A blood sample efficient detection method for laboratory

Abstract

The present invention relates to the technical field of blood test data processing, and specifically to an efficient blood sample detection method for a laboratory. The present invention obtains the medical file similarity between re-examination patients and obtains multiple re-examination patient clusters based on the correlation coefficient of medical file data of different dimensions between re-examination patients; screens out necessary re-examination patient clusters based on the consistency of the indicator data set between the automated detection and manual re-examination of the re-examination patients in each re-examination patient cluster; for any necessary re-examination patient cluster, analyzes the data distribution of the automated detection indicator data set of different re-examination patients, obtains the re-examination importance and equipment abnormality of each indicator; obtains the re-examination corrected importance of each indicator in each necessary re-examination patient cluster. The present invention determines the detection order of re-examination by obtaining the accurate re-examination importance of each indicator in the cluster, thereby improving the efficiency of blood sample detection.

Description

Technical Field

[0001] The present invention relates to the technical field of blood test data processing, and in particular to a blood sample efficient detection method for a laboratory. Background Art

[0002] Blood testing technology is a very important part of the laboratory department. It is based on hematological theory and uses laboratory experimental methods to analyze multiple indicators in blood samples to help diagnose and monitor various diseases. With the development of biotechnology and information technology, blood sample testing mostly uses a full laboratory automation system, which helps to reduce the waiting time for sample processing and analysis and increase the detection rate.

[0003] Considering that the automated system may encounter some abnormal or suspicious results when processing samples, which may affect the accuracy of the test, in the prior art, blood samples with abnormal or suspicious performance are manually retested to verify the accuracy of the automated system's test results. However, due to the large number of blood samples, it is not considered that the test indicators may be affected by external factors such as equipment failure and produce abnormal changes, resulting in an extension of the test cycle of the indicators that need to be retested due to the actual abnormal changes and delayed results, and the efficiency of blood sample testing is poor. Summary of the invention

[0004] In order to solve the technical problem that the detection index may be affected by external factors such as equipment failure and produce abnormal changes, and the efficiency of blood sample detection is poor, the purpose of the present invention is to provide a blood sample efficient detection method for the laboratory, and the technical solution adopted is as follows:

[0005] The present invention proposes a method for efficiently detecting blood samples for a laboratory, the method comprising:

[0006] Obtain multi-dimensional medical archive data of patients who were manually re-examined during the historical blood sample testing process of each batch, as well as the indicator data sets under the corresponding automated testing and manual re-examination;

[0007] According to the correlation coefficients of the medical file data of different dimensions between the re-examined patients, the similarity of the medical files between the re-examined patients is obtained; according to the similarity of the medical files between the re-examined patients, all the re-examined patients are clustered to obtain multiple clusters of re-examined patients; according to the consistency of the indicator data sets between the automated detection and manual re-examination of the re-examined patients in each cluster of re-examined patients, the necessary clusters of re-examined patients are screened out;

[0008] For any necessary re-examination patient cluster, the re-examination importance of each indicator is obtained based on the data distribution of the automated detection indicator data sets of different re-examination patients; the device abnormality of each indicator in each necessary re-examination patient cluster is obtained based on the similarity of changes in the automated detection indicator data sets between each re-examination patient in the necessary re-examination patient cluster and different patients in the same batch;

[0009] The re-examination importance is corrected according to the device abnormality of each indicator in each necessary re-examination patient cluster, so as to obtain the re-examination corrected importance of each indicator in each necessary re-examination patient cluster.

[0010] Furthermore, the method for obtaining the medical record similarity includes:

[0011] Obtain a word segmentation array of each dimension of the medical file data of the re-examined patient, and obtain a correlation coefficient of the word segmentation array of the medical file data of the same dimension between the re-examined patients as the dimension array correlation;

[0012] The dimensional array correlations of all medical file data of the same dimensions between the re-examined patients are averaged to serve as the medical file similarity between the re-examined patients.

[0013] Furthermore, the method for obtaining the clustering of re-examination patients includes:

[0014] According to the similarity of medical records between re-examination patients, K-means clustering was performed on all re-examination patients to obtain multiple re-examination patient clusters.

[0015] Furthermore, the method for obtaining the necessary re-examination patient clusters includes:

[0016] For any cluster of re-examination patients, the number of re-examination patients with the same indicator data set between automated testing and manual re-examination is counted as the number of re-examination matches;

[0017] According to the difference characteristics between the number of re-examination matches and the number of all re-examination patients, the possibility of no need for re-examination for each re-examination patient cluster was obtained, and the difference characteristics were negatively correlated with the possibility of no need for re-examination.

[0018] If the possibility of a re-examination patient cluster not requiring re-examination is less than a preset re-examination threshold, the corresponding re-examination patient cluster will be regarded as a re-examination patient cluster that requires re-examination.

[0019] Furthermore, the method for obtaining the re-examination importance includes:

[0020] If the indicator data in the patient's automated detection indicator data set does not fall within the preset indicator threshold range, the corresponding indicator is regarded as an abnormal indicator;

[0021] For any cluster of patients who need to be re-examined, the ratio between the number of re-examined patients with abnormal indicators for each indicator and the number of all re-examined patients is obtained as the re-examination emphasis of each indicator.

[0022] Furthermore, the method for obtaining the device abnormality degree includes:

[0023] For any necessary re-examination patient cluster, the possibility of equipment failure for each re-examination patient is obtained based on the similarity of changes in the automated detection indicator data set between each re-examination patient and each patient in the same batch;

[0024] If the probability of equipment failure of the re-examined patient is greater than the preset failure threshold, the intersection of abnormal indicators between the re-examined patient and all other patients in the same batch is obtained; for each indicator in the intersection, the probability of equipment failure of the corresponding re-examined patient is used as the local abnormality of the equipment of the re-examined patient under each indicator; for each indicator outside the intersection, the local abnormality of the equipment of the re-examined patient under each indicator is set to 0;

[0025] The mean of the local abnormality of the equipment of different re-examination patients under each indicator is obtained as the overall abnormality of the equipment for each indicator in the necessary re-examination patient cluster.

[0026] Furthermore, the method for obtaining the possibility of equipment failure includes:

[0027] For any cluster of patients who need to be re-examined, obtain the correlation coefficient of the set of all abnormal indicators between each re-examined patient and each patient in the same batch as the local abnormal correlation;

[0028] The mean of the local abnormality correlation between each re-examination patient and all patients in the same batch was calculated as the device failure probability of each re-examination patient in each required re-examination patient cluster.

[0029] Furthermore, the method for obtaining the importance of re-examination and correction includes:

[0030] Calculate the difference between the positive integer 1 and the device abnormality of each indicator in each cluster of patients who need to be re-examined as the adjustment weight;

[0031] The re-examination importance of each indicator in each cluster of patients who need re-examination is weighted according to the adjustment weight to obtain the re-examination revised importance of each indicator in each cluster of patients who need re-examination.

[0032] Furthermore, the re-examination importance of each indicator in each cluster of patients who need to be re-examined is weighted according to the adjustment weight to obtain the re-examination corrected importance of each indicator in each cluster of patients who need to be re-examined, including:

[0033] The product of the re-examination importance and the adjustment weight of each indicator in each cluster of patients who need re-examination is obtained as the re-examination corrected importance of each indicator in each cluster of patients who need re-examination.

[0034] Furthermore, the correlation coefficient is obtained by using the Jaccard correlation coefficient.

[0035] The present invention has the following beneficial effects:

[0036] The present invention obtains the medical file similarity between re-examination patients based on the correlation coefficient of medical file data of different dimensions between re-examination patients, and quantifies the similarity of medical file data between different re-examination patients; clusters all re-examination patients based on the medical file similarity between re-examination patients to obtain multiple re-examination patient clusters, which helps to identify patient groups with similar medical conditions or disease characteristics through clustering; based on the consistency of the indicator data set between the automated detection and manual re-examination of the re-examination patients in each re-examination patient cluster, screens out the necessary re-examination patient clusters, and identifies the re-examination patient clusters that need to be focused on manually Cluster; for any necessary re-examination patient cluster, according to the data distribution of the automated detection indicator data set of different re-examination patients, the re-examination importance of each indicator is obtained, and by analyzing the data distribution, the importance of different indicators in the re-examination can be understood; according to the similarity of the change of the automated detection indicator data set between each re-examination patient in the necessary re-examination patient cluster and different patients corresponding to the same batch, the device abnormality of each indicator in each necessary re-examination patient cluster is obtained; the re-examination correction importance of each indicator in each necessary re-examination patient cluster is obtained, and the actual importance of each indicator in the re-examination is more accurately reflected. The present invention determines the detection order of re-examination by obtaining the accurate re-examination importance of each indicator in the cluster, thereby improving the efficiency of blood sample detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 The present invention provides a flowchart of a method for efficiently detecting blood samples in a laboratory, according to an embodiment of the present invention.

[0039] Figure 2 A flow chart of a method for obtaining clusters of patients who need to be re-examined provided by one embodiment of the present invention;

[0040] Figure 3A flow chart of a method for obtaining device abnormality provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, characteristics and effects of a blood sample efficient detection method for a laboratory according to the present invention, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0042] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0043] The specific scheme of a method for efficiently detecting blood samples for a laboratory provided by the present invention is described in detail below with reference to the accompanying drawings.

[0044] See also Figure 1 , which shows a flow chart of a method for efficient blood sample detection for a laboratory provided by an embodiment of the present invention, the specific method includes:

[0045] Step S1: Obtain multi-dimensional medical file data of patients who were manually re-examined during the historical blood sample testing process of each batch, as well as the indicator data sets under the corresponding automated testing and manual re-examination.

[0046] In an embodiment of the present invention, in order to improve the efficiency of blood sample testing, it is necessary to analyze the historical medical data of the manually re-examined patients, which is helpful to discover and correct errors in automated testing and improve the accuracy of the overall test. First, for blood sample testing, based on the indicator data set of the patient's blood sample under automated testing obtained in the full laboratory automated testing, for patients who are determined to need manual re-examination in the full laboratory automated testing, the indicator data set under manual re-examination is obtained and stored in the hospital database; therefore, the multi-dimensional medical file data of the manually re-examined patients in the process of each batch of historical blood sample testing is obtained from the hospital database, as well as the indicator data set under the corresponding automated testing and manual re-examination. Among them, the medical file data involves multiple dimensions, including chief complaint, diagnosis opinion, current medical history, test examination, past history, etc., and the indicator data set includes detection data under indicators such as red blood cell count, hemoglobin concentration, white blood cell count, and platelet count.

[0047] It should be noted that, in one embodiment of the present invention, the time range of the historical blood sample detection process is set to half a year, that is, patient data within half a year is obtained for analysis; in other embodiments of the present invention, the detection time range can be set according to specific circumstances, which is not limited or elaborated here.

[0048] It should be noted that, in order to facilitate the subsequent data processing and avoid differences between the units and magnitudes of the data, in the embodiments of the present invention, the data is standardized to eliminate the influence of the dimension in the data calculation. The specific means are well known to those skilled in the art and will not be described in detail here.

[0049] Step S2: according to the correlation coefficients of the medical file data of different dimensions between the re-examined patients, obtain the similarity of the medical files between the re-examined patients; according to the similarity of the medical files between the re-examined patients, cluster all the re-examined patients to obtain multiple re-examined patient clusters; according to the consistency of the indicator data sets between the automated detection and manual re-examination of the re-examined patients in each re-examined patient cluster, screen out the necessary re-examination patient clusters.

[0050] Medical file data may differ between patients, but they may also have similarities. By calculating the correlation coefficient, these differences and similarities can be quantified, and the commonalities and differences between patients can be better understood, thereby revealing the similarities of medical files between patients. Based on the correlation coefficients of medical file data of different dimensions between re-examined patients, the similarities of medical files between re-examined patients can be obtained.

[0051] Preferably, in one embodiment of the present invention, the method for obtaining medical file similarity includes:

[0052] Obtain a word segmentation array of each dimension of the medical file data of the re-examined patient, and obtain a correlation coefficient of the word segmentation array of the medical file data of the same dimension between the re-examined patients as the dimension array correlation;

[0053] The local data correlation of all medical file data of the same dimension between the re-examined patients was averaged and used as the medical file similarity between the re-examined patients.

[0054] In one embodiment of the present invention, the formula for medical file similarity is expressed as:

[0055] ;

[0056] in, Indicates and Similarity of medical records between re-examined patients; Indicates The first patient to be re-examined Dimensional word segmentation array of medical archive data; Indicates The first patient to be re-examined Dimensional word segmentation array of medical archive data; The number of dimensions representing the medical archive data; represents the Jaccard correlation coefficient.

[0057] In the formula for medical record similarity, It indicates the correlation coefficient of the word segmentation array of the same dimensional medical file data between the re-examined patients, that is, the dimensional array is correlated. The greater the dimensional array correlation, the more similar the word segmentation arrays of the medical file data of the same dimensionality are, the more likely they are patients with the same disease information characteristics, and the greater the similarity of the medical files.

[0058] It should be noted that, in one embodiment of the present invention, after obtaining the word segmentation array of each dimension of medical file data through Jieba word segmentation, the Jaccard correlation coefficient is used to calculate the similarity between the two arrays to obtain the corresponding correlation coefficient, and the value range is between 0 and 1. The specific Jieba word segmentation and Jaccard correlation coefficient are technical means well known to those skilled in the art and will not be elaborated here.

[0059] By clustering the re-examination patients with similar medical records into one category, we can discover the differences in diagnosis and treatment and common characteristics between different patient groups, which is helpful for targeted analysis of the patient's condition characteristics, thereby simplifying the analysis process and improving analysis efficiency; according to the similarity of the medical records between the re-examination patients, all the re-examination patients are clustered to obtain multiple re-examination patient clusters.

[0060] Preferably, in one embodiment of the present invention, the method for obtaining the cluster of re-examination patients includes:

[0061] According to the similarity of medical records between re-examination patients, K-means clustering was performed on all re-examination patients to obtain multiple re-examination patient clusters.

[0062] The K-means clustering algorithm clusters multiple cluster objects into specified K clusters based on the similarity between the cluster objects. Each cluster object belongs to and only belongs to one cluster with the smallest distance to the center of the cluster of re-examination patients.

[0063] It should be noted that, in one embodiment of the present invention, when the K-means clustering algorithm is used to cluster all re-examination patients, the method for obtaining the K value is: the K value is determined by the elbow rule to obtain a corresponding number of re-examination patient clusters. The specific K-means clustering algorithm and the elbow rule are technical means well known to those skilled in the art and will not be described in detail here.

[0064] By comparing the indicator data sets between manual re-examination and automated testing, the accuracy of automated testing can be reflected in a timely manner. Through consistency verification, clusters with poor consistency between automated testing results and manual re-examination results can be screened out. These clusters are regarded as patients who need re-examination, and targeted manual re-examinations are performed to reduce the amount of calculations. Based on the consistency of the indicator data sets between the automated test and manual re-examination of the re-examined patients in each cluster of re-examined patients, clusters of patients who need re-examination are screened out.

[0065] Preferably, in one embodiment of the present invention, the method for obtaining the cluster of patients who need to be re-examined can be found in Figure 2 , which shows a flow chart of a method for obtaining clusters of patients who need to be re-examined, including:

[0066] Step S201: For any re-examination patient cluster, the number of re-examination patients with the same indicator data set between the automated test and the manual re-examination is counted as the number of re-examination matches.

[0067] By counting the number of re-examination matches, we can identify the number of patients with the same indicator data set between automated testing and manual re-examination. The greater the number of re-examination matches, the more it indicates that the indicator data results of manual re-examination and automated testing are consistent, and the better the accuracy of automated testing technology.

[0068] Step S202: According to the difference characteristics between the number of re-examination matches and the number of all re-examination patients, the possibility of not needing re-examination for each re-examination patient cluster is obtained, and the difference characteristics are negatively correlated with the possibility of not needing re-examination.

[0069] The difference between the number of re-examination matches and the number of all re-examination patients reflects the degree of consistency between automated testing and manual re-examination. It can be inferred that automated testing is more reliable in clustering and reduce unnecessary re-examinations.

[0070] In one embodiment of the present invention, the formula for the possibility of not requiring recheck is expressed as:

[0071] ;

[0072] in, Indicates The probability that a cluster of re-examination patients does not need re-examination; Indicates The number of re-examination patients with the same indicator data set between automated testing and manual re-examination in the re-examination patient cluster, that is, the number of re-examination matches; Indicates The number of all re-examination patients in the re-examination patient cluster.

[0073] It should be noted that, in one embodiment of the present invention, the difference characteristics can be expressed by calculating the ratio between the number of re-examination matches and the number of all re-examination patients. The larger the ratio, the larger the number of re-examination matches, the smaller the difference between the number of re-examination matches and the number of all re-examination patients, and the greater the possibility that no re-examination is required. In other embodiments of the present invention, the difference can also be expressed by calculating the absolute value of the difference between the number of re-examination matches and the number of all re-examination patients. The smaller the absolute value of the difference, the smaller the difference between the number of re-examination matches and the number of all re-examination patients, the smaller the difference, and the greater the possibility that no re-examination is required.

[0074] Among them, if the difference characteristics are expressed by calculating the absolute value of the difference, it is necessary to perform a negative correlation mapping on the absolute value of the difference, so that the smaller the absolute value of the difference, the greater the possibility of not requiring re-inspection; the specific means are technical means well known to technical personnel in this field and will not be elaborated here.

[0075] Step S203: If the possibility of the re-examination patient cluster not requiring re-examination is less than a preset re-examination threshold, the corresponding re-examination patient cluster is used as a re-examination patient cluster that requires re-examination.

[0076] It should be noted that, in one embodiment of the present invention, the size of the preset re-inspection threshold is 0.5; in other embodiments of the present invention, the size of the preset re-inspection threshold can be set according to specific circumstances, which is not limited or elaborated here.

[0077] Step S3: For any necessary re-examination patient cluster, the re-examination importance of each indicator is obtained according to the data distribution of the automated detection indicator data sets of different re-examination patients; the device abnormality of each indicator in each necessary re-examination patient cluster is obtained according to the similarity of changes in the automated detection indicator data sets between each re-examination patient in the necessary re-examination patient cluster and different patients in the same batch.

[0078] The data distribution of the automated detection indicator data sets of different re-examination patients may be different, reflecting the changes in the disease manifestations of the indicators among different patients and evaluating the re-examination importance of each indicator; for any necessary re-examination patient cluster, the re-examination importance of each indicator is obtained based on the data distribution of the automated detection indicator data sets of different re-examination patients.

[0079] Preferably, in one embodiment of the present invention, the method for obtaining the re-examination importance level includes:

[0080] If the indicator data in the patient's automated detection indicator data set does not fall within the preset indicator threshold range, the corresponding indicator is regarded as an abnormal indicator;

[0081] For any cluster of patients who need to be re-examined, the ratio between the number of re-examined patients with abnormal indicators for each indicator and the number of all re-examined patients is obtained as the re-examination emphasis of each indicator.

[0082] In one embodiment of the present invention, the formula for the re-examination importance is expressed as:

[0083] ;

[0084] in, Indicates The first in the cluster of patients who need reexamination The importance of re-examination of each indicator; Indicates The number of all re-examination patients in the cluster of patients who need re-examination; Indicates The first in the cluster of patients who need reexamination The number of patients re-examined with abnormal indicators.

[0085] In the formula for re-examination importance, Indicates the amount of calculation; Indicates The first in the cluster of patients who need reexamination The first indicator is the ratio of the number of re-examination patients with abnormal indicators to the number of all re-examination patients. The larger the ratio, the The more patients with abnormal indicators are re-examined, the more the corresponding indicators need to be re-examined.

[0086] It should be noted that, in one embodiment of the present invention, the preset indicator threshold range is pre-acquired by the implementers based on relevant professional information and experience. If the indicator data falls within the preset indicator threshold range, the indicator is considered to be in a normal state; if the indicator data exceeds the preset indicator threshold range, that is, it does not fall within the preset indicator threshold range, the indicator is considered to be in an abnormal state.

[0087] Taking into account the situation that when there is equipment failure under automated testing, the equipment failure may cause similar abnormal changes in the indicator data of all patients in the same batch. By analyzing the similarity of changes in the automated testing indicator data sets between patients, the consistency of abnormal changes in indicators is reflected, which is helpful for the evaluation of equipment abnormalities. According to the similarity of changes in the automated testing indicator data sets between each re-examination patient in the necessary re-examination patient cluster and different patients in the same batch, the equipment abnormality of each indicator in each necessary re-examination patient cluster is obtained.

[0088] Preferably, in one embodiment of the present invention, the method for obtaining the device abnormality degree is as follows: Figure 3 , which shows a flow chart of a method for obtaining device abnormality, including:

[0089] Step S301: For any necessary re-examination patient cluster, the possibility of equipment failure for each re-examination patient is obtained according to the similarity of changes in the automated detection index data set between each re-examination patient and each patient in the same batch.

[0090] The similarity of changes in the automated detection indicator data sets between patients can reflect the overall abnormal similarity of different patients in the same batch. The greater the similarity, the greater the possibility of equipment failure.

[0091] Preferably, in one embodiment of the present invention, the method for obtaining the possibility of equipment failure includes:

[0092] For any cluster of patients who need to be re-examined, obtain the correlation coefficient of the set of all abnormal indicators between each re-examined patient and each patient in the same batch as the local abnormal correlation;

[0093] The mean of the local abnormality correlation between each re-examination patient and all patients in the same batch was calculated as the device failure probability of each re-examination patient in each required re-examination patient cluster.

[0094] In one embodiment of the present invention, the formula for the probability of equipment failure is expressed as:

[0095] ;

[0096] in, Indicates The first in the cluster of patients who need reexamination The probability of equipment failure for each re-examination patient; Indicates The first in the cluster of patients who need reexamination The set of all abnormal indicators of the re-examination patients; Indicates The first in the cluster of patients who need reexamination The same batch of re-examination patients The set of all abnormal indicators of a patient; Indicates The number of patients in the same batch of re-examination patients; represents the Jaccard correlation coefficient.

[0097] In the formula for the probability of equipment failure, Indicates the calculation The first in the cluster of patients who need reexamination The re-examination patients are from the same batch The correlation coefficient of the set of all abnormal indicators between patients, that is, the local abnormality correlation, the larger the correlation coefficient, the greater the local abnormality correlation, and the more consistent the abnormalities. If the mean value of the local abnormality correlation between the re-examined patient and all patients in the same batch is larger, all patients in the same batch have consistent abnormalities in their indicators, and are more likely to be affected by the overall impact of equipment failure, the possibility of equipment failure for each re-examined patient is greater.

[0098] Step S302: If the possibility of equipment failure of the re-examined patient is greater than the preset failure threshold, obtain the intersection of the abnormal indicators between the re-examined patient and all patients in the same batch; for each indicator in the intersection, the possibility of equipment failure of the corresponding re-examined patient is used as the local abnormality of the equipment of the re-examined patient under each indicator; for each indicator outside the intersection, the local abnormality of the equipment of the re-examined patient under each indicator is set to 0.

[0099] For the indicators of the same batch that are more consistently abnormal, that is, the indicators within the intersection, they are more likely to be affected by equipment failure, the smaller the possibility of true abnormality, the greater the possibility of equipment failure, and the greater the local abnormality of the equipment for the indicators affected by the equipment. The possibility of equipment failure for the corresponding re-examined patients is used as the local abnormality of the equipment for the re-examined patients under each indicator. The more inconsistent the abnormal indicators, that is, the indicators outside the intersection, the less affected they are by equipment failure, and the local abnormality of the equipment for the re-examined patients under each indicator is set to 0.

[0100] It should be noted that, in one embodiment of the present invention, the size of the preset fault threshold is 0.5. In other embodiments of the present invention, the size of the preset fault threshold can be set according to specific circumstances, which is not limited or elaborated herein.

[0101] Step S303: obtaining the mean of the local abnormality of the equipment of different re-examination patients under each indicator as the overall abnormality of the equipment for each indicator in the necessary re-examination patient cluster.

[0102] By calculating the mean, an overall analysis of the local abnormality of the equipment of all re-examined patients under each indicator in the necessary re-examination patient cluster is performed, which intuitively reflects the overall abnormality level of the corresponding cluster on this indicator, and helps to identify which clusters may have a higher risk of equipment abnormality. The greater the local abnormality of the equipment, the greater the overall abnormality of the equipment, and the more affected by the equipment, resulting in abnormal results.

[0103] In one embodiment of the present invention, the formula for the overall abnormality of the equipment is expressed as:

[0104] ;

[0105] in, Indicates The first in the cluster of patients who need reexamination The overall abnormality of the equipment for each indicator; Indicates The first in the cluster of patients who need reexamination Under the indicator The local abnormality of the device in each re-examination patient; Indicates The number of re-examination patients in the cluster of patients who need re-examination.

[0106] Step S4: Correct the re-examination importance according to the device abnormality of each indicator in each necessary re-examination patient cluster, and obtain the re-examination corrected importance of each indicator in each necessary re-examination patient cluster.

[0107] The device abnormality reflects the degree to which each indicator is affected by the device abnormality. The greater the device abnormality, the more likely the re-examination will be affected by the device abnormality. By correcting the re-examination importance of the indicator by the device abnormality, it is possible to more accurately determine which indicators need to be re-examined first, thereby improving the efficiency and accuracy of the re-examination work. Therefore, according to the device abnormality of each indicator in each necessary re-examination patient cluster, the re-examination importance is corrected to obtain the re-examination corrected importance of each indicator in each necessary re-examination patient cluster.

[0108] Preferably, in one embodiment of the present invention, the method for obtaining the importance of re-inspection and correction includes:

[0109] Calculate the difference between the positive integer 1 and the device abnormality of each indicator in each cluster of patients who need to be re-examined as the adjustment weight;

[0110] The re-examination importance of each indicator in each cluster of patients who need re-examination is weighted according to the adjustment weight to obtain the re-examination revised importance of each indicator in each cluster of patients who need re-examination.

[0111] The weighted acquisition method is as follows: obtain the product of the re-examination importance of each indicator in each cluster of patients who need re-examination and the adjustment weight as the re-examination corrected importance of each indicator in each cluster of patients who need re-examination.

[0112] In one embodiment of the present invention, the formula for rechecking and correcting the importance is expressed as:

[0113] ;

[0114] in, Indicates The first in the cluster of patients who need reexamination The importance of rechecking and correcting each indicator; Indicates The first in the cluster of patients who need reexamination The importance of rechecking each indicator; Indicates The first in the cluster of patients who need reexamination The abnormality of the device.

[0115] In the formula for rechecking the corrected importance, It means to calculate the positive integer 1 and the The first in the cluster of patients who need reexamination The difference in the device abnormality of the first indicator, The first in the cluster of patients who need reexamination The adjustment weight of each indicator is used to make a weighted adjustment on the importance of re-inspection. The greater the equipment abnormality, the more likely it is that the re-inspection is caused by the influence of the equipment. The smaller the adjustment weight, the smaller the importance of re-inspection is, and the smaller the importance of re-inspection correction is.

[0116] Based on this, after obtaining the re-examination and correction importance of all indicators in each cluster of patients who need to be re-examined, all indicators will be sorted in order from large to small according to the re-examination and correction importance, which will be used as the indicator order for subsequent re-examination of patients' blood samples to improve the efficiency of manual re-examination.

[0117] It should be noted that, in addition to obtaining the order of indicators detected in different clusters of patients who need re-examination, in another embodiment of the present invention, it is also possible to determine whether the real-time patient's blood sample needs manual re-examination, including: obtaining the real-time patient's automated detection indicator data set under full laboratory automated detection, and for patients judged to need manual re-examination, obtaining the mean of the medical record similarity between the patient and all patients in each re-examination patient cluster as the average medical record similarity; selecting the patient with the largest average medical record similarity value between the patient and different re-examination patient clusters, and using the corresponding re-examination patient cluster as the patient's matching cluster; if the matching cluster does not correspond to the re-examination patient cluster, it is finally determined that the patient does not need manual re-examination; if the matching cluster corresponds to the re-examination patient cluster, it is finally determined that the patient needs manual re-examination, and the blood sample is tested according to the indicator order in the corresponding re-examination patient cluster.

[0118] In summary, the present invention obtains the medical file similarity between re-examination patients based on the correlation coefficient of medical file data of different dimensions between re-examination patients, and obtains multiple re-examination patient clusters; according to the consistency of the indicator data set between the automated detection and manual re-examination of the re-examination patients in each re-examination patient cluster, the necessary re-examination patient clusters are screened out; for any necessary re-examination patient cluster, the data distribution of the automated detection indicator data set of different re-examination patients is analyzed to obtain the re-examination importance and equipment abnormality of each indicator; and the re-examination corrected importance of each indicator in each necessary re-examination patient cluster is obtained. The present invention determines the detection order of re-examination by obtaining the accurate re-examination importance of each indicator in the cluster, thereby improving the efficiency of blood sample detection.

[0119] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0120] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A blood sample efficient detection method for laboratory, characterized in that: The method comprises: Obtain multi-dimensional medical archive data of patients who were manually re-examined during the historical blood sample testing process of each batch, as well as the indicator data sets under the corresponding automated testing and manual re-examination; According to the correlation coefficients of the medical file data of different dimensions between the re-examination patients, the similarity of the medical files between the re-examination patients is obtained; according to the similarity of the medical files between the re-examination patients, all the re-examination patients are clustered to obtain multiple re-examination patient clusters; according to the consistency of the indicator data sets between the automated detection and manual re-examination of the re-examination patients in each re-examination patient cluster, the necessary re-examination patient clusters are screened out; For any necessary re-examination patient cluster, the re-examination importance of each indicator is obtained based on the data distribution of the automated detection indicator data sets of different re-examination patients; the device abnormality of each indicator in each necessary re-examination patient cluster is obtained based on the similarity of changes in the automated detection indicator data sets between each re-examination patient in the necessary re-examination patient cluster and different patients in the same batch; Correcting the re-examination importance according to the device abnormality of each indicator in each necessary re-examination patient cluster, and obtaining the re-examination corrected importance of each indicator in each necessary re-examination patient cluster; The multi-dimensional medical file data includes chief complaint, diagnosis, current medical history, test results, and past medical history; The method for obtaining the medical record similarity includes: Obtain a word segmentation array of each dimension of the medical file data of the re-examined patient, and obtain a correlation coefficient of the word segmentation array of the medical file data of the same dimension between the re-examined patients as the dimension array correlation; The dimensional array correlations of all medical file data of the same dimension between the re-examined patients are averaged as the medical file similarity between the re-examined patients; The method for obtaining the re-examination importance level includes: If the indicator data in the patient's automated detection indicator data set does not fall within the preset indicator threshold range, the corresponding indicator is regarded as an abnormal indicator; For any cluster of patients who need to be re-examined, the ratio between the number of re-examined patients with abnormal indicators for each indicator and the number of all re-examined patients is obtained as the re-examination emphasis of each indicator.

2. The method for efficiently detecting blood samples used in a laboratory according to claim 1, characterized in that: The method for obtaining the re-examination patient clusters comprises: According to the similarity of medical records between re-examination patients, K-means clustering was performed on all re-examination patients to obtain multiple re-examination patient clusters.

3. The method for efficient blood sample detection for laboratory according to claim 1, characterized in that: The method for obtaining the necessary re-examination patient clusters includes: For any cluster of re-examination patients, the number of re-examination patients with the same indicator data set between automated testing and manual re-examination is counted as the number of re-examination matches; According to the difference characteristics between the number of re-examination matches and the number of all re-examination patients, the possibility of no need for re-examination for each re-examination patient cluster was obtained, and the difference characteristics were negatively correlated with the possibility of no need for re-examination. If the possibility of a re-examination patient cluster not requiring re-examination is less than a preset re-examination threshold, the corresponding re-examination patient cluster will be regarded as a re-examination patient cluster that requires re-examination.

4. The method for efficiently detecting blood samples used in a laboratory according to claim 1, characterized in that: The method for obtaining the device abnormality degree includes: For any necessary re-examination patient cluster, the possibility of equipment failure for each re-examination patient is obtained based on the similarity of changes in the automated detection indicator data set between each re-examination patient and each patient in the same batch; If the probability of equipment failure of the re-examined patient is greater than the preset failure threshold, the intersection of abnormal indicators between the re-examined patient and all other patients in the same batch is obtained; for each indicator in the intersection, the probability of equipment failure of the corresponding re-examined patient is used as the local abnormality of the equipment of the re-examined patient under each indicator; for each indicator outside the intersection, the local abnormality of the equipment of the re-examined patient under each indicator is set to 0; The mean of the local abnormality of the equipment of different re-examination patients under each indicator is obtained as the overall abnormality of the equipment for each indicator in the necessary re-examination patient cluster.

5. The method for efficiently detecting blood samples used in a laboratory according to claim 4, characterized in that: The method for obtaining the possibility of equipment failure includes: For any cluster of patients who need to be re-examined, obtain the correlation coefficient of the set of all abnormal indicators between each re-examined patient and each patient in the same batch as the local abnormal correlation; The mean of the local abnormality correlation between each re-examination patient and all patients in the same batch was calculated as the device failure probability of each re-examination patient in each required re-examination patient cluster.

6. The method for efficient blood sample detection for laboratory according to claim 1, characterized in that: The method for obtaining the importance of rechecking and correcting comprises: Calculate the difference between the positive integer 1 and the device abnormality of each indicator in each cluster of patients who need to be re-examined as the adjustment weight; The re-examination importance of each indicator in each cluster of patients who need re-examination is weighted according to the adjustment weight to obtain the re-examination revised importance of each indicator in each cluster of patients who need re-examination.

7. The method for efficiently detecting blood samples used in a laboratory according to claim 6, characterized in that: The step of weighting the re-examination importance of each indicator in each cluster of patients who need to be re-examined according to the adjustment weight to obtain the re-examination corrected importance of each indicator in each cluster of patients who need to be re-examined includes: The product of the re-examination importance and the adjustment weight of each indicator in each cluster of patients who need re-examination is obtained as the re-examination corrected importance of each indicator in each cluster of patients who need re-examination.

8. The method for efficient detection of blood samples for laboratory according to claim 1, characterized in that: The method for obtaining the correlation coefficient is the Jaccard correlation coefficient.

Images