Instrument out-of-control judgment method and system, electronic equipment and storage medium
By weighting the sample measurement values at the out-of-control level of different data models, comprehensively judging the out-of-control condition of the detection instrument corresponding to the sample measurement values, the problems of high false positive alarm rate and cumbersome judgment results in the prior art are solved, and the accuracy and sensitivity of out-of-control judgments are improved.
Patent Information
- Application Number
- CN202311491445.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-11-09
- Publication Date
- 2025-05-09
AI Technical Summary
In the prior art, when judging the loss of control of the instrument, the sample real-time quality control method has problems such as high false positive alarm rate, complicated judgment results, and conflicting results of different data models.
By obtaining the preset number of sample measurement values, at least two preset functional models are used to detect the sample measurement values in turn, the out-of-control score is calculated based on the control limit, the total out-of-control score is calculated based on the preset weight coefficient, and the detection instrument corresponding to the sample measurement value is comprehensively judged.
It improves the accuracy of out-of-control judgments, while improving the false positive alarm rate, taking into account sensitivity, and reducing conflicts in the judgment results of different data models.
Smart Images
Figure CN119959560A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of quality control of medical diagnosis, and in particular to a method, system, electronic device and storage medium for determining if an instrument is out of control. Background Art
[0002] In-house quality control is designed to monitor the precision of routine laboratory work and improve the consistency of sample testing within and between batches in routine laboratory work. Therefore, in-house quality control can help the laboratory test results to be accurate and reliable. However, due to the inability to continuously monitor quality, the matrix effect, and the high cost of quality control products, in-house quality control cannot accurately reflect the instrument detection performance in a timely manner.
[0003] Sample real-time quality control (PBRTQC) uses the results of patient clinical specimens to continuously monitor the stability of the detection system and determine out-of-control situations in real time without adding additional manpower and material costs. PBRTQC selects appropriate single or multiple data models to monitor sample values in real time based on the test items and historical results, and then different data models determine whether the sample values are out of control based on the judgment criteria. This processing method is cumbersome, and the judgment results of different data models often conflict with each other. Summary of the invention
[0004] In order to solve or partially solve the problems existing in the related art, the present application provides a method, system, electronic device and storage medium for judging instrument out of control, which can improve the false positive alarm rate while taking into account the sensitivity, thereby improving the accuracy of out of control judgment.
[0005] A first aspect of the present application provides a method for determining if an instrument is out of control, wherein the instrument is used to detect a sample and generate a sample measurement value, comprising:
[0006] Obtain a preset number of sample measurements;
[0007] At least two preset function models are used to detect the sample measurement values in sequence; the parameters of the function model include control limits;
[0008] Based on the control limit, sequentially calculate the out-of-control fraction of each sample measurement value in each function model;
[0009] Calculating the total out-of-control score of each sample measurement value in the function model in turn according to the preset weight coefficient and the out-of-control score;
[0010] Calculate the sum of the total out-of-control scores of the preset number of sample measurement values according to the total out-of-control score;
[0011] The possibility of out-of-control of the detection instrument corresponding to the sample measurement value is generated based on the sum of the out-of-control total scores.
[0012] Preferably, the method further comprises:
[0013] Determining the test items that the instrument is used to test the sample;
[0014] At least two of the function models are determined according to the detection items.
[0015] Preferably, the steps before obtaining a preset number of samples to be tested include:
[0016] Acquire historical data of the test items of the sample detected by the instrument;
[0017] Cleaning the historical data, and converting the cleaned historical data into normally distributed data;
[0018] The control limits corresponding to each of the function models are calculated according to the normal distribution data.
[0019] Preferably, the step of sequentially calculating the out-of-control fraction of each sample measurement value in each function model based on the control limit comprises:
[0020] For each of the function models, determining a control limit corresponding to the function model;
[0021] The out-of-control fraction of each of the sample measurements in the function model is calculated in sequence based on the control limit.
[0022] Preferably, the step of sequentially calculating the total out-of-control score of each sample measurement value in the function model according to the preset weight coefficient and the out-of-control score includes:
[0023] For each of the function models, determining a weight coefficient corresponding to the function model;
[0024] According to the weight coefficient corresponding to the function model, the out-of-control total score of each sample measurement value in the function model is calculated in turn.
[0025] Preferably, the step of generating the possibility of out-of-control of the detection instrument corresponding to the sample measurement value based on the sum of the out-of-control total scores comprises:
[0026] comparing the sum of the out-of-control total scores with a preset critical value;
[0027] The possibility of out-of-control of the detection instrument corresponding to the sample measurement value is generated according to the comparison result.
[0028] A second aspect of the present application provides a system for determining an instrument out of control, wherein the instrument is used to detect a sample and generate a sample measurement value, including:
[0029] A first acquisition module, used to acquire a preset number of sample measurement values;
[0030] A detection module, used to detect the sample measurement values in sequence using at least two preset function models; the parameters of the function model include control limits;
[0031] An out-of-control score module, used for sequentially calculating the out-of-control score of each sample measurement value in each function model based on the control limit;
[0032] An out-of-control total score module, used to sequentially calculate the out-of-control total score of each sample measurement value in the function model according to a preset weight coefficient and the out-of-control score;
[0033] An out-of-control total score sum module, used to calculate the sum of the out-of-control total scores of the preset number of sample measurement values according to the out-of-control total score;
[0034] A generating module is used to generate the possibility of out-of-control of the detection instrument corresponding to the sample measurement value based on the sum of the out-of-control total scores.
[0035] Preferably, the device further comprises:
[0036] A test item module, used to determine the test items that the instrument uses to test the sample;
[0037] The function model module is used to determine at least two function models according to the detection items.
[0038] Preferably, the device further comprises:
[0039] A second acquisition module is used to acquire historical data of the detection items of the sample detected by the instrument;
[0040] A conversion module, used for cleaning the historical data and converting the cleaned historical data into normal distribution data;
[0041] The control limit module is used to calculate the control limits corresponding to each of the function models according to the normal distribution data.
[0042] A third aspect of the present application provides an electronic device, including:
[0043] Processor; and
[0044] The memory stores executable codes thereon, and when the executable codes are executed by the processor, the processor is caused to execute the method as described above.
[0045] A fourth aspect of the present application provides a computer-readable storage medium having executable code stored thereon. When the executable code is executed by a processor of an electronic device, the processor is caused to execute the method as described above.
[0046] The technical solution provided by the present application may include the following beneficial effects: by obtaining a preset number of sample measurements, at least two preset function models are used to detect the sample measurements in turn; the parameters of the function model include control limits, based on the control limits, the out-of-control score of each sample measurement in each function model is calculated in turn, according to the preset weight coefficient and the out-of-control score, the total out-of-control score of each sample measurement in the function model is calculated in turn, the sum of the total out-of-control scores of the preset number of sample measurements is calculated according to the total out-of-control score, and the possibility of out-of-control of the detection instrument corresponding to the sample measurement is generated based on the sum of the total out-of-control scores, so that by weighting the out-of-control levels of the sample measurements in different data models, the out-of-control situation of the detection instrument corresponding to the sample measurement is comprehensively judged, which can improve the false positive alarm rate while taking into account the sensitivity, thereby improving the accuracy of out-of-control judgment.
[0047] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] The above and other objects, features and advantages of the present application will become more apparent by describing in more detail the exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.
[0049] Figure 1 It is a flow chart of a method for judging an instrument out of control shown in an embodiment of the present application;
[0050] Figure 2 is another flow chart of a method for determining if an instrument is out of control as shown in an embodiment of the present application;
[0051] Figure 3 is an optional function model shown in the embodiment of the present application;
[0052] Figure 4 is a schematic diagram of the detection results shown in the embodiment of the present application;
[0053] Figure 5 yes Figure 4 A partial enlarged schematic diagram of a part of the invention;
[0054] Figure 6 yes Figure 4 A partial enlarged schematic diagram of a part of the invention;
[0055] Figure 7 It is a structural schematic diagram of a system for judging an instrument out of control shown in an embodiment of the present application;
[0056] Figure 8 It is a schematic diagram of the structure of an electronic device shown in an embodiment of the present application. DETAILED DESCRIPTION
[0057] The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.
[0058] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.
[0059] It should be understood that although the terms "first", "second", "third", etc. may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.
[0060] In the related art, PBRTQC uses different data models. The method of judging whether the sample measurement value is out of control according to the judgment criteria is rather cumbersome, and the judgment results of different data models often conflict with each other, which may result in false positives.
[0061] In response to the above problems, an embodiment of the present application provides a method for determining if an instrument is out of control. The method can weight the out-of-control levels of samples in different data models and comprehensively determine the out-of-control situation of the samples. This method can improve the false positive alarm rate while taking into account sensitivity, thereby improving the accuracy of out-of-control judgment.
[0062] The technical solution of the embodiments of the present application is described in detail below with reference to the accompanying drawings.
[0063] Figure 1 It is a flow chart of a method for judging whether an instrument is out of control as shown in an embodiment of the present application.
[0064] See also Figure 1The instrument is used to detect the sample and generate the sample measurement value. The sample analysis instrument can be, for example, a chemiluminescence analyzer, a biochemical analyzer, a blood cell analyzer, a blood gas analyzer, a coagulation analyzer, a urine analyzer, a secretion analyzer, a blood glucose analyzer, etc. for analyzing body fluid samples. The method includes:
[0065] Step 101, obtaining a preset number of sample measurement values;
[0066] The preset number of sample measurements can be a plurality of sample measurements that are continuous in time. In the embodiment of the present application, the sensitivity and false alarm rate corresponding to different numbers of sample measurements when out of control can be calculated based on the data of the test instrument, so as to select the appropriate number of sample measurements. It can be understood by those skilled in the art that the larger the number of statistically weighted out-of-control sample measurements, the more stringent the out-of-control condition, the lower the false alarm rate, and the lower the sensitivity.
[0067] Step 102, using at least two preset function models to detect sample measurement values in sequence; the parameters of the function model include control limits;
[0068] Using a single model to detect sample measurements may result in false positives. In an embodiment of the present application, at least two different function models may be used to detect sample measurements. In one example, the function model may be a movemedian (moving median) model, a move yin-yang ratio model, a movemean (moving mean) model, a moveEMA (moving exponential moving average) model, a moveEWA (moving exponential weighted average) model, a movesd (moving standard deviation) model, a moveIQRM model, a stock model, etc. The appropriate function model may be selected according to the items to be detected and the way the data is processed.
[0069] Step 103, calculating the out-of-control fraction of each sample measurement value in each function model in sequence based on the control limit;
[0070] For each sample measurement, according to the control limits corresponding to each function model, the state of the sample measurement in each function model can be determined, and different values can be assigned to different states, so that the out-of-control fraction of each sample measurement in each function model can be calculated.
[0071] Step 104, calculating the total out-of-control score of each sample measurement value in the function model in turn according to the preset weight coefficient and the out-of-control score;
[0072] For different function models, different weight coefficients can be set for each function model, and the total out-of-control score of each sample value in different function models can be calculated according to the state of each sample value in the function model. The weight coefficient can be predetermined, and the sensitivity and false detection rate of different function models can be calculated based on a large amount of historical data, and the sensitivity and false detection rate can be combined through formulas, and then different weights can be given to different function models according to the results.
[0073] Step 105, calculating the sum of the total out-of-control scores of the preset number of sample measurements according to the total out-of-control scores;
[0074] After calculating the out-of-control total scores of all sample measurements, all out-of-control total scores are added together to obtain the sum of the out-of-control total scores of a preset number of sample measurements.
[0075] Step 106, generating the possibility of out-of-control of the detection instrument corresponding to the sample measurement value based on the sum of the out-of-control total scores.
[0076] Based on the sum of the out-of-control scores, the possibility of out-of-control of the testing instrument corresponding to the sample measurement value can be generated. It can be understood that the larger the sum of the out-of-control scores, the greater the possibility of out-of-control; the smaller the sum of the out-of-control scores, the smaller the possibility of out-of-control.
[0077] The embodiment of the present application discloses a method for judging an out-of-control of an instrument. The method obtains a preset number of sample measurements and uses at least two preset function models to detect the sample measurements in sequence. The parameters of the function model include control limits. The out-of-control score of each sample measurement in each function model is calculated in sequence based on the control limits. The total out-of-control score of each sample measurement in the function model is calculated in sequence according to the preset weight coefficient and the out-of-control score. The sum of the total out-of-control scores of the preset number of sample measurements is calculated according to the total out-of-control score. The possibility of out-of-control of the detection instrument corresponding to the sample measurement is generated based on the sum of the total out-of-control scores. Therefore, by weighting the out-of-control levels of the sample measurements in different data models, the out-of-control situation of the detection instrument corresponding to the sample measurement is comprehensively judged. This can improve the false positive alarm rate while taking into account the sensitivity, thereby improving the accuracy of out-of-control judgment.
[0078] Figure 2 This is another flow chart of a method for determining if an instrument is out of control as shown in an embodiment of the present application.
[0079] See also Figure 2 , the instrument is used to detect the sample to generate a sample measurement value, the method comprising:
[0080] Step 201, obtaining historical data of test items of test samples by the instrument;
[0081] For the detection items of the sample to be tested, continuous relevant historical data of the detection items within a period of time can be obtained, and the time can be determined according to the actual needs of technical personnel in this field.
[0082] Step 202, cleaning the historical data, and converting the cleaned historical data into normally distributed data;
[0083] After obtaining historical data, the data needs to be cleaned, and irrelevant data such as sample type, test submission time, test completion time, whether it is abnormal, whether to export data, etc. need to be deleted, experimental failure data need to be eliminated, and quality control data need to be deleted. Quality control data does not belong to sample data. Quality control data includes a low value and a high value. Before each test item is used for testing, the quality control data needs to be tested once. The test frequency is high. If it is included in the statistics, it will have a significant impact on the final calculation results, so it needs to be eliminated.
[0084] In the present application, the historical data stored in the database is equipped with different identifiers to mark the data, and data cleaning can be completed with the help of data identifiers.
[0085] After cleaning the data, determine whether the data is normally distributed. If not, convert it to a normal distribution so that the data as a whole is approximately normally distributed, thereby obtaining normally distributed data. In one example, the normal transformation methods include box-cox transformation, logarithmic transformation, inverse transformation, etc., which are not limited in the embodiments of the present application, and those skilled in the art can select one of the methods according to actual needs.
[0086] Step 203, calculating the control limits corresponding to each function model according to the normal distribution data.
[0087] Control limits refer to the control range specified when implementing quality control procedures for analytical tests, including upper and lower control limits. The reasonable selection of upper and lower control limits is of great significance to the detection results of the controlled equipment. If the upper and lower control limits are set too wide, the detection results may fall within the range of the upper and lower control limits, so that abnormal results cannot be detected, resulting in missed alarms; if the upper and lower control limits are set too narrow, normal results may also exceed the upper and lower control limits, causing false alarms and increasing the false alarm rate. In one example, the data of an instrument within one year was counted, and the historical records were checked, and a total of 9 instrument out-of-control cases were found. Referring to Table 1, the out-of-control detection rate and out-of-control false alarm rate of different weighted out-of-control judgment control limits can be compared.
[0088] Weighted out-of-control judgment control limits Out-of-control detection rate Out-of-control false alarm rate 14 56％ 0％ 12 78％ 0％ 10 100％ 0％ 8 100％ 22％ 6 100％ 56％
[0089] Table 1
[0090] After converting the historical data into normally distributed data, the control limits corresponding to each function model can be calculated, and the status of the sample measurement value in the function model can be determined based on the control limits.
[0091] Step 204, obtaining a preset number of sample measurement values;
[0092] The preset number of sample measurements can be a plurality of sample measurements that are continuous in time. In the embodiment of the present application, the sensitivity and false alarm rate corresponding to different numbers of sample measurements when out of control can be calculated based on the data of the test instrument, so as to select the appropriate number of sample measurements. It can be understood by those skilled in the art that the larger the number of statistically weighted out-of-control sample measurements, the more stringent the out-of-control condition, the lower the false alarm rate, and the lower the sensitivity.
[0093] In the present application, the collection time of the preset number of sample measurements may be completely unrelated to the collection time of the historical data in step 201. For example, the preset number of sample measurements are the sample measurements of the samples to be tested detected by the sample analyzer today. The historical data are the historical test data of the sample analyzer for 90 days starting from yesterday.
[0094] Step 205, using at least two preset function models to detect sample measurement values in sequence; the parameters of the function model include control limits;
[0095] Using a single model to detect sample measurements may result in false positives. In an embodiment of the present application, at least two different function models may be used to detect sample measurements. In one example, the function model may be a movemedian model, a move yin-yang ratio model, a movemean model, a moveEMA model, a moveEWA model, a movesd model, a moveIQRM model, a stock model, etc. The appropriate function model may be selected based on the items to be detected and the way the data is processed.
[0096] In an optional embodiment of the present application, the method further includes:
[0097] Step S11, determining the test items that the instrument is used to test the sample;
[0098] Different function models are usually used for different instruments and different test items of the instruments. For example, the FT4 project is more suitable for the moving mean model, the TSH project is more suitable for the moving median model, and the ATG project is more suitable for the moving standard deviation model. Therefore, you can first determine the test items used when the instrument detects the sample.
[0099] Step S12, determining at least two function models according to the detection items.
[0100] According to the determined test items, at least two function models required for the test sample measurement values can be determined. Figure 3As shown, it is a selectable function model. If the data does not need to be truncated in the subsequent processing process, the movemedian model or the move yin-yang ratio model can be selected. If the data needs to be truncated in the subsequent processing process, you can choose from the movemean model, moveEMA model, moveEWA model, movesd model, moveIQRM model, and stock model. In general, data that is too high or too low can be considered as abnormal data. Abnormal data will affect the accuracy of the control limits of subsequent calculations. Therefore, the abnormal data must be truncated first and removed from the training data set. In an embodiment of the present application, the data is truncated by truncation replacement, and the data above the upper truncation limit is replaced with the upper truncation limit, and the data below the lower truncation limit is replaced with the lower truncation limit. Among them, the functions of the moveEMA model and the moveEWA model are:
[0101] Y t =k*y t +(1-k)*Y (t-1)
[0102] In the moveEMA model, the moving weighting coefficient k = 2 / (n+1), where n is the number of samples to be tested; in the moveEWA model, the moving weighting coefficient k = 0.1 / 0.02...
[0103] The function of the moveIQRM model is:
[0104] Y t = Upper and lower quartile difference / median In one example, the historical data of the FT4 test item of a certain instrument is normally distributed. The movemean model is selected, where the cutoff limits are 12 and 22, and the mean (mean) ± 3 SD (standard deviation) of the historical data movemean results are used as control limits, and the sliding window N is selected as 15, 30, and 60. Among them, the parameters of the model are built-in by default and can also be adjusted manually. The default parameters are derived from the pre-debugging results.
[0105] In one example, the historical data of the FT4 test project of a certain instrument is normally distributed. The movemean model, movemedian model, and moveEMA model are selected, and the cutoff limits are 12 and 22. The mean±3 SD of the historical data movemean result is used as the control limit, and the sliding window is 30. Among them, the parameters of the model are built-in by default and can also be adjusted manually. The default parameters are derived from the pre-debugging results.
[0106] In one example, the historical data of the TSH test project of a certain instrument is skewed. The historical data is first transformed into normal distribution data by box-cox transformation, and the movemean model is selected. The cutoff limits are 0.4 and 4, and the mean±3 SD of the historical data movemean results are used as control limits. The sliding window is selected as 50, 100, and 200. Among them, the parameters of the model are built-in by default and can also be adjusted manually. The default parameters are derived from the pre-debugging results.
[0107] Step 206, for each function model, determining the control limit corresponding to the function model;
[0108] It is understandable that the control limits corresponding to different function models may be different. Therefore, it is necessary to determine the control limits corresponding to each function model.
[0109] Step 207, calculating the out-of-control score of each sample measurement value in the function model based on the control limit.
[0110] According to the control limit of the function model and the sample measurement value, the out-of-control score of each sample measurement value in different function models can be calculated. If the sample measurement value exceeds the control limit corresponding to the function model, it can be judged that the sample measurement value is out of control in the function model, and the out-of-control score in the function model is recorded as 1; if the sample measurement value does not exceed the control limit corresponding to the function model, it can be judged that the sample measurement value is not out of control in the function model, and the out-of-control score in the function model is recorded as 0. During the calculation, the detection result of the current sample and the N-1 historical detection results closest to the current moment are selected through the sliding window N as a batch of real-time detection data to calculate the sample measurement value of the sample to be tested. When the sliding window is relatively small, the response is more sensitive, but it is easily disturbed by the sample distribution; when the sliding window is relatively large, it is not easily disturbed by the sample distribution, but the response is not sensitive, so the sliding window can be set based on the pre-debugging results.
[0111] Step 208, for each function model, determining a weight coefficient corresponding to the function model;
[0112] The weight coefficient can be predetermined, and the sensitivity and false detection rate of different function models are calculated based on a large amount of historical data, the sensitivity and false detection rate are combined through a formula, and then different weights are given to different functions based on the results. It is understandable that different function models have different weight coefficients in different detection items.
[0113] Step 209, according to the weight coefficient corresponding to the function model, calculate the total out-of-control score of each sample measurement value in the function model in turn.
[0114] When calculating the total out-of-control score of each sample measurement, it is necessary to consider the weight coefficient corresponding to the function model, multiply the out-of-control score of each sample measurement in each function model by the weight coefficient, and add up the results of the multiplication of each function model to get the total out-of-control score of each sample measurement in the function model.
[0115] In one example, for example, the FT4 detection project of a certain instrument, the movemean model, movemedian model, and moveEMA model are selected, and a sample measurement value is calculated respectively. If the sample measurement value exceeds the control limit corresponding to the model, the sample to be tested is judged to be out of control in this model, and out of control is recorded as 1, and no out of control is recorded as 0. Then the out-of-control results of different models are weighted. The weight coefficients of the movemean model, movemedian model, and moveEMA model in the FT4 detection project are set to 1, 1, and 1 respectively. If the sample measurement value is out of control in all three models, the total out-of-control score is recorded as 3; if the sample measurement value is out of control in only 2 models, the total out-of-control score is recorded as 2; if the sample measurement value is out of control in only 1 model, the total out-of-control score is recorded as 1; if the sample measurement value is out of control in 0 models, the total out-of-control score is recorded as 0. It can be understood by those skilled in the art that in this example, the value of the weight coefficient and the choice of the function model are only used as an example, and this application does not limit this.
[0116] Step 210, calculating the sum of the total out-of-control scores of the preset number of sample measurements according to the total out-of-control scores;
[0117] After calculating the total out-of-control score of each sample measurement value, the sum of the total out-of-control scores of a preset number of sample measurement values may be calculated.
[0118] In one example, five consecutive sample measurements may be selected, the out-of-control total score of each sample measurement may be calculated respectively, and then the out-of-control total scores of the five sample measurements may be added together to obtain the sum of the out-of-control total scores.
[0119] Step 211, comparing the sum of the out-of-control total score and the preset critical value;
[0120] It is understandable that the greater the sum of the out-of-control scores, the greater the possibility of out-of-control; the smaller the sum of the out-of-control scores, the smaller the possibility of out-of-control. Therefore, a critical value can be set to judge the possibility of out-of-control, and compare the sum of the out-of-control scores with the critical value. The critical value is set based on the pre-debugging results, and the appropriate value is selected by comparing the sensitivity and false alarm rate.
[0121] Step 2012, generating the possibility of the detection instrument corresponding to the sample measurement value being out of control according to the comparison result.
[0122] According to the comparison result of the sum of out-of-control scores and the critical value, the possibility of out-of-control of the detection instrument corresponding to the generated sample measurement value can be judged. In one example, for the FT4 detection project of a certain instrument, the movemean model, movemedian model, and moveEMA model are selected, and the function model weight coefficients are set to 1, 1, and 1 respectively. Five consecutive sample measurements are selected, and the critical values can be set to 5 and 10. If the sum of the total out-of-control scores of the sample measurements is greater than or equal to 10, it is considered that the detection instrument is likely to be out of control; if the sum of the total out-of-control scores of the sample measurements is greater than or equal to 5 and less than 10, it is considered that the detection instrument is likely to be out of control; if the sum of the total out-of-control scores of the sample measurements is less than 5, it is considered that the detection instrument is likely to be out of control.
[0123] In one example, if Figure 4 As shown, it is a schematic diagram of the detection results of the detection method in the embodiment of the present application. In this example, the FT4 detection project of a certain instrument is used, and the movemean model, movemedian model, and moveEMA model are selected to perform multi-model joint detection on 2012 batches of samples to be tested. The detection results can be displayed in the results of any function model. Figure 4 The real-time quality control results of the samples in the circled parts 401 and 402 exceed the lower control limit. If the common out-of-control judgment method is used, it can be judged as out-of-control. If this method is used, it is considered that the possibility of out-of-control is small. After verification, the instrument is normal on that day, the quality control product is under control, and there is no out-of-control situation. Therefore, the detection method in the embodiment of this application can improve the accuracy of out-of-control judgment.
[0124] Reference Figure 5 ,for Figure 4 The middle circle 401 is a partial enlarged schematic diagram.
[0125] Reference Figure 6 ,for Figure 4 The middle circle 402 shows a partial enlarged schematic diagram.
[0126] The embodiment of the present application discloses a method for judging an out-of-control of an instrument, which comprises the following steps: obtaining historical data of a detection item of a sample detected by the instrument, cleaning the historical data, converting the cleaned historical data into normal distribution data, calculating control limits corresponding to each function model according to the normal distribution data, and obtaining a preset number of sample measurements; using at least two preset function models to detect the sample measurements in sequence; the parameters of the function model include control limits, for each function model, determining the control limits corresponding to the function model, calculating the out-of-control score of each sample measurement value in the function model in sequence based on the control limits, for each function model, determining the weight coefficient corresponding to the function model, calculating the total out-of-control score of each sample measurement value in the function model in sequence according to the weight coefficient corresponding to the function model, calculating the sum of the total out-of-control scores of a preset number of sample measurements according to the total out-of-control scores, comparing the sum of the total out-of-control scores and a preset critical value, and generating the possibility of out-of-control of the detection instrument corresponding to the sample measurement value according to the comparison result, thereby weighting the out-of-control levels of the sample measurement values in different data models, and comprehensively judging the out-of-control situation of the detection instrument corresponding to the sample measurement value, which can improve the false positive alarm rate while taking into account the sensitivity, thereby improving the accuracy of out-of-control judgment.
[0127] Corresponding to the aforementioned application function implementation method embodiment, the present application also provides a system for determining if an instrument is out of control, an electronic device, and corresponding embodiments.
[0128] Figure 7 It is a structural schematic diagram of a system for determining if an instrument is out of control as shown in an embodiment of the present application.
[0129] See also Figure 7 The instrument is used to detect the sample and generate the sample measurement value. The system includes:
[0130] The first acquisition module 701 is used to acquire a preset number of sample measurement values;
[0131] The preset number of sample measurements may be a plurality of sample measurements that are continuous in time. In the embodiment of the present application, the sensitivity and false alarm rate corresponding to different numbers of sample measurements out of control may be calculated based on the data of the test instrument, thereby selecting an appropriate number of samples.
[0132] The detection module 702 uses at least two preset function models to detect the sample measurement values in sequence; the parameters of the function model include control limits;
[0133] In the embodiment of the present application, at least two different function models may be used to detect sample measurement values.
[0134] An out-of-control score module 703 is used to sequentially calculate the out-of-control score of each sample measurement value in each function model based on the control limit;
[0135] For each sample measurement, according to the control limits corresponding to each function model, the state of the sample measurement in each function model can be determined, and different values can be assigned to different states, so that the out-of-control fraction of each sample measurement in each function model can be calculated.
[0136] The out-of-control total score module 704 is used to calculate the out-of-control total score of each sample measurement value in the function model in turn according to the preset weight coefficient and the out-of-control score;
[0137] For different function models, different weight coefficients can be set for each function model, and the total out-of-control score of each sample measurement value in different function models can be calculated according to the state of each sample measurement value in the function model.
[0138] The out-of-control total score sum module 705 is used to calculate the sum of the out-of-control total scores of a preset number of sample measurement values according to the out-of-control total score;
[0139] After calculating the out-of-control total scores of all sample measurements, all out-of-control total scores are added together to obtain the sum of the out-of-control total scores of a preset number of sample measurements.
[0140] The generating module 706 is used to generate the possibility of the detection instrument being out of control corresponding to the sample measurement value based on the sum of the out-of-control total scores.
[0141] Based on the sum of the total out-of-control scores, the possibility of the detection instrument corresponding to the sample measurement value can be generated. It can be understood that the larger the sum of the out-of-control scores, the greater the possibility of out-of-control; the smaller the sum of the out-of-control scores, the smaller the possibility of out-of-control.
[0142] In an optional embodiment of the present application, the system further includes:
[0143] A test item module is used to determine the test items that the instrument uses to test the sample;
[0144] The function model module is used to determine at least two function models according to the detection items.
[0145] In an optional embodiment of the present application, the system further includes:
[0146] The second acquisition module is used to acquire historical data of the test items of the instrument test samples;
[0147] For the detection items of the samples to be tested, continuous relevant historical data of the detection items within a period of time can be obtained, and the time can be determined according to the actual needs of technical personnel in this field. In the present application, the collection time of the preset number of sample measurement values of the samples to be tested may be completely unrelated to the collection time of the historical data. For example, the preset number of sample measurement values of the samples to be tested are the sample measurement values of the samples to be tested detected by the sample analyzer today. The historical data is the historical detection data of the sample analyzer for 90 days starting from yesterday.
[0148] A conversion module is used to clean the historical data and convert the cleaned historical data into normal distribution data;
[0149] After obtaining historical data, it is necessary to clean the data. After cleaning the data, determine whether the data is normally distributed. If it is not normally distributed, convert it to a normally distributed state so that the data as a whole is approximately normally distributed, thereby obtaining normally distributed data. The historical data stored in the database is equipped with different identifiers to mark the data, and data cleaning can be completed with the help of data identifiers.
[0150] The control limit module is used to calculate the control limits corresponding to each function model based on the normal distribution data.
[0151] After converting the historical data into normal distribution data, the control limits corresponding to each function model can be calculated, and the sample measurement values can be detected based on the control limits.
[0152] In an optional embodiment of the present application, the out-of-control score module 703 includes:
[0153] A first determination submodule is used to determine the control limit corresponding to each function model;
[0154] The out-of-control score submodule is used to calculate the out-of-control score of each sample measurement value in the function model based on the control limit.
[0155] In an optional embodiment of the present application, the out-of-control total score module 704 includes:
[0156] The second determination submodule is used to determine the weight coefficient corresponding to each function model;
[0157] The out-of-control total numerator module is used to calculate the out-of-control total score of each sample measurement value in the function model in turn according to the weight coefficient corresponding to the function model.
[0158] In an optional embodiment of the present application, the generating module 706 includes:
[0159] A comparison submodule, used to compare the sum of the out-of-control total score and the preset critical value;
[0160] The generation submodule is used to generate the possibility of the detection instrument corresponding to the sample measurement value being out of control according to the comparison result.
[0161] The embodiment of the present application discloses a system for judging whether an instrument is out of control. The system obtains a preset number of sample measurements and uses at least two preset function models to detect the sample measurements in sequence. The parameters of the function model include control limits. The out-of-control score of each sample measurement in each function model is calculated in sequence based on the control limits. The total out-of-control score of each sample measurement in the function model is calculated in sequence according to the preset weight coefficient and the out-of-control score. The sum of the total out-of-control scores of the preset number of sample measurements is calculated according to the total out-of-control score. The possibility of out-of-control of the detection instrument corresponding to the sample measurement is generated based on the sum of the total out-of-control scores. Therefore, by weighting the out-of-control levels of the sample measurements in different data models, the out-of-control situation of the detection instrument corresponding to the sample measurement is comprehensively judged. This can improve the false positive alarm rate while taking into account the sensitivity, thereby improving the accuracy of out-of-control judgment.
[0162] Regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated again here.
[0163] Figure 8 It is a schematic diagram of the structure of an electronic device shown in an embodiment of the present application.
[0164] See also Figure 8 , the electronic device 800 includes a memory 810 and a processor 820 .
[0165] The processor 820 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.
[0166] The memory 810 may include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage. Among them, ROM can store static data or instructions required by the processor 820 or other modules of the computer. The permanent storage device may be a readable and writable storage device. The permanent storage device may be a non-volatile storage device that does not lose the stored instructions and data even after the computer is powered off. In some embodiments, the permanent storage device uses a large-capacity storage device (such as a magnetic or optical disk, flash memory) as a permanent storage device. In some other embodiments, the permanent storage device may be a removable storage device (such as a floppy disk, optical drive). The system memory may be a readable and writable storage device or a volatile readable and writable storage device, such as a dynamic random access memory. The system memory may store some or all instructions and data required by the processor at run time. In addition, the memory 810 may include any combination of computer-readable storage media, including various types of semiconductor storage chips (such as DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), and disks and / or optical disks may also be used. In some embodiments, the memory 810 may include a readable and / or writable removable storage device, such as a laser disc (CD), a read-only digital versatile disc (e.g., DVD-ROM, double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (e.g., SD card, mini SD card, Micro-SD card, etc.), a magnetic floppy disk, etc. The computer-readable storage medium does not include carrier waves and transient electronic signals transmitted wirelessly or wired.
[0167] The memory 810 stores executable codes, and when the executable codes are processed by the processor 820 , the processor 820 can execute part or all of the methods described above.
[0168] In addition, the method according to the present application may also be implemented as a computer program or a computer program product, which includes computer program code instructions for executing some or all of the steps in the above method of the present application.
[0169] Alternatively, the present application can also be implemented as a computer-readable storage medium (or non-transitory machine-readable storage medium or machine-readable storage medium) on which executable code (or computer program or computer instruction code) is stored. When the executable code (or computer program or computer instruction code) is executed by a processor of an electronic device (or server, etc.), the processor executes part or all of the steps of the above-mentioned method according to the present application.
[0170] The embodiments of the present application have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.
Claims
1. A method for determining if an instrument is out of control, characterized in that: The instrument is used to detect a sample and generate a sample measurement value, and the method includes: Obtain a preset number of sample measurements; At least two preset function models are used to detect the sample measurement values in sequence; the parameters of the function model include control limits; Based on the control limit, sequentially calculate the out-of-control fraction of each sample measurement value in each function model; Calculating the total out-of-control score of each sample measurement value in the function model in turn according to the preset weight coefficient and the out-of-control score; Calculate the sum of the total out-of-control scores of the preset number of sample measurement values according to the total out-of-control score; The possibility of out-of-control of the detection instrument corresponding to the sample measurement value is generated based on the sum of the out-of-control total scores.
2. The method according to claim 1, characterized in that The method further comprises: Determining the test items that the instrument is used to test the sample; At least two of the function models are determined according to the detection items.
3. The method according to claim 2, characterized in that The steps before obtaining a preset number of sample measurement values include: Acquire historical data of the test items of the sample detected by the instrument; Cleaning the historical data, and converting the cleaned historical data into normally distributed data; The control limits corresponding to each of the function models are calculated according to the normal distribution data.
4. The method according to claim 1, characterized in that The step of sequentially calculating the out-of-control fraction of each sample measurement value in each function model based on the control limit comprises: For each of the function models, determining a control limit corresponding to the function model; The out-of-control fraction of each of the sample measurements in the function model is calculated in sequence based on the control limit.
5. The method according to claim 1, characterized in that The step of sequentially calculating the total out-of-control score of each sample measurement value in the function model according to the preset weight coefficient and the out-of-control score includes: For each of the function models, determining a weight coefficient corresponding to the function model; According to the weight coefficient corresponding to the function model, the out-of-control total score of each sample measurement value in the function model is calculated in turn.
6. The method according to claim 1, characterized in that The step of generating the possibility of out-of-control of the detection instrument corresponding to the sample measurement value based on the sum of the out-of-control total score comprises: comparing the sum of the out-of-control total scores with a preset critical value; The possibility of out-of-control of the detection instrument corresponding to the sample measurement value is generated according to the comparison result.
7. A system for judging an instrument out of control, characterized in that: The instrument is used to detect a sample and generate a sample measurement value. The system includes: A first acquisition module, used to acquire a preset number of sample measurement values; A detection module, used to detect the sample measurement values in sequence using at least two preset function models; the parameters of the function model include control limits; An out-of-control score module, used for sequentially calculating the out-of-control score of each sample measurement value in each function model based on the control limit; An out-of-control total score module, used to sequentially calculate the out-of-control total score of each sample measurement value in the function model according to a preset weight coefficient and the out-of-control score; An out-of-control total score sum module, used to calculate the sum of the out-of-control total scores of the preset number of sample measurement values according to the out-of-control total score; A generating module is used to generate the possibility of out-of-control of the detection instrument corresponding to the sample measurement value based on the sum of the out-of-control total scores.
8. The system according to claim 7, characterized in that The system further comprises: A test item module, used to determine the test items that the instrument uses to test the sample; The function model module is used to determine at least two function models according to the detection items.
9. An electronic device, characterized in that: include: processor; as well as A memory having executable codes stored thereon, which, when executed by the processor, causes the processor to execute the method according to any one of claims 1 to 6.
10. A computer-readable storage medium having executable codes stored thereon, which, when executed by a processor of an electronic device, causes the processor to execute the method according to any one of claims 1 to 6.