| Abstract|| |
Aims: (a) To evaluate the types and frequencies of preanalytical errors occurring in a tertiary care hematology diagnostic center and (b) To evaluate differences if any, across groups [outpatient data (OPD) vs inpatient data (IPD), type of test requested [complete blood count (CBC) vs coagulation] and laboratory (routine vs emergency). Settings and Design: A prospective study was conducted over a period of nine months (August 2017–April 2018) to address the above objectives. All samples received in the clinical hematology division of our institute were included in the analysis. Materials and Methods: Categories of preanalytical errors were defined. This included insufficient, clotted, diluted, and lipemic samples. Clerical errors such as wrong/absent sample labeling, requisition form-sample mismatch, and wrong vacutainer selection were also documented. IPD and OPD data, as well as data pertaining to samples sent for different tests [complete blood count (CBC)/coagulation] and in the routine and emergency laboratories, were segregated. Statistical Analysis Used: All errors in each category were recorded as numbers and corresponding percentages (proportions). The two-tailed z-test was applied to assess the significance of the difference in proportions across all groups. Statistical significance was kept at P < 0.05. Results: A total of 189,104 samples were received in the clinical hematology laboratory during the aforementioned period, out of which preanalytical errors were found in 4052 (2.14%) samples. Inadequate sample quantity (ISQ) comprised the bulk of preanalytical errors in our laboratory (1.11% of total samples) followed by sample clots (0.88%). There was no significant difference in the error frequencies in OPD and IPD (P = 0.1031). The proportion of errors was higher in routine vis-à -vis emergency samples and also in samples sent for coagulation analysis vis-à -vis CBC.
Keywords: Error, hematology, laboratory testing, phase, preanalytical
|How to cite this article:|
Gaur K, Puri V, Shukla S, Sharma S, Suman S, Singh R. Finish before the start: Analyzing preanalytical sample errors in a tertiary care hematology laboratory. Indian J Pathol Microbiol 2020;63:435-40
|How to cite this URL:|
Gaur K, Puri V, Shukla S, Sharma S, Suman S, Singh R. Finish before the start: Analyzing preanalytical sample errors in a tertiary care hematology laboratory. Indian J Pathol Microbiol [serial online] 2020 [cited 2020 Dec 3];63:435-40. Available from: https://www.ijpmonline.org/text.asp?2020/63/3/435/291660
| Introduction|| |
The clinical hematology laboratory plays a pivotal role in the delivery of efficient and high-quality healthcare services to patients. More often than not the baseline complete blood count (CBC) marks the initiation of diagnostic assessment. A successful test result may be defined as one that is accurate, precise, reproducible, timely delivered, and error-free. With the advent of automated sample analyzers, errors in laboratory testing have registered a dramatic near 300-fold reduction., However, despite the significant progress made, test errors still exist despite conscientious efforts made to weed out the same.
The process of testing patient samples coined as the “total testing process” (TTP) has traditionally been described by Lundberg in 1981, as all processes falling within the “brain–brain loop”, i.e., from the clinician's selection of the test (brain-mediated) to the analysis of the final result (again brain-mediated). The TTP incorporates three phases –preanalytical, analytical, and postanalytical.
Out of the three, the preanalytical phase is the most complex, having multiple stakeholders and hence is the most vulnerable to error. Previous work has described errors in the preanalytical phase to be as high as 62 %. Studies focussing exclusively on errors in laboratory hematology services however are extremely few. Prior studies from Asia have largely been from private laboratories/institutions.,, To the best of our knowledge, this is the largest such work from a teaching-based medical center in Southeast Asia.
We undertook the present study to evaluate the types and frequencies of preanalytical errors occurring at our center, examining the differences if any across groups (OPD vs IPD), routine vs emergency laboratories, and type of test requested (routine CBC vs coagulation).
| Materials and Methods|| |
This study was conducted in a 1200-bedded teaching hospital catering to all clinical specialties. This was a prospective work conducted over a period of nine months (August 2017–April 2018) at the clinical pathology unit of our institute, comprising routine clinical hematology and emergency pathology laboratories. Both laboratories adhere to internal and external quality control checks, and participate in the National external quality assurance program. The emergency laboratory caters to samples taken out after the noon from either inpatients, outpatients, or casualty. All 189,104 samples received in this time frame were included in the study. This included 133,142 IPD and 55,962 OPD samples. Phlebotomy for inpatients was carried out by interns/nursing staff and by trained laboratory personnel for outpatients in the centralized sample collection facility. As part of the routine laboratory workflow, the samples from various wards/OPD collection facility were transported in the morning for analysis by paramedical staff. Blood samples were collected in standard 3 ml Becton Dickinson (BD™) vacutainer tubes—K2EDTA for hemogram and sodium citrate (3.2%) for coagulation testing. It was ensured that expired vacutainers were not used and samples were analyzed using five-part automated hematology analyzers (Model: Sysmex™ XT-2000i). As the study was conducted on documenting the quality of patient samples without any identification/experimentation/invasive procedures/additional sampling whatsoever, it was exempted from ethical clearance, as per the guidelines of the institutional ethics committee.
Various categories of sample-related preanalytical errors were delineated for the purpose of this study. This included inadequate sample quantity (ISQ) defined as all samples having a sample volume less than the manufacturer specified mark, clotted, hemolyzed, diluted, lipemic, and overfilled samples (with sample volume above the manufacturer specified mark). Other errors pertaining to the identification or receiving included incorrect/absent sample labeling, requisition form-sample mismatch, incorrect vacutainer, and wrong test selection.
Documentation of errors: All technical staff working in the unit were educated about the categories of sample and or receiving errors as well as their possible implications. Accordingly, errors detected at all stations in the laboratories before the process of actual analysis were recorded in the daily workflow log book. Data on the routine and emergency laboratories were tabulated. IPD and OPD data, as well as data of samples sent for different tests (CBC/coagulation/ESR), were segregated. However, erythrocyte sedimentation rate (ESR) data were not used for the final analysis as all samples coming to our unit for ESR invariably also had a requisition for CBC, hence ESR did not constitute a mutually exclusive category.
Statistical methods: Analysiswas done using statistical software R 3.5.0. All errors were recorded as numbers and corresponding percentages. The z-test (two-tailed) was applied to assess the significance of the difference in percentages across groups of interest. A P < 0.05 was statistically significant.
| Results|| |
Over a period of nine months during which this study was conducted, a total of 189,104 patient samples were analyzed.
Out of the total samples studied, 133,142 (70.41%) comprised of outpatient samples, while the remaining 55,962 (29.59%) samples were of inpatients. Most of the samples (78.8%) were received for baseline hematology testing (CBC/ESR) whereas coagulation tests (prothrombin time, activated prothrombin time) were requested in 21.2% of the total samples. Most of the samples in this work were from adult patients (96.7%) with the remaining comprised of pediatric samples. Errors were documented in a total of 4,052 samples accounting for 2.14% of the total samples received [Figure 1].
The most frequent errors noted in our work were insufficient samples (52.2%), followed by clotted samples (41.26%) and cases in which there was a mismatch between the details labeled on the samples and the attached requisition forms (2.42%) [Figure 2]. In terms of the total errors noted, there was no significant difference in the error percentages seen in OPD vs IPD samples (P = 0.1031). Comparing the error subtypes across OPD and IPD groups, there was a statistically significant difference (P < 0.001) in the proportion of clotted, insufficient, and mismatched samples. The proportion of mismatched and insufficient sample quantity was more in the IPD group, clotted samples being more in the OPD group [Table 1]. There was a significant difference (P < 0.001) in the various preanalytical errors in samples run in the routine vis-a vis emergency laboratories [Table 2]. Total error proportion was significantly more in the routine laboratory subset, however, the proportion of lipemic, hemodilute, and wrong tube errors were more in emergency samples. Coagulation samples displayed a higher total error proportion vis-a-vis CBC. A statistically significant difference (P < 0.001) was noted while comparing the data of samples run for CBC and coagulation tests in all categories of errors [Table 3]. On analyzing the data further, all errors were seen with a significant proportion in the coagulation subset except clotted, diluted, and incorrect vacutainers, which were more in the CBC group. [Table 4] highlights a composite comparative analysis of emergency, routine, CBC, and coagulation subsets. All emergency samples and routine CBC samples showed more errors in admitted patients. Routine coagulation samples showed no significant difference in error proportion between IPD and OPD.
|Table 2: Comparative data of pre-analytical errors: Emergency vs Routine laboratories|
Click here to view
|Table 4: Complete blood count vs coagulation in routine and emergency samples|
Click here to view
| Discussion|| |
The preanalytical phase comprises all steps preceding the performance of the actual laboratory test, i.e., requisition, patient identification, specimen collection, labeling, transportation, accession, and pre-test processing. This phase is complex, dependent on human skills (or the lack of them!) and is characterized by multiple stakeholders and divested accountability. Hence, despite progress in technology and implementation of quality control measures, it remains the “Achilles heel” of the total testing processes. Its vulnerability can be gauged by the fact that mistakes in any of the components of the pre-testing phase may comprise up to 75% of total laboratory errors. Preanalytical errors affect the accuracy and timely delivery of clinical care with some authors citing that 26% of errors afflict overall patient management and may account for up to 1.2% of total hospital operating costs.,
The types of errors occurring in the preanalytical phase may broadly be categorized as those pertaining to the pre-preanalytical phase, mostly outside the realm of the laboratory (requisition, patient identification and preparation, phlebotomy, transport, and accession) and the actual preanalytical phase referring to the processes involved in sample preparation prior to testing. In the hematology laboratory, this would specifically refer to the final mixing of all specimens sent, checking for sample quality, and the processes of centrifugation and aliquot preparation specifically in the case of coagulation tests. In the present study, data have been retrieved from all parameters that were assessable within the laboratory, hence, primarily focussing on requisition forms and sample quality.
In our study, the total error ascertained in the preanalytical phase was 2.14% (4052/189,104). This is comparable to a prior study done by Chawla et al. at a large super-specialty center in India. However, it is more than similar hematology-based studies reported from the country [Table 5].,, This difference may be ascribed to the fact that this work is the largest study done to date from an Indian academic center, with a large number of both medical and paramedical trainees. Phlebotomy is a complex yet under-rated skill, a recent data study highlighting 11 phases and as many as 77 steps for the perfect performance of alluded to the procedure. The same work also validated an apparently obvious fact—errors reduce commensurate with the duration of training experience.
|Table 5: A comparison of preanalytical error in previous studies in hematology laboratories in India|
Click here to view
The most frequent error noted in our study was insufficient sample volume. This corroborates with another recent study by Arul et al. A cause for this may be suboptimal quality and proficiency of phlebotomy. Further analysis of our results revealed this error to be more frequent in the IPD samples. Sample-form mismatch was also more frequent in this group. This may reflect a need for imparting better teaching and training to both interns and nursing staff carrying out the sampling in the wards. Interns, especially represent a heterogenous workforce, each batch being posted with different units, sometimes for varying time intervals and coming in at different stages of their training, some with previous clinical postings, and others without. Continuous monitoring of phlebotomy performance is an extremely essential complement to the training of the work personnel involved in blood sampling.,
The actual effect of sample volume on hemogram test results remains a subject of study. Chen et al. reported that hematocrit and mean corpuscular volume of samples with 2 ml volume analyzed on the Sysmex™ NE-8000 were significantly lower than samples with 5 ml blood volume. They found values of MCHC with 2 ml blood volume samples to be significantly higher than those with 5 ml blood volume. Xu et al., on the other hand, found that acceptable blood count values may be obtained from samples with volumes as less as 1 ml using the Sysmex™ XE-2100 analyzer.
The second most frequently noted error in this work was clotted samples, accounting for 41.26% of all sample rejections, lower than a large Chinese study in this subject (57%). As a fraction of errors of total samples seen, however, the error percent (0.88%) was more than a similar work from an Indian academic center. This again may be due to the larger sample size taken in our study. It may also highlight suboptimal post phlebotomy mixing of the samples. Our data show this error to be more in the OPD samples. Thus, though sampling by laboratory personnel posted at the outpatient collection center may be adequate, other aspects such as proper mixing of the samples may be overlooked. This reiterates that persistent performance monitoring of paramedical staff, trained by the laboratory is essential.
In our work, a greater proportion of errors were noted in samples processed routinely. All samples processed in the routine laboratory were under the direct vigilance of the consultant pathologist manning the laboratory, conceivably leading to a better detection rate of errors.
While comparing the data of samples sent for routine CBC versus coagulation, ISQ remained the major cause of preanalytical error in the coagulation group whereas clotted samples were more frequently seen in CBC samples. As illustrated in [Table 4], the total number of errors both in routine and emergency samples taken for both CBC and coagulation were more frequent from admitted patients. It is interesting to note the findings of a study by Cadumaro et al. in which hemolysis rates reduced on delegating phlebotomy from untrained junior doctors to trained nurses. As our results have indicated ISQ as the most frequent error, previous comments and work by Dale et al. have highlighted the perils of switching to smaller volume collection tubes, in turn translating to a “re-engineering” of the testing process.
A more pragmatic approach would be to improve the training practices of interns and junior resident doctors. Given the anxiety of patients regarding phlebotomy, tangible efforts must be made in this direction, for both patient safety and timely diagnosis. This could be accomplished by the usage of a mannequin training arm or virtual phlebotomy simulator to train students early, somewhat similar to models used in basic life support training modules. In the wake of the ongoing change in undergraduate medical curriculum in India and the emphasis on skill development, the time may be right to execute a complete shift at the initiation of training from direct patient-based sampling to simulated learning before dealing with patients, for better outcomes.
Study and Limitations and future directions for research:
The current work has a few limitations. First, a postintervention analysis has not been included in this work. This is especially important as studies in literature have shown conflicting results in regard to intervention. Second, the effect of all variables classically described in the pre-preanalytical phases such as fasting, exercise, stress, drugs, diet, and tourniquet usage have not been analyzed in this study as also the clinical correlates/profiles of the patients whose samples showed errors. The present work dealt primarily with adult samples hence, subsequent work comparing errors with an exclusively pediatric cohort may be devised.
| Conclusions|| |
Preanalytical errors though inevitable, are definitely avoidable. Continuous communication is the key to minimizing such errors. A prudent way forward [Figure 3] would involve intra and extra laboratory teaching programs directed at paramedical and medical staff involved at all stages of the preanalytical phase. Such continuous laboratory education (CLEs) programs should stress strict adherence to standard operating procedures, continuous monitoring, and evaluation of technical proficiency of phlebotomy (issuing of competency certificates) to ensure high standards of laboratory services. A shift in the phlebotomy training methodology of junior doctors is also suggested, using virtual and simulation models. Minimizing human involvement wherever possible via computerization at the sample reception desk (checkpoint monitoring) is desirable.
Regular meetings between diagnosticians and clinical colleagues as well as feedback from the clinicians and patients must be encouraged. Further studies on the impact of intervention on the preanalytical phase would aid in reducing error frequencies and improve the standard of laboratory care.
The authors would like to thank Deepa Dhiman, Nikhil Mishra, Aarti Tyagi for their valuable technical contributions in assisting the completion of this work.
The work described herein adheres to ethical guidelines as outlined in the revised Helsinki guidelines and our institutional committee. Patient confidentiality has been maintained throughout the course of the work. The study was conducted on documenting the quality of patient samples without any identification/experimentation/invasive procedures/additional sampling whatsoever, hence was exempted from ethical clearance.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Witte VanNess SA, Angstadt DS, Pennell BJ. Errors, mistakes, blunders, outliers, or unacceptable results: How many? Clin Chem 1997;43:1352-6.
Plebani M. Laboratory-associated and diagnostic errors: A neglected link. Diagnosis (Berl) 2014;1:89-4.
Lundberg GD. Acting on significant laboratory results. JAMA 1981;245:1762-3.
Hammerling JA. A review of medical errors in laboratory diagnostics and where we are today. Lab Med 2012;43:41-4.
Carraro P, Plebani M. Errors in a stat laboratory: Types and frequencies 10 years later. Clin Chem 2007;53:1338-4.
Upreti S, Upreti S, Bansal R, Jeelani N, Bharat V. Types and frequency of preanalytical errors in haematology lab. J Clin Diagn Res 2013;7:2491-3.
Narang V, Kaur H, Selhi PK, Sood N, Singh A. Preanalytical errors in hematology laboratory- An avoidable incompetence. Iranian J Pathol 2016;11:151-4.
Arul P, Pushparaj M, Pandian K, Chennimalai L, Rajendran K, Selvaraj E, et al.
Prevalence and types of preanalytical error in hematology laboratory of a tertiary care hospital in South India. J Lab Physicians 2018;10:237-40.
] [Full text]
Carraro P, Plebani M. Errors in a stat laboratory: Types and frequencies 10 years later. Clin Chem 2007;53:1338-2.
Plebani M. Errors in clinical laboratories or errors in laboratory medicine? Clin Chem Lab Med 2006;44:750-9.
Green SF. The cost of poor blood specimen quality and errors in preanalytical processes. Clin Biochem 2013;46:1175-9.
Chawla R, Goswami B, Tayal D, Mallika V. Identification of the types of preanalytical errors in the clinical chemistry laboratory: 1-year study at G.B. Pant Hospital. Lab Med 2010;41:89-2.
Seemann TL, Nybo M. Continuous quality control of the blood sampling procedure using a structured observation scheme. Biochem Med (Zagreb) 2016;26:337-45.
Chen BH, Fong JF, Chiang CH. Effect of different anticoagulant, underfilling of blood sample and storage stability on selected hemogram. Kaohsiung J Med Sci 1999;15:87-93.
Xu M, Robbe VA, Jack RM, Rutledge JC. Under-filled blood collection tubes containing K2EDTA as anticoagulant are acceptable for automated complete blood counts, white blood cell differential, and reticulocyte count. Int J Lab Hematol 2010;32:491-7.
Ye Y, Wang W, Zhao H, He F, Zhong K, Yuan S, et al.
Haematology specimen acceptability: A national survey in Chinese laboratories. Biochem Med (Zagreb) 2018;28:420-9.
Cadamuro J, von Meyer A, Wiedemann H, Klaus Felder T, Moser F, Kipman U, et al.
Hemolysis rates in blood samples: Differences between blood collected by clinicians and nurses and the effect of phlebotomy training. Clin Chem Lab Med 2016;54:1987-92.
Dale JC, Ruby SG. Specimen collection volumes for laboratory tests. Arch Pathol Lab Med 2003;127:162-8.
Wandell H. Using a virtual reality simulator in phlebotomy training. Lab Med 2010;41:463-6.
Li HY, Huang XN, Yang YC, Huang WF, Chen L, Song P, et al.
Reduction of preanalytical errors in clinical laboratory through multiple aspects and whole course intervention measures. J Evid Based Med 2014;7:172-7.
Department of Pathology, Lady Hardinge Medical College, Shaheed Bhagat Singh Marg, New Delhi - 110 001
Source of Support: None, Conflict of Interest: None
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]