Detection mechanism detection quality determination method based on combination of clustering and data calibration

Abstract

The invention provides a method for determining the detection quality of a detection mechanism based on combination of clustering and data calibration, and can be applied to the technical field of medical treatment and public health. The method comprises the steps of obtaining a medical detection data set; integrating the medical detection data belonging to the same detection sample in the plurality of medical detection data to obtain sample detection data corresponding to each detection sample; clustering the sample detection data corresponding to the plurality of detection samples to obtain a plurality of data clusters, and removing outlier data clusters from the plurality of data clusters to obtain a plurality of target data clusters; for the same detection item, performing data calibration processing on the multiple pieces of target medical detection data corresponding to the detection item to obtain multiple target detection results; and determining the detection quality corresponding to the target detection mechanisms according to the target detection results. According to the method, the adopted patient data fits the medical and health scene, and the accuracy of the determined detection quality is relatively high by clustering and calibrating the data.

Description

Technical Field

[0001] This application relates to the field of medical and health technologies, and particularly to a method for determining the detection quality of testing institutions based on the combination of clustering and data calibration. Background Art

[0002] External quality assessment (EQA) is a process in which multiple laboratories analyze the same sample, and an external independent agency collects and provides feedback on the test results reported by the laboratories to evaluate the detection quality of the laboratories. External quality assessment (referred to as "EQA" for short) is an important means to achieve the mutual recognition of test results and an objective evidence for evaluating the detection quality. The purpose of EQA is to evaluate the accuracy of laboratory test results and establish the comparability of results between laboratories. In the field of medical and health technologies, EQA technology is used to evaluate the detection quality of each medical testing institution to be evaluated, so as to provide a basis for whether the test results of each medical testing institution can be mutually recognized.

[0003] The current EQA uses EQA samples for testing. EQA samples are generally the same sample manually prepared by an external independent agency. However, in the medical and health scenario, the samples tested by medical testing institutions are real samples taken from the test sites of patients, and such samples will vary according to different patients, so there are some differences between them and EQA samples. Therefore, the traditional method of using EQA samples to evaluate the detection quality of medical testing institutions has limitations. Summary of the Invention

[0004] This application provides a method for determining the detection quality of testing institutions based on the combination of clustering and data calibration. The accuracy of the determined detection quality is relatively high and has great reference significance, and thus can accurately provide an effective basis for whether the test results of each medical testing institution can be mutually recognized.

[0005] In a first aspect, this application provides a method for determining the detection quality of testing institutions based on the combination of clustering and data calibration, including:

[0006] Obtain a medical test data set; the medical test data set includes medical test data of multiple testing institutions, the medical test data is obtained after testing the corresponding test samples, the test samples are obtained after sampling the corresponding patient test sites, and the medical test data includes test items, test results, and test basic information; the test basic information includes at least one of information for describing the test process and patient visit information.

[0007] Integrate the medical test data belonging to the same test sample among multiple medical test data to obtain the sample test data corresponding to each test sample;

[0008] Cluster the sample test data corresponding to the multiple test samples to obtain multiple data clusters, and remove the outlier data clusters from the multiple data clusters to obtain multiple target data clusters;

[0009] For the same test item, perform data calibration processing on the multiple target medical test data corresponding to the test item to obtain multiple target test results; the target medical test data are the data in the target data clusters;

[0010] Determine the detection quality corresponding to multiple target detection institutions according to the multiple target test results; the target detection institutions are the detection institutions that provide the target medical test data.

[0011] Optionally, for the same test item, performing data calibration processing on the test results in the multiple target medical test data corresponding to the test item to obtain multiple target test results includes:

[0012] Perform the following operations for the same test item:

[0013] Construct a multiple regression model corresponding to the test item with the test results in the target medical test data corresponding to the test item as the dependent variable and the basic test information in the target medical test data as the independent variable;

[0014] Use the multiple regression model corresponding to the test item to perform data calibration processing on the test results in the target medical test data corresponding to the test item to obtain multiple target test results.

[0015] Optionally, using the multiple regression model corresponding to the test item to perform data calibration processing on the test results in the target medical test data corresponding to the test item to obtain multiple target test results includes:

[0016] Perform the following operations for each of the target medical test data corresponding to the test item:

[0017] Input the basic test information in the medical test data into the multiple regression model and output the fitting result corresponding to the basic test information;

[0018] Determine the difference between the test result in the target medical test data and the fitting result as the target test result.

[0019] Optionally, after obtaining the medical test data set, the method further includes:

[0020] Perform data structuring and data normalization on each of the medical test data to obtain multiple processed medical test data;

[0021] From multiple processed medical test data for the same test item, eliminate the medical test data with outlier test results to obtain multiple remaining medical test data.

[0022] Optionally, clustering the sample test data respectively corresponding to the multiple test samples to obtain multiple data clusters, including:

[0023] Taking one sample test data as one sample point, respectively plot each of the sample test data in a high-dimensional space; the dimension of the high-dimensional space is equal to the data dimension of the sample test data;

[0024] Using a preset clustering algorithm, cluster the sample test data in the high-dimensional space; the preset clustering algorithm is a density-based clustering algorithm;

[0025] Determine the region with the highest sample density in the high-dimensional space, label the sample test data within the region as the same data cluster, and remove the samples in the region from the high-dimensional space;

[0026] Using the preset clustering algorithm, cluster the remaining sample test data in the high-dimensional space;

[0027] And so on, until the remaining sample test data in the high-dimensional space accounts for a preset percentage of the total number of sample test data, terminate the iterative process to obtain multiple data clusters.

[0028] Optionally, the eliminating the outlier data clusters from the multiple data clusters to obtain multiple target data clusters includes:

[0029] Determine the distances between each of the data clusters and other data clusters;

[0030] From the multiple data clusters, eliminate the data cluster with the largest sum of distances to other data clusters to obtain the multiple target data clusters.

[0031] Optionally, the determining the detection quality respectively corresponding to the multiple target detection institutions according to the test results respectively included in the multiple target medical test data includes:

[0032] For the same test item, according to the multiple test results respectively corresponding to each of the target detection institutions, determine the detection reference target values and preset offset percentages respectively corresponding to at least one quantile;

[0033] Determine the detection benchmark numerical range corresponding to each of the at least one quantile according to the detection benchmark target value and the preset offset percentage corresponding to each of the at least one quantile;

[0034] For the same detection item, determine the detection quality score of each target detection agency in the corresponding detection item according to the detection benchmark numerical range corresponding to each of the at least one quantile and the multiple detection results corresponding to each target detection agency.

[0035] Optionally, the determining the detection quality score of each target detection agency in the corresponding detection item according to the detection benchmark numerical range corresponding to each of the at least one quantile and the multiple detection results corresponding to each target detection agency includes:

[0036] For the same detection item, obtain the quantile value of each target detection agency at each quantile according to the multiple detection results corresponding to each target detection agency;

[0037] For each quantile, if it is determined that the quantile value of the target detection agency at the quantile is within the detection benchmark numerical range corresponding to the quantile, determine that the score of the target detection agency at the quantile is the first score;

[0038] If it is determined that the quantile value of the target detection agency at the quantile is not within the detection benchmark numerical range corresponding to the quantile, determine that the score of the target detection agency at the quantile is the second score; the first score is greater than the second score;

[0039] Determine the sum of the scores of the target detection agency at each quantile as the detection quality score of the target detection agency.

[0040] In a second aspect, the present application provides a detection agency detection quality determination device based on the combination of clustering and data calibration, including:

[0041] An acquisition module, configured to acquire a medical detection data set; the medical detection data set includes a plurality of medical detection data, the medical detection data is obtained after detecting a corresponding detection sample, the detection sample is obtained after sampling a detection part of a corresponding patient, and the medical detection data includes a detection item, a detection result, and detection basic information; the detection basic information includes at least one of information for describing the detection process and patient visit information;

[0042] An integration module, configured to integrate the medical detection data belonging to the same detection sample among the plurality of medical detection data to obtain sample detection data corresponding to the plurality of detection samples respectively;

[0043] A clustering module, configured to cluster the sample detection data respectively corresponding to the multiple detection samples to obtain multiple data clusters, and remove the outlier data clusters from the multiple data clusters to obtain multiple target data clusters;

[0044] A calibration module, configured to perform data calibration processing on the detection results in the multiple target medical detection data corresponding to the same detection item to obtain multiple target detection results; the target medical detection data are the data in the target data clusters;

[0045] A determination module, configured to determine the detection quality respectively corresponding to multiple target detection institutions according to the multiple target detection results; the target detection institutions are the detection institutions that provide the target medical detection data.

[0046] In a third aspect, the present application provides an electronic device, including: a processor and a memory communicatively connected to the processor;

[0047] The memory stores computer-executable instructions;

[0048] The processor executes the computer-executable instructions stored in the memory to implement the method for determining the detection quality of a detection institution based on the combination of clustering and data calibration as described in any item of the first aspect.

[0049] In a fourth aspect, the present application provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, they are used to implement the method for determining the detection quality of a detection institution based on the combination of clustering and data calibration as described in any item of the first aspect.

[0050] In a fifth aspect, the present application provides a computer program product, including computer-executable instructions, and when the computer-executable instructions are executed by a processor, they implement the method for determining the detection quality of a detection institution based on the combination of clustering and data calibration as described in any item of the first aspect.

[0051] The method for determining the detection quality of a testing institution based on the combination of clustering and data calibration provided by this application uses a medical testing dataset as the data source. Since the medical testing data in the medical testing dataset is obtained by the testing institution testing the test samples, and the test samples are taken from the test sites of patients, unlike the external quality assessment samples used in related technologies which are artificially configured, the medical testing data is real patient data, which is more in line with the real medical and health scenarios. On this basis, by integrating the medical testing data, sample testing data is obtained, and then, taking the test samples as the main body, the sample testing data is clustered to obtain multiple data clusters. As a preliminary diagnosis, the sample testing data with the same disease tendency or population characteristics is gathered together. By removing the outlier data clusters, the automatic screening of the data is realized, making the data concentrated and stable. On this basis, by performing data calibration on the test results in the target medical testing data, the test results corresponding to each target testing institution can reach the same level. The accuracy of the detection quality corresponding to each target testing institution determined according to the target test results is relatively high and has great reference significance. Furthermore, it can accurately provide an effective basis for whether the test results can be mutually recognized among medical testing institutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The drawings herein are incorporated into the specification and form a part of this specification, showing embodiments consistent with this application, and are used together with the specification to explain the principles of this application.

[0053] Figure 1 is a schematic diagram of an application scenario provided by this application;

[0054] Figure 2 is a schematic flowchart of a method for determining the detection quality of a testing institution based on the combination of clustering and data calibration provided by this application;

[0055] Figure 3 is a schematic flowchart of another method for determining the detection quality of a testing institution based on the combination of clustering and data calibration provided by this application;

[0056] Figure 4 is a schematic structural diagram of a device for determining the detection quality of a testing institution based on the combination of clustering and data calibration provided by this application;

[0057] Figure 5 is a schematic structural diagram of an electronic device provided by this application.

[0058] Through the above drawings, the clear embodiments of this application have been shown, and there will be more detailed descriptions later. These drawings and text descriptions are not intended to limit the scope of the concept of this application in any way, but to explain the concept of this application to those skilled in the art by referring to specific embodiments. Detailed implementation manners

[0059] Here, exemplary embodiments will be described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0060] Terms such as "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. In the description of the following embodiments, "a plurality of" means two or more unless otherwise specifically defined.

[0061] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use, and processing of relevant data need to comply with relevant laws, regulations, and standards, and corresponding operation entrances are provided for users to select authorization or rejection.

[0062] To clearly understand the technical solution of the present application, the solutions of the prior art will be introduced in detail first.

[0063] External quality assessment (EQA for short), also known as inter-laboratory quality assessment, is a process in which multiple laboratories analyze the same sample, and an external independent agency collects and feeds back the test results reported by the laboratories to evaluate the test quality of the laboratories. External quality assessment (abbreviated as "EQA") is an important means to achieve the mutual recognition of test results and an objective evidence for evaluating test quality. The purpose of EQA is to evaluate the accuracy of laboratory test results and establish the comparability of results among laboratories. In the field of medical and health technologies, EQA technology is used to evaluate the test quality of each medical testing institution to be evaluated, so as to provide a basis for whether the test results can be mutually recognized among medical testing institutions.

[0064] The current EQA uses EQA samples for testing. EQA samples are generally the same sample artificially prepared by an external independent agency. However, in the medical and health scenario, the samples tested by medical testing institutions are real samples taken from the test sites of patients, and such samples will vary according to different patients.

[0065] Among them, the external quality assessment samples are not real patient samples, which can only simulate the characteristics of patient samples. There may be matrix effects that affect the results of external quality assessment, and there is a lack of direct correlation with clinical impacts. The interchangeability of external quality assessment samples themselves is poor. Moreover, due to the selection of external quality assessment samples, which involves the types, concentration levels, quality assessment frequencies, result judgment rules, etc. of the selected external quality assessment samples, not only are there differences between external quality assessment samples and patient samples, but there are also limitations in the traditional method of using external quality assessment samples to evaluate the detection quality of medical testing institutions. In addition, due to the influence of subjective factors of laboratory personnel on technical parameter settings, non-standardized evaluation criteria, instrument calibration, and method selection in this traditional method, to a certain extent, the quality assessment results cannot reflect the true capabilities of the laboratory, affecting the objectivity of the external quality assessment results.

[0066] In view of the technical problems in the prior art, the inventors found in their research that in the medical and health scenarios, medical testing institutions will test a large number of test samples obtained by sampling from the patient's testing sites to obtain corresponding test results. To solve this problem, the test samples can be used to replace the external quality assessment samples. Using real patient data as the basis, it can be directly related to the patient's clinical status and is not affected by matrix effects. In this way, there is no need to manually configure external quality assessment samples, and the detection quality of the medical testing institution itself can also be determined based on the test results obtained by the medical testing institution.

[0067] On this basis, the inventors considered that the data volume of the test samples is large, and the test samples are affected by factors such as the patient himself, the disease, the season, and the testing means, resulting in large differences between the test samples. Then, the differences between the obtained test results are also large. If a large number of test results with large differences are directly used for the evaluation of detection quality, the accuracy may not be high. The inventors' research found that the data can be processed by clustering, so as to automatically screen the data, making the screened data tend to be stable and concentrated, and being more able to represent the true detection situation of medical testing institutions. Moreover, considering the large differences between the test results obtained by different medical testing institutions, the data can also be processed by calibration, so that the test results of each medical testing institution are on the same level line, and the accuracy of the determined detection quality based on this will also be relatively high.

[0068] Specifically, the data source adopted in this application is a medical test data set. Since the medical test data in the medical test data set is obtained by a testing institution testing a test sample, and the test sample is obtained by sampling a patient's test site, different from the external quality assessment samples used in the related art which are artificially configured, the medical test data is real patient data, which is more in line with the real medical and health scenarios. On this basis, sample test data is obtained by integrating the medical test data, and then multiple data clusters are obtained by clustering the sample test data with the test sample as the main body. As a preliminary diagnosis, the sample test data with the same disease tendency or population characteristics is clustered together. By removing the outlier data clusters, automatic screening of the data is achieved, making the data set and stable. On this basis, by calibrating the test results in the target medical test data, the test results corresponding to each target testing institution can reach the same level. The accuracy of the test quality corresponding to each target testing institution determined according to the target test results is relatively high and has great reference significance, and thus can accurately provide an effective basis for whether the test results can be mutually recognized among medical testing institutions.

[0069] The application scenario of the method for determining the test quality of a testing institution based on the combination of clustering and data calibration provided in the embodiments of the present application will be introduced below.

[0070] Figure 1 is a schematic diagram of an application scenario provided by the present application. As Figure 1 shown, this application scenario includes: electronic device 1 and electronic device 2. Each device is connected through a network.

[0071] Among them, electronic device 1 is the device corresponding to a medical testing institution, and electronic device 1 is used to provide the medical test data of the medical testing institution. Electronic device 2 is used to determine the test quality of the medical testing institution according to the medical test data provided by electronic device 1.

[0072] See Figure 1 , in an application scenario, electronic device 2 obtains medical test data from the electronic devices 1 corresponding to multiple medical testing institutions respectively to form a medical test data set, then integrates the medical test data in the medical test data set to obtain the sample test data corresponding to multiple test samples, then clusters the sample test data, and removes the outlier data clusters to obtain multiple target data clusters, and then calibrates the test results, so as to determine the test quality corresponding to each target detection structure according to the calibrated target test results.

[0073] In another application scenario, the electronic device 1 corresponding to each medical testing institution can be regarded as a computing node. This computing node obtains the medical testing data of the corresponding medical testing institution, integrates the medical testing data to obtain the sample testing data corresponding to multiple test samples, then clusters the sample testing data, eliminates the outlier data clusters to obtain multiple target data clusters, and then calibrates the data of the test results to obtain multiple target test results of this medical testing institution; the electronic device 2 obtains multiple target test results of the corresponding medical testing institutions sent by each computing node, and thereby determines the testing quality corresponding to each medical testing institution.

[0074] It should be noted that for the convenience of representation, Figure 1 in the following, an example is given with the number of electronic devices 1 being 1, but this does not limit the present application. In the actual medical and health scenarios, the number of medical testing institutions is large, and each medical testing institution has its own corresponding electronic device.

[0075] The following uses specific embodiments to elaborate in detail on the technical solution of the present application and how the technical solution of the present application solves the above technical problems. These several specific embodiments below can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below with reference to the accompanying drawings.

[0076] Figure 2 is a schematic flowchart of a method for determining the testing quality of a testing institution based on the combination of clustering and data calibration provided by the present application. The execution subject of the method for determining the testing quality of a testing institution based on the combination of clustering and data calibration provided by the present application is a device for determining the testing quality of a testing institution based on the combination of clustering and data calibration, and this device for determining the testing quality of a testing institution based on the combination of clustering and data calibration is integrated in an electronic device. As Figure 2 shown, the method for determining the testing quality of a testing institution based on the combination of clustering and data calibration provided in this embodiment includes the following steps:

[0077] Step S101, obtain a medical testing data set; the medical testing data set includes the medical testing data of multiple testing institutions. The medical testing data is obtained after testing the corresponding test samples, and the test samples are obtained after sampling the corresponding patient's testing site. The medical testing data includes test items, test results, and test basic information; the test basic information includes at least one of the information for describing the testing process and the patient's medical treatment information.

[0078] In the medical and health care scenario, when a patient visits a hospital, medical staff take samples from the patient's test site to obtain test samples, and then a medical testing institution tests the test samples to obtain test results and records the medical testing data. The medical testing institution can be an in-hospital medical laboratory or an external testing laboratory that cooperates with the hospital. Among them, the medical testing data set includes a large amount of medical testing data, and this large amount of medical testing data comes from multiple medical testing institutions, and the medical testing data of each medical testing institution includes the medical testing data obtained by separately testing a large number of test samples.

[0079] Optionally, the test items include clinical biochemistry items, clinical blood items, clinical body fluid items or clinical immunology items. Optionally, the information used to describe the test process in the test basic information includes the test date and test time, and the patient visit information includes the patient's gender, age, department visited and type of visit. Optionally, the type of visit is emergency, outpatient or inpatient. For example, a medical test data includes: test item "white blood cells", test date "2023 / 10 / 01", test time "01:20", test result "10×10 9 cells / L", patient gender "male", patient age "30 years old", department visited "cardiovascular surgery" and type of visit "outpatient".

[0080] In one possible implementation, the medical testing institution sends the medical testing data of its own institution to the email of the quality assessment personnel by email, and then the quality assessment personnel summarize the medical testing data of each medical testing institution received by the email through an electronic device to obtain a medical testing data set, and trigger the electronic device to store the medical testing data set, so that the electronic device can execute the method provided in this application according to the medical testing data set to determine the testing quality of the medical testing institution. Exemplarily, the medical testing data set is stored in the form of an Excel file. In another possible implementation, the medical testing institution provides a data interface externally, and the electronic device automatically grabs the medical testing data of the medical testing institution through the data interface and stores the grabbed medical testing data, so that the electronic device can execute the method provided in this application according to the medical testing data set to determine the testing quality of the medical testing institution.

[0081] In this embodiment, the electronic device is the terminal device used by the quality assessment personnel responsible for the external quality assessment of medical testing institutions. Correspondingly, the terminal device determines the testing quality of medical testing institutions locally. Alternatively, since the amount of medical testing data is large, it requires a large amount of storage resources, and the computing resources required for testing quality determination are also large. However, the storage resources and computing resources of the terminal devices used by the quality assessment personnel are limited. In this case, the terminal device can determine the testing quality of medical testing institutions with the help of the quality assessment platform, and the quality assessment platform is provided with background services by the server. Correspondingly, the electronic device in this embodiment is provided as the server, and the terminal device used by the quality assessment personnel is communicatively connected to the server. Correspondingly, the quality assessment personnel log in to the quality assessment platform through the terminal device and upload the medical testing data set to the quality assessment platform in the form of an Excel file. Alternatively, the quality assessment platform automatically grabs the medical testing data of medical testing institutions through the data interface, so that the server of the quality assessment platform can determine the testing quality of medical testing institutions. This method saves the storage resources and computing resources of the devices used by the quality assessment personnel through data interaction between devices. This embodiment is described by taking the electronic device as the terminal device used by the quality assessment personnel responsible for the external quality assessment of medical testing institutions as an example, and the implementation method for the server to determine the testing quality of medical testing institutions is the same and will not be elaborated here.

[0082] Step S102: Integrate the medical testing data belonging to the same test sample among multiple medical testing data to obtain the sample testing data corresponding to each of the multiple test samples.

[0083] In this step, the integration method can be to directly splice at least one medical testing data belonging to the same test sample to obtain a sample testing data. Alternatively, the integration method can also be to delete the duplicate data among at least one medical testing data belonging to the same test sample and splice the remaining data to obtain a sample testing data.

[0084] Among them, the sample testing data corresponding to the test sample can characterize the sample characteristics corresponding to the test sample and can represent the disease tendency and population characteristics of the patient to whom the test sample belongs.

[0085] Step S103: Cluster the sample testing data corresponding to each of the multiple test samples to obtain multiple data clusters, and eliminate the outlier data clusters from the multiple data clusters to obtain multiple target data clusters.

[0086] Among them, the multiple sample testing data in one data cluster have similar characteristics, and the disease tendency and population characteristics of the patients are relatively close. The outlier data cluster refers to the data cluster with a relatively large difference compared with other data clusters among the multiple data clusters.

[0087] Optionally, a preset clustering algorithm can be used for clustering, and the preset clustering algorithm can be a density-based clustering algorithm. Exemplarily, the preset clustering algorithm is an algorithm such as OPTICS (Ordering Points to Identify the Clustering Structure) algorithm or DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm.

[0088] In this step, by taking the detection samples as the main body, the sample detection data is clustered to obtain multiple data clusters. As a preliminary diagnosis, the sample detection data with the same disease tendency or population characteristics is clustered together. On this basis, the inventor considered that the outlier data clusters may interfere with the determination of the detection quality. Therefore, by removing the outlier data clusters, the interference is reduced, and the automatic screening of the data is realized, making the data set and stable.

[0089] Step S104: For the same detection item, perform data calibration processing on the detection results in the multiple target medical detection data corresponding to the detection item to obtain multiple target detection results; the target medical detection data is the data in the target data cluster.

[0090] Among them, the target detection result is the detection result obtained after performing data calibration processing on the original detection result.

[0091] In this step, the inventor found in the research that when different medical detection institutions perform detections on the same detection item, the obtained detection results will be affected by multiple factors, such as patient gender, patient age, detection time, etc. That is to say, the detection basic information will affect the detection result, and in addition to the detection basic information, there may also be unknown information that affects the detection result. Then the detection basic information and this unknown information may also affect the detection quality of the medical detection institution. Correspondingly, the detection results can be first subjected to data calibration processing to try to exclude the influence of the detection basic information on the detection results, so that the detection results of different medical detection institutions are on the same level, which is convenient for measuring the detection quality of the medical detection institution. Correspondingly, the operation of performing data calibration processing on the detection results in the multiple target medical detection data corresponding to the detection item can specifically be to first establish a model that can represent the correlation between the detection basic information and the detection results, and then calibrate the detection results through this model, so that the obtained target detection results after calibration exclude the influence of the detection basic information on the detection results.

[0092] Step S105: According to the multiple target detection results, determine the detection quality corresponding to multiple target detection institutions respectively; the target detection institution is the detection institution that provides the target medical detection data.

[0093] Among them, the detection quality can measure the detection ability of a detection institution. The higher the detection quality, the higher the detection ability, and the higher the accuracy of the detection results detected by the detection institution. Then the greater the possibility that other detection institutions recognize the detection results. On the contrary, the lower the detection quality, the lower the detection ability, and the lower the accuracy of the detection results detected by the detection institution. Then the smaller the possibility that other detection institutions recognize the detection results. In this step, multiple target detection results can represent the true detection situation of the corresponding target detection institution, so that the detection quality corresponding to each target detection institution can be determined according to these multiple target detection results.

[0094] In this step, the operation of determining the detection quality corresponding to multiple target detection institutions according to multiple target detection results can be as follows: for each target detection institution, first determine at least one data that can represent the true detection situation of the target detection institution according to the target detection result corresponding to the target detection institution, and then determine the detection quality of the target detection institution according to the difference between this at least one data and the reference range, where the reference range can be a preset numerical range or a numerical range determined according to these multiple target detection results.

[0095] This embodiment provides a detection quality determination scheme for medical detection institutions. The data source used in this detection quality determination scheme is a medical detection data set. Since the medical detection data in the medical detection data set is obtained by detecting detection samples, and the detection samples are obtained by sampling the detection parts of patients, different from the external quality assessment samples used in the related art which are artificially configured, the medical detection data is real patient data, which is more in line with the real medical and health scenarios. On this basis, the sample detection data is obtained by integrating the medical detection data. Thus, taking the detection samples as the main body, the sample detection data is clustered to obtain multiple data clusters. As a preliminary diagnosis, the sample detection data with the same disease tendency or population characteristics is gathered together. By removing the outlier data clusters, the automatic screening of the data is realized, making the data set and stable. On this basis, by calibrating the detection results in the target medical detection data, the detection results corresponding to each target detection institution can reach the same level. The accuracy of the detection quality corresponding to each target detection institution determined according to the target detection results is relatively high and has great reference significance. Furthermore, it can accurately provide an effective basis for whether the detection results can be mutually recognized among medical detection institutions.

[0096] Figure 3 It is a schematic flowchart of another method for determining the detection quality of a detection institution based on the combination of clustering and data calibration provided by this application. As Figure 3As shown in the figure, the method for determining the detection quality of the detection mechanism based on the combination of clustering and data calibration provided in this embodiment further refines steps S103 to S105 on the basis of the method for determining the detection quality of the detection mechanism based on the combination of clustering and data calibration provided in the previous embodiment of this application. The method for determining the detection quality of the detection mechanism based on the combination of clustering and data calibration provided in this embodiment includes the following steps:

[0097] Step S201, obtain a medical detection data set.

[0098] Among them, the medical detection data set includes multiple medical detection data. The medical detection data is obtained after detecting the corresponding detection samples, and the detection samples are obtained after sampling the detection parts of the corresponding patients. The medical detection data includes detection items, detection results, and detection basic information; the detection basic information includes at least one of the information for describing the detection process and the patient visit information.

[0099] In this embodiment, the implementation manner of step S201 is the same as that of step S101 in the previous embodiment, and will not be elaborated here.

[0100] In a possible design, after obtaining the medical detection data set, the electronic device executes the following steps S202 and S203 to preprocess the data.

[0101] Step S202, perform data structuring processing and data normalization processing on each medical detection data to obtain multiple processed medical detection data.

[0102] Data structuring processing: Numerically process each data included in the medical detection data to form structured data that can be read by a computer and has a preset data format.

[0103] Data standardization processing: Perform normal standardization on each data included in the medical detection data to unify the measurement scale, that is, unify the numerical unit, and reduce the adverse effects brought by data with different measurement scales during the internal operation process of the algorithm.

[0104] It should be noted that in addition to the above preprocessing, the medical detection data can also be desensitized to achieve reliable protection of sensitive privacy data.

[0105] Step S203, eliminate the medical detection data with outlier detection results from multiple processed medical detection data for the same detection item to obtain multiple remaining medical detection data.

[0106] Among them, the outlier detection result refers to the detection result that is significantly different from other detection results among multiple detection results corresponding to the same detection item. Since there may be outlier detection results among the detection results in multiple medical detection data, and the outlier detection results may interfere with the determination of the detection quality, it is necessary to eliminate the medical detection data with outlier detection results. Optionally, the implementation methods for eliminating the medical detection data with outlier detection results include MAD (Median Absolute Deviation), standard deviation method, percentile method, etc.

[0107] In this embodiment, by performing data structuring processing and data normalization processing on the medical detection data, and eliminating the medical detection data with outlier detection results, the preprocessing of the data is realized, providing effective and accurate data for the subsequent processing process.

[0108] Step S204: Integrate the medical detection data belonging to the same detection sample among multiple medical detection data to obtain the sample detection data corresponding to each of the multiple detection samples.

[0109] In this embodiment, the implementation method of step S204 is the same as that of step S102 in the previous embodiment, and will not be elaborated here.

[0110] Step S205: Cluster the sample detection data corresponding to each of the multiple detection samples to obtain multiple data clusters.

[0111] Among them, the multiple sample detection data in one data cluster have similar characteristics, and the disease tendencies and population characteristics of the patients are relatively close.

[0112] In a possible design, the implementation method of step S205 includes the following steps S2051 - S2055.

[0113] S2051: Take one sample detection data as one sample point, and plot each sample detection data in a high-dimensional space respectively; the dimension of the high-dimensional space is equal to the data dimension of the sample detection data.

[0114] Among them, since the sample detection data contains multiple data and belongs to high-dimensional features, it is necessary to plot the sample detection data in the high-dimensional space in the form of sample points.

[0115] S2052: Use a preset clustering algorithm to cluster the sample detection data in the high-dimensional space; the preset clustering algorithm is a density-based clustering algorithm.

[0116] S2053: Determine the region with the highest sample density in the high-dimensional space, mark the sample detection data in the region as the same data cluster, and remove the samples in the region from the high-dimensional space.

[0117] Among them, the range of the region with the highest sample density can be set as needed, and this embodiment does not limit it.

[0118] S2054, Use a preset clustering algorithm to cluster the remaining sample detection data in the high-dimensional space.

[0119] S2055, and so on, until the remaining sample detection data in the high-dimensional space accounts for a preset percentage of the total number of sample detection data, and terminate the iterative process to obtain multiple data clusters.

[0120] Among them, the preset percentage can be set as needed, and this embodiment does not limit it. For example, the preset percentage is 5%.

[0121] In this embodiment, each sample detection data can be regarded as a sample point in the high-dimensional space. The closer the distance between sample points, the more similar the two sample detection data are. Therefore, the denser the space, the more concentrated the sample points and the more similar the sample detection data. By iteratively clustering the sample points in the space, multiple relatively accurate data clusters can be obtained.

[0122] Step S206, Remove the outlier data clusters from the multiple data clusters to obtain multiple target data clusters.

[0123] Among them, the outlier data cluster refers to the data cluster with a large difference compared with other data clusters among the multiple data clusters.

[0124] In a possible design, the implementation manner of step S206 includes the following steps S2061-S2062.

[0125] Step S2061, Determine the distances between each data cluster and other data clusters.

[0126] Specifically, for any data cluster, the distance between the center point of the data cluster and the center points of other data clusters is determined as the distance between the data cluster and other data clusters. Among them, the center point refers to the sample point located at the central position of the data cluster in the high-dimensional space. Optionally, the methods for calculating the distance between two sample points include: Euclidean distance, Manhattan distance, Chebyshev distance, or Minkowski distance.

[0127] Step S2062, Remove the data cluster with the largest sum of distances from other data clusters from the multiple data clusters to obtain multiple target data clusters.

[0128] Among them, for each data cluster, calculate the sum of the distances between this data cluster and other data clusters. The larger the sum of the distances, it indicates that the distances between this data cluster and other data clusters are relatively far, which means that the disease tendencies or population characteristics represented by this data cluster are quite different from those represented by other data clusters. Then this data cluster needs to be removed.

[0129] In this embodiment, the distance between data clusters can represent the similarity degree between data clusters. The larger the distance, the lower the similarity degree. Since the outlier data clusters may interfere with the determination of the detection quality, by removing the data cluster with the largest sum of the distances from other data clusters, the target data clusters with similar disease tendencies or population characteristics are retained, the interference is reduced, the automatic screening of data is realized, and the data set is made concentrated and stable.

[0130] Step S207: For the same detection item, perform data calibration processing on the detection results in the multiple target medical detection data corresponding to the detection item to obtain multiple target detection results; the target medical detection data are the data in the target data cluster.

[0131] Among them, the target detection result is the detection result obtained after performing data calibration processing on the original detection result.

[0132] In a possible design, the implementation manner of step S207 includes the following steps S2071 - step S2072.

[0133] For the same detection item, perform the following operations:

[0134] Step S2071: Using the detection result in the target medical detection data corresponding to this detection item as the dependent variable and the basic detection information in this target medical detection data as the independent variable, construct the multiple regression model corresponding to this detection item.

[0135] Optionally, the multiple regression model is a multiple linear regression equation. Exemplarily, the information used to describe the detection process in the basic detection information includes the detection date and detection time, and the patient visit information includes the patient's gender, patient's age, visit department, and visit type, a total of 6 parameter items. Therefore, the number of independent variables is 6. Correspondingly, the multiple regression model can be:

[0136] Y = β 0 + β 1 X 1 + β 2 X 2 + β 3 X 3 + β 4 X 4 + β 5 X 5 + β6 X 6 + ε

[0137] Wherein, Y is the dependent variable, i.e., the detection result, and β 0 is the intercept, and β i is the coefficient of the independent variable X i where i ∈ [1, 6], and ε is the residual term.

[0138] Optionally, β i can be determined according to the least squares method. Optionally, the intercept is the average value of a preset number of detection results for this detection item, and the preset number is the daily average detection volume of this detection item.

[0139] Step S2072: Use the multiple regression model corresponding to the detection item to perform data calibration processing on the detection results in the target medical detection data corresponding to the detection item, so as to obtain multiple target detection results.

[0140] In this embodiment, by establishing a multiple regression model to simulate the correlation between the basic detection information and the detection result in the target medical detection data, the detection result can be calibrated according to the multiple regression model, making the result of the data calibration processing more accurate.

[0141] In a possible design, the implementation manner of step S2072 includes:

[0142] Perform the following operations on each piece of target medical detection data corresponding to the detection item:

[0143] Input the basic detection information in the medical detection data into the multiple regression model, and output the fitting result corresponding to the basic detection information;

[0144] Determine the difference between the detection result in the target medical detection data and the fitting result as the target detection result.

[0145] Wherein, the fitting result is the detection result fitted according to the basic detection information, and there is a difference between this fitting result and the true detection result, and this difference cannot be explained by the basic detection information, and this difference is also the residual.

[0146] In this embodiment, by determining the residual between the detection result and the fitting result as the processed target detection result, the influence of the basic detection information on the detection result is excluded, so that the obtained target detection results are at the same level.

[0147] Step S208: Use a preset data processing algorithm to perform corresponding processing on the multiple target detection results respectively corresponding to each target detection institution.

[0148] Optionally, the preset data processing algorithms include: Exponentially Weighted Moving Average algorithm (EWMA), Moving Average algorithm (MA), or Moving Standard Deviation algorithm (MovSD). In this step, for each target detection institution, the preset data processing algorithm is used to process the multiple target detection results corresponding to the target detection institution, so as to implement filtering processing of the target detection results, making the stability of the multiple processed target detection results corresponding to the target detection institution stronger.

[0149] Step S209: Determine the detection quality corresponding to each of the multiple target detection institutions according to the multiple target detection results; the target detection institution is an institution that provides target medical detection data.

[0150] In the related art, external quality assessment is carried out through artificially configured external quality assessment samples. Since the true detection values corresponding to the detection items in the configured external quality assessment samples are known, by comparing the detection results obtained by the detection institution with the true detection values, the detection quality of the detection institution can be determined.

[0151] However, the medical detection data used in this application is obtained by detecting the detection samples taken from the patient's detection site, and there is no "true detection value" that can be referenced. To solve this technical problem, the inventor found that the detection benchmark value range can be determined according to the multiple detection results corresponding to each target detection institution. This detection benchmark value range can be regarded as a reference range, providing a measurement standard for the determination of detection quality.

[0152] In a possible design, the implementation manner of step S209 includes the following steps S2091 - S2093.

[0153] Step S2091: For the same detection item, determine the detection benchmark target value and the preset offset percentage corresponding to at least one quantile according to the multiple detection results corresponding to each target detection institution.

[0154] Among them, the number of quantiles can be set as needed, and this embodiment does not limit it. Optionally, the number of quantiles is 1, 3, or 5. If the number of quantiles is 1, then this one quantile includes the 50% quantile; if the number of quantiles is 3, then the 3 quantiles include the 25% quantile, 50% quantile, and 75% quantile; if the number of quantiles is 5, then the 5 quantiles include the 5% quantile, 25% quantile, 50% quantile, 75% quantile, and 95% quantile.

[0155] Each quantile has a corresponding detection reference target value and a preset offset percentage. The detection reference target value and the preset offset percentage are used to determine the detection reference numerical range corresponding to the quantile. Among them, the preset offset percentage can be set as needed, and this embodiment does not limit it. For example, the preset offset percentage is 5%, 10%, etc.

[0156] Step S2092: Determine the detection reference numerical ranges corresponding to at least one quantile according to the detection reference target values and the preset offset percentages respectively corresponding to at least one quantile.

[0157] Optionally, the detection reference numerical range is a numerical set composed of multiple discrete points, and each discrete point represents a discrete numerical value. For example, the detection reference numerical range is {2, 3, 4, 5}. Or, the detection reference numerical range can also be a numerical interval composed of continuous points. For example, the detection reference numerical range is [2, 5].

[0158] Optionally, the implementation method for determining the detection reference target values respectively corresponding to at least one quantile includes:

[0159] For the same detection item, according to the multiple detection results respectively corresponding to each target detection agency, calculate the quantile values of each target detection agency at each quantile respectively;

[0160] For each quantile, calculate the quantile values of each target detection agency at this quantile according to a preset calculation method to obtain the detection reference target value corresponding to this quantile, and the preset calculation method is to calculate the mean value or the median.

[0161] Exemplarily, taking any one of at least one quantile as an example, the quantile values of multiple target detection agencies at this quantile are 4, 5, 4, and 6 respectively. When the preset calculation method is to calculate the mean value, the detection reference target value corresponding to this quantile is 4.75. When the preset calculation method is to calculate the median, the detection reference target value corresponding to this quantile is 4.5.

[0162] By determining the quantile values of each target detection agency at each quantile according to the multiple target detection results corresponding to each target detection agency, and the quantile value can represent the true detection situation of the target detection agency, and then calculating the detection reference target values corresponding to each quantile according to the quantile values of each target detection agency at each quantile. It can be seen that the determined detection reference target values are jointly determined according to the target detection results of multiple target detection agencies. Therefore, the credibility of the detection reference target value is higher and the reference is stronger. In addition, the preset calculation method can be set as needed to calculate the mean value or the median, and the flexibility is higher.

[0163] Optionally, the implementation method of step S2092 includes:

[0164] Determine the upper and lower boundaries of the detection reference value range corresponding to at least one quantile according to the detection reference target value corresponding to at least one quantile and the preset offset percentage. The upper boundary is the sum of the detection reference target value corresponding to the quantile and the offset amount, and the lower boundary is the difference between the detection reference target value corresponding to the quantile and the offset amount; the offset amount is the product of the detection reference target value and the preset offset percentage;

[0165] Based on the upper and lower boundaries of the detection reference value range corresponding to at least one quantile, determine the corresponding detection reference value range.

[0166] Exemplarily, taking any one of the at least one quantiles as an example, the detection reference target value corresponding to this quantile is 4, and the preset offset percentage is 10%, then the offset amount is 0.4, the upper boundary is 4.4, and the lower boundary is 3.6. Correspondingly, the detection reference value range is [3.6, 4.4].

[0167] By taking the detection reference target value as the benchmark and taking the product of the detection reference target value and the preset offset percentage as the offset amount, determine the upper and lower boundaries of the detection reference value range, so that the size of the determined detection reference value range is more appropriate.

[0168] It should be noted that the preset offset percentage corresponding to any quantile can also be replaced by the standard deviation of the quantile value of each target detection agency at this quantile.

[0169] Step S2093, for the same detection item, determine the detection quality scores of each target detection agency in the corresponding detection item according to the detection reference value ranges corresponding to at least one quantile respectively and the multiple detection results corresponding to each target detection agency respectively.

[0170] Among them, the detection quality score is used to measure the detection quality. The higher the score, the higher the quality, and the lower the score, the lower the quality.

[0171] In this embodiment, by determining the detection reference value ranges corresponding to each quantile according to the multiple target detection results corresponding to each target detection agency respectively, when determining the detection reference value range, taking at least one quantile as a reference point, by determining the detection reference target value and the preset offset percentage corresponding to each quantile respectively, determine the detection reference value range corresponding to each quantile respectively. In this way, the determined detection reference value range has greater reference significance. On this basis, the detection reference value range can be regarded as a reference range, thus providing a measurement standard for the determination of detection quality, and the determined detection quality score has greater reference significance.

[0172] In a possible design, the implementation manner of step S2093 includes:

[0173] For the same detection item, according to the multiple detection results respectively corresponding to each target detection agency, obtain the quantile values of each target detection agency at each quantile point;

[0174] For each quantile point, if it is determined that the quantile value of the target detection agency at the quantile point is within the detection benchmark numerical range corresponding to the quantile point, then determine that the score of the target detection agency at the quantile point is the first score;

[0175] If it is determined that the quantile value of the target detection agency at the quantile point is not within the detection benchmark numerical range corresponding to the quantile point, then determine that the score of the target detection agency at the quantile point is the second score; the first score is greater than the second score;

[0176] Determine the sum of the scores of the target detection agency at each quantile point as the detection quality score of the target detection agency.

[0177] Among them, since the quantile values of each target detection agency at each quantile point have been determined in the previous steps by the electronic device, therefore, the previously determined quantile values can be directly obtained in this step.

[0178] Among them, if the quantile value of the target detection agency at a certain quantile point is within the detection benchmark numerical range corresponding to the quantile point, it indicates that the detection level of the target detection agency at this quantile point is relatively stable and the detection quality is relatively high. On the contrary, if the quantile value of the target detection agency at a certain quantile point is not within the detection benchmark numerical range corresponding to the quantile point, it indicates that the detection level of the target detection agency at this quantile point is not very stable and the detection quality is not very high.

[0179] For example, the first score is 20, the second score is 0, and the number of at least one quantile point is 5, then the full score is 100.

[0180] In this embodiment, taking the detection benchmark numerical ranges respectively corresponding to each quantile point as the reference basis, using the quantile values of the target detection agency at each quantile point to represent the true detection situation of the target detection agency, by judging whether the quantile values of the target detection agency at each quantile point are within the detection benchmark numerical ranges respectively corresponding to each quantile point, to determine the detection level and detection quality of the target detection agency, and determining the sum of the scores as the detection quality score of the target detection agency, realizing the quantification of the detection quality of the target detection agency, with relatively high accuracy.

[0181] The above embodiment is described by taking one preset data processing algorithm as an example. In other embodiments, there are multiple preset data processing algorithms. Correspondingly, the method provided in this application further includes:

[0182] Respectively determine the detection quality scores of the target detection agency under each preset data processing algorithm;

[0183] Fuse multiple detection quality scores determined for the target detection agency to obtain the final detection quality score of the target detection agency.

[0184] Exemplarily, the fusion process is to calculate the mean or weighted average.

[0185] Considering various preset data processing algorithms, by fusing the detection quality scores determined under each preset data processing algorithm, the final detection quality score of the target detection agency is obtained, and the accuracy of the determined final detection quality score is higher.

[0186] It should be noted that in order to evaluate the detection quality score of the target detection agency, a score threshold can be set. Correspondingly, if the detection quality score of the target detection agency exceeds the score threshold, it indicates that the detection quality of the target detection agency is high; if the detection quality score of the target detection agency does not exceed the score threshold, it indicates that the detection quality of the target detection agency is low. For example, the full score is 100 and the score threshold is 80.

[0187] Figure 4 It is a schematic structural diagram of a detection agency detection quality determination device based on the combination of clustering and data calibration provided by this application. As Figure 4 shown, in this embodiment, the detection agency detection quality determination device 300 based on the combination of clustering and data calibration can be set in an electronic device. The detection agency detection quality determination device 300 based on the combination of clustering and data calibration includes:

[0188] An acquisition module 301, configured to acquire a medical detection data set; the medical detection data set includes medical detection data of multiple detection agencies. The medical detection data is obtained after detecting the corresponding detection samples, and the detection samples are obtained after sampling the detection parts of the corresponding patients. The medical detection data includes detection items, detection results, and detection basic information; the detection basic information includes at least one of information for describing the detection process and patient visit information;

[0189] An integration module 302, configured to integrate the medical detection data belonging to the same detection sample among the multiple medical detection data to obtain sample detection data corresponding to multiple detection samples respectively;

[0190] A clustering module 303, configured to cluster the sample detection data corresponding to multiple detection samples respectively to obtain multiple data clusters, and remove the outlier data clusters from the multiple data clusters to obtain multiple target data clusters;

[0191] The calibration module 304 is configured to perform data calibration processing on the detection results in multiple target medical detection data corresponding to a detection item for the same detection item, so as to obtain multiple target detection results; the target medical detection data is the data in the target data cluster;

[0192] The determination module 305 is configured to determine the detection quality corresponding to multiple target detection institutions according to the multiple target detection results; the target detection institution is the detection institution that provides the target medical detection data.

[0193] Optionally, the calibration module 304 includes:

[0194] The construction sub-module is configured to, for the same detection item, construct a multiple regression model corresponding to the detection item with the detection results in the target medical detection data corresponding to the detection item as the dependent variable and the basic detection information in the target medical detection data as the independent variable;

[0195] The calibration sub-module is configured to, for the same detection item, perform data calibration processing on the detection results in the target medical detection data corresponding to the detection item by using the multiple regression model corresponding to the detection item, so as to obtain multiple target detection results.

[0196] Optionally, the calibration sub-module is specifically configured to:

[0197] Perform the following operations on each target medical detection data corresponding to the detection item:

[0198] Input the basic detection information in the medical detection data into the multiple regression model, and output the fitting result corresponding to the basic detection information;

[0199] Determine the difference between the detection result in the target medical detection data and the fitting result as the target detection result.

[0200] Optionally, the device 300 further includes:

[0201] The preprocessing module is configured to perform data structuring processing and data normalization processing on each medical detection data to obtain multiple processed medical detection data;

[0202] The elimination module is configured to eliminate the medical detection data with outlier detection results from multiple processed medical detection data for the same detection item, so as to obtain multiple remaining medical detection data.

[0203] Optionally, the clustering module 303 includes a clustering sub-module, and the clustering sub-module is configured to:

[0204] Taking one sample detection data as one sample point, respectively plot each sample detection data in a high-dimensional space; the dimension of the high-dimensional space is equal to the data dimension of the sample detection data;

[0205] Use a preset clustering algorithm to cluster the sample detection data in the high-dimensional space; the preset clustering algorithm is a density-based clustering algorithm;

[0206] Determine the region with the highest sample density in the high-dimensional space, mark the sample detection data within the region as the same data cluster, and remove the samples in the region from the high-dimensional space;

[0207] Use a preset clustering algorithm to cluster the remaining sample detection data in the high-dimensional space;

[0208] And so on, until the remaining sample detection data in the high-dimensional space accounts for a preset percentage of the total number of sample detection data, terminate the iterative process to obtain multiple data clusters.

[0209] Optionally, the clustering module 303 includes an elimination sub-module, and the elimination sub-module is used for:

[0210] Determine the distances between each data cluster and other data clusters respectively;

[0211] Eliminate the data cluster with the largest sum of distances from other data clusters from the multiple data clusters to obtain multiple target data clusters.

[0212] Optionally, the determination module 305 includes:

[0213] The first determination sub-module is used to determine the detection reference target value and the preset offset percentage corresponding to at least one quantile according to the multiple detection results respectively corresponding to each target detection agency for the same detection item;

[0214] The second determination sub-module is used to determine the detection reference numerical range corresponding to at least one quantile according to the detection reference target value and the preset offset percentage corresponding to at least one quantile;

[0215] The third determination sub-module is used to determine the detection quality score of each target detection agency in the corresponding detection item according to the detection reference numerical range corresponding to at least one quantile and the multiple detection results respectively corresponding to each target detection agency for the same detection item.

[0216] Optionally, the third determination sub-module is specifically used for:

[0217] For the same detection item, obtain the quantile values of each target detection agency at each quantile according to the multiple detection results respectively corresponding to each target detection agency;

[0218] For each quantile, if it is determined that the quantile value of the target detection agency at the quantile is within the detection reference numerical range corresponding to the quantile, determine that the score of the target detection agency at the quantile is the first score;

[0219] If it is determined that the quantile value of the target detection agency at the quantile is not within the detection reference numerical range corresponding to the quantile, then determine that the score of the target detection agency at the quantile is the second score; the first score is greater than the second score;

[0220] Determine the sum of the scores of the target detection agency at each quantile as the detection quality score of the target detection agency.

[0221] The detection agency detection quality determination device based on the combination of clustering and data calibration provided in this embodiment can execute the technical solutions of the corresponding method embodiment, and its implementation principle and technical effects are similar to those of the corresponding method embodiment, and will not be elaborated here one by one.

[0222] This application embodiment also provides an electronic device. The electronic device can be provided as a terminal device or can also be provided as a server. Such as, a smart phone, a digital broadcast terminal, a messaging device, a tablet device, a medical device, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers.

[0223] Figure 5 It is a schematic structural diagram of an electronic device provided by this application. As Figure 5 shown, the electronic device 400 includes: a processor 401 and a memory 402 communicatively connected to the processor 401.

[0224] The memory 402 stores computer execution instructions; the processor 401 executes the computer execution instructions stored in the memory 402 to implement the method for determining the detection quality of a detection agency based on the combination of clustering and data calibration provided by this application.

[0225] Among them, in this application embodiment, the memory 402 and the processor 401 are connected through a bus. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity of representation, only a thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0226] The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application as described herein and / or claimed. The various components are interconnected using different buses and can be mounted on a common motherboard or otherwise mounted as needed.

[0227] In an exemplary embodiment, there is also provided a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, are used to implement the method for determining the detection quality of a detection mechanism based on the combination of clustering and data calibration provided by the present application.

[0228] In an exemplary embodiment, there is also provided a computer program product including computer-executable instructions that, when the computer-executable instructions in the computer program product are executed by a processor, are used to implement the method for determining the detection quality of a detection mechanism based on the combination of clustering and data calibration provided by the present application.

[0229] It should be noted that, for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to the present application.

[0230] Furthermore, it should be noted that although the steps in the flowchart are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps has no strict order limitation, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps.

[0231] It should be understood that the above device embodiments are merely illustrative, and the devices of the present application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units, modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed.

[0232] In addition, unless otherwise specified, each functional unit / module in the various embodiments of the present application may be integrated into one unit / module, may exist as individual physical units / modules, or may be integrated together with two or more units / modules. The above integrated unit / module may be implemented in the form of hardware or in the form of a software program module.

[0233] When the integrated unit / module is implemented in the form of hardware, the hardware can be a digital circuit, an analog circuit, etc. The physical implementation of the hardware structure includes but is not limited to transistors, memristors, etc. Unless otherwise specified, the processor can be any suitable hardware processor, such as a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a controller, a microcontroller, a microprocessor, or other electronic components. Unless otherwise specified, the memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a USB flash drive, a Random-Access Memory (RAM), a Static Random-Access Memory (SRAM), a Dynamic Random-Access Memory (DRAM), an Enhanced Dynamic Random-Access Memory (EDRAM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), an Erasable Programmable Read-Only Memory (EPROM), a Programmable Read-Only Memory (PROM), a Read-Only Memory (ROM), a Resistive Random Access Memory (RRAM), a High-Bandwidth Memory (HBM), a Hybrid Memory Cube (HMC), a magnetic memory, a flash memory, a disk, an optical disk, a mobile hard disk, or a magnetic disk, and other media that can store program code.

[0234] When the integrated unit / module is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on such understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present application.

[0235] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as within the scope described in this specification.

[0236] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any variations, uses, or adaptations of the present application, which follow the general principles of the present application and include the common general knowledge or conventional technical means in the technical field not disclosed in the present application. The specification and the embodiments are only regarded as exemplary, and the true scope and spirit of the present application are pointed out by the following claims.

[0237] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present application is only limited by the appended claims.

Claims

1. A method for determining the detection quality of a detection mechanism based on the combination of clustering and data calibration, characterized in that, it includes: Obtain a medical detection data set; the medical detection data set includes medical detection data of multiple detection mechanisms, the medical detection data is obtained after detecting the corresponding detection samples, the detection samples are obtained after sampling the detection parts of the corresponding patients, and the medical detection data includes detection items, detection results and detection basic information; the detection basic information includes at least one of information for describing the detection process and patient visit information; Integrate the medical detection data belonging to the same detection sample among the multiple medical detection data to obtain sample detection data corresponding to each detection sample; Cluster the sample detection data corresponding to the multiple detection samples respectively to obtain multiple data clusters, and eliminate the outlier data clusters from the multiple data clusters to obtain multiple target data clusters; For the same detection item, perform data calibration processing on the multiple target medical detection data corresponding to the detection item to obtain multiple target detection results; The target medical detection data is the data in the target data cluster; Determine the detection quality corresponding to multiple target detection mechanisms according to the multiple target detection results; the target detection mechanism is the detection mechanism that provides the target medical detection data.

2. The method according to claim 1, characterized in that, The step of performing data calibration processing on the detection results in the multiple target medical detection data corresponding to the same detection item to obtain multiple target detection results includes: Perform the following operations for the same detection item: Construct a multiple regression model corresponding to the detection item with the detection result in the target medical detection data corresponding to the detection item as the dependent variable and the detection basic information in the target medical detection data as the independent variable; Use the multiple regression model corresponding to the detection item to perform data calibration processing on the detection results in the target medical detection data corresponding to the detection item to obtain multiple target detection results.

3. The method according to claim 2, characterized in that, The step of using the multiple regression model corresponding to the detection item to perform data calibration processing on the detection results in the target medical detection data corresponding to the detection item to obtain multiple target detection results includes: Perform the following operations for each of the target medical detection data corresponding to the detection item: Input the detection basic information in the medical detection data into the multiple regression model and output the fitting result corresponding to the detection basic information; Determine the difference between the detection result in the target medical detection data and the fitting result as the target detection result.

4. The method according to claim 1, characterized in that, After obtaining the medical detection data set, the method further includes: Perform data structuring processing and data normalization processing on each of the medical detection data to obtain multiple processed medical detection data; From multiple processed medical test data for the same test item, eliminate the medical test data with outlier test results to obtain multiple remaining medical test data.

5. The method according to claim 1, wherein, the clustering of the sample test data respectively corresponding to the multiple test samples to obtain multiple data clusters includes: Taking one sample test data as one sample point, respectively plotting each of the sample test data in a high-dimensional space; the dimension of the high-dimensional space is equal to the data dimension of the sample test data; Using a preset clustering algorithm to cluster the sample test data in the high-dimensional space; the preset clustering algorithm is a density-based clustering algorithm; Determining the region with the highest sample density from the high-dimensional space, marking the sample test data within the region as the same data cluster, and removing the samples in the region from the high-dimensional space; Using the preset clustering algorithm to cluster the remaining sample test data in the high-dimensional space; And so on, until the remaining sample test data in the high-dimensional space accounts for a preset percentage of the total number of sample test data, terminating the iterative process to obtain multiple data clusters.

6. The method according to claim 1, wherein, the eliminating of the outlier data clusters from the multiple data clusters to obtain multiple target data clusters includes: Determining the distances between each of the data clusters and other data clusters; Eliminating the data cluster with the largest sum of distances to other data clusters from the multiple data clusters to obtain the multiple target data clusters.

7. The method according to any one of claims 1-6, wherein, the determining of the detection quality respectively corresponding to the multiple target detection institutions according to the test results respectively included in the multiple target medical test data includes: For the same test item, determining the detection reference target value and the preset offset percentage respectively corresponding to at least one quantile according to the multiple test results respectively corresponding to each of the target detection institutions; Determining the detection reference numerical range respectively corresponding to the at least one quantile according to the detection reference target value and the preset offset percentage respectively corresponding to the at least one quantile; For the same test item, determining the detection quality score of each of the target detection institutions in the corresponding test item according to the detection reference numerical range respectively corresponding to the at least one quantile and the multiple test results respectively corresponding to each of the target detection institutions.

8. The method according to claim 7, wherein, the determining of the detection quality score of each of the target detection institutions in the corresponding test item according to the detection reference numerical range respectively corresponding to the at least one quantile and the multiple test results respectively corresponding to each of the target detection institutions includes: For the same test item, obtaining the quantile value of each of the target detection institutions at each of the quantiles according to the multiple test results respectively corresponding to each of the target detection institutions. For each of the quantiles, if it is determined that the quantile value of the target detection institution at the quantile is within the detection benchmark numerical range corresponding to the quantile, then determine that the score of the target detection institution at the quantile is the first score; If it is determined that the quantile value of the target detection institution at the quantile is not within the detection benchmark numerical range corresponding to the quantile, then determine that the score of the target detection institution at the quantile is the second score; the first score is greater than the second score; Determine the sum of the scores of the target detection institution at each of the quantiles as the detection quality score of the target detection institution.

9. A device for determining the detection quality of a detection institution by combining clustering and data calibration, characterized in that, it includes: An acquisition module, configured to acquire a medical detection data set; the medical detection data set includes medical detection data of multiple detection institutions, the medical detection data is obtained after detecting the corresponding detection samples, the detection samples are obtained after sampling the detection parts of the corresponding patients, and the medical detection data includes detection items, detection results and detection basic information; the detection basic information includes at least one of information for describing the detection process and patient visit information; An integration module, configured to integrate the medical detection data belonging to the same detection sample among the multiple medical detection data to obtain sample detection data corresponding to multiple detection samples respectively; A clustering module, configured to cluster the sample detection data corresponding to the multiple detection samples respectively to obtain multiple data clusters, and remove the outlier data clusters from the multiple data clusters to obtain multiple target data clusters; A calibration module, configured to perform data calibration processing on the detection results in the multiple target medical detection data corresponding to the detection item to obtain multiple target detection results; The target medical detection data is the data in the target data cluster; A determination module, configured to determine the detection quality corresponding to multiple target detection institutions respectively according to the multiple target detection results; the target detection institution is the detection institution that provides the target medical detection data.

10. An electronic device, characterized in that, it includes: A processor and a memory communicatively connected to the processor; The memory stores computer execution instructions; The processor executes the computer execution instructions stored in the memory to implement the method for determining the detection quality of a detection institution by combining clustering and data calibration according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer execution instructions, and when the computer execution instructions are executed by a processor, they are used to implement the method for determining the detection quality of a detection institution by combining clustering and data calibration according to any one of claims 1 to 8.

12. A computer program product, including computer execution instructions, characterized in that, When the computer execution instructions are executed by a processor, they implement the method for determining the detection quality of a detection institution by combining clustering and data calibration according to any one of claims 1 to 8.

Images