Laboratory quality control
Laboratory quality control (QC) refers to the set of procedures and practices within a laboratory's quality management system designed to monitor the accuracy, precision, and reliability of test results, particularly during the analytical phase of testing, to detect, evaluate, and correct errors before patient or client reports are issued.[1] It ensures that laboratory outputs meet established standards for timeliness and validity, thereby supporting informed decision-making in healthcare, research, and industrial applications.[2] A core element of laboratory QC is internal quality control (IQC), which involves the routine use of control samples with known analyte values to verify the performance of analytical instruments, reagents, and operators on a daily basis.[3] IQC employs statistical tools such as Levey-Jennings control charts and Westgard rules to identify systematic or random errors, allowing laboratories to maintain process stability and take corrective actions promptly.[4] Complementing IQC is external quality assessment (EQA), also known as proficiency testing, where laboratories receive blind samples from an external provider and compare their results against peer laboratories or reference values to assess overall comparability and identify biases.[5] These combined approaches form the backbone of QC, with EQA providing an independent benchmark for laboratory competence.[6] The implementation of laboratory QC is mandated by international and national regulations to safeguard patient safety and ensure regulatory compliance, such as under the Clinical Laboratory Improvement Amendments (CLIA) in the United States, which requires ongoing monitoring of testing accuracy and precision.[7] Similarly, the ISO 15189 standard emphasizes QC as integral to accreditation, requiring laboratories to define strategies for control materials, frequency of testing, and continuous evaluation to minimize risks of erroneous results that could impact diagnosis, treatment, or public health outcomes.[8] Effective QC programs also incorporate staff training, equipment maintenance, and documentation to foster a culture of quality, ultimately enhancing trust in laboratory services across clinical, pharmaceutical, and environmental testing domains.[9]Fundamentals
Definition and Scope
Laboratory quality control encompasses the systematic processes and procedures implemented in laboratories to monitor, evaluate, and maintain the reliability, accuracy, and precision of analytical results throughout the testing process. This involves the use of statistical methods and control materials to detect and mitigate errors or variations in laboratory operations before results are released, ensuring that outputs meet predefined standards of quality.[9][10] The scope of laboratory quality control applies broadly across diverse settings, including clinical laboratories where it supports diagnostic testing, such as clinical chemistry assays for patient blood analysis; analytical laboratories focused on environmental or chemical testing to verify regulatory compliance; research laboratories developing and validating new analytical methods; and industrial laboratories ensuring the consistency of product testing in manufacturing processes. For instance, in clinical settings, quality control safeguards patient care by minimizing diagnostic errors, while in environmental testing, it upholds standards for pollutant detection to protect public health.[11][12] Historically, laboratory quality control emerged in the mid-20th century, with its foundational principles rooted in the application of statistical process control to biological data during the 1950s; a seminal contribution was the 1950 introduction of control charts by Levey and Jennings, which enabled ongoing monitoring of analytical performance.[13] This evolution built on earlier industrial quality practices but adapted them specifically for laboratory environments to address variability in biological and chemical analyses.[14] At its core, laboratory quality control comprises four key components: planning, which involves establishing quality objectives and selecting appropriate control measures; implementation, where standardized procedures and materials are applied in daily operations; monitoring, through which performance is continuously assessed using tools like control charts; and corrective actions, taken to investigate and resolve any identified deviations to restore system reliability.[9]Importance in Laboratory Operations
Laboratory quality control (QC) plays a pivotal role in safeguarding patient safety by minimizing errors that could lead to misdiagnosis and subsequent harm. Inadequate QC practices contribute significantly to diagnostic inaccuracies, with pre-analytical errors—such as improper sample collection or handling—accounting for 60-70% of all laboratory errors, often resulting in delayed or incorrect treatments that exacerbate patient conditions.[15] For instance, false positives in cancer screening tests, sometimes arising from laboratory contamination or processing mistakes, can trigger unnecessary biopsies, emotional distress, and invasive procedures for patients who do not have the disease. These errors not only endanger lives but also incur substantial financial losses, with inappropriate laboratory testing alone estimated to cost Medicare between $1.95 billion and $3.28 billion in 2019 in excess expenses due to repeat tests and unwarranted interventions.[16] Furthermore, failures in QC expose laboratories to legal liabilities, including malpractice lawsuits when diagnostic errors lead to adverse outcomes, as seen in cases where mislabeled specimens result in wrongful treatments.[17] Beyond immediate risks, robust QC enhances operational efficiency and fosters trust in laboratory results essential for clinical decision-making. Effective QC reduces error rates, leading to faster turnaround times for test results, which improves patient service and enables timely medical interventions. This error reduction also yields cost savings by avoiding the need for retesting and mitigating downstream healthcare expenditures, thereby optimizing resource allocation in laboratory workflows. By ensuring reliable data, QC bolsters clinician confidence in using lab outputs for diagnosis and treatment planning, ultimately contributing to better patient outcomes and reduced variability in care delivery.[18] QC is integral to total laboratory quality management systems (LQMS), which prioritize prevention over reactive correction to maintain consistent performance across all testing phases. LQMS frameworks integrate QC into broader processes, including personnel training and equipment maintenance, to proactively identify and mitigate risks before they impact results. This preventive approach aligns with established quality control strategies that emphasize ongoing monitoring and validation to sustain high standards, distinguishing modern laboratory operations from error-prone corrective measures.[19]Quality Control Procedures
Internal Quality Control
Internal quality control (IQC) refers to the systematic processes implemented within a laboratory to monitor the ongoing accuracy and precision of analytical measurements by incorporating control samples with established target values into routine testing workflows.[20] These control materials, typically representing low, normal, and high concentration levels relevant to decision points, are analyzed concurrently with specimens to detect deviations in performance that could affect result reliability.[21] The primary components of IQC include the selection of appropriate control materials, defined measurement frequency, statistical evaluation rules, and predefined thresholds for acceptability, all designed to ensure consistent analytical quality without unnecessarily interrupting testing operations.[20] IQC protocols generally require running controls at least once per analytical run or daily for high-volume tests, with laboratories tailoring frequency based on instrument stability, test complexity, and risk assessment to balance detection sensitivity and operational efficiency.[21] Acceptance criteria are evaluated using multirule systems, such as the Westgard rules, which apply statistical patterns to flag potential errors; for instance, a single control value exceeding the mean by 3 standard deviations (1_{3s}) or two consecutive values exceeding 2 standard deviations (2_{2s}) triggers rejection of the run to prevent erroneous reporting.[22] These rules, originally developed for efficient error detection in quantitative assays, emphasize a combination of warning and rejection limits to minimize false alarms while identifying systematic issues like random errors or shifts. Control materials for IQC fall into two main categories: commercial products, which are pre-formulated by manufacturers or third-party vendors with assigned values and stabilizers for broad analyte coverage, and in-house preparations, such as pooled specimen aliquots customized to specific laboratory needs but requiring validation for stability and commutability. Commercial controls offer convenience and traceability but may introduce matrix effects differing from routine samples, whereas in-house options can better mimic testing specimens at lower cost, though they demand rigorous characterization to establish target values and shelf life. Laboratories must select controls that span the analytical measurement range, ideally including at least two levels per run, to effectively monitor performance across variations.[20] Shifts, trends, or drifts in IQC data—indicating gradual changes in instrument calibration or reagent performance—are identified through sequential plotting and statistical analysis, prompting adjustments such as recalculating control limits from at least 20 consecutive measurements once stability is restored.[21] For example, a persistent upward trend might signal reagent deterioration, while abrupt shifts could arise from environmental factors, requiring immediate investigation to differentiate random variation from true out-of-control events.[22] Visualization tools like Levey-Jennings charts facilitate this detection by graphing control results against mean and standard deviation limits.[21] When IQC fails, corrective actions prioritize result integrity by halting reporting of affected results, followed by targeted troubleshooting such as instrument recalibration, reagent replacement, or preventive maintenance to address the root cause.[20] Laboratories then verify the fix through repeat QC runs until acceptability is confirmed, and may conduct retrospective retesting of specimens from the prior stable run to mitigate any impact on decisions.[21] Documentation of all actions, including the failure investigation and resolution, is essential for compliance and continuous improvement, ensuring traceability and informing future preventive strategies.[20]External Quality Assessment
External Quality Assessment (EQA), also known as proficiency testing (PT), is a systematic process where laboratories receive blinded samples from an external provider to evaluate their testing performance against peer laboratories or reference standards. This approach allows for inter-laboratory comparisons to ensure the accuracy and reliability of results in laboratory settings. Organizations such as the College of American Pathologists (CAP) and the World Health Organization (WHO) administer these programs, distributing samples that mimic routine specimens to detect systematic errors or biases not identifiable through internal processes alone.[23][24][25] The process begins with the preparation and distribution of stable, commutable samples—often lyophilized human or bovine sera—by the EQA provider, ensuring they reflect real matrices without identifiable characteristics. Laboratories treat these samples as routine specimens, performing analyses with standard methods and personnel, then submit results electronically within specified deadlines, typically via online portals. Results are evaluated against target values derived from peer group consensus or assigned references, using statistical metrics such as z-scores, where a participant's deviation is calculated as z = \frac{x - \mu}{\sigma} (with x as the participant's result, \mu as the target mean, and \sigma as the standard deviation), flagging scores outside ±2 as potential issues. Peer group comparisons follow, with feedback reports detailing individual performance, trends, and educational insights to guide corrective actions.[25][26][27] EQA programs offer significant benefits by benchmarking laboratory performance, identifying method-specific biases, and supporting continuous quality improvement, which enhances result reliability and standardization across networks. They complement internal quality control by revealing inter-laboratory variations and promoting best practices through educational resources. However, limitations include dependence on the accuracy of participant submissions, potential matrix effects from sample preparation that may not fully replicate testing scenarios, and a focus primarily on analytical phases, overlooking pre- and post-analytical errors. Additionally, while EQA maintains quality levels, it may not drive substantial improvements beyond established thresholds without integrated internal monitoring.[6][24][27] Participation in EQA is typically conducted quarterly—three to four times per year—to provide regular feedback, with more frequent cycles possible for high-risk analytes. For accreditation under standards like ISO 15189 or CLIA, enrollment is mandatory, requiring successful participation in proficiency testing events as defined by the standards.[24][25][28]Monitoring Tools
Control Charts Overview
Control charts are graphical statistical tools employed in laboratory quality control to monitor the stability of analytical processes over time by plotting measurement data against established limits, enabling the differentiation between common cause variation—random, inherent fluctuations expected in a stable process—and special cause variation, which signals assignable, non-random shifts requiring corrective action.[29] Developed by Walter A. Shewhart in 1924 at Bell Laboratories, these charts provide a visual means to assess whether a laboratory's measurement system remains in a state of statistical control, thereby supporting timely detection of process deviations that could compromise result reliability.[30] Their primary purpose in laboratories is to ensure consistent performance of analytical methods, reducing the risk of erroneous patient results through proactive identification of instability.[31] Several types of control charts are utilized in laboratory settings, each suited to detecting specific patterns of variation. Shewhart control charts, the foundational type named after their inventor, are effective for monitoring individual measurements or small subgroups, focusing on larger shifts in process means or variances. For detecting smaller, gradual shifts, cumulative sum (CUSUM) charts accumulate deviations from the target mean, offering greater sensitivity to sustained changes than Shewhart charts.[32] Exponentially weighted moving average (EWMA) charts, in contrast, emphasize recent data points through a weighting factor, making them particularly useful for identifying trends or small drifts in laboratory processes.[33] The basic construction of control charts involves plotting sequential quality control data points on a graph with time or run order on the x-axis. A central line represents the process mean, calculated from an initial stable period of data, while upper and lower control limits are typically set at ±3 standard deviations from the mean to encompass approximately 99.7% of expected variation under common cause conditions.[29] These limits are derived empirically from the laboratory's own control material data rather than theoretical specifications, ensuring relevance to the specific analytical system.[31] Points falling outside these limits or exhibiting non-random patterns, such as runs or trends, indicate potential special causes warranting investigation. In laboratory applications, control charts are routinely applied to track the stability of analyte measurements, such as glucose concentrations in blood samples or hemoglobin levels in hematology assays, allowing technicians to verify instrument calibration and reagent performance before reporting patient results.[34] This ongoing surveillance helps maintain the precision and accuracy of diagnostic testing, with charts updated daily or per run to reflect real-time process behavior. In clinical laboratories, variants like the Levey-Jennings chart adapt these principles specifically for internal quality control data.[35]Levey-Jennings Chart
The Levey-Jennings chart, a graphical tool for monitoring laboratory quality control, was developed in 1950 by Stanley Levey and E.R. Jennings specifically for clinical chemistry applications.[36] This chart adapts the Shewhart control chart principles to plot individual quality control (QC) measurements from reference samples, enabling laboratories to detect variations in analytical performance over time. It remains a foundational method in laboratory medicine for ensuring the reliability of test results. Construction of a Levey-Jennings chart involves plotting daily or per-run QC values on the y-axis against sequential runs or dates on the x-axis. The chart features a central horizontal line representing the target mean value, calculated from an initial set of at least 20-30 stable QC measurements, along with parallel lines at ±1 standard deviation (SD), ±2 SD, and ±3 SD from the mean. These SD lines are derived from the same initial data set, providing visual boundaries for acceptable variation based on the normal distribution of QC results.[37] Interpretation of the Levey-Jennings chart typically employs the Westgard multirule system, introduced in 1981, which combines multiple statistical rules to distinguish between random and systematic errors while minimizing false rejections. Key rules include 1_{3s} (a single QC value exceeding ±3 SD, indicating a possible random error like an outlier), 2_{2s} (two consecutive values exceeding ±2 SD on the same side of the mean, suggesting a systematic shift), and R_{4s} (the range between consecutive values exceeding 4 SD, flagging potential random error in paired controls). For example, a single point beyond the ±3 SD line might signal random error from instrument noise, prompting immediate investigation, whereas a series of points trending above the mean could indicate systematic error due to reagent deterioration, requiring corrective action before resuming patient testing.[38]Validation and Verification
Method Validation
Method validation in laboratory quality control refers to the systematic process of confirming that a new or modified analytical method is suitable for its intended purpose prior to routine implementation, encompassing assessments of key performance characteristics such as accuracy, precision, linearity, and reportable range. According to CLSI guidelines, this evaluation ensures the method meets predefined criteria for reliability and robustness in clinical or research settings.[39] Similarly, ISO 15189:2022 specifies that laboratories must validate examination procedures to verify their fitness for use, particularly when developing or significantly altering methods, through objective evidence of performance. This one-time comprehensive study distinguishes validation from ongoing verification, focusing on establishing baseline method reliability. Key parameters evaluated during method validation include precision, assessed via within-run (repeatability) and between-run (reproducibility) coefficients of variation (CV%), where typical targets aim for CV values below 5-10% depending on the analyte and medical decision point. Accuracy is examined through recovery studies, involving the addition of known analyte amounts to samples and measuring the percentage recovered, ideally 95-105% for most assays. Interference testing screens for potential disruptions from substances like hemoglobins, lipids, or drugs, using protocols such as CLSI EP07 to quantify effects via paired-difference analysis. Linearity confirms proportional response across the analytical measurement range (AMR), while the reportable range defines the clinically relevant span, often verified with dilutions or spikes.[40][41] The validation process begins with protocol design, outlining sample types, concentrations (e.g., low, normal, high), number of replicates (typically 20-40 per level), and statistical methods, as recommended by CLSI EP series documents. Data collection spans multiple days and operators to capture variability, with at least 5-20 runs for precision estimates per CLSI EP05-A3. Statistical analysis follows, including calculation of means, standard deviations, CV%, regression for linearity, and t-tests or Deming regression for bias assessment against a reference method per CLSI EP09-A3.[42] Acceptance criteria require that the total analytical error (TE = bias + 1.96 × imprecision) does not exceed the allowable total error (TEa) derived from biological variation data, such as those in the EFLM database, where TEa for analytes like glucose might be set at 7.1% based on within-subject variation. For example, if biological variation yields a desirable imprecision goal of CV < 2.5% and bias < 2.6%, the method passes if total error remains within these limits across tested conditions. Failure prompts method refinement or rejection, ensuring patient safety and diagnostic utility.[43][44]Method Verification
Method verification serves as an abbreviated validation process for laboratories implementing quantitative measurement procedures that have already been validated by the manufacturer, focusing on confirming that the method performs acceptably under the laboratory's specific conditions. This process is guided by CLSI EP15-A3, which provides a protocol to verify precision and estimate bias (a measure of trueness) in a single experiment typically completed over five working days using two or three quality control materials at different concentrations. The protocol involves running multiple replicates (e.g., five per day) to assess within-run and between-run imprecision, ensuring the laboratory's results align with the manufacturer's claims, such as coefficient of variation limits. Key parameters evaluated include precision, trueness through spiking experiments, and comparability to a reference method. Precision is determined by calculating repeatability and reproducibility from replicate measurements, while trueness is assessed by comparing results from spiked samples or certified reference materials against expected values to estimate bias. For method comparison, patient samples are analyzed on both the new method and a reference procedure, with linear regression used to evaluate agreement; the regression line's slope and intercept indicate proportional and constant bias, respectively. Unlike the full method validation required for new or modified methods (detailed in Method Validation), verification emphasizes these targeted checks for transferred assays.[45] Verification is conducted at initial implementation of the method, following major changes such as new reagent lots, instrument upgrades, or significant environmental shifts, and periodically, such as annually, to ensure sustained performance.[46] Statistical acceptance criteria for method comparison studies typically require a slope between 0.95 and 1.05 (indicating minimal proportional error) and an intercept near zero (indicating negligible constant bias), with confidence intervals confirming the method's equivalence to the reference.[41] These criteria help laboratories confirm that systematic errors remain within clinically acceptable limits, supporting reliable routine use.[47]Performance Metrics
Analytical Sensitivity and Specificity
Analytical sensitivity in laboratory testing refers to the ability of an assay to detect low concentrations of the target analyte, often quantified as the limit of detection (LoD), which is the lowest concentration reliably distinguishable from background noise. According to CLSI guidelines, the LoD is calculated as the limit of blank (LoB) plus a multiple of the standard deviation of low-concentration samples, ensuring the assay's capability to identify true positives at minimal levels. Analytical specificity, conversely, measures the assay's ability to selectively detect the intended analyte without interference from similar substances or cross-reacting materials, evaluated through interference and cross-reactivity studies. In laboratory quality control, these analytical performance characteristics are distinct from diagnostic metrics. For qualitative laboratory tests, diagnostic sensitivity and specificity assess the test's ability to correctly identify diseased and non-diseased individuals, respectively, using a 2x2 contingency table that categorizes results based on the true disease status and test outcome. Diagnostic sensitivity, also known as the true positive rate, is calculated as the proportion of true positives (TP) among all actual positives: \text{Sensitivity} = \frac{\text{TP}}{\text{TP} + \text{FN}} where FN represents false negatives.[48] Diagnostic specificity, or the true negative rate, is the proportion of true negatives (TN) among all actual negatives: \text{Specificity} = \frac{\text{TN}}{\text{TN} + \text{FP}} with FP denoting false positives.[48] These metrics assess a test's discriminatory power in laboratory quality control, independent of disease prevalence.| Disease Present | Disease Absent | |
|---|---|---|
| Test Positive | TP | FP |
| Test Negative | FN | TN |