiling of the samples demonstrated that this was due to the enzyme-linked immunosorbent assay test instead of sample mix-up.

CONCLUSION

We have developed an SNP profiling assay that provides a way to link human serum samples to a source, without post-PCR processing. The chance for two randomly chosen individuals to have an identical profile is 1 in 18 000. Solving potential serum sample mix-ups will secure downstream evaluations and critical decisions concerning the patients involved.

ABO incompatible transfusions--experience from the UK Serious Hazards of Transfusion (SHOT) scheme Transfusions ABO incompatible

The Serious Hazards of Transfusion (SHOT) scheme has now accumulated 8 years' data. The most frequently occurring adverse event, accounting for 1832/2628 (70%) of incidents reported, is 'incorrect blood component transfused' (IBCT) in which the patient receives a blood component that did not meet the correct specification or that was intended for another patient. Errors can occur at all stages of the transfusion chain and, in about half of cases, multiple system failures can be identified. Analysis of 130 ABO incompatible transfusions reported between 1999 and 2003 identified 221 separate errors, 68% of which took place in clinical areas and 29% in hospital laboratories. The commonest single error is failure to check at the bedside that the right blood is being given to the right patient. Certain patients, e.g. neonates, those transfused at night and in critical care situations, appear to be particularly vulnerable. It is encouraging that, against a background of ever increasing numbers year-on-year of reports of IBCT, the incidence of ABO incompatible transfusions already shows a downward trend, suggesting the emergence of a safety culture.

Reducing adverse events in blood transfusion

Against a background of ever increasing expenditure on blood safety, less attention has been paid to improving the safety of the transfusion chain within hospitals. Based on reports to the Serious Hazards of Transfusion (SHOT scheme) between 1996 and 2003, the risk of an error occurring during transfusion of a blood component is estimated at 1:16 500, an ABO incompatible transfusion at 1:100 000 and the risk of death as a result of an 'incorrect blood component transfused' (IBCT) is around 1:1 500 000. There are opportunities for error at a number of critical points in the transfusion chain, starting with the decision to transfuse, prescription and request, patient sampling, pretransfusion testing and finally the collection of the component from the blood refrigerator and administration to the patient, consistently the commonest error in successive SHOT reports. Successive 'Better Blood Transfusion' initiatives and the 2003 Annual Report of the Chief Medical Officer for England have drawn welcome attention to the importance of safe and appropriate transfusion and the National Patient Safety Agency has now set a target of reducing the number of ABO incompatible transfusions by 50% over 3-5 years.

Errors in transfusion medicine

In the past two decades, health services in the developed world have directed much resource toward improving the safety of the blood supply. Blood is collected, tested, and processed within a carefully controlled environment, and quality is assured by rigorous donor selection procedures and increasingly sensitive and sophisticated testing for transfusion-transmitted pathogens. Additional safety strategies implemented by some blood services include leukocyte reduction, bacterial screening, and pathogen inactivation. Thus, the transfusion chain from the donor to the point of issue from the blood center is highly regulated and secure, and transfusion-transmitted infection is an increasingly rare event.

Resolution of a serum sample mix-up through the use of short tandem repeat DNA typing

BACKGROUND

A sample mix-up occurred in a tissue procurement laboratory in which aliquots of serum from two tissue donors were accidentally mislabeled. The clues to the apparent mixup involved discrepant Hepatitis C test results. In an attempt to resolve the apparent mix up, DNA typing was performed using serum samples as a possible source of genomic DNA.

STUDY DESIGN AND METHODS

Two hundred microliter aliquots of two reference sera and aliquots prepared from them were subjected to DNA extraction. PCR amplification of 9 STR loci was performed on the extracts and amplicons were analyzed by capillary electrophoresis.

RESULTS

About 1 microg/ml of DNA was recovered from all serum samples and was of sufficient quality to direct the amplification of most, if not all STR loci allowing the mislabeled specimens to be traced to the proper tissue donor.

CONCLUSIONS

Serum is a useful source of genomic DNA for STR analysis in situations in which such samples are the only source of DNA for testing. Interestingly, one of the tissue donors on life support and repeatedly receiving blood products, exhibited a mixed DNA profile indicative of the presence of DNA from multiple individuals in the bloodstream.

Providing feedback to users on unacceptable practice in the delivery of a hospital transfusion service--a pilot study

The Serious Hazards of Transfusion Scheme's (SHOT) annual report continues to emphasize the importance of investigating serious transfusion errors. It is now recognized that lessons can also be learnt from near-miss events as these occur more frequently than serious errors in transfusion. One of the key features in developing a culture that can promote safety in relation to transfusion is providing feedback to staff on what is acceptable and unacceptable practice. We have developed a scoring system based on the number of serious errors and near misses that occur in the transfusion service to provide feedback to staff on their performance in relation to the administration of blood components and blood products. This was developed as part of an ongoing error logging system that we have previously described. The implementation of this feedback has resulted in an increased awareness within our organization of safety issues in relation to transfusion, and has highlighted the importance of the detection and correction of less serious errors.