PATIENT SAFETY/ORIGINAL RESEARCH

PATIENT SAFETY/ORIGINAL RESEARCH Significant Reduction of Laboratory Specimen Labeling Errors by Implementation of an Electronic Ordering System Paired With a Bar-Code Specimen Labeling Process Peter M. Hill, MD, MS, Darren Mareiniss, MD, JD, Paula Murphy, RN, NC III, Heather Gardner, RN, Yu-Hsiang Hsieh, PhD, Frederick Levy, MD, JD, Gabor D. Kelen, MD From the Department of Emergency Medicine, Johns Hopkins University School of Medicine, Baltimore, MD (Hill, Mareiniss, Hsieh, Levy, Kelen); and Johns Hopkins Hospital, Baltimore, MD (Hill, Mareiniss, Murphy, Gardner, Levy, Kelen). Study objective: We measure the rate of emergency department (ED) specimen processing error reduction after implementation of an electronic physician order entry system paired with a bar-coded specimen labeling process. Methods: A cohort pre- and postintervention study was conducted in the ED during a 61-month period ending September 2008 in a large urban teaching hospital. Historically, laboratory order and requisition processing was done by hand. Interventions included implementing an ED-specific electronic documentation and information system, which included physician order entry with patient verification through bar-coded wristbands and barcoded specimen labels. The main outcome measure was processing error rate, defined as unlabeled/mislabeled/wrong patient specimen or requisition. Pre- and postimplementation data were tabulated monthly and compared in aggregate by 2 test. The contribution of ED error to total institution specimen error was also calculated. Results: Of the 724,465 specimens collected preintervention, 3,007 (0.42%) were recorded as errors versus 379 errors (0.11%) of 334,039 specimens collected postintervention, which represents a 74% relative and 0.31% absolute decrease (95% confidence interval 0.28% to 0.32%). The proportion of institutional errors contributed by the ED was reduced from 20.4% to 11.4%, a 44% relative and 9.0% absolute reduction (95% confidence interval 7.7% to 10.3%). Subanalysis revealed that the majority of continued errors occur when the physician order entry/bar-code system could not be used (eg, blood bank or surgical pathology specimens). Conclusion: Combining an electronic physician order entry with bar-coded patient verification and electronic documentation and information system generated specimen labels can significantly reduce ED specimen-related errors, with sizable influence on institutional specimen-related errors. Continued use of hand labeling and processing for special specimens appears inadvisable, though the cost-effectiveness of this intervention has not been established. [Ann Emerg Med. 2010;56:630-636.] Please see page 631 for the Editor s Capsule Summary of this article. Provide feedback on this article at the journal s Web site, www.annemergmed.com. 0196-0644/$-see front matter Copyright 2010 by the American College of Emergency Physicians. doi:10.1016/j.annemergmed.2010.05.028 INTRODUCTION The Institute of Medicine report To Err Is Human 1 estimated that between 44,000 and 98,000 people die annually in the United States from preventable medical errors in hospitals alone. As a result of this report, there is increasing interest and focus on issues of quality and patient safety. In addition, reducing medical errors and improving patient safety has become a major focus of the hospital accrediting body, The Joint Commission. Patient identification errors occur in about 0.005% to 1% of laboratory samples. 2-4 Approximately 1 in 18 sample identification errors leads directly to an adverse event, 3 and 5% to 27% of all laboratory errors can either affect patient care or cause inappropriate intervention and cause significant clinical effect. 4-6 Accurately collecting and processing laboratory specimens is critical to safe, efficient patient care. 3 The consequences of misidentifying a specimen can range from annoyance, inconvenience, and delayed care to initiating patient therapy based on incorrect data, with potentially serious consequences. There are many opportunities for error in the process of receiving an order for a laboratory specimen, collecting and then labeling the specimen, filling out requisition 630 Annals of Emergency Medicine Volume 56, NO. 6 : December 2010

Hill et al Editor s Capsule Summary What is already known on this topic Specimen processing in the emergency department (ED) is imperfect, and errors occasionally lead to important negative consequences for patients. What question this study addressed Can computerized physician order entry combined with a bar-code system for verifying specimen and patient identity decrease errors in laboratory specimen processing? What this study adds to our knowledge In this single-site, pre-post observational study, reported errors decreased from 0.42% to 0.11% and the percentage of hospital-wide errors attributable to the ED decreased from 20% to 11%. How this is relevant to clinical practice The success of this intervention suggests that other sites can reduce their error rate by using similar strategies. However, this study may have failed to fully capture the kinds of errors more likely to occur in a physician order entry system, and the study did not assess the cost-effectiveness of this intervention. forms for the tests indicated, and then submitting the tests to the laboratory for processing. 7,8 Much work has been done studying this process, and many interventions have been used to enhance safety. 9-12 However, little attention has been paid to addressing error during the process of specimen collection and handling in the emergency department (ED). 13,14 We describe the implementation of a 2-component intervention to enhance the process of ED specimen collection and decrease potential error. Accordingly, the purpose of this study was to determine the effect of a comprehensive informatics-reliant ED-based laboratory specimen procurement system on specimen processing errors. MATERIALS AND METHODS Study Design A cohort pre- and postintervention study was conducted during a 61-month period from September 2004 through September 2009, with a 2-component structured intervention implemented in the ED on May 2008, with retrospective data analysis of data concurrently collected. This represents 44 months of preintervention data and 17 months of postintervention data. The preintervention period was chosen to take advantage of the complete data set available to us, to establish an appropriate baseline, and to observe any trending over time before the intervention. Seventeen months worth of postintervention data was available at data analysis, which we Reduction of Laboratory Specimen Labeling Errors believe is sufficient time to ensure that any effect the intervention has on specimen labeling error reduction is sustainable. The study was considered exempt from review by the institutional review board. Setting The study was conducted in a large, urban, university-based, academic, adult ED with an emergency medicine residency program and an annual census of 57,000. This facility enjoyed a satellite specimen laboratory within the ED that could run simple studies on site or further process specimens for analysis in the main hospital laboratories. Selection of Participants All laboratory specimens collected in the ED and sent to any hospital laboratory from September 2004 through September 2009 were included in the study. These included all hematology, chemistry, microbiology, immunology, pathology, critical care stat tests, urinalysis, drug assays toxicology, pathology, and blood bank specimens. In addition, all total laboratory requests and cancellation rates caused by labeling errors from the entire institution were available for comparison. Interventions Before the intervention, there was a manual specimen ordering and labeling process. Providers would circle or write the name of the requested laboratory study on an official hospital order sheet. The nurse would then take that sheet, stamp a laboratory requisition form, and transcribe the orders by selecting the appropriate study or studies from the form. The nurse would then stamp blank labels, using the embosser with the plastic patient identification card made on registration. The specimen would then be collected, labeled, placed in a bag along with the requisition, and delivered to the ED satellite laboratory. The nurse would verify on the original order form that the specimens had been collected. Implementation of a comprehensive ED electronic documentation and information system including physician order entry combined with bar-code technology was initiated on April 29, 2008 (for purposes of analysis, May 2008 was chosen as the intervention date). The electronic documentation and information system (AllscriptsED, Carey, NC) was integrated with the laboratory information system. Previous standard patient wristbands were changed to one electronically generated, printed, and placed on initiation of the patient s ED visit. These wristbands contain a bar code linked to the patient s identity. Provider laboratory orders through physician order entry generate a message sent to the hospital laboratory information system, containing the patient s name, medical record number, account number, and test code. The laboratory computer system processes that information and sends the information back to a label printer (Datamax, Orlando, FL) proximate to the patient s room. The label that is printed contains the patient s name, medical record number, laboratory test requested, type of Volume 56, NO. 6 : December 2010 Annals of Emergency Medicine 631

Reduction of Laboratory Specimen Labeling Errors specimen container required, and a bar code. Concurrently, a patient-specific red icon displays on the electronic documentation and information system tracking board. This icon alerts the nurse that a specimen has been ordered by the provider. The nurse gathers the phlebotomy supplies, the labels associated with the patient, and a handheld, wireless bar-code reader (Symbol Technologies, Holtsville, NY). The phlebotomist (who is either a nurse or a clinical technician) first scans a bar-code label on the back of his or her own identification badge that records his or her identity in the system as the person collecting the sample, next scans the labels, and finally scans the patient. The electronic documentation and information system then verifies that the labels and studies are as ordered for the intended patient. The specimens are collected, labels are placed on the specimens, and the labels are scanned a second time to verify that the specimens have been collected for the correct person. The red icon on the tracking board then turns yellow. The specimens are then delivered to the laboratory, where they are processed and the studies are performed. When the specimen results return, the yellow icon on the tracking board turns green. The physician order entry system reduces the risk of illegible orders, transcription errors, and misreading of the orders. It also reduces the risk of ordering for the wrong patient by using an active verification process, requiring the providers to verify that they selected the correct patient by prompting them to enter the first 3 letters of the patient s last name before placing the order. This procedure forces the providers to consider whether or not the patient they selected was the one they intended. In addition, this system eliminates the need for paper requisitions required in the manual system. The electronic labeling intervention could not be used for blood bank specimens (institutional policy) or tissue specimens sent to clinical pathology (technical limitations). The intervention was also not used in the critical care bays where Level I trauma and severely ill critical care patients are treated. In these situations, the standard manual labeling process is used. Finally, the manual system is used when the electronic system goes down unexpectedly or for planned maintenance. Data Collection and Processing As part of their normal quality assurance program, the clinical pathology information system keeps a record of each specimen processed in all of their laboratories. The system also keeps a record of how many specimens had to be cancelled because of loss, hemolysis, and specimen labeling error. We acquired the total number of specimens, total number of mislabeled specimens, and percentage of unlabeled specimens for both the ED and the institution from the Department of Pathology Clinical Laboratory monthly monitoring reports. We also recorded, from the same source, the type of mislabeled specimen problem: unlabeled specimen, unlabeled requisition, mismatch between the specimen and requisition label, and a wrong patient specimen. The wrong patient specimen means the label and requisition (if applicable) match, but it is Hill et al determined that the specimen is not from the intended patient. Detection of this type of error occurs in 3 ways: (1) results are recorded for a patient for whom laboratory tests were not ordered, (2) results are not consistent with previous laboratory values for a particular patient, and (3) results do not return for a patient who had samples drawn, and they were never recorded as being received by the laboratory. In each of these instances, detection depends on the recognition by the ordering provider, the phlebotomist, or the laboratory technician that something is wrong. Obviously, there remains the possibility that some of these errors pre- and postintervention were not detected. The same method of data collection was used during both phases of the study period. There was no change in the clinical pathology information system or change in the definition of a specimen processing error during the study period (see next section). Primary Data Analysis The main outcome measure was processing error rate, defined as unlabeled/mislabeled/wrong patient specimen or requisition. Data from the 2 periods were tabulated monthly and compared in aggregate by 2 test. The contribution of reduced ED specimen error to total institution specimen processing error was also calculated. For a subanalysis, for the third and fifth weeks in July and the second week of August 2009 every mislabeled specimen in the ED was investigated to determine the source of error postintervention, ie, either from the electronic or remaining manual process. RESULTS There were 724,465 ED specimens collected during the preintervention period. Of these, 3,007 (0.42%) were reported as mislabeled specimen errors (average of 68 specimens a month). During the postintervention phase, of the 334,039 specimens collected, 379 (0.11%) were identified as errors (average of 22 specimens a month), which represents an absolute error reduction of 0.31% (95% confidence interval [CI] 0.28% to 0.32%) and a relative error reduction of 74% (Table 1). Figure 1 displays monthly data and shows an abrupt error reduction immediately after the intervention, and the results of the intervention have been sustained. The total hospital error rate preintervention was 0.21%. Although ED-generated samples accounted for only 10.3% of specimens handled by the laboratory, the ED accounted for 20.4% of the institutional specimen errors (Tables 1 and 2). After the intervention, the total hospital laboratory labeling error rate was 0.10% (95% CI 0.10% to 0.11%), a 52% reduction. ED-based specimens accounted for 10.2% of all institutional requests postintervention, but ED errors accounted for only 11.4% (44% proportionate reduction) of all errors (Tables 1 and 2). Figure 2 suggests that the institutional error rate had been steadily decreasing during the entire period studied, coincident with the slow introduction of physician order entry and bar-code labeling throughout the hospital. However, the ED intervention accounts for 0.03% of the 632 Annals of Emergency Medicine Volume 56, NO. 6 : December 2010

Hill et al Reduction of Laboratory Specimen Labeling Errors Table 1. ED and institutional error rates before and after intervention. Location Preintervention, September 2004 April 2008 Postintervention, May 2008 September 2009 Errors Specimens % Errors Errors Specimens % Errors Absolute Difference, % (95% CI)* ED 3,007 724,465 0.42 379 334,039 0.11 0.31 (0.28 032) Institution (including ED) 14,755 7,067,197 0.21 3,328 3,259,951 0.10 0.11 (0.10 0.11) Institution (excluding ED) 11,748 6,342,732 0.19 2,949 2,925,912 0.10 0.08 (0.08 0.09) * 2 Test. Figure 1. Total and percentage of ED errors from September 2004 through September 2009. Table 2. Proportion of ED and institutional error rates for specimen processing.* Specimen Preintervention Postintervention ED, N Institution, N % ED ED, N Institution, N % ED Risk Difference, % (95% CI) Errors 3,007 14,755 20.4 379 3,328 11.4 9.0 (7.7 10.3) Specimen total 724,485 7,067,197 10.3 334,039 3,260,003 10.2 *Institution figures are inclusive of ED data. 2 Test. Figure 2. Total and percentage of institutional errors from September 2004 through September 2009. 0.11% absolute decrease in error realized by the institution during the study period. The more dramatic decrease in error rate observed in the ED compared with the institution as a whole suggests a larger benefit in error reduction when the more rigorous physician order entry and bar-code system is applied to a more chaotic, error-prone environment. Distribution of types of errors seen in the ED before and after intervention is shown in Table 3. The proportion of unlabeled requisitions and mislabeled specimens had absolute reductions of 4% (95% CI 1.4% to 6.6%) and 12.8% (95% CI 7.6% to 17.9%), respectively, whereas the proportion of unlabeled and wrong-patient specimens had absolute increases of 8.3% (95% CI 3.1% to 13.6%) and 8.4% (95% CI 4.6% to 12.3%), respectively. The majority (60%) of errors that still occur are during processes that continue to rely on manual processing (blood bank requests, critical care) (Table 4). LIMITATIONS The study was conducted at a single site, and manual specimen processing and electronic documentation and Volume 56, NO. 6 : December 2010 Annals of Emergency Medicine 633

Reduction of Laboratory Specimen Labeling Errors Hill et al Table 3. Distribution of error rate types before and after intervention. Preintervention Postintervention Error % Error Error % Error Risk Difference, % (95% CI) Unlabeled specimen 1,011 33.6 159 42.0 8.3 ( 13.6 to 3.1) Unlabeled requisition 303 10.1 23 6.1 4.0 (1.4 to 6.6) Mislabeled specimen/requisition 1,455 48.4 135 35.6 12.8 (7.6 to 17.9) Wrong patient specimen 238 7.9 62 16.4 8.4 ( 12.3 to 4.6) Specimen total 3,007 379 Table 4. Three-week sampling for source of error postimplementation.* Source of Error Number (%) Transfusion Medicine system 13 (34) Critical care 10 (26) Pathology 0 Electronic system 15 (40) Total 38 *Dates: Third and fifth weeks of July and first week of August 2009. information-system-enhanced processing may differ in other institutions. There were other patient safety initiatives both before and after the implementation of the intervention, aimed at improving many other ED processes. However, these initiatives appear to have slight if any effect before the intervention (Figure 1), and given low effectiveness of general measures before study intervention, these other safety initiatives are unlikely to account for error reduction experienced. There were ongoing initiatives to reduce the specimen-related error rate throughout the hospital at the same time, such as slowly implementing physician order entry and bar-coded specimen labeling, but again they appear to have slight to no effect on ED performance before the intervention, and none of the other hospital interventions involved the ED. Detection of most errors depends on the laboratory, and it is possible that some errors were not detected. Wrong-patient specimen errors can also be detected by the phlebotomist and the practitioner who ordered the specimen. The method of detecting these errors (see Data Collection and Processing and Discussion ) would be unaffected by the intervention, and there is no reason to think that the rate of error detection would change coincident with the date of the intervention. Finally, an electronic system can introduce new source of error. Some of these errors are related to workarounds within the electronic system (see Discussion ). DISCUSSION Specimen identification errors have the potential to cause serious harm and injury to patients. Such injuries have been reported in both the media and peer-reviewed literature. 15,16 In addition, these types of errors have resulted in legal action. In an analysis of 272 surgical pathology legal claims, 13 (5%) involved allegations that the specimen had been incorrectly identified. 17 In a study by Bonini et al, 4 the majority of laboratory errors resulted from specimen misidentification. One of the few large multi-institutional studies to examine specimen misidentification found similar results. 3 In a study by Valenstein et al, 3 laboratories at 120 participating institutions prospectively monitored specimen misidentification for 5 weeks. Of the 6,705 errors reported, the cause of the error was identified for 72% (4,852) of the samples. Of these, 55.5% were due to primary specimen label error. Valenstein et al 3 estimated that as many as 160,900 adverse events a year could result from specimen errors. 3 Accordingly, it appears that proper identification and verification of specimens would address many of the reported causes of error and potentially result in improved patient safety. One method that has been shown to improve the accuracy of specimen labeling and patient identification is the use of computerized bar-coding systems to match specimens to patients. 12 Such a system was used by Hayden et al 12 within a large oncology hospital and had been shown to significantly reduce clinical laboratory specimen identification error. In this 3-year study, investigators observed a decrease in the mislabeled specimen rate from 0.03% to 0.005% after the implementation of a bar-code-based verification process. In addition, 2 other studies have used similar bar-coding systems and also significantly reduced patient identification error while using point-of-care testing devices. 18,19 ED application of this technology is not well described in the literature. Killeen et al 14 presented an abstract in 2005 describing a significant error reduction associated with an electronic documentation and information system physician order entry system combined with a patient and specimen bar-coding process in the ED. 14 Although consistent with our findings, the study was limited in its duration and sample size and did not evaluate the contribution the ED had to the institutional error rate pre- and postintervention. Further, we were unable to find a corresponding peer-reviewed publication. As the gateway to care in the hospital, the ED presents an added level of danger for mislabeling errors. The chaotic and hectic ED environment contributes to more labeling errors at baseline (Table 1). Targeting the ED for methods of error reduction is therefore important. In addition, a mislabeled sample and incorrect result in the ED has the added risk of resulting in wrong diagnosis or treatment, or at least delaying care and increasing the patient s length of stay. There is a direct relationship between longer lengths of stay of ED patients and ED crowding, causing increasing wait times and an increasing 634 Annals of Emergency Medicine Volume 56, NO. 6 : December 2010

Hill et al number of patients leaving without being seen. 20 Therefore, the potential for harm of a mislabeled specimen is not just to the involved patient but also to the pool of patients waiting for emergency evaluation and care. Before the intervention in our ED, specimen labeling was a manual process of visually comparing the identity on the stamped labels containing 2 patient identifiers, with the identity on the patient s wristband. During preintervention data collection, there were many initiatives in our ED related to improved specimen processing, including direct bedside labeling, in an attempt to reduce the mislabeled specimen rate. These included signs posted throughout the ED (including at the window of the laboratory specimen drop-off counter), reminding people about the importance of proper identification and bedside labeling, reminders during nursing meetings and by e-mail, multiple in-services, and education campaigns. However, as our data showed, these initiatives had minimal measurable and sustainable effect on the error rate. The use of the new electronic process that incorporates physician order entry forces the bar-code verification process that must occur at the bedside. Potential for human error is significantly eliminated. Of interest is that much (60%) of the residual error postintervention is due to those specimens that by either institutional policy or due to workflow problems must be processed manually. Blood bank requisitions and critical care specimens are those that can least tolerate error or patient delay. Given our findings, the practice of manual processing of any specimen must be reexamined. 21 The data represent only those errors that were identified and reported. We do not know how many mislabeled specimens went undetected and how many patients were harmed as a result during the study period. This problem as it relates to studying laboratory specimen error rates is well described in the literature. 3 However, it is reasonable to assume that reduction in detected errors would also reflect a reduction in undetected errors. The system we describe still produces patient labeling error. As seen in Table 3, practitioners can still obtain a specimen and submit it to the laboratory without placing a label on the container. Because the computer-based system eliminates the need for a requisition, errors related to unlabeled or mislabeled requisitions are substantially eliminated and can only be associated with specimens not collected with the physician order entry and bar-code system. Shortcuts and workarounds are ubiquitous in medicine, and in this case the label printer offers a workaround for the phlebotomist. A bar-code scanner was attached to the label printer that can be used to scan the specimen labels instead of at the bedside. Extra bar-code patient labels are printed at registration to be used for various purposes during the patient s visit. The phlebotomist can use one of these labels to act as the patient s armband and also scan it at the label printer. Therefore, specimen labels can be scanned and verified remotely and not by using the patient s wristband. In addition, Reduction of Laboratory Specimen Labeling Errors we also learned that the laboratory can still process specimens collected when the patient s wristband was not scanned. Through case-by-case investigation, these workarounds are chiefly responsible for the wrong patient specimen errors that occur when the new physician order entry and bar-code system are used. The bar-code scanners have now been removed from the label printer. The ED will pilot the next phase of this technology: bedside label printing generated only after scanning the patient s bar-coded wristband. This eliminates labels being printed outside patient rooms, further ensuring that the label matches the patient. In addition, we have eliminated the practice of printing extra labels, forcing the phlebotomist to scan the patient s wristband to produce the specimen labels. Finally, it is certainly possible that electronic systems will increase other types of ordering errors not detected or reported to the laboratory. For example, although it is easy to hand write an order for the wrong patient, especially if you do not remember the patient s name, this error may be compounded in an electronic system when the names are close together on a computer screen and selecting the wrong patient is fairly easy. In addition, if patients were switched from one room to another without the provider s being notified and the provider does not remember the patient s name, he or she may simply be ordering tests for the patient in room 15 without paying attention to the name associated with that bed. In this last case, the dialog box prompting the provider to verify that the patient selected is the patient for whom they intend to write the order may not be enough to stop the ordering process. These errors would be recognized only when the expected test result does not return for the intended patient. In summary, in this environment, electronic documentation and information system physician order entry integrated with the central hospital laboratory, as well as bar coding and barcode scanning, significantly reduced laboratory specimen processing errors in the ED, with influence on overall institutional errors. Further, these data suggest that this solution should be applied to all specimen handling in the ED. However, the cost-effectiveness of this intervention has not been established. Supervising editor: David L. Schriger, MD, MPH Author contributions: PMH, PM, and HG conceived and designed the study. PMH supervised the study conduct. PM and HG were responsible for data collection and acquisition. Y-HH conducted statistical analysis. PMH, Y-HH, and GDK analyzed the data. PMH, DM, FL, and GDK drafted the article and were responsible for critical article revision. PMH takes responsibility for the paper as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article that might create any potential conflict of interest. The authors have stated that no such relationships exist. See the Volume 56, NO. 6 : December 2010 Annals of Emergency Medicine 635

Reduction of Laboratory Specimen Labeling Errors Hill et al Manuscript Submission Agreement in this issue for examples of specific conflicts covered by this statement. Publication dates: Received for publication December 2, 2009. Revisions received March 26, 2010, and April 14, 2010. Accepted for publication May 4, 2010. Available online September 6, 2010. Presented at the Society of Academic Emergency Medicine annual meeting, May 2009, New Orleans, LA. Address for correspondence: Peter M. Hill, MD, MS, Department of Emergency Medicine, Johns Hopkins Hospital, Marburg B-186, Baltimore, MD 21287; 410-955-8708, fax 410-614-0141; E-mail phill@jhmi.edu. REFERENCES 1. Kohn LT, Corrigan JM, Donaldson MS, eds. To Err Is Human: Building a Safer Health System. Committee on Quality of Health Care in America, Institute of Medicine. Washington, DC: National Academy Press; 1999. 2. Valenstein PN, Sirota RL. Identification errors in pathology and laboratory medicine. Clin Lab Med. 2004;24:979-996. 3. Valenstein PN, Raab SS, Walsh MK. Identification errors involving clinical laboratories: a College of American Pathologists Q-Probes study of patient and specimen identification errors at 120 institutions. Arch Pathol Lab Med. 2006;130:1106-1113. 4. Bonini P, Plebani M, Ceriotti F, et al. Errors in laboratory medicine. Clin Chem. 2002;48:691-698. 5. Plebani M, Carraro P. Mistakes in a stat laboratory: types and frequency. Clin Chem. 1997;43:1348-1351. 6. Nutting PA, Main DS, Fischer PM, et al. Problems in laboratory testing in primary care. JAMA. 1996;8:635-639. 7. Astion ML, Shojania KG, Hamill TR, et al. Classifying laboratory incident reports to identify problems that jeopardize patient safety. Am J Clin Pathol. 2003;120:18-26. 8. Wagar EA, Tamashiro L, Yasin B, et al. Patient safety in the clinical laboratory: a longitudinal analysis of specimen identification errors. Arch Pathol Lab Med. 2006;130:1662-1668. 9. Wagar EA, Stankovic AK, Raab S, et al. Specimen labeling errors: a Q-Probes analysis of 147 clinical laboratories. Arch Pathol Lab Med. 2008;132:1617-1622. 10. Dock B. Improving the accuracy of specimen labeling. Clin Lab Sci. 2005;18:210-212. 11. Francis DL, Prabhakar S, Sanderson SO. A quality initiative to decrease pathology specimen-labeling errors using radiofrequency identification in a high-volume endoscopy center. Am J Gastroenterol. 2009;104:972-975. 12. Hayden RT, Patterson DJ, Jay DW, et al. Computer-assisted barcoding system significantly reduces clinical laboratory specimen identification errors in a pediatric oncology hospital. J Pediatr. 2008;152:219-224. 13. Quillen K, Murphy K. Quality improvement to decrease specimen mislabeling in transfusion medicine. Arch Pathol Lab Med. 2006; 130:1196-1198. 14. Killeen JP, Chan TC, Jones KGDA. Impact of bar coding technology and computerized physician order entry on reducing laboratory specimen misidentification errors in the emergency department. Acad Emerg J. 2005;12(suppl 1):49. 15. CBS Evening News Early Show. Mastectomy mistake fuels debate. January 21, 2003. Available at: http://www.cbsnews. com/stories/2003/01/18/health/main537085.shtml. Accessed November 30, 2009. 16. Sazama K. Reports of 355 transfusion-associated deaths: 1976 through 1985. Transfusion. 1990;30:583-590. 17. Troxel DB. Error in surgical pathology. Am J Surg Pathol. 2004; 28:1092-1095. 18. Colard DR. Reduction of patient identification errors using technology. Point Care. 2005;4:61-63. 19. Rao AC, Burke DA, Dighe AS. Implementation of bar-coded wristbands in a large academic medical center: impact on pointof-care error rates. Point Care. 2005;4:119-122. 20. Kyriacou DN, Ricketts V, Dyne PL, et al. A 5-year time study analysis of emergency department patient care efficiency. Ann Emerg Med. 1999;34:326-335. 21. Dzik WH. New technology for transfusion safety. Br J Haematol. 2006;136:181-190. Did you know? Annals accepts audio and video files as ancillaries to the main article. Visit http://www.annemergmed.com/content/instauth/ for more details! 636 Annals of Emergency Medicine Volume 56, NO. 6 : December 2010