From errors to excellence: the pre-analytical journey to improved quality in diagnostics. A scoping review

Introduction

Pre-analytical errors (PAE) are not random occurrences. They occur during specimen collection, handling, processing, and transportation. These errors can significantly impact the quality and accuracy of test results, underscoring the need for vigilance and the crucial role of healthcare professionals in the pre-analytical phase of the laboratory testing process [1], [2], [3]. However, they can be prevented and reduced with the right approach. By being aware of these errors and implementing prevention and detection strategies, healthcare professionals can play a pivotal role in improving the quality of laboratory testing.

Recognition and understanding of PAE in medical laboratories has changed over the years, evolving significantly to keep healthcare professionals well informed. In the United States of America (USA), Clinical Laboratory Improvement Amendments (CLIA), as part of national standards addressed errors at all stages of the testing process, including the pre-analytical phases [4], 5]. In Australia, the equivalent standard is AS:ISO 15189, regulated by the National Association of Testing Authorities (NATA). In Europe, the same standard is used and regulated by European authorities.

Figure 1:

PRISMA flow chart of publications included in the review.

In the 1970s when, the term “preanalytical phase” was introduced in medical laboratory diagnostics, highlighting awareness surrounding this phase as a key contributor to laboratory errors [6]. Laboratories were initially focused on analytical errors with minimal attention paid to the pre-analytical and post-analytical phases, so it was not until the 2000s that full acknowledgment of how pre-analytical variables affect results in laboratory medicine was given [7]. The total testing process (TTP) incorporates all phases of the sample journey, and errors occurring at any stage can impact patient care, diagnosis and treatment. With health care providers and institutions emphasising patient-centred approaches to health care, greater controls across pre-analytical, analytical, and post-analytical phases [7] of the TTP are necessary to ensure optimum patient care, diagnosis, and treatment. It is well understood that errors in the TTP can negatively impact patients’ health outcomes [8]. This implies a need to address defects within the TTP to enhance patient safety without compromising quality of care or timely medical interventions [9].

PAEs in laboratory analysis, including patient preparation, specimen collection, labelling, order entry into the Laboratory Information Management System (LIMS), transportation, and storage, can significantly impact test results and patient management, potentially leading to misdiagnosis or inappropriate treatment [10], 11]. Despite technological advances, these errors, which occur primarily in the pre-analytical phase, persist (ranging from 0.1 % to 9.3 %) compared to the pre-pre-analytical (test-selection phase), analytical and post analytical phases, influencing patient outcomes and costs [1], 12]. Quality indicators (QIs) are measures in place to detect and quantify errors, so continuous improvement processes may be introduced to enhance laboratory testing and patient care [3]. This highlights the importance of comprehensive frameworks designed to enhance the overall reliability of laboratory processes toward achieving optimal clinical outcomes. For example, the “Five Rights” approach in various process stages emphasises the systematically improved laboratory testing quality and patient safety [13].

Traditional pre-analytical errors, comprising 61.9–68.2 % of all laboratory errors, occur along the overall pathway from test ordering up to sample preparation. These involve patient misidentification, labelling, and handling. On the other hand, pre-pre-analytical errors involve inappropriate test selection and are more frequent due to overutilisation, underutilisation, or misinterpretation of requests. This highlights the differences and similarities between the two classes of errors, which indeed necessitates broad strategies for improving patient safety [14]. An Australian hospital study recently commented on a PAE rate of almost 25 %, with about 75 % being avoidable, thus highlighting that improved processes and training are required for clinical and administrative staff involved in pre-analytical processes [15], 16].

This scoping review focuses on root causes to better understand and establish best-practice guidelines for PAEs, highlights the lack of standardised definitions and measurement units for PAEs and advocates for QIs to detect and reduce errors. It discusses the historical timeline of PAE, the potential of AI to improve laboratory efficiency, and the importance of a patient-centred approach. Education, training, and meeting ISO 15189 pre-analytical requirements for accurate laboratory results are also highlighted. Overall, the review showcases the crucial role of healthcare professionals in managing PAEs and the need for best-practice guidelines and interventions to address the issue. A comprehensive literature review was conducted through various sources, including PubMed, Scopus and other relevant publications. The publications were selected for review based on their input and evidence to support and critique the findings of pre-analytical errors in medical laboratories per the criteria depicted in the PRISMA flow chart in Figure 1 to collate evidence from a historical perspective and gain an understanding of this field.

A global perspective on the historical timeline of PAE

A 50-year timeline of the evolution of studies is depicted in Figure 2, and the flow on-commentary demonstrates how the field has evolved over the past five decades.

Figure 2:

A 50-year timeline of the evolution of studies on pre-analytical errors (PAEs) [7], 10], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41].

The TTP concept emerged in the 1970s, encompassing steps from considering tests to diagnosis and error reduction strategies. It included the need for quality control beyond analytical error monitoring and addressed the potential of AI and machine learning in managing pre-analytical errors, suggesting future directions in laboratory computer systems in addition to the potential role of technology in addressing these challenges [17], 19]. George Lundberg’s model in the 1980s highlighted the impact of events before and after laboratory results in reducing errors [20]. In 1995, Goldschmidt demonstrated the need for sophisticated tools to transform raw data into valuable information, emphasising interdisciplinary collaboration and context definition. This process is crucial for informed decision-making and lays the groundwork for intelligent systems in PAE management [21].

In the early 2000s, the emphasis shifted to more rigorous methodologies and technological solutions for error detection [10]. Total quality management (TQM) gained prominence, emphasising optimisation across pre-analytical, analytical, and post-analytical phases. Accurate interpretation of the TQM framework emphasised the appropriateness of including test selection and sample handling [22]. As a continuation of TQM, standardised sample handling and processing protocols as part of PAE reduction to avoid inaccurate measurements were also highlighted [18]. A pivotal study focused on measurement uncertainty, advocating strict, standardised protocols and education for managing pre-analytical variability to enhance diagnostic testing reliability and patient care quality [23].

As evidence-based practice (EBP) paradigms emerged, error mitigation strategies were shifted in clinical laboratory science (CLS), towards a comprehensive approach that included pre-analytic and post-analytic processes, where educators play a pivotal role in implementing evidence-based practices, influencing accreditation, quality improvement, and patient safety [24]. Further studies showed pre-analytical (46–68.2 % of errors) and post-analytical (18.5–47 % of errors) phases, surpassing those in the analytical phase, emphasising a patient-centred, quality-driven comprehensive and standardised approach to minimise these errors [25]. Further seminal work paved the way for providing practical algorithms for the early detection of potential PAEs, before proceeding to the testing phase. This proactive approach was crucial for maintaining the integrity of laboratory results and ensuring patient safety. By implementing these algorithms, laboratories potentially reduce the incidence of erroneous results due to sample mishandling, ultimately contributing to better clinical outcomes for patients with coagulation disorders [26].

The Institute of Medicine (IOM) report highlighted how medical errors are linked to an event that may be preventable and might lead to or cause inappropriate medication use or even patient harm while medication is within the control of the health care professional, the patient, or consumer. While they occur less frequently (0.1–9.3 %), these pre-analytical errors remain critical given the wide usage in healthcare, including those happening within clinical laboratories. The report emphasised that they may compose up to 75 % of the total laboratory errors, thus highly influencing both patient care and hospital costs [27]. Sciacovelli echoed this and emphasised that 50–75 % of laboratory errors occur in the pre-analytical phase, necessitating rigorous quality controls to ensure accurate and reliable laboratory services [28], also emphasising the importance of pre-analytical steps in maintaining laboratory quality by addressing potential errors in blood specimen collection, highlighting proper techniques, accurate additives, and controlled transportation, preparation, and storage to ensure accurate and reliable test results [29].

The efforts by organisations like the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) and the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) to focus on pre-analytical guidelines during the early 21st century is reassuring. These initiatives recognise that pre-analytical errors, can significantly impact patient outcomes, leading to increased healthcare costs and resource utilisation [7], 31], 42]. An Australian study on PAEs in a hospital laboratory highlighted their diversity, noting that the sample collection ward played a critical role, with recommendations to standardise definitions and introduce targeted interventions based on error categories, to significantly improve healthcare cost savings [32].

ISO 15189 is an International Standard specifically designed for the accreditation of medical laboratories. The need for education and training on pre-analytical issues, challenges in laboratory automation, and the significance of precise sampling time for managing pre-analytical variables and meeting ISO 15189 pre-analytical requirements are amplified by the significance of quality assurance and standardised procedures for accurate and reliable laboratory results in patient care and research [33], 34].

In the recent shift in PAE quality improvement, there is a newfound emphasis on collaborating with clinicians, integrating diagnostics, and using AI technology. This shift is expected to revolutionise disease diagnosis and management, ultimately enhancing patient care [35], 36]. While AI technology might be one part of the solution, education and training on pre-analytical issues, acknowledging the challenges in laboratory automation, and meeting ISO 15189 pre-analytical requirements are essential to ensure accurate and reliable laboratory results for patient care and research [35], 36]. Further, as AI technology is integrated into reducing PAE, it is important also to contextualise bias comprehension of the analytical aspects of the test, which are crucial in defining measurement uncertainty [37] in laboratory test results (e.g. from metrological, statistical, laboratory, and clinical perspectives), emphasising the importance of estimating and treating bias to reduce errors, improve patient safety, and enhance clinical decision-making. There have been recent moves in the field of PAE reduction to understand the significance of bias, types of bias, measurement conditions, and impact. Several models for acceptable bias limits and the role of external quality assessment schemes and advocating for a pragmatic approach to handling bias in clinical laboratories have been proposed recommending future research to focus on eliminating bias for healthcare efficiency [38]. All of these factors distil down to quality management involving continuous analysis, improvement, and the re-evaluation of resources, processes, and services, where the pre-analytical and post-analytical phases of laboratory testing are prone to specific errors and challenges, highlighting also the use of AI models for management of PAEs [39].

Machine learning approaches have demonstrated high accuracy in identifying blood clots in specimens, and software platforms have been developed to standardise the recording and analysis of PAEs, allowing for global benchmarking. Challenges remain in ensuring the widespread adoption of these technologies and addressing concerns related to error evaluation and sample handling [40].

From a current and future perspective, the potential of AI to improve laboratory efficiency, patient safety, and cost-effectiveness highlights the need for further validation and consideration of regulatory, safety, and privacy concerns. Figure 3 indicates that AI and machine learning techniques are advancing very rapidly, and the potential to harness these AI advancements in the context of PAE management is vast and untapped.

Figure 3:

A 50-year timeline of the number of studies on pre-analytical error (PAE).

Classifications of PAE

There is an urgent need for comprehensive reporting systems encompassing errors within the TTP by incorporating representative pre-analytical performance measures and criteria for specimen acceptability [1]. This highlights the importance of grading laboratory errors based on their seriousness, the importance of pre- and post-analytical errors, and the need for reliable QIs to identify areas for improvement in laboratory testing. The significance of standardising data collection, collaborating with different care operators, and using QIs to ensure the quality and safety of laboratory information is also important [9], 41]. The International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) developed a comprehensive model of QI’s encompassing 53 measurements to monitor 27 QI’s which has been widely adopted in Europe [2], 43]. The Key Incident Monitoring and Management System (KIMMS) and Failure Mode Effects Analysis (FMEA) are also programs that work together to identify and manage PAEs in laboratories. KIMMS, an Australasian program initiated in 2008, records incidents and episodes to calculate incident rates. It defines risk using incident frequency, harm rating, and detection difficulty score. FMEA assigns quantified risk to each incident type, helping prioritise errors based on harm. Together, these systems aid in identifying, monitoring, and reducing errors in laboratory testing, focusing improvement efforts and resources on the most critical incident types [44], 45].

Types of errors

The most common errors are related to incorrect procedures for sample collection, leading to haemolysis and clotting. Evidence-based quality specifications, process analysis, a strong safety culture, and continuous monitoring are needed to enhance the quality of care and improve patient safety in laboratory testing. Additionally, challenges posed by staff shortages and the acritical diffusion of point-of-care testing have been highlighted by several authors. Actions suggested included addressing diagnostic errors, improving patient identification, and ensuring the transmission of critical test results. Furthermore, the lack of a universal definition of “laboratory error” and an “allowable error rate” has been noted, indicating the need for standardisation across various studies [1], 8], 10], 25].

Figure 4 summarises the broad PAE categories in the literature and synthesises them into a table format, noting that this is not an exhaustive list.

Figure 4:

Types of PAEs.

Specific factors causing PAEs

Kalra emphasises the importance of evidence-based quality indicators, preventive and corrective actions, and integrating laboratory services with healthcare to improve patient safety, and highlights the potential impact of technological solutions and the appropriateness of test requesting and interpretation in reducing diagnostic errors and enhancing patient outcomes [8]. Green provided practical insights into the high incidence of medication errors and critical incidents in Intensive Care Units (ICUs), attributing these issues to human error, poor communication, a lack of standardisation, and understaffing [27]. Green suggests potential solutions, including using daily goals forms, closed-model ICUs managed by intensivists, critical incident reporting systems, and the inclusion of pharmacists in ICU teams [27].

Figure 5 attempts to classify the broad categories of PAE based on the expansion of Figure 4 from the synthesis of existing literature into the specific factors causing PAEs.

Figure 5:

Specific PAE classification of types [2], 3], 16], 28], 32], 45], 46].

Current strategies for minimising pre-analytical errors

Patient safety is a key issue in laboratory medicine. There must be an improvement in error rates in all phases of the testing process. The cornerstone of any error reduction strategy is effective communication. The multidisciplinary approach, standardisation, simplification, and teamwork are part of the key issues. It is necessary to develop QIs for all steps of the testing process and to establish quality specifications to monitor and improve laboratory performance. The KIMMS program aims to reduce errors; indeed, while there has been a significant risk reduction associated with this program, the cost of error is still high, indicating that a clear need exists to define the root cause better and develop best practices [8], 9]. The cost of recollections due to errors in an Australian study is estimated to be 27 million AUD per year, which increases to 66 million AUD when considering the need to understand root causes better and establish best-practice guidelines [45]. In this context some of the current strategies used in contemporary practice are briefly outlined in 1.4.1–1.4.4. Further “Application of AI in reducing pre-analytical errors” section discusses how with technological advancements like AI, PAE error minimisation can be better handled.

LIMS compatibility

Laboratory Information Management Systems (LIMS) are crucial in reducing errors in medical laboratories, ensuring accurate sample tracking, automated data entry, and standardisation of procedures. They contribute to improved accuracy, efficiency, patient safety, and regulatory compliance [30], 55], 56]. In our locality, the New South Wales (NSW) Government prioritises establishing uniform data standards and dictionaries in the healthcare system to enhance decision-making capabilities for patient care and health system management. The standardised data structures are aiding in identifying care variability and driving improvements in clinical operations and resource utilisation. The interoperability of ICT systems is being streamlined by adopting health information standards and reducing integration costs by focusing on consistent standards, technology solutions, stakeholder engagement, and data exchange solutions while addressing policy, privacy, and security requirements [57].

Application of AI in reducing pre-analytical errors

In the last three decades, adopting process analysis, education, and safety culture has been recommended to minimise errors in laboratory medicine. AI is promising to drive improvement in managing PAE. The misuse of laboratory testing and the role of AI model support in addressing this issue is an encouraging development, though still in its nascent stage. Technology and automation, interdepartmental collaboration, and clinical audits have been advocated as means of limiting the occurrence of errors. All these approaches require healthcare professionals, laboratory managers, and policymakers to be proactive and supportive. Furthermore, in the above highlighted approaches different AI applications in the pre-analytical phase such as improvement in test accuracy and a decrease in manual work with a reduction in error possibility is discussed [9], 10], 40], 58]. Table 2 provides an overview of the application in areas like phlebotomy: development of automated blood drawing robots, smart blood tubes, barcodes, radio-frequency identification (RFID), voice recognition devices, transport systems, and instrumentation tools for reducing PAE and ensuring sample integrity. Moreover, predictive analytics and safety devices can increase healthcare workers’ safety in the context of continuous optimisation of workflows and protocols, and real-time specimen monitoring is also detailed. This is also further summarised in Figure 6, showing the selected application of AI in reducing pre-analytical errors in the different processes that occur in the specimen handling and processing workflow.

Figure 6:

Summary of selected application of AI in reducing pre-analytical errors in specimen handling and processing workflow and Table 2 [3], 10], 41], 47], 53], 60], 61], 63], 64], [66], [67], [68], [69], [70], [71], [72], [73], [74], [75], [76], [77].

Quality control and anomaly detection

Introducing a ‘Quality Query’ reporting system into a clinical biochemistry laboratory and prospective tracking of quality failures over 19 months led to 397 reports being completed, with most reporting focused on the pre-analytical phase, and few having significant impacts on patient care. However, the potential for adverse outcomes was noted, which puts a comprehensive approach to quality control into perspective. Risk management and standardisation of practices were being emphasised to reduce error rates and enhance patient safety with respect to laboratory testing [77]. Another study revealed that as few as 27.6 % samples were error-free, the most frequent errors being associated with the tests ordered without clinical requisition/indication, submitted requisition forms without physician identification and unsatisfactory sample volumes [78]. The identification, frequency of inappropriate laboratory test utilisation, the errors in clinical laboratories, the assessment of the impact of computer systems on test ordering and interpretation to improve patient safety, all stems from understanding and addressing these issues. It underlines the need for a patient-centred approach, efficient communication systems, proactive tools like FMEA and HAZOP, and the development of QIsfor all the steps in testing processes. This would enable the clinical laboratory to evaluate and monitor their performance with the aim of minimising error rates and improving patient safety by underlining the need for teamwork and multidisciplinary cooperation, and by providing a patient-centred approach that reduces errors in laboratory medicine. Including risk management, total quality systems and standardisation of practices will help reduce pre-analytical variability and enhance patient safety in laboratory testing [10], 46], 77], 79].

Staff training and decision support

In a comprehensive analysis of pre-analytical errors in a clinical laboratory, the importance of reducing errors, particularly in the pre-analytical phase, the use of technology and automation, such as pre-analytical workstations, information technology, and robotics, to enhance accuracy and efficiency in specimen collection and handling were key initiatives that needed staff training and decision support. Successful initiatives include using wireless networks, automated labelling systems, bar-coded ID wristbands, standardised collection procedures, and pre-analytic workstations interfaced with analysers [52]. The impact of pre-analytical errors on laboratory testing quality and patient safety, was also a critical factor highlighting the need for ongoing improvement in pre-analytical processes to enhance laboratory efficiency and patient safety. One study showed an error rate of 0.047 %, with a Six Sigma value of 4.9, indicating well-controlled processes. However, the total rate of significant errors was 13.54 %, highlighting the importance of pre-analytical error management in the clinical laboratory by Staff Training and Decision Support [80].

Continuous process improvement

The crucial role of clinical laboratories in enhancing patient safety by managing the pre- and post-analytical phases using technology and QIsto minimise errors and improve efficiency in specimen collection and handling is the core of continuous process improvement. Standardising processes and continual monitoring are needed to implement quality improvement in the total testing process. Implementing computerised order entry systems, bar-coded patient identification, and automated sample labelling systems are effective strategies to prevent errors (60–70 % of issues in laboratory diagnostics often result from mishandling during specimen collection, handling, preparing, or storing) outside the laboratory. The importance of continuous process improvement in laboratory diagnostics, specifically in the pre-analytical phase, focusing on reducing operational costs and enhancing patient safety, is an inevitable part of the operational management of PAE [46], 81].

Conclusions

While the recognition of PAEs has been part of the broader evolution of laboratory quality assurance over several decades, concerted efforts to specifically address these errors have become more prominent in the last 20–30 years. Increased emphasis on PAEs reflects the ongoing effort in understanding and improving the whole laboratory testing process in pursuit of better patient care. However, from an unbiased perspective, there are still major obstacles to their adoption, involving a lack of available digital tools, inadequate access to health data, and personnel with skills in AI. Notably, there were those laboratories in regional/remote/rural areas where a lack of stable internet connections was noted, which emphasises other infrastructure challenges and increased training being principal among them will be important if the full potential of AI is to be realised. Future initiatives aimed at the development of online training resources were postulated to address such needs [82]. Further there are integration challenges of generative AI tools in laboratory medicine. Where queries to AI tools (e.g ChatGPT, Perplexity, Google Gemini, Cohere, and You.com) showed marked variation in their responses to various marker sensitivity (ranging from 70–100 %) and specificity (7–80 %), highlighting the rather low level of heterogeneity in diagnostic information obtained with the help of generative AI tools, hence questioning the potential clinical decision-making benefits using these AI tools reiterating the importance of critical evaluation for effective integration of generative AI into laboratory medicine [83].

This review underlines several critical gaps in the literature about PAEs in clinical laboratories based on the scoping of the literature.

1. The lack of standard definitions, classification, and measurement units is identified as a serious barrier to comparing studies and defining best-practice guidelines, and subsequent recommendations are difficult to implement [27], 74].

2. Education and training in pre-analytical issues are required; however, the current literature does not provide, in general, detailed strategies or programs for such education and training [9], 23], 34], 35], 37], 68], 73].

3. Even though the role of AI and machine learning in PAE management is recognised, further research on Risk Probability Indexation models is required, with a focus on regulatory, safety, and privacy concerns [8], 9], [45], [46], [47, 52], 67], 78], 80], 82].

4. There is a need for more extensive efficacy studies and actual implementation of these technologies in real life, thereby setting the scene for future research on how to avoid bias of laboratory test results in healthcare efficiency [39], 49], 51], 83].

5. The need for discussions on quality management through continuous analysis, improvement, and re-evaluation of resources, processes, and services was also observed [10], 40].

6. Although software platforms have been developed to standardise recording and analysis for global benchmarking, there is an evident lack of information concerning translation in the adoption and usage of these platforms across diverse healthcare settings [10], 41].

In conclusion, these gaps thus provide further opportunities for research and exploration to better manage PAEs and enhance patient care.