Prevention of extra-analytical phase errors by non-analytical automation in clinical laboratory Klinik laboratuvarda non-analitik otomasyon ile extra-analitik faz hatalar ı n ı n önlenmesi

During previous decades, significant improvements in laboratory errors have become a substantial part of reducing preventable diagnostic errors. In clinical laboratory practice, the errors in the testing process are primarily associated with extra-analytical phase error sources, influencing the test result quality profoundly. Thus, the management of these critical error sources makes their effects preventable thanks to automation and computer sciences. The implementation of non-analytical automated systems requires a risk management strategy based on laboratory ’ s workflow and bottlenecks. Then, the improvements can be measured and evaluated by the usage of quality indicators (QI). Consequently, the total quality of laboratory diagnostics and higher patient safety is closely dependent on this type of automation. This review will help laboratory professionals, managers, and directors improve the total testing processes (TTP). The automation technologies have added a serious impact on the proficiency of laboratory medicine. Several ins-trumentations have now partially or entirely automated many manual tasks to improve standardization, organization, efficiency, and TTP quality. The implementation of non-analytical automation has made them manageable. As a result, non-analytical automation within and outside the clinical laboratory will necessarily lessen the error sources ’ effect on the total test process, enhancing the quality of the test results.

Automation comprises extra-analytical and analytical processes, with the latter being developed earlier than the former [1]. Clinical laboratory automation has improved such laboratory processes as specimen labeling, sorting, transport, processing, loading on the analyzers, analysis, storage, and archiving [2]. Although the analytical and non-analytical automation has developed asynchronously, clinical laboratory performances have dramatically changed to a satisfactory level today [3].
Previously, the term automation was used for clinical chemistry analyzers to describe the processes with only minimal involvement of an analyst. The automation, for the past two decades, has also covered extra-analytical as well as analytical procedures. It is well known that extra-analytical error sources dominate those of analytical systems and have become substantial for the efficiency of clinical laboratories [1,4,5]. Since a high-quality test result can be defined as the fact that it must be accurate and be reported faster, i.e., short turnaround time (TAT), the value of extra-analytical technologies besides the automated analyzers may be appreciated well. Consequently, the implementation of automatable processes to the extra-analytical phase will become crucial, considering the improvement of clinical laboratory performance and patient's sample safety [6].
Extra-analytical and analytical components have recently been integrated or interfaced with each other to construct fully integrated modular laboratory automation. The current developments in non-analytical automation provide error elimination, improve the quality, and reduce the labor, costs, and TAT [3]. On the one hand, a wide range of variables is included in the extra-analytical phase, such as test order and sample collection, delivery to the laboratory, handling, and processing before analysis. On the other hand, the pre-analytical part of the extraanalytical-phase covers all processes outside and inside the

Laboratory automation
The general term used for automated clinical laboratory (instrumentation + LIS/HIS) Total laboratory automation (TLA) Laboratory automation composed of heterogeneous, physically-integrated analytical and extra-analytical systems (many analyzers performing different types of tests on different sample matrices) Analytical automation Automated analytical systems or workstations Non-analytical automation Part of laboratory automation other than analyzers; signifies pre-and post-analytical automation Intra-laboratory nonanalytical automation Pre-analytical automation within the clinical laboratory (pre-and post-analytical automation) Extra-laboratory nonanalytical automation Pre-analytical automation outside the clinical laboratory; signifies prepreanalytical phase automation Extra-analytical phase Intra-and extra-laboratory processes other than analytic processes On-line analyzer An analyzer interfacing the integrated laboratory automation system Off-line analyzer An analyzer interfacing with no integrated laboratory automation system; stand-alone analyzer Integrated/modular laboratory automation An automation composed of pre-analytical, analytical and post-analytical component linked by a conveyor Modular analytical system An analyzer designed as modules; module addition or exclusion possibility depending on the need Modular non-analytical system A non-analytical system designed as modules; module addition or omission possibility depending on the need Stand-alone preanalytical system A pre-analytical specimen processor interfacing no analyzer Turnaround time (TAT) The duration in which the laboratory reports the test result Analytical TAT Testing duration beginning with sample load to the analyzer and ending with readout.

Intra-laboratory TAT
Reporting duration beginning with sample submission by laboratory and ending with report verification. Total TAT Reporting duration beginning with either test ordering or specimen collection and ending with report verification.
clinical laboratory or begins with test order and ends with analyzer entrance of the processed samples. Similarly, the post-analytical part of the extra-analytical phase comprises auto-verification, recapping, automated specimens archiving and storage, decapping in the case of a rerun, and secondary specimen sorting for off-line analyzers [1,3,7]. In short, extra-analytical processes summarized in Table 2 can be automated by non-analytical instrumentation in the clinical laboratory together with some limitations depending on the facilities. Either of the following automation can automate some of these processes: (a) stand-alone pre-analytical systems (single-function or multifunction) [24][25][26][27][28] and (b) integrated pre-and post-analytical systems [1,3,27,[29][30][31][32]. Consequently, we think that many extraanalytical processes, within and outside the clinical laboratory, have nowadays been or may in the future, be automated. Resultantly, the extra-analytical phase management may prevent some errors, resulting in the total quality of laboratory diagnostics and high-quality test results, as considered in the following sections. For this reason, the review has been intended especially for laboratory professionals, managers, and directors for the sake of patient safety and will help improve the TTP in the laboratory.
(1) Error sources/errors in the extra-analytical phase of clinical laboratory. In the clinical laboratory, each step in total testing processes (TTP) may be affected by laboratory errors. However, a large body of evidence has shown that the extra-analytical phase (both intra-and extra-laboratory) are more prone to errors than the analytical phase, as mentioned earlier [7,[25][26][27][28][29][30][31]. Fortunately, those kinds of errors have been able to be prevented partly by the introduction of some technological and computer-aided informatic solutions to that phase. Several studies have shown that after the implementation of stand-alone pre-analytical systems to clinical laboratories, the errors in the conventional pre-analytical steps have been considerably reduced [32][33][34].
Similarly, some of the technological and computer-aided bioinformatic solutions have been able to be implemented to the extra-laboratory part of the non-analytical phase, resulting in partial automation in that phase. On the other hand, the extra-laboratory part of the extra-analytical phase is not under the control of laboratory staff since other nonlaboratory health care stuff execute the processes at this phase, covering several processes, including test order and specimen collection, identification, labeling, handling, tracing, and transporting. Consequently, laboratory professionals and staff cannot directly control and are not responsible for that phase's errors. However, it is of critical importance for the laboratory professionals to partly standardize and control the extra-laboratory steps [29,[34][35][36]. By using bioinformatic technology and robotics, thousands of clinical laboratories worldwide could have the opportunity to reduce errors caused by pre-analytical processes, resulting in improvement in accuracy and TTP. Unfortunately, there are limited numbers of studies focusing on the error improvements by implemented tools. Before the implementation of such automation, each laboratory does need justification for demonstrable improvements. However, it is clear that non-analytical automation effectively reduces the number of laboratory errors occurring during TTP [31,32].
Lippi et al. [7] summarized major sources of pre-analytical variability, associated with patient preparation, sample collection, sample transportation, sample preparation for analysis, and sample storage. Under these main categories, Lippi et al. detailed 20 different sub-variables. The clinicians and the patients are generally interested in total testing error and timeliness (i.e., TAT), which are the basis of test quality or the measure of performance of the clinical laboratory. Automation of analytical and extra-analytical phases of laboratory processes enhanced clinician and patient satisfaction, which is because prompt, electronic reporting of test results and critical value alerting by bioinformatics as well as test quality are of critical importance [28,37]. Thus, TAT improvement may be considered as a quality indicator (QI), a measure of laboratory performance. The implementation of performance measurements must evaluate TTP.
For this reason, the guidelines and regulations on the clinical laboratory test performance encourage the laboratories to improve all steps of the TTP as well as the analytical phase [38]. International Standard for Accreditation of Clinical Laboratories (ISO 15189:2012) prompts the laboratorians to establish QIs to assess clinical laboratory performance and to reduce the errors caused by the extra-analytical phase. Non-analytical automation in integrated/modular laboratory automation systems Specimen preparation Automated specimen storage and retrieval The standard similarly states that the laboratory shall establish QIs to monitor and evaluate performance throughout critical aspects of pre-examination, examination, and post-examination processes. It also suggests the process of monitoring QIs be planned, which includes establishing the objectives, methodology, interpretation, limits, action plan, and duration of measurement [39][40][41][42][43].
In the application of QIs to all clinical laboratories thoroughly, one can have difficulties or discrepancies since a laboratory's QIs assessment belongs to its establishment and not applicable to other laboratories. Resultantly, a harmonization of the usage of IQs is needed. For this purpose, the Working Group on Laboratory Errors and Patient Safety (WG-LEPS) of the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) planned a project in 2008 for establishing the Model of QI (MQI) [44,45]. The project was targeted to define a list of MQIs and to determine the data collection procedures and quality specifications for laboratory result evaluation. In succeeding years, the MQIs were discussed and approved in the 2013 Padova Consensus Conference and were tested on different occasions. Thus, the resultant MQI covers 53 QIs, which are generally key processes. Of them, 39 are related to the extra-analytical (pre-and post-analytical) phase, which is of critical importance. The MQIs have been listed by classification based on TTP phases. Besides, the significance of a given QI and data collection difficulty determine its order of priority, grading from 1 to 4 (one the highest, four the lowest priority). The QIs, for example, with priority 1 is mandatory and the first to be put into practice [46]. MQIs of critical processes associated with the pre-analytical phase include misidentification errors, test transcription errors, incorrect sample type, incorrect fill level, unsuitable samples for transportation and storage problems, sample hemolysed, and inappropriate time in sample collection, and almost all of which have priority 1. On the other hand, MQIs of critical processes associated with post-analytical phase processes encompass inappropriate TATs, incorrect laboratory reports, notification of critical values, and results notification (TAT), with the majority having priority 1 as well.
After the implementation of automation to the extraanalytical phase, the MQI ratios will become low when compared with those of non-automated processes, since MQIs have rather good match with extra-analytical phase errors. A study describing the comparison of the pre-and post-automation QI evaluation of the same clinical laboratory would be valuable and show the importance of non-analytical automation.
(2) Improvements in extra-analytical phase errors by non-analytical automation.
The non-analytical automation, within and outside the clinical laboratory, will substantially lessen the effect of the error sources in TTP and enhance the quality of the test results, resulting in an accurate result(s) and faster test result reporting. For the implementation of an MQI, a QI form is prepared first, which includes many characteristics such as identification code, method of data collection, a method for data processing, the goal for improvement activities, and so on. The classification of the QI in that form is also marked as Efficiency, Structure, Effectiveness, Activity/process, Timeliness (TAT), Results, Pre-analytical phase, Intra-analytical phase, Post-analytical phase, Safety of the staff, Outcome, Support processes, and competence. There may be a relationship between some of these classification items (e.g., Timeliness, Preanalytical phase, Post-analytical phase, and Safety of the staff) and the improvements executed by non-analytical automation tools. On the other hand, there is a good match between pre-analytical error types and their amelioration by non-analytical automation. Similarly, there is a considerable overlap of pre-analytical error types with MQIs. The majority of the non-analytical automation steps are closely associated with increased patient and healthcare staff safety. Lippi, G. et al. reported the pre-analytical error types [41]. We divided these error types into two categories: intraand extra-laboratory extra-analytical errors. Two tables were instructed (Tables 3 and 4), describing the errors and their prevention. The extra-laboratory errors occur in the pre-pre-analytical phase, and intra-laboratory errors in the intra-aboratory pre-and post-analytical phases. Each error type mentioned in Tables 3 and 4 were illuminated by a detailed comment or explanation about the probable improvements provided by non-analytical automation at the respective steps for lessening the error ratio. We can summarize this situation as follows.
Ten of 12 extra-laboratory extra-analytical errors in Table 3 and all intra-laboratory non-analytical errors in Table 4 may be prevented by automation. İmprovements can be summarized as follows, which are detailed in two tables. Missing sample and/or test request is prevented by computerized physician order entry and query-host communication. Positive patient identification by bioinformatics prevents wrong/missing identification of the patients. In vitro hemolysis may partly be averted by adaptor usage, using automated venous sampling systems, and using devices to find the veins or real-time digital imagers for sampling. All are closely associated with both patient and healthcare staff safety and high-quality testing. Similarly, the tube labeler and preparer tool hampers the wrong container issue, related to correct patient ID and tube labeling for phlebotomy and meaning

Extra-laboratory error types or inconveniences
Comments and explanation of the improvement manner instance, while a  mL-vacuum is enough for complete blood cell counting tubes at sea level, at least a  mL-vacuum volume is necessary for that kind of sampling at , m altitude (by our experience). It is possible to make the manufacturer manufacture high volume-vacuum tubes when ordered. . Inappropriate blood to anticoagulant ratio The same solution as that of insufficient sample error may be suggested for improper blood to anticoagulant ratio. The phlebotomist must use an adaptor for sampling and wait for complete filling of the tube vacuum. . Insufficient mixing of the sample No automation is possible for this error. The mixing is a staffdependent situation.

. Inappropriate transport and storage conditions
It is associated with transport systems for specimen delivery to the laboratory. Pneumatic tube system (PTS) (or conveyors) and robots, which are automated specimen transport systems, can get rid of it. Transportation monitoring systems can help trace the samples in this traveling. The PTSs are fast transporter all laboratory specimens (outpatients and inpatients). For safe transportation by PTS, a carrier with a high radius (about  mm), soft material support within the carrier, low or optimized acceleration and deceleration, non-sharpened corners or bending points in the tubing network may be critical.

. Inadequate label alignment on the specimen container
The tube labeler and preparer may lessen the errors associated with inadequate label alignment on specimen containers, especially for outpatients. Inappropriate label alignment means no reading by the bar-code readers located in the total laboratory automation system. Removal of this error source by automated devices shortens total TAT. . The low-resolution print quality of the bar-code label The tube labeler and preparer may lessen the errors associated with the bar-code label's low-resolution print quality, especially for outpatients. Inappropriate label print means no reading by the bar-code readers located in the total

Extra-laboratory error types or inconveniences
Comments and explanation of the improvement manner laboratory automation system. Removal of this error source by automated devices improves total TAT. TTP, total testing processes; PTS, pneumatic tube system; TAT, Turnaround time. Table : Intra-laboratory extra-analytical errors and their prevention by automation.

Intra-laboratory error types or inconveniences
Comments and explanation of the improvement manner

. Inappropriate centrifugation conditions
In automated centrifugation, the standardization of the requirements such as balancing is possible, which is necessary for specimen (naturally patient) safety. This system prevents broken samples. Generally, the capacity of automated system centrifuges is remarkably high, providing shortened TAT and staff safety. . Inappropriate storage condition in the posttest phase It is associated with automated specimen storage and retrieval, which makes it easy that the storage conditions are standardized, and that the retrieval is fastened. The robotics of the system finds a specimen within seconds for a rerun, dilution, add-on tests, etc. Appropriate storage conditions are obtained, such as dark storage, refrigeration, closed samples, meaning sample safety, and improved TAT. . Delay in sample processing steps in laboratory The tube loader and sorter can submit the specimen containers and their sorting for laboratory devices based on defined criteria. It is a very fast (throughput higher than , samples per hour) and an efficient tool, improving the workflow in clinical laboratories, particularly in the pre-analytical phase. It is a tasktargeted modular system. The inappropriate specimens are rejected and are not submitted to the laboratory. Thus, the sample sorting and accepting of the laboratory is critical and must be accelerated to make TAT reduce and achieve staff safety. . Delayed submission and inefficient specimen traceability The tube loader and sorter execute the submission function and provide efficient specimen traceability.

Intra-laboratory error types or inconveniences
Comments and explanation of the improvement manner . Inability to process primary tube Similar to analytical systems, non-analytical automation devices can also process primary tubes, which eliminates aliquoting steps, and the on-line analyzers can directly use these tubes for sampling, which shortens TAT. These tubes can also be directly stored in the storage cabinet.

. Errors from inadequate sample integrity
This module of the pre-analytical system checks sample integrity. Specimen quality is critical for producing a high-quality test result. Unnecessary repetitions of tests are prevented by this module, improving the test results and TAT. . Errors from different manual processes Different mechanical tasks are carried out using this module, and all processes involved are accelerated, reducing TAT. Decapper, aliquoter, bar-code labeler, recapper units of modular pre-analytical prevent these errors. . Impairment in secondary sample traceability Secondary tubes generated by the aliquoting unit are sorted to off-line analyzers's racks. Possible chaos and sample loss for rare tests are prevented. Specimen safety is provided, and resampling is prevented.

. Delayed test reporting
Auto-verification is essential for clinical laboratory automation and is based on computer sciences, with no manual intervention. The auto-verification either sends the patients' reports directly to the clinician's desktop or blocks them for the laboratorian to concentrate on the pathological results. In direct reporting, the results bear defined criteria for acceptance of the test results as non-pathological, meaning greatly decreased inspection time for test results or improved TAT. . The deceleration in total laboratory automation Process-managing software connects the LIS to the fullyautomated system composed of pre-analytical, analytical, and post-analytical components. It considers the whole system as if it were a single device. A single operator can manage the entire system. It is a kind of integrator, providing the integration of TTP with maximal effectiveness. All factors improved result in high-quality test reporting.
TAT, Turnaround time. patient safety. Adaptor usage and using the containers with a pre-determined vacuum volume, which is possible, hampers insufficient sample. During sampling, an adaptor is used for sampling, and complete filling of the tube vacuum lasted, precluding inappropriate blood to anticoagulant ratio. Inappropriate transport and storage conditions may be put an end to by PTS, which is closely associated with improved TAT and high-quality testing. The tube labeler and preparer restrains poor label alignment on the specimen container, especially in outpatients, reducing TAT. Automated centrifugation prevents inappropriate centrifugation conditions, shortening TAT, and providing staff safety.
The automated specimen storage and retrieval system anticipates the inappropriate storage condition of the samples in the posttest phase, meaning sample safety and improved TAT. An automatic tube loader and sorter usage can avert the delay in sorting and submission of the specimens, reducing TAT, achieving staff safety, and providing efficient specimen traceability. A delay in sample processing steps in the laboratory may be restricted by primary tube usage, which shortens TAT and provides direct storage within the automated cabinet. Volume/clotting/bubble sensors contained in pre-analytical automation systems counteract the errors originated from inadequate sample integrity, improving the test results, and TAT. Some errors coming from different manual processes are prevented by the module of decapper, aliquoter, bar-code labeler, and recapper contained in pre-analytical automation systems, which is associated with staff safety and TAT reduction. Impairment in secondary sample traceability is avoided by secondary tube sorter module of intra-laboratory non-analytical automation, relating to patient safety and reduced TAT. Delayed test reporting is counteracted by automatic verification, meaning improved TAT. Process-controlling software managing the system precludes the deceleration in total laboratory automation, resulting in high-quality test reporting. Table 5 summarizes extra-analytical errors, improvements by automation, and improved parameters, including TAT, test result quality, patient safety, and staff safety. These four parameters are principal factors determining the performance of a clinical laboratory, proving the importance of non-analytical automation for high-quality laboratory performance.
The management and standardization of non-analytical laboratory processes, especially of pre-analytical ones, have been primarily achieved thanks recently to automation, either by stand-alone devices or those integrated with analytical automation, as well as by informatics and computer sciences. Non-analytical automated systems should be implemented in the laboratory processes. For this purpose, a risk management strategy should first be developed for the systematic analysis of the present laboratory's workflow and bottlenecks. Then, the measurements of improvements can be executed by the pathways of QIs and outcome measures. Non-analytical automation within and outside the clinical laboratory will standardize as possible as the processes in that phase. The standardization of a process within TTP will necessarily make it possible to control, to trace, to measure, to evaluate, and to improve it, providing the integration of TTP with maximal effectiveness, and all factors improved result in a high-quality test result reporting.

Conclusion
The considerations mentioned above suggest that the clinical laboratory automation and computer sciences have rendered non-analytical laboratory processes mostly manageable and standardized the automation technologies have added a serious impact on the proficiency of laboratory medicine. The laboratory processes need the implementation of a non-analytical automated system. A risk management strategy, for this reason, should first be developed for the systematic analysis of the present laboratory's workflow and bottlenecks. Then, the application of QIs for different steps allows measuring the improvements in TTP. Several manual processes have been automated by instrumentations, resulting in improvements in standardization, organization, efficiency, and quality of TTP. Consequently, non-analytical automation, within and outside the clinical laboratory, will permanently reduce the effect of the error sources in TTP and enhance the test result quality. More studies, or QI-oriented measurements, should be executed on the improvements in extra-analytical steps by non-analytical automation improvements, achieving accurate and faster test result reporting.
Research funding: None declared. Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission. Competing interests: Authors state no conflict of interest.