Dedalus.eu

The role of 2D bar code and electronic cross-matching
in the reduction of misidentification errors
in a pathology laboratory. A safety system assisted
by the use of information technology
U.O. Anatomia Patologica e Citologia, Ospedale Infermi, Rimini, Italy Mismatch errors • Risk management • 2D Barcode technology • Safety management of patient specimens Introduction. Mismatching of patients and specimens can lead
being used; in addition, the system performs an electronic cross- to incorrect histopathological diagnoses. Most misidentification check of tissue blocks and slides, which is managed by the labora- errors in laboratories occur during the manual pre-laboratory and laboratory phases. In the past few years, we have examined this Results. The present system permits full sample traceability
vital and challenging issue in our unit and introduced appropriate from the moment samples reach the laboratory to the issuing procedures. Recently, we have paid special attention to the prob- of the final report. Indeed, the LIS records samples, blocks and lem of specimen mix-ups in the gross examination phase and the slides in real time throughout the entire procedure, as well as mismatching of blocks and slides in the cutting phase.
the operator’s name, and the date and time each individual pro- Objective. We have focused on the reduction of the potential
cedure is done. This facilitates later monitoring of the entire sources of mismatching of specimen containers, tissue blocks and slides, focusing in particular on the most critical steps which are Conclusions. The introduction of 2D bar code and electronic
gross cutting and preparation of microtome sections.
cross-checking represents a crucial step in significantly increas- Design. A 2D bar code directly printed on the labels of specimen
ing the safe management of cases and improving the quality of containers, and directly printed onto cassettes and slides, is now Introduction
incorrectly-recorded laterality and anatomical sites. Another two steps in the procedure that are particularly Since the publication of “To err is human” in 1999 1, prone to error are the gross and cutting phases, which substantial work has been done to reduce factors that are characterized by sample mix-ups and block and slide contribute to errors in medical and surgical pathology practice. Procedures in the histopathology unit involve A significant reduction in the number of misidentifica- multistep processes with several handoffs of materi- tion errors on accession was achieved in 2008 with the als, which are all potential sources of error 2. Errors that elimination of handwritten requests and handwritten may occur at any stage of processing vary in frequency, labels, and by the introduction of an order entry with depending on the laboratory. Several papers have been electronic requests and labels. In addition, direct print- published that analyze and propose solutions 3-5. Over ing of cassettes and slides by automatics printers inter- the past five years, we have approached this challenging faced with the laboratory information system produced issue in our laboratory, with particular focus on the pre- a considerable reduction in block and slide mismatching The most critical step is the accession phase, which is However, data analysis in 2009 revealed continuing characterized by incorrect patient identifications and block and slide mismatching. For this reason, at the be- The author is grateful to the entire Pathology Unit Staff; particular Giovanna Fabbretti, U.O. Anatomia Patologica e Citologia, thanks are given to Luigi Santucci, Director of the Information Ospedale Infermi, via Settembrini 2, 47900 Rimini, Italy - E-mail: Technology Department, and to Francesco Graziani, Rina Velati ginning of 2010, a 2D bar code was introduced, which Fig. 1. cassettes and slides with a directly printed 2D bar code is directly printed onto container labels, cassettes and and accession code number. Slides also show readable text: name slides, in order to reduce mismatching in the gross ex- of institution, type of stain (he: yellow slides and immunostains amination and cutting phases. This new technology is for Ki-67, progesterone and oestrogen receptors: white charged slides) name and surname of patient.
also an effective means of improving sample traceability The purpose of the present work is to discuss the highly reliable work procedure we have developed, which fully utilizes the benefits of information technology.
Materials and methods
The entire process was reorganised in May 2008 when a new laboratory information system (LIS, Armonia Dedalus, SpA, Italy) was integrated with the Hospital Information System (HIS; Trak-care, Traksystem, Aus- tralia), with an HL7 interface for receiving orders from physicians through HIS order entry. This eliminated the need for handwritten requests and handwritten container labels. At the same time, the LIS was interfaced with the cassette and slide printers (Leica Microsystems, Ban- matching of a block and its associated slides. It is im- nockburn, IL) to handle cassette and slide printing case- possible for two identically identified blocks or slides to by-case during the gross and cutting phases; this avoids exist. For example, if a slide is printed and then the same the need for manual code transcription. All of the above slide is printed again, the first slide printed is identified has been described in detail in a previous publication 6. in the 2D bar code as 11-I-13500A21 and the second one
Since 2010, the LIS has used a 2D bar code and has been 11-I-13500A22.
interfaced with both cassette and slides printers (a Lei- This is of fundamental importance and is a key point ca printer in the Cytology Lab and Slide Mate printers regarding matching of blocks and slides.
[Thermo Fisher Scientific, Waltham, MA] at the Cutting Each workstation in our unit is equipped with a PC, Station); the LIS also has been integrated with a Leica monitor and scanner. We have also equipped each cut- BOND-III instrument, which fully automates immuno- ting station with small slide printers to avoid the need to histochemistry work; 2D bar codes are directly printed preprint slides. The LIS manages each individual step onto immunohistochemistry slides at the cutting station: via the 2D bar code regarding the processing of samples, the BOND-III reads the 2D slide bar codes. Extensive blocks and slides by recording the name of the operator bar code printing testing and validation for cassettes and and the date and time of the step; in this way, each single slides was conducted by Leica, Thermo Fisher Scien- case is traceable during the entire work procedure.
tific and Dedalus, and for scanner configuration by the The LIS furthermore records any error or problem de- Dedalus Company and Metrologic Instruments Inc. We tected at any stage in the workflow. This function is chose the Metrologic MS1690 Focus, which is an om- quick and easy to access by using a keyboard; in this nidirectional scanner capable of reading all standard 1D function, a list of predetermined parameters are dis- played: e.g. error or problem type, possible corrective During set up, we carried out ping testing on cassettes action, date, time and operator. Cases where an error has and slides. No input failure occurred. Bar code misread- been detected are marked by a special icon, so that the ing may be caused by poor quality cassette and slide ma- pathologist is alerted and can check the validity of the terials, which can cause variations in printing quality. corrective actions taken before diagnosis.
We always test any new material that will be used.
Errors and problems are subdivided in the following The following are printed on cassettes: the accession way: accession errors, specimen errors or problems, and code (e.g. 11-I-11340), specimen container letter (e.g. misidentification during the processing procedure. Each A, B, C), subpart block number (e.g. 1, 2) and a 2D bar subgroup is further divided into other sub-categories (e.g. code, which includes a progressive printing number misidentification during gross examination, embedding, cutting, etc.). This system permits rapid analysis of col- The following is printed on the slides as human readable lected data. Once a month, a specially trained technical text: accession code (e.g. 11-I-13800), patient name and surname, type of stain (e.g. HE, PAS), the name of our The unit’s workflow, which is bar code based, is de- unit (Anat Pat, RN); in addition there is a 2D bar code, scribed in a consistent and easy-to-read manner.
which also encodes a progressive printing number.
1. Accession phase: after a double check to verify that The progressive printing number, found in both slide data on the electronic request corresponds to that on and cassette 2D bar codes, is essential for the univocal the medical report that accompanies specimens (e.g. role of 2D bar coDe anD electronic croSS-matching in the reDuction of miSiDentification errorS Fig. 2. Downloaded request form and adhesive labels attached to for sentinel lymph nodes, yellow for small biopsies, specimen containers. above: patient data; middle: the two sub- blue for lymph nodes, pink for skin biopsies and mitted specimens: 1) skin from the lumbo-sacral region; 2) skin green for surgical specimens), section number, and from the patient’s hip and clinical information; below: the space the routine stains or immunostains, if provided. The default setting may be modified at any time during the process. Cassettes are directly printed (Leica Mi- crosystems, Bannockburn, IL) case-by-case during gross examination. The printing process is quick and 3. Tissue embedding phase: after processing each cas- sette is read by the scanner before embedding the tis- sue. The LIS displays the following: code number, tissue type, fragment number and notes, if recorded during gross examination, including operator name, date, time and status (Fig. 4); after reading, the cas- sette’s status is changed from processing to executed. When all samples related to a single case are embed- Fig. 4. tissue embedding station: the cassette is read by the scan- ner. the liS displays all relevant information and shows a list of bronchoscopy, endoscopic report, etc.), the case is entered into the LIS by scanning a bar code on the paper copy of the electronic request, determining the recovery of the request from HIS (Fig. 2). The LIS provides a lab worksheet with number (e.g. 11-I-14500) both as readable text and as a bar code (Fig. 3), and also provides labels for specimen con- tainers in readable text as well as a 2D bar code. Once a misidentification error is detected, the case is rejected and it will be processed after the error has 2. Gross examination phase: the specimen containers are moved to the gross bench for sectioning and re- cording of macroscopic findings. Our LIS provides many predetermined parameters for each anatomi- cal site and each medical procedure; for example, the topographic code (SNOMED), the number and colour of the cassette (orange for urgent cases, white Fig. 5. cutting station: the cassette is read by the scanner, the liS shows all tissue block information (patient name and surname, Fig. 3. an internal lab worksheet: accession number and 1-di- code number, tissue type, number of fragments, embedded mensional bar code and adhesive labels for specimen containers status, operator and date) and the slide mate printer prints the ded in cassettes, the case status is changed from gross because the clinical information was not concordant executed to embedded. The LIS sequentially shows a with histological appearance. In other cases, the slide list of all embedded cassettes on the monitor in the samples clearly did not correspond with the anatomical work session, and, if required, supplies a printed site indicated in the request when viewed under the mi- croscope. Another particularly important result achieved 4. Cutting phase: just before cutting, the operator reads by the introduction of 2D bar coding is the introduction the block’s bar code with the scanner, and the slide of automated tracing; it is now possible in real time, to printer prints all the associated slides; after section trace a specimen container or missing block and locate cutting (and only at this time - before it is picked up) the slide is read by the scanner. If the slide does Indeed, the LIS manages the workflow, step by step, re- not match the block, a message error on the moni- cording the operator’s name, date and time of each single tor alerts the operator (Fig. 5). The LIS displays the step. We are now able to know what is happening in real changing status of the slide from requested to vali- time, and to take immediate action to locate a misplaced dated only if the slide matches correctly. When all slides related to a single case are validated, the case’s status is changed from embedded to cut.
5. Checkout phase: at the end of the entire work flow Discussion
procedure, there is the final check before delivering slides to the referring pathologist. Each slide is read The case-by-case direct printing of bar code numbers by the scanner, and when all slides of a single case on cassettes and slides by automated printers managed (routine stain, special stains and immunostains) are by the LIS prevents errors caused by handwritten labels ‘pinged’ the case is ready to be sent for medical ex- and by transcription. Checking correspondence between the code number on container labels and the cassette at the gross station and between block and the slide at the cutting station was previously done visually and was therefore subject to error caused by fatigue and lack of The results achieved have been particularly good and of Even if the mismatch rate was low in the gross exami- significant importance. Since the introduction in 2010 nation and cutting phases, and in keeping with data re- of 2D bar codes on container labels, we have not had a ported in recent literature 7, an error that mismatches a single case of sample mix up in the gross examination slide to the wrong patient can have serious consequences phase in a total of 26,964 histological cases. In the gross examination phase, each case begins with a reading For this reason, we worked closely with the LIS provider of the 2D bar code on the container label, and the LIS to design a system that would prevent this type of error. makes it impossible for a code number that is different to The result is that we have up-graded our LIS with the the case number in question to be printed on a cassette. introduction of 2D bar codes on labels of specimen con- In contrast, in 2009 we had 10 errors in a total of 26,961 tainers, and direct printed on cassettes and slides. The (0.03%) cases that involved mismatch of samples from biggest leap in improved quality was achieved by the introduction of electronic cross-match managed by LIS. Additionally, in the cutting phase we have had no mis- Another important advance is that there is now sample match since automatic cassette and slide cross checking traceability throughout the entire workflow.
was made possible by the introduction of 2D bar codes In a recent paper, Zarbo et al. 8 describes a workflow in 2010 (26,964 histological cases; 80,571 tissue blocks). dependent on bar code reading and illustrates the use of In contrast in the same period in 2009, we had 32 mis- traditional bar codes on specimen container labels, in matches from a total of 26,961 cases (0.11%) (80,361 specific labels for slides and use 2D bar code only for tissue blocks) caused by the transfer of sections from one block to a mismatched slide. Data analysis showed that Unfortunately, in their laboratory, electronic requests mismatch errors were more or less equally distributed are not yet employed and cases are accessioned manu- between routine cutting (14 cases) and re-cutting. There ally from handwritten requisitions, which are often in- was a slightly greater error prevalence for re-cutting (18 complete and unclear, as noted by Dimenstein 9. The cases), where the errors involved cases with similar code labelling of slides represents an additional manual step numbers (e.g. 09-I-23715 and 09-I-23915); 12 of 18 er- that is time consuming, prone to error and finally more rors involved specimens from different patients, and 4 of expensive than directly printing on them.
18 involved different specimens from the same patient.
The electronic checking introduced in the cutting sta- Of the 14 routine cutting mismatch errors, 10 involved tion overcomes the problem of operators failing to fol- different patients. None of the errors for either the gross low standard procedures, which was an issue that Zarbo examination or the cutting phase resulted in adverse con- emphasized in his report. The LIS prevents proceeding sequences for the patient, as they were detected during to the next case and alerts the operator is procedures are subsequent steps. The errors were noticed in some cases not followed. Furthermore, if the slide’s bar code is not role of 2D bar coDe anD electronic croSS-matching in the reDuction of miSiDentification errorS read by the scanner, the case is not validated. The intro- procedure, and not later in the pathology lab by a member duction of electronic cross checking of blocks and slides of administration or technical staff. During gross tissue is an effective means of preventing inevitable human er- examination, LIS case data can be accessed by reading rors in the cutting phase caused by fatigue, lack of con- the 2D bar code on container labels, avoiding mix-up of specimens; the direct printing of cassettes one case at a During the development of this project, the only con- time avoids the need for them to be prepared in advance cern was the possible increase in processing times. How- and eliminates the risk of confusing cassettes from dif- ever, during the first three weeks after the adoption of ferent patients. The direct printing of slides, one block at the new workflow we experienced only a small delay a time, at the moment of cutting of sections, eliminates in slide delivery, which was caused by the need to train the need for labelling, which is a time consuming step. all operators; such training is obviously necessary when More importantly, it also eliminates a potential source of introducing new organizational procedures. All techni- error because traditional labelling is a manual procedure cal staff have very positively accepted this new working that is visually checked. Furthermore, labelling is more procedure. In addition, in recent years much has been expensive than direct printing of slides. The introduction accomplished in training all operators in risk manage- of electronic cross-checking using 2D bar codes directly ment, and on-going work has been done with the entire printed onto blocks and slides represents a very important team to identify the causes of mismatching and improv- qualitative leap. In our experience, it represents the best ing workflow. The knowledge of when, where and why method for avoiding block and slide mismatching.
misidentification errors occur, which is a fundamental The redesigned workflow with 2D bar codes has an- prerequisite for their successful reduction, has been fa- other advantage: real time case traceability throughout cilitated by the LIS, which allows quick, easy and com- the entire procedure. Gradually we redesigned the entire plete error reporting at each step of the work flow, as workflow procedure over a period of years. The support we received from top management was crucial for its In summary, the work over the last few years has been success. In our experience, no single piece of technology focused on simplifying workflow procedures as much as can eliminate errors in a complex system such as a pa- possible by utilizing information technology, and the em- thology work flow composed of multiple handoffs. Each ployment of bar coding to minimize operator caused error. laboratory has to consider the individual requirements of The process was streamlined by eliminating some poten- tially error prone procedures, most importantly eliminat- The LIS and bar code technology play a leading role in ing manual accession input in the LIS by using a direct making the entire process far safer. However, there is electronic request entry. It is important to note that in this also the need for standard operating procedures for each manner, the patient and his or her samples are correctly step, accompanied by an efficient system of recording identified at the time they are taken, in the place they are errors for every phase (pre-lab, lab and post-lab) and rig- taken and by the clinician who performed the medical orous daily compliance with all procedures.
References
gical pathology: an 18-mounth experience. Am J Clin Pathol 2010;134:466-70.
Kohn LT, Corrigan JM, Donaldson MS, eds. To err is human: Fabbretti G. Risk management: correct patient and specimen iden- building a safer health system. Washington, DC: National Acad- tification in a surgical pathology laboratory. The experience of Infermi Hospital, Rimini, Italy. Pathologica 2010;102:96-101.
Nakhleh RE. Error reduction in surgical pathology. Arch Pathol Nakhleh RE, Idowu MO, Souers RJ, et al. Mislabeling of cases, specimens, blocks, and slides: a college of American pathologists Zarbo RJ, D’Angelo R. The Henry Ford Production System: effec- study of 136 institutions. Arch Pathol Lab Med 2011;135:969-74.
tive reduction of process defects and waste in surgical pathology. Zarbo RJ, Tuthill JM, D’Angelo R, et al. The Henry Ford Produc- tion System: reduction of surgical pathology in process misidenti- D’Angelo R, Zarbo RJ. The Henry ford Production System: mea- fication defects by bar code-specified work process standardisa- sures of process defects and waste in surgical pathology. Am J tion. Am J Clin Pathol 2009;131:468-77.
Dimenstein IB. Letter to the Editor. Am J Clin Pathol Layfield LJ, Anderson GM. Specimen labelling errors in sur-

Source: http://www.dedalus.eu/upload/Patho-06-11.pdf