Sample self-relay method for laboratory real-time quality control and automation system

Abstract

The invention discloses a sample self-relay method for laboratory real-time quality control and an automatic system, and belongs to the technical field of clinical laboratory automatic quality control. According to the method, an independent quality control cycle is established for a plurality of preset key concentrations, and a concentration-bias and concentration-inaccuracy mapping table is established based on historical quality control data. Independent counting is maintained for each target concentration, when the detection number reaches the length L of the current analysis batch, quality control is automatically triggered, two patient samples with the concentrations meeting the requirements are screened from the current detection batch to serve as a self-relay quality control sample pair, and judgment is conducted according to the rule of 1-3 s after returning and re-detection. And meanwhile, querying a mapping table according to the actual concentration of the sample to calculate a sigma value, querying a preset sigma-L mapping table based on the sigma value, and dynamically updating the next analysis batch length of the target concentration. Each cycle operates independently and asynchronously to form a closed loop, so that accurate, automatic and real-time quality control of different concentrations and different frequencies is realized, and the implementation bottleneck of personalized 6 sigma quality control in cost and reliability is overcome.

Description

Technical Field

[0001] This invention belongs to the field of automated quality control technology in clinical laboratories, specifically relating to a sample self-relay method and automated system for real-time quality control in laboratories. Background Technology

[0002] Internal quality control in clinical laboratories is a core element in ensuring accurate test results and preventing medical risks. Currently, driven by the Six Sigma quality management philosophy, the industry consensus requires that quality control strategies must be tailored to the actual performance level of the tested items, and that key concentrations such as those at medical decision levels and treatment cut-off points be closely monitored. However, the existing technical system, which heavily relies on commercially available quality control materials and fixed sampling rules, has exposed a series of profound and interconnected systemic bottlenecks when implementing this advanced goal, seriously hindering the effective implementation of Six Sigma management.

[0003] First, the existing model faces the dual constraints of economic cost and data source reliability. To achieve differentiated, high-frequency monitoring of multiple key concentration points, the consumption and procurement costs of quality control materials will increase exponentially, placing a heavy economic burden on large laboratories. More fundamentally, quality control materials repeatedly used online to adapt to automated production lines face the long-term risks of evaporation, degradation, and microbial contamination, leading to uncontrollable drift in their target values. This instability in physicochemical properties inherently compromises the reliability of quality control materials as the data foundation for high-frequency, long-term monitoring, undermining the cornerstone of building precise quality models.

[0004] Secondly, the quality control strategy suffers from two rigidities, failing to meet dynamic and precise clinical needs. Firstly, the monitoring frequency is rigid. Current triggering rules based on fixed intervals or fixed batch lengths cannot be dynamically adjusted according to fluctuations in the real-time performance of the testing system, nor can they respond to significant performance differences of the same item across different concentration ranges. For example, the precision of cardiac troponin in low-value and high-value ranges may differ by several times, resulting in quality control resources being unable to be precisely deployed at critical concentration points and weak performance ranges. Secondly, the monitoring targets themselves are rigid. The concentrations of commercially available quality control products are preset by manufacturers and the selection is limited. When clinical guidelines are updated or new treatment thresholds emerge, laboratories cannot flexibly and promptly adjust the target concentrations to be monitored, requiring a lengthy cycle of quality control product procurement, validation, and standard traceability, causing a severe lag in the clinical response of the quality control strategy.

[0005] Secondly, existing quality control models lack an effective mechanism for real-time process control using patient samples. Traditional methods rely on inserting quality control samples before testing begins or at fixed intervals. This discrete sampling method lacks the ability to verify the ongoing testing process itself in real time. More importantly, the value of massive amounts of patient samples, as the most readily available and authentic quality control material, is severely underestimated. Although repeatable testing of patient samples can theoretically directly reflect the current precision of the system, in existing processes, it is only used as a passive, post-hoc means of troubleshooting, such as for confirmation after a loss of control. It has never been systematically integrated into an active, automatically triggered, real-time quality control signal source that can guide the frequency of subsequent quality control. This leads to two consequences: first, quality control activities are disconnected from the actual testing process in time; second, the richest real-time data resources are idle and cannot be used to build a self-adjusting, self-optimizing dynamic quality control closed loop.

[0006] Finally, the scalability of the entire system is limited by the external supply chain. Adjustments to quality control strategies and the addition of concentration points are strictly constrained by the availability of commodity quality control materials and fixed concentration configurations. The laboratory lacks the ability to respond quickly and autonomously according to local needs.

[0007] In summary, existing quality control systems centered on commercially available quality control products face insurmountable structural contradictions when implementing the goals of 6σ refined management: a balance between cost and data source reliability is difficult to achieve; their static rules are severely out of sync with the clinical need for dynamic and precise monitoring; and the most valuable, time-sensitive patient sample data remains idle, unable to be transformed into real-time quality control decisions. Therefore, there is an urgent need in this field for an innovative technological solution that must break free from absolute dependence on commercial quality control products, allowing laboratories to directly set monitoring targets for key clinical concentrations, and innovatively utilize patient samples themselves to achieve real-time, closed-loop process validation and frequency self-adjustment. This would transform advanced 6σ quality management from a theoretical framework into an economically feasible, responsive, and self-scalable daily practice system. This invention is a systematic solution proposed to address these closely related and pressing needs. Summary of the Invention

[0008] This invention aims to address the technical challenges faced by clinical laboratories in implementing personalized, high-frequency 6σ quality control, including high costs, insufficient data source reliability, and rigid strategies. To this end, it provides a sample self-response method and automated system for real-time quality control in laboratories. This method establishes independent quality control cycles for multiple preset key concentrations, and builds concentration-bias and concentration-imprecision mapping tables based on historical quality control data. The system maintains independent counts for each target concentration. When the number of tests reaches its dedicated current analysis batch length L, quality control is automatically triggered, and two patient samples with the required concentrations are intelligently selected from the current batch as a self-response quality control sample pair. After the sample pair is retested, it is judged according to the 1-3s rule. Simultaneously, the σ value is calculated based on the actual concentration of the current sample, and the next analysis batch length L for that target concentration is dynamically updated. This achieves precise and automated quality control at different frequencies for different concentrations, effectively overcoming the cost and reliability bottlenecks in personalized 6σ quality control.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] A sample self-relay method for real-time quality control in the laboratory is characterized by establishing independent quality control cycles for at least one preset quality control target concentration, each cycle including the following steps:

[0011] S1: Maintain an independent counting channel for each of the quality control target concentrations. When the cumulative number of samples tested for the target concentration reaches its dedicated current analysis batch length L, real-time quality control is automatically triggered. From the patient samples tested in the current analysis batch, two samples with concentrations that meet the target concentration requirements are intelligently selected and identified as a self-relay quality control sample pair.

[0012] S2: The self-relay quality control samples are scheduled back to the detection unit for immediate retesting, the deviation between the first test result and the retest result of each sample is calculated, and the control cycle is determined to be in control or out of control according to the 1-3s quality control rule.

[0013] S3: Based on the actual concentrations of the two self-relay samples determined in S1, query the preset concentration-bias mapping table and concentration-imprecision mapping table, calculate the performance value using the 6σ metric theory, and dynamically update the length L of the next analysis batch for the target concentration cycle accordingly.

[0014] S4: After completing S2 and S3, reset the count for the target concentration cycle and continue to run independently based on the new length L. Each target concentration cycle repeats S1 to S3 asynchronously and in parallel.

[0015] Specifically, the step of calculating the performance value using the 6σ metric theory and dynamically updating the length L of the next analytical batch for the target concentration cycle includes:

[0016] Calculate the σ value for each of the two self-relay samples using the following formula: , where TEa is the total allowable error of the project, B is the bias obtained from the concentration-bias mapping table, and CV is the coefficient of variation obtained from the concentration-imprecision mapping table;

[0017] Take the smaller of the two σ values ​​as ;

[0018] Will Round to one decimal place. If the result is greater than 6, it will be rounded to 6.

[0019] Using the final processed σ value, query the σ-L mapping table pre-set based on N=2 and the 1-3s rule to determine and update the next analysis batch length L.

[0020] Specifically, the concentration-bias mapping table is established based on historical interlaboratory quality assessment data, and the concentration-imprecision mapping table is established based on historical internal quality control data. It contains at least the standard deviation (SD) information at different concentration levels and supports linear interpolation calculation.

[0021] Specifically, the intelligent screening identifies two samples with concentrations that meet the target concentration requirements, including:

[0022] S1.1: Using the preset quality control target concentration as the center, set the first allowable concentration deviation range, and select samples whose concentration falls within this range from the tested samples in the current analytical batch to form a primary candidate list;

[0023] S1.2: Select the two samples with the earliest detection time from the primary candidate list as pending candidates;

[0024] If samples with the same detection time exist, they are selected from among them in ascending order of absolute deviation between their concentration and the target concentration.

[0025] S1.3: If the number of specimens in the primary candidate list is less than two, expand the allowable concentration deviation range to form a new candidate list and repeat the screening logic of S1.2.

[0026] S1.4: For the candidate specimens identified by S1.2 or S1.3, verify whether they are still in an accessible location of the pipelined online storage unit through the pipeline specimen location system. If the candidate specimen is not in an accessible location of the pipelined online storage unit, remove it from the candidate list and return to S1.2 to select supplementary specimens from the remaining candidate specimens until two accessible specimens are identified.

[0027] Specifically, if the candidate specimen to be determined is not in an accessible location of the pipeline's online storage unit, then execute...

[0028] S1.4.1: From the currently valid candidate list, select the next sample as a new candidate based on the priority of the detection time from earliest to latest;

[0029] S1.4.2: If there are multiple candidates with the same detection time, select them from among them according to the priority of the absolute deviation between the concentration and the target concentration from small to large.

[0030] S1.4.3: For the newly selected candidate specimens, re-execute the physical location verification in S1.4;

[0031] S1.4.4: If two specimens in an accessible location cannot be determined after traversing the candidate list during the replacement process, an exception log is recorded and this quality control is skipped.

[0032] Specifically, determining the controlled or uncontrolled state of the quality control cycle based on the 1-3s quality control rule includes:

[0033] S2.1: Retrieve the cumulative standard deviation (SD) corresponding to the actual concentration of the self-relay sample or the concentration of the quality control target from the concentration-imprecision mapping table;

[0034] S2.2: Set 3SD as the allowable deviation limit;

[0035] S2.3: When the absolute value of the intra-batch deviation of any self-relay sample exceeds the allowable deviation limit, it is determined to be out of control.

[0036] Specifically, the automated system for real-time quality control in the laboratory is integrated into the testing pipeline and includes:

[0037] The dynamic analysis batch management module is used to maintain an independent counting channel for each quality control target concentration and to issue an independent trigger signal when the count reaches its analysis batch length L.

[0038] The self-relay sample decision module is used to respond to trigger signals and select and lock two self-relay quality control samples from the current analysis batch for the corresponding target concentration.

[0039] The real-time backtesting scheduling and execution module is used to control the pipeline to send locked sample pairs back to the detection unit for retesting;

[0040] The controllable state determination module is used to calculate the intra-batch deviation of the sample and make a controllable / out-of-control determination based on the 1-3s rule;

[0041] The adaptive frequency calculation module is used to calculate and update the next analysis batch length L for the corresponding target concentration cycle based on the locked sample data.

[0042] Compared with the prior art, the present invention has the following significant advantages:

[0043] 1. This invention fundamentally overcomes the economic and reliability bottlenecks in implementing personalized, high-frequency 6σ quality control. Through its core mechanism of "patient sample self-relay," it completely eliminates reliance on commercial quality control products, thereby avoiding the exponential cost increases caused by high-frequency, multi-concentration monitoring. It also thoroughly avoids the data source reliability risks caused by the poor online stability of quality control products, providing a new, economically feasible, and data-reliable technical path for advanced quality management practices in laboratories.

[0044] 2. This invention represents a breakthrough in quality control paradigms, moving from "uniform monitoring of all projects" to "personalized and precise monitoring of concentration ranges." By establishing independent cycles for different clinically critical concentrations, this invention effectively achieves "different quality control frequencies for different concentrations." The system can dynamically allocate monitoring resources based on real-time performance evaluation (σ value) at each concentration point: automatically extending the batch length for high-performance or stable concentration areas to improve efficiency; and automatically increasing monitoring density for critical or key medical decision points to ensure safety, thereby achieving optimal and intelligent allocation of quality control resources.

[0045] 3. A fully automated, real-time intelligent quality control closed loop was constructed. This method achieves unmanned operation throughout the entire process, from independent multi-concentration triggering, intelligent dual-sample screening and scheduling, and immediate retest determination to dynamic adaptive frequency planning. Crucially, it utilizes readily available real-time patient sample data to establish an online feedback closed loop of "detection-real-time evaluation-instant adjustment," shortening the response delay in quality control and more effectively preventing batch errors.

[0046] 4. Significantly enhances the clinical relevance and flexibility of quality control strategies. Laboratories can independently set one or more key monitoring concentrations based on clinical needs, such as medical decision levels or pathological high values, allowing quality control activities to directly focus on the core ranges affecting diagnostic and treatment decisions, forming a comprehensive monitoring network. This solves the pain point of traditional quality control strategies, which are limited by fixed concentrations of quality control materials and cannot quickly respond to changes in clinical needs.

[0047] 5. Synergistic Enhancement with Existing Quality Systems. This invention specializes in "real-time independent monitoring of multi-concentration precision," filling the gap between traditional quality control and patient data quality control. The combination of these three elements forms a more sensitive, comprehensive, and complete new modern laboratory quality assurance system. Crucially, under the economically feasible premise of "patient sample self-relay," this invention establishes quality objectives superior to traditional fixed-frequency models: the system prioritizes extremely high error detection sensitivity and extremely low patient risk, rather than simply pursuing maximum efficiency. Through real-time, high-frequency monitoring, the system can issue alerts at early stages when errors cannot be identified by traditional models, thereby substantially reducing the number of expected unreliable patient outcomes to a lower level, achieving a fundamental improvement in quality assurance capabilities.

[0048] 6. This invention achieves a fundamental breakthrough over traditional approaches in terms of technical reliability and overall cost structure. Compared to the traditional approach that relies on commercially available quality control products, the "patient sample self-relay" mechanism adopted in this invention avoids the risk of data source distortion caused by online degradation of quality control products, ensuring the authenticity of performance evaluation data. Simultaneously, it shifts the main cost of implementing high-frequency, personalized quality control from expensive external quality control product procurement to internal routine reagent consumption, achieving a breakthrough in economic feasibility. This simultaneously solves the dual constraints of insufficient data reliability and high economic costs faced in implementing advanced 6σ quality control. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the automated system architecture for independent real-time quality control of multi-target concentrations implemented in this invention.

[0050] Figure 2 This is the overall flowchart of the multi-concentration independent cyclic sample self-relay real-time quality control method described in this invention;

[0051] Figure 3 This is an example diagram illustrating the σ-L mapping used in this invention to dynamically determine the independent batch length of each target concentration analysis. Detailed Implementation

[0052] Compared to the traditional approach that relies on commercially available quality control products to achieve personalized, high-frequency 6σ quality control, the "patient sample self-relay" and "dynamic analysis batch management" approach proposed in this invention has the following decisive advantages in practice:

[0053] In traditional approaches, commercially available quality control materials used repeatedly online are susceptible to target value drift due to evaporation, degradation, and contamination, potentially distorting the quality control signal. This invention directly uses patient samples from the current analytical batch for immediate retesting, with a matrix completely identical to that of routine test samples. This fundamentally avoids the aforementioned risks, ensuring that the data source used for performance evaluation (σ-value calculation) accurately reflects the real-time status of the testing system.

[0054] In the traditional approach, achieving high-frequency, multi-concentration monitoring based on real-time σ values ​​would lead to an exponential increase in the cost of commodity quality control materials, making it economically infeasible. This invention shifts the core cost of implementing advanced quality control from externally procured quality control materials to internal consumption of routine retesting reagents, thus economically removing the obstacle to large-scale implementation.

[0055] Based on the above two points, this invention simultaneously solves the dual constraints of insufficient data reliability and high economic cost faced by the implementation of advanced 6σ quality control, and provides clinical laboratories with an economical, feasible and data-reliable real-time adaptive quality control path.

[0056] Example 1: Real-time quality control process based on dual-concentration independent cycles

[0057] To make the objectives, technical solutions, and advantages of this invention clearer, the embodiments of this invention will be described in detail below with reference to the accompanying drawings, taking the blood glucose (GLU) test in a clinical biochemistry laboratory as an example. This invention is not limited to this embodiment; it can be applied to any project suitable for automated automated testing.

[0058] See Figures 1-3 This embodiment demonstrates the complete application of the core logic of the method of the present invention in an automated production line. The method establishes an independent and parallel "monitoring-evaluation-planning" closed loop for each preset clinical key concentration (target concentration) on an analysis batch basis.

[0059] (1) System initialization and performance database establishment:

[0060] The system establishes a core performance database for each detection item, which serves as the basis for all calculations and judgments.

[0061] 1) Establishing a concentration-bias mapping table: The system receives and imports historical interlaboratory quality assessment data for this test item. Based on this, the system calculates the relative bias B between the laboratory's test results and the target value at each concentration level. The system sorts these (concentration, bias) data points and generates a mapping table that supports linear interpolation calculation.

[0062] 2) Establish a concentration-imprecision mapping table: The system retrieves historical internal quality control data, calculates the cumulative mean and standard deviation for different concentration levels, sorts these (concentration, standard deviation) data points, and generates a mapping table that supports linear interpolation calculation. This table is used for quality control judgment, and the coefficient of variation CV calculated by CV = SD / Mean is used to measure the σ value.

[0063] 3) Parameter preset: The operator enters the allowable total error TEa for the item, such as blood glucose ±7%, and presets the target concentration that needs to be monitored independently, such as the medical decision level of 6.0 mmol / L and the pathological high value of 12.0 mmol / L;

[0064] 4) Advanced implementation of performance data: The discrete mapping table mentioned above can generate a continuous concentration-performance function through linear regression or spline interpolation, thereby providing a smoother and more accurate performance parameter estimate for any concentration point.

[0065] (2) Determination of the first-round quality control frequency (analytical batch length L):

[0066] Upon system startup, an initial σ value is calculated using a preset target concentration to determine the length L of the first analytical batch. For example, with a target concentration of 6.0 mmol / L:

[0067] 1) The system uses interpolation to find the example bias B=1.0% and the example cumulative CV=1.5% at this concentration from the two mapping tables in (1);

[0068] 2) Calculate the σ value according to the 6σ metric theory formula: σ = (TEa - |B|) / CV = (7 - 1) / 1.5 = 4.0;

[0069] 3) Query the pre-defined σ-L mapping table, such as... Figure 3 As shown, the corresponding batch length L is obtained. It should be noted that the σ-L mapping relationship used in this invention is calculated based on the widely accepted Westgard model in the field of clinical quality control. The Westgard model uses the constraint of ensuring a sufficient error detection rate and controlling the number of expected unreliable patient results to an acceptable low level. In the high-frequency quality control mode achieved by the patient sample self-relay of this invention, a more stringent quality objective than the traditional mode is adopted, aiming to achieve a more sensitive response to system performance through a shorter batch length. Figure 3 A set of exemplary mapping relationships based on N=2 and the 1-3s rule is shown, for example, L=21 when σ=4.0;

[0070] 4) The system sets this L value, for example 21, as the initial trigger threshold for the target concentration A (6.0 mmol / L) cycle. Similarly, it sets an independent initial L value for the target concentration B (12.0 mmol / L) cycle.

[0071] (3) Real-time detection, triggering, and dual-sample relay:

[0072] 1) After the test begins, the independent counting modules for target concentration cycle A and target concentration cycle B are activated to accumulate the number of patient specimens that meet the GLU test requirements.

[0073] 2) When the count of target concentration A cycle reaches its current L value, for example 21, the system automatically triggers a quality control event for that target concentration.

[0074] 3) The self-relay sample decision module is activated, intelligently selecting two samples from the current L tested samples that have a concentration closest to 6.0 mmol / L for target concentration A. The selection logic specifically includes:

[0075] Preliminary selection: Set an allowable deviation range centered on the target concentration, screen out samples whose concentration falls within this range, and form a preliminary candidate list;

[0076] Priority screening: Select the two samples with the earliest detection time from the list first; if the times are the same, select according to the absolute deviation of the concentration from the target concentration from the smallest to the largest.

[0077] Range expansion: If there are fewer than two candidates, the deviation range is expanded and the screening is repeated;

[0078] Physical location verification: Verify whether the candidate specimen is still in an online accessible location using the pipeline positioning system; if not, replace the candidates according to priority until two accessible specimens are identified.

[0079] 4) The module will identify the two specimens that are finally locked, such as Glu-001 and Glu-002, as the self-relay quality control sample pair, and send their identification along with the initial test data to the quality control judgment module.

[0080] (4) Immediate retest and in-control decision

[0081] 1) The real-time backtesting scheduling execution module receives instructions and controls the pipeline to automatically schedule the locked sample pairs to the original detection unit;

[0082] 2) The detection unit uses the exact same parameters as the initial test, such as reagent batch number, temperature, and time, to retest the sample to ensure that the test conditions are consistent for the two tests.

[0083] 3) The quality control judgment module calculates the deviation between the first result and the retest result for each sample. This deviation is the intra-batch deviation measured under strictly consistent testing conditions during the same continuous testing process.

[0084] 4) The quality control judgment module then uses the 1-3s rule to make a judgment. The specific judgment process is as follows:

[0085] a) From the concentration-imprecision mapping table, retrieve the cumulative standard deviation (SD) corresponding to the actual concentration of the sample (or the target concentration).

[0086] b) Set 3SD as the allowable deviation limit;

[0087] c) If the absolute value of the intra-batch deviation of any self-relay sample exceeds this allowable deviation limit, the cycle is determined to be "out of control"; otherwise, it is determined to be "in control".

[0088] The system will only continue to allow subsequent patient reports when the system is in "under control" status.

[0089] (5) Dynamic updates to the analysis batch length are performed in parallel with the decision-making process, and the batch length for the next analysis will be updated regardless of the decision-making result. Specifically, this includes:

[0090] 1) The adaptive frequency calculation module queries two mapping tables based on the actual concentrations of the two locked samples, and obtains their respective biases (B1, B2) and coefficients of variation (CV1, CV2) through interpolation.

[0091] 2) Calculate the σ values ​​for the two samples respectively: σ1 = (TEa - |B1|) / CV1, σ2 = (TEa - |B2|) / CV2.

[0092] 3) Following the principle of conservatism, take the smaller of σ1 and σ2 as... .

[0093] 4) To Perform engineering standardization: First, round to one decimal place; if the result is greater than 6, treat it as 6 to obtain the final σ value;

[0094] 5) Using this final σ value, query the σ-L mapping table to dynamically determine the length L' of the next analytical batch for the target concentration cycle, and immediately update it to its independent counting channel.

[0095] (6) Continuous operation of independent loops

[0096] 1) After the target concentration A cycle is completed in this quality control, immediately reset its count to zero and start a new round of independent counting and monitoring with the newly calculated L' value.

[0097] 2) The target concentration B cycle runs completely independently and asynchronously, unaffected by the A cycle, and is triggered, filtered, judged and updated according to its own rhythm;

[0098] 3) Based on this, each target concentration cycle forms multiple independent detection-evaluation-planning closed loops to achieve all-weather, adaptive real-time quality control.

[0099] (7) Maintenance and dynamic adjustment of system parameters

[0100] To ensure the timeliness of the model, the system supports dynamic adjustment of the performance database and quality control frequency:

[0101] 1) Mapping table update: The system can periodically, such as monthly / quarterly, automatically import the latest interlaboratory quality assessment and internal quality control data, recalculate and update the concentration-bias and concentration-imprecision mapping tables;

[0102] 2) Analysis Batch Length Recalculation: When the mapping table is updated or the total allowable error for a project changes, resulting in a change in the calculated σ value, the analysis batch length calculation module will automatically recalculate the L value for each target concentration cycle. If the difference between the new L value and the original value exceeds a certain threshold, such as 20%, the system will prompt the operator to confirm and update.

[0103] (8) Example of running effect

[0104] For clarity, let's assume the system has been running for a period of time:

[0105] For low-concentration cycling, such as target concentration ~6.0 mmol / L, after calculation and normalization based on its self-relay sample, σ=4.8 is obtained, and the length of the next analytical batch L1=127 is obtained from the table;

[0106] For high-concentration cycling, such as a target concentration of ~12.0 mmol / L, σ is calculated to be 5.3, and L2 is calculated to be 387 from the table.

[0107] Subsequently, quality control was triggered approximately every 127 samples in the low-concentration cycle, while it was triggered every 387 samples in the high-concentration cycle. This clearly demonstrates the precise management capability of this invention, which allows for different optimal monitoring frequencies at different concentrations due to performance differences, ensuring safety in the critical low-concentration range while improving detection efficiency in the high-concentration range.

[0108] Example 2:

[0109] This embodiment provides an automated system for real-time quality control in a laboratory to implement the above method. The system is integrated into a testing production line and includes:

[0110] The dynamic analysis batch management module is used to maintain an independent counting channel for each quality control target concentration and to issue an independent trigger signal when the count reaches its analysis batch length L.

[0111] The self-relay sample decision module is used to respond to trigger signals and select and lock two self-relay quality control samples from the current analysis batch for the corresponding target concentration.

[0112] The real-time backtesting scheduling and execution module is used to control the pipeline to send locked sample pairs back to the detection unit for retesting.

[0113] The controllable state determination module is used to calculate the intra-batch deviation of the sample and make a controllable / out-of-control determination based on the 1-3s rule.

[0114] The adaptive frequency calculation module is used to query the mapping table, calculate the σ value, and update the next analysis batch length L of the corresponding target concentration cycle based on the locked sample data.

[0115] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments under the guidance of the present invention without departing from the spirit and scope of the present invention. All of these variations are within the protection scope of the present invention.

Claims

1. A sample self-relay method for real-time quality control in the laboratory, characterized in that, Establish independent quality control cycles for at least one preset quality control target concentration, with each cycle including the following steps: S1: Maintain an independent counting channel for each of the quality control target concentrations. When the cumulative number of samples tested for the target concentration reaches its dedicated current analysis batch length L, real-time quality control is automatically triggered. From the patient samples tested in the current analysis batch, two samples with concentrations that meet the target concentration requirements are intelligently selected and identified as a self-relay quality control sample pair. S2: The self-relay quality control samples are scheduled back to the detection unit for immediate retesting, the deviation between the first test result and the retest result of each sample is calculated, and the control cycle is determined to be in control or out of control according to the 1-3s quality control rule. S3: Based on the actual concentrations of the two self-relay samples determined in S1, query the preset concentration-bias mapping table and concentration-imprecision mapping table, calculate the performance value using the 6σ metric theory, and dynamically update the length L of the next analysis batch for the target concentration cycle accordingly. S4: After completing the quality control events (including S2 and S3), reset the count of the target concentration cycle and continue to run independently based on the new length L. Each target concentration cycle repeats S1 to S3 asynchronously and in parallel.

2. The sample self-relay method for real-time quality control in the laboratory as described in claim 1, characterized in that, The process of calculating performance values ​​using the 6σ metric theory and dynamically updating the length L of the next analytical batch for the target concentration cycle includes: Calculate the σ value for each of the two self-relay samples using the following formula: , where TEa is the total allowable error of the project, B is the bias obtained from the concentration-bias mapping table, and CV is the coefficient of variation obtained from the concentration-imprecision mapping table; Take the smaller of the two σ values ​​as ; Will Round to one decimal place. If the result is greater than 6, it will be rounded to 6. Using the final processed σ value, query the σ-L mapping table pre-set based on N=2 and the 1-3s rule to determine and update the next analysis batch length L.

3. The sample self-relay method for real-time quality control in the laboratory as described in claim 1, characterized in that, The concentration-bias mapping table is established based on historical interlaboratory quality assessment data, and the concentration-imprecision mapping table is established based on historical internal quality control data. It contains at least the standard deviation (SD) information at different concentration levels and supports linear interpolation calculation.

4. The sample self-relay method for real-time quality control in the laboratory as described in claim 1, characterized in that, The intelligent screening identifies two samples with concentrations that meet the target concentration requirements, specifically including: S1.1: Using the preset quality control target concentration as the center, set the first allowable concentration deviation range, and select samples whose concentration falls within this range from the tested samples in the current analytical batch to form a primary candidate list; S1.2: Select the two samples with the earliest detection time from the primary candidate list as pending candidates; If samples with the same detection time exist, they are selected from among them in ascending order of absolute deviation between their concentration and the target concentration. S1.3: If the number of specimens in the primary candidate list is less than two, expand the allowable concentration deviation range to form a new candidate list and repeat the screening logic of S1.

2. S1.4: For the candidate specimens identified by S1.2 or S1.3, verify whether they are still in an accessible location of the pipelined online storage unit through the pipeline specimen location system. If the candidate specimen is not in an accessible location of the pipelined online storage unit, remove it from the candidate list and return to S1.2 to select supplementary specimens from the remaining candidate specimens until two accessible specimens are identified.

5. The sample self-relay method for real-time quality control in the laboratory as described in claim 4, characterized in that, If the candidate specimen to be determined is not in an accessible location of the pipeline's online storage unit, then execute: S1.4.1: From the currently valid candidate list, select the next sample as a new candidate based on the priority of the detection time from earliest to latest; S1.4.2: If there are multiple candidates with the same detection time, select them from among them according to the priority of the absolute deviation between the concentration and the target concentration from small to large. S1.4.3: For the newly selected candidate specimens, re-execute the physical location verification in S1.4; S1.4.4: If two specimens in an accessible location cannot be determined after traversing the candidate list during the replacement process, an exception log is recorded and this quality control is skipped.

6. The sample self-relay method for real-time quality control in the laboratory as described in claim 1, characterized in that, The determination of the in-control or out-of-control state of the quality control cycle based on the 1-3s quality control rule includes: S2.1: Retrieve the cumulative standard deviation (SD) corresponding to the actual concentration of the self-relay sample or the concentration of the quality control target from the concentration-imprecision mapping table; S2.2: Set 3SD as the allowable deviation limit; S2.3: When the absolute value of the intra-batch deviation of any self-relay sample exceeds the allowable deviation limit, it is determined to be out of control.

7. An automated system for real-time quality control in a laboratory for performing the method of any one of claims 1-6, characterized in that, The system is integrated into the testing production line and includes: The dynamic analysis batch management module is used to maintain an independent counting channel for each quality control target concentration and to issue an independent trigger signal when the count reaches its analysis batch length L. The self-relay sample decision module is used to respond to trigger signals and select and lock two self-relay quality control samples from the current analysis batch for the corresponding target concentration. The real-time backtesting scheduling and execution module is used to control the pipeline to send locked sample pairs back to the detection unit for retesting; The controllable state determination module is used to calculate the intra-batch deviation of the sample and make a controllable / out-of-control determination based on the 1-3s rule; The adaptive frequency calculation module is used to calculate and update the next analysis batch length L for the corresponding target concentration cycle based on the locked sample data.

Images