Specimen labeling errors: a Q-Probes analysis of 147 clinical laboratories.
* Context.--Accurate specimen identification is critical for
quality patient care. Improperly identified specimens can result in
delayed diagnosis, additional laboratory testing, treatment of the wrong
patient for the wrong disease, and severe transfusion reactions.
Specimen identification errors have been reported to occur at rates of
0.1% to 5%.
Objective.--To determine the frequency of labeling errors in a multi-institutional survey.
Design.--Labeling errors were categorized as: (1) mislabeled, (2) unlabeled, (3) partially labeled, (4) incompletely labeled, and (5) illegible label. Blood specimens for routine or stat chemistry, hematology, and coagulation testing were included. Labeling error rates were calculated for each participant and tested for associations with institutional demographic and practice variable information.
Results.--More than 3.3 million specimen labels were reviewed by 147 laboratories. Labeling errors were identified at a rate of 0.92 per 1000 labels. Two variables were statistically associated with lower labeling error rates: (1) laboratories with current, ongoing quality monitors for specimen identification (P = .008) and (2) institutions with 24/7 phlebotomy services for inpatients (P = .02). Most institutions had written policies for specimen labeling at the bedside or in outpatient phlebotomy areas (96% and 98%, respectively). Allowance of relabeling of blood specimens by primary collecting personnel was reported by 42% of institutions.
Conclusions.--Laboratories actively engaged in ongoing specimen labeling quality monitors had fewer specimen labeling errors. Also, 24/7 phlebotomy services were associated with lower specimen error rates. Establishing quality metrics for specimen labeling and deploying 24/7 phlebotomy operations may contribute to improving the accuracy of specimen labeling for the clinical laboratory.
Medical laboratories (Management)
Medical errors (Surveys)
Wagar, Elizabeth A.
Stankovic, Ana K.
Nakhleh, Raouf E.
Walsh, Molly K.
|Publication:||Name: Archives of Pathology & Laboratory Medicine Publisher: College of American Pathologists Audience: Academic; Professional Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2008 College of American Pathologists ISSN: 1543-2165|
|Issue:||Date: Oct, 2008 Source Volume: 132 Source Issue: 10|
|Topic:||Event Code: 200 Management dynamics Computer Subject: Company business management|
|Product:||Product Code: 7397030 Biomedical Testing Labs NAICS Code: 54138 Testing Laboratories SIC Code: 8071 Medical laboratories; 8731 Commercial physical research|
|Organization:||Organization: College of American Pathologists|
|Geographic:||Geographic Scope: United States Geographic Code: 1USA United States|
In early surveys, laboratory errors were classified in several
ways, including cause, phase of testing, responsible party, and impact
on the patient. (1) Data from these studies and other sources have
shaped our thinking and caused a shift in the approach to laboratory
errors. We have now come to recognize that most errors occur outside of
the analytic phase and are often beyond the immediate control of the
Patient identification and specimen handling are error-prone steps that occur primarily outside the laboratory. In a recent survey of error types associated with invasive procedures, 10 of 17 were related to patient identification failure. (4) The College of American Pathologists has surveyed wristband identification errors in a Q-Tracks format and shown that continuous quality monitoring improves patient identification practices. (5,6) Both the College of American Pathologists and the Joint Commission have implemented patient identification as a major patient safety goal for 2008. (7,8) Specimen labeling errors are a recognized preanalytic source of concern for appropriate patient identification management. In several studies, it has been determined that specimen identification errors occur at a rate ranging from 0.1% to 5%. (9-12)
The focus of this survey was to determine the frequency of specimen labeling errors from multiple institutions for routine and stat priority chemistry, hematology, and coagulation blood specimens. The goals were to: (1) better define the rate of specimen identification errors by using data from multiple institutions, and (2) identify statistically significant associations with demographics and laboratory practices that may assist individual laboratories in improving their specimen labeling procedures.
MATERIALS AND METHODS
This study was conducted from April 1, 2007, through May 31, 2007, with final data submission by June 8, 2007. Participants prospectively reviewed labels from chemistry, hematology, and coagulation specimens until a total of 30 labeling errors were identified or until specimen labels were reviewed for 30 days, whichever came first. Labels were reviewed to determine whether the required information on them was complete (contained 2 patient identifiers and any additional information required by the laboratory) and corresponded to the specimen requisition. For those labels for which the information was either incomplete or did not correspond to the requisition, the labeling errors were categorized into the following types: (1) mislabeled, (2) unlabeled, (3) partially labeled, (4) incompletely labeled, and (5) illegible label. If more than 1 error type occurred per label, all error types were recorded. The definitions for the error types and terminology used in this study are shown in Table 1.
All blood specimens received for routine or stat chemistry, hematology, and coagulation testing were included in the study. All secondary specimens (whether previously aliquoted at another laboratory or within the laboratory) were excluded. All nonblood specimens, esoteric testing specimens (ie, molecular pathology), and specimens destined for transfusion medicine, microbiology, and anatomic pathology laboratory sections were also excluded.
A total of 147 laboratories participated in this survey. Not every laboratory responded to every question in the demographics profile and in the practices questionnaire In some cases, information provided in earlier surveys was used to complete the demographics information. Most (95.2%) of the 147 participating institutions were located in the United States, with the remaining institutions located in Canada (4), Australia (1), South Korea (1), and Spain (1). Of the participating institutions, 42.0% were teaching hospitals, and 29.9% had a pathology residency program. Within the past 2 years, the College of American Pathologists inspected 85.0% of the laboratories, and inspections were conducted by the Joint Commission at 38.3% of the institutions. Table 2 displays the self-reported characteristics of the participating institutions.
The rate of specimen labeling errors was tested for associations with the institutional demographics and practice variable information, as well as with each of the 5 defined labeling error types. This was done using nonparametric Kruskal-Wallis tests for categorical variables and rank-transformed linear regression analysis for quantitative variables. A P value of less than .05 was considered statistically significant.
Participants from 147 institutions reviewed a total of 3324888 specimen labels and identified 3043 labeling errors (0.92 errors per 1000 labels) during the study period. The median number of labeling errors identified per institution was 24, with a range of 0 to 30. The median number of specimen labels reviewed was 13301, with a range of 53 to 402 976. The distribution of the quality indicator is shown in Table 3.
Aspects of Practice and Their Effect on Specimen Labeling Errors
Demographic and practice variable data were evaluated to identify associations with lower rates of specimen labeling errors. Significant associations (P < .05) were found with several practice variables. These relationships are listed in the following text and in Table 4. A lower rate of specimen labeling errors occurred in laboratories that have current, ongoing quality monitors for specimen identification in addition to this Q-Probes study (P = .008) and in institutions that have 24/7 phlebotomy services for in patients (P = .02). No other significant associations were found with the demographics or practice variables.
Other Study Findings
Participants categorized labeling errors into 5 error types and, if necessary, more than 1 error type per label was allowed. For all institutions combined, the greatest number of errors resulted from mislabeled specimens (29.9%), and the fewest were from illegible labels (6.1%). Data are listed in Table 5.
For each specimen label containing 1 or more errors, participants recorded the following information: patient type, requisition type, if the requisition was complete, if the label was handwritten, and if a barcode was included on the primary label. Labeling errors were reported about equally for inpatients versus other collection sites. There was no statistically significant difference in reported errors based on collection site. Also, 50.8% of all specimen requisitions were paper requisitions. Of these, 29.5% were received as hand-labeled specimens, and 39.4% contained a barcode on the primary specimen label. Table 6 lists these data.
The 50th (median), 10th, and 90th percentiles for specimen volumes, and the associated estimates of percentage of tests performed stat are shown in Table 7. These are self-reported data sets (compared with study collected data) from each of the participating laboratories. Also reported is the type of personnel collecting inpatient and outpatient specimens: (1) phlebotomists, (2) nurses, (3) physicians, and (4) other health care personnel (ie, care partners). Most phlebotomy was performed by phlebotomists and nurses, regardless of patient location.
Table 8 shows the self-reported practices of the participating clinical laboratories. A high percentage of institutions had a written policy for specimen labeling at the bedside and in outpatient phlebotomy areas (95.8% and 97.9%, respectively). Laboratories are still receiving a mixture of paper and electronic order and labeling mechanisms. For example, most institutions still only produce barcode labels for specimens at the laboratory information system level (59.3%). Paper requisitions are received from both inpatient and outpatient areas by 49.6% of the participants.
Of the reporting institutions, 80.6% had a current ongoing quality monitor for specimen identification that was primarily managed within the laboratory, in addition to this Q-Probes study. Of the institutions maintaining a quality monitor, 81.7% reported their results to other institutional administrators.
Interestingly, laboratories showed more variability in their procedures for relabeling of specimens. A total of 42% of reporting laboratories (60 laboratories) allowed relabeling of blood specimens by the primary collecting personnel. Of these, 68.3% allowed relabeling if patient identifiers were missing or incorrect. About 93% had a written procedure for relabeling specimens. The majority of participating laboratories (86.2%) had 24/7 phlebotomy for inpatients, most of which managed the phlebotomy services from within the laboratory (79.2%; Table 8).
Additional statistical analysis was also performed examining the relationship of lower percentages of specimen errors to each of the 5 labeling error types. Lower specimen error reporting was noted with a lower rate of partially labeled specimens and a lower percentage of incomplete labels, with P < .001 in both circumstances (Table 9). These findings are not surprising, since the specimen error rate is expected to be lower in clinical laboratories that are actively managing specimen labeling problems in their service environments.
Of the demographics and practice variables tested for association with lower specimen labeling error rates, 2 relationships were statistically significant: (1) laboratories that have current, ongoing quality monitors for specimen identification in addition to the Q-Probes study, and (2) institutions that have 24/7 phlebotomy services for inpatients. Laboratories showed an awareness of the importance of specimen labeling and were regularly monitoring accurate specimen labeling practices. It is perhaps not surprising that laboratories that are actively engaged in a specimen-labeling quality monitor have fewer specimen labeling errors. When attention is paid to a problem, results improve, as demonstrated previously in a Q-Tracks study of wristband identification errors. (4) Also, 82% of the laboratories engaged in a quality monitor reported their results to a hospital or other institutional administration in addition to internal laboratory reporting. Although these activities intuitively agree with our expectations, the results of this survey statistically support the use of quality monitors in improving laboratory performance and provide an added argument for using them in the management of specimen labeling errors.
The correlation between 24/7 inpatient phlebotomy services and lower specimen labeling errors is an important finding. Phlebotomists are focused on specimens, compared with nurses and others who have multiple tasks. Phlebotomists are, therefore, likely to be aware of all aspects of specimen collection at a higher level, including specimen identification. Laboratories prefer to control phlebotomy in inpatient settings based on an assumption that with control, the quality can be better managed. More than 86% of the survey participants maintained 24/7 phlebotomy services.
Proof that 24/7 phlebotomy improves specimen identification has remained elusive, especially when arguing for increased commitment of resources with hospital or other administration representatives. A recent study in a single institution associated implementation of 24/7 phlebotomy services with a reduction in laboratory specimen labeling errors by trend analysis. (13) Although we have analyzed data so as to minimize the effect of the large amount of variability (by using nonparametric methods), it was difficult to separate other variables, such as educational efforts, in confirming the value of the 24/7 phlebotomy service implementation. Also, this study did not include transfusion medicine specimens, since the labeling of specimens for the blood bank may follow different requirements. This Q-Probes study suggests, however, that there is significant value for 24/7 phlebotomy and provides an added justification for the resource allocation in institutions considering this option for specimen collection.
Also noted in this survey was the allowance of relabeling of blood specimens by primary collecting personnel at 42% of reporting institutions. Whether or not a clinical laboratory should allow specimens to be relabeled is a somewhat controversial issue. Clearly, for irretrievable specimens (eg, cerebrospinal fluid, blood cultures taken prior to antibiotic therapy), there are typically provisions for relabeling. Where relabeling was allowed by primary collecting personnel, 68.3% of participating laboratories permitted it when the patient identifiers were missing or incorrect. Identifying best practices regarding the relabeling of specimens may be an important topic for individual laboratories to review. Similarly, given the prevalence of relabeling in this survey, a future study may be of value.
The range of error reporting was wide in this survey. As noted in Table 3, the rate of specimen labeling errors per 1000 labels at the 10th percentile was 52 per 1000. The median (50th percentile) was 1.31 errors per 1000 labels. Only 0.22 specimen labeling errors per 1000 labels occurred at the 90th percentile (higher percentile ranks indicate better relative performance). The wide range of error reporting may indicate variability in laboratory practices.
The most common error type was the mislabeled specimen (29.9% of reported specimen labeling errors). Partially labeled, unlabeled, and incomplete labels were 22.7%, 21.9%, and 20.7% of reported errors, respectively. Illegible labels accounted for only 6.1% of all reported labeling errors (Table 5). Mislabeled specimens are perhaps the most serious errors, since they may be associated with subsequent problems for patient care. A consequence of mislabeled specimens can be a wrong result for a patient, which may, in turn, cascade into wrong therapy. Mislabeled specimens can have a significant impact on patient safety and health care costs. Laboratories may wish to focus on mislabeled specimens for a quality monitor, since this category is the most frequent error type.
Given that only 39.4% of specimens had barcoded labels, information technology solutions may not be easy to implement. The penetration of electronic solutions remains problematic and expensive for many institutions. To assure prevention of specimen identification errors, such conversions may be facilitated by accrediting organizations, such as the Joint Commission.
Coordination of laboratory quality activities with institutional dashboards and metrics is part of many ongoing institutional quality improvement initiatives. As patient safety goals become more prominent drivers in health care, laboratories may be asked to participate in an ever-widening circle of performance measurements.14 Specimen identification error analysis is an important metric that contributes not only to institutional goals but to improvements in patient safety.
We thank Bushra Yasin, PhD, for her assistance in reviewing this manuscript, and Christine Bashleban for her valuable assistance in assembling the data.
Accepted for publication March 1 1, 2008.
(1.) Bonini P, Plebani M, Ceriotti F, et al. Errors in laboratory medicine. Clin Chem. 2002;48:691-698.
(2.) Plebani M. Errors in clinical laboratories or errors in laboratory medicine? Clin Chem Lab Med. 2006;44:750-759.
(3.) Plebani M, Carraro P. Mistakes in a stat laboratory; types and frequency. Clin Chem. 1997;43:1348-1351.
(4.) Howanitz PJ, Renner SW, Walsh MK. Continuous wristband monitoring over 2 years decreases identification errors: a College of American Pathologists Q-Tracks study. Arch Pathol Lab Med. 2002;126:809-815.
(5.) Zarbo RJ, Jones BA, Friedberg RC, et al. Q-Tracks: a College of American Pathologists program of continuous laboratory monitoring and longitudinal tracking. Arch Pathol Lab Med. 2002;126:1036-1044.
(6.) Chassin M, Becher EC. The wrong patient. Ann Intern Med. 2002;136:826-833.
(7.) 2008 Joint Commission national patient safety goals. Available at: http://www.jcpatientSafety.org. Accessed April 30, 2008.
(8.) CAP patient safety goals. Available at: http://www.cap.org/ docs/laboratory_accreditation/checklists/laboratory_general_sep07.pdf. Accessed April 30, 2008.
(9.) Valenstein PN, Sirota RL. Identification errors in pathology and laboratory medicine. Clin Lab Med. 2004;24:979-996.
(10.) Ihojie J, Rbaniak SJ. Comparing near misses with actual mistransfusion events: a more accurate reflection of transfusion errors. Br J Haematol. 2000;108: 458-460.
(11.) Novis DA. Detecting and preventing the occurrence of errors in the practices of laboratory medicine and anatomic pathology: 15 years of experience with the College of American Pathologists Q-PROBES and Q-TRACKS programs. Clin Lab Med. 2004;24:965-978.
(12.) Howanitz PJ. Errors in laboratory medicine: practical lessons to improve patient safety. Arch Pathol Lab Med. 2005;129:1 252-1261.
(13.) 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.
(14.) Institute of Medicine. To Err is Human: Building a Safer Health System. Washington, DC: National Academy Press; 2000.
Elizabeth A. Wagar, MD, FCAP; Ana K. Stankovic, MD, PhD, FCAP; Stephen Raab, MD, FCAP; Raouf E. Nakhleh, MD, FCAP; Molly K. Walsh, PhD
From the Department of Pathology and Laboratory Medicine, University of California Los Angeles Clinical Laboratories, Los Angeles (Dr Wagar); Preanalytical Systems, BD Diagnostics, Franklin Lakes, NJ (Dr Stankovic); the Department of Pathology, University of Colorado, University of Colorado Denver Health Sciences Center, Anschutz Medical Campus, Aurora (Dr Raab); the Department of Pathology, St Luke's Hospital, Jacksonville, Fla (Dr Nakhleh); and the Department of Biostatistics, College of American Pathologists, Northfield, Ill (Dr Walsh).
The authors have no relevant financial interest in the products or companies described in this article.
Reprints: Elizabeth A. Wagar, MD, FCAP, Department of Pathology and Laboratory Medicine, UCLA Clinical Laboratories, AL-206 CHS, 1 0833 Le Conte Ave, Los Angeles, CA 90095-1 732 (e-mail: firstname.lastname@example.org).
Table 1. Definition of Terms Properly labeled specimen: A properly labeled specimen includes 2 correct patient identifiers (name and birth date, or name and unique institutional identifier) attached to a tube of the identified patient's blood, as well as any additional information required by the individual laboratory. Mislabeled specimen: A mislabeled specimen is one on which 1 or both patient identifiers are incorrect. A specimen with a label from a different patient, 2 contradictory labels on 1 specimen, or labels that did not correspond to the requisition were considered mislabeled. Unlabeled specimen: A blood specimen received in the clinical laboratory without a label or without any patient identifiers on the label. Partially labeled specimen: A specimen with a label that lists only 1 patient identifier. Incomplete requisition: A specimen received with a requisition that does not include 2 patient identifiers. For the purpose of this study, specimens with an incomplete requisition were categorized as partially labeled specimens. Incomplete specimen label: A specimen that is labeled correctly with 2 patient identifiers but lacks information that the laboratory requires, or the additional information is incorrect. For example, the date and time of blood draw, sex, or specimen type may be missing from the label. Criteria in addition to 2 patient identifiers may vary from institution to institution and were not limited to the examples provided above. Illegible specimen label: A specimen label that has illegible patient identifiers that cannot be read electronically or manually. Specimen/requisition mismatch: A specimen received with a requisition that does not match the request and/or patient identified on the specimen tube. For the purpose of this study, specimens with a requisition mismatch were categorized as mislabeled specimens. Table 2. Institution Demographics No. of Percentage of Institutions Institutions Institution type Private, nonprofit 60 56.6 State, county, or city 10 9.4 hospital Private, profit 7 6.6 University hospital 7 6.6 Independent laboratory 6 5.7 Governmental, federal 3 2.8 Other 13 12.3 Total 106 100.0 Occupied bed size 0-150 35 34.3 151-300 37 36.3 301-450 18 17.6 451-600 5 4.9 >600 7 6.9 Total 102 100.0 Institution location City 55 51.4 Suburban 34 31.8 Rural 18 16.8 Total 107 100.0 Governmental affiliation Nongovernmental 90 84.9 Nonfederal governmental 13 12.3 Federal governmental 3 2.8 Total 106 100.0 Table 3. Distribution of the Rate of Specimen Labeling Errors (Quality Indicator) All Institutions' Percentiles * 50th 10th 25th (Median) 75th 90th Rate of specimen labeling 52.27 7.15 1.31 0.41 0.22 errors per 1000 labels * Higher percentile ranks indicate better relative performance. Table 4. Relationship Between Practice Variables and the Rate of Specimen Labeling Errors Median Specimen Institutional No. of Labeling Error Practice Institutions Rate/1000 Labels Ongoing quality monitor for specimen identification Yes 116 1.08 No 28 2.41 Total 144 24/7 phlebotomy service Yes 125 1.09 No 20 2.74 Total 145 Table 5. Frequency of Labeling Error Types (n = 3011 Labels) No. of Percentage Labeling Error Type Errors of Errors * Mislabeled 901 29.9 Partially labeled 685 22.7 Unlabeled 659 21.9 Incomplete labels 622 20.7 Illegible labels 185 6.1 Total 3052 101.3 * Multiple error types were allowed per label. Table 6. Patient, Requisition, and Label Types for Specimens Containing Labeling Errors No. Percent * Patient type Inpatient 1470 48.5 Other 1563 51.5 Total 3033 100.0 Requisition type Paper 1515 50.8 Electronic 1470 49.2 Total 2985 100.0 Complete requisition Yes 2636 89.6 No 307 10.4 Total 2943 100.0 Handwritten label Yes 870 29.5 No 1474 50.1 Not applicable 600 20.4 (ie, unlabeled specimen) Total 2944 100.0 Barcode on primary label Yes 1182 39.4 No 1233 41.1 Not applicable 587 19.5 (ie, unlabeled specimen) Total 3002 100.0 Table 7. Distributions of Specimen Volume, Percentage of Stat Tests, Collection Personnel Type, and Estimated Rate of Barcode Misalignment All Institutions' Percentiles 50th 10th (Median) 90th n * Total No. of specimens 90 720 427 251 1 500 000 141 received in the clinical laboratory during the previous 12 mo Estimated percentage of 5 25 45 143 tests performed stat Collection personnel type for inpatient specimens, % Phlebotomists 0 75 98 141 Nurses 2 15 60 141 Physicians 0 0 5 141 Other health care 0 0 50 141 personnel (ie, care partners) Collection personnel type for outpatient specimens, % Phlebotomists 23 76 98 125 Nurses 0 10 50 125 Physicians 0 0 1 125 Other health care 0 1 45 125 personnel (ie, care partners) Estimated percentage of 0 1 10 61 barcode misalignment, if barcode reader is employed in specimen processing equipment or other laboratory instrumentation * n represents the number of reporting institutions. Table 8. Laboratory Practices With Respect to Specimen Labeling No. of Percentage of Institutions Institutions Handwritten, paper requisitions accompany specimens from: Inpatient area only 3 2.2 Outpatient area only 65 48.2 Both inpatient and 67 49.6 outpatient areas Total 135 100.0 Electronic patient wristband identification systems (barcoding) used for order entry Yes 14 9.8 No 129 90.2 Total 143 100.0 Barcode labels produced by laboratory or hospital information system (LIS or HIS) for phlebotomy * LIS only 86 59.3 HIS only 12 8.3 Both LIS and HIS 35 24.1 Not applicable 12 8.3 Total 145 100.0 Written policy for specimen labeling at the bedside Yes 139 95.8 No 2 1.4 Not applicable 4 2.8 Total 145 100.0 Written policy for specimen labeling in outpatient phlebotomy areas Yes 142 97.9 No 1 0.7 Not applicable 2 1.4 Total 145 100.0 Laboratory has a current ongoing quality monitor for specimen identification in addition to this Q-Probes study Yes 116 80.6 No 28 19.4 Total 144 100.0 If laboratory has a current ongoing specimen identification quality monitor in addition to this Q-Probes study, is it primarily managed within the laboratory? Yes 113 97.4 No 3 2.6 Total 116 100.0 If laboratory has a current ongoing quality monitor in addition to this Q-Probes study, is it routinely reported to hospital or other institutional administration in addition to laboratory personnel? Yes 94 81.7 No 21 18.3 Total 115 100.0 Relabeling of blood specimens by primary collecting personnel allowed Yes 60 42.0 No 83 58.0 Total 143 100.0 If relabeling is allowed by primary collecting personnel, is relabeling permitted when the patient identifiers are missing or incorrect? Yes 41 68.3 No 19 31.7 Total 60 100.0 If relabeling is allowed by primary collecting personnel, does laboratory have a written procedure for relabeling specimens? Yes 54 93.1 No 4 6.9 Total 58 100.0 24/7 phlebotomy service for inpatients Yes 125 86.2 No 20 13.8 Total 145 100.0 Laboratory management of phlebotomy services (ie, phlebotomists are laboratory employees) Yes 114 79.2 No 27 18.7 Not applicable 3 2.1 Total 144 100.0 * LIS indicates laboratory information system; HIS, hospital information system. Table 9. Relationship of Percentage of Specimen Errors With Partially Labeled and Incompletely Labeled Specimens Median Specimen No. of Labeling Error Institutions Rate/1000 Labels Percentage of errors with partially labeled specimens 0-9 83 0.69 10-39 35 2.74 40-74 16 3.46 75-100 12 4.75 Percentage of errors with incompletely labeled specimens 0 76 0.65 1-9 19 2.18 10-49 25 4.17 50-100 26 4.55
|Gale Copyright:||Copyright 2008 Gale, Cengage Learning. All rights reserved.|