Clinical immunodetection automatic auditing rule construction method, system and equipment

Abstract

The invention relates to the field of intelligent medical treatment, in particular to a method, equipment and a system for constructing a clinical immunodetection automatic auditing rule. Comprising the following steps: S1, acquiring a clinical immunodetection report; s2, performing first rule extraction on the immune report to obtain a first rule label; s3, performing automatic auditing based on the first rule label to obtain an auditing missing number; and S4, obtaining a missed checking report, performing second rule extraction to obtain a second rule label, performing automatic checking based on the first rule label and the second rule label, and repeating the steps S3-S4 until the number of missed checking is smaller than a preset threshold value to obtain a final checking rule label. The traditional Chinese medicine composition has good clinical value.

Description

Technical Field

[0001] This application relates to the field of intelligent medicine, and specifically relates to a method, device, program product, and computer-readable storage medium for constructing an automatic review rule for clinical immunological tests. Background Art

[0002] Automatic audit means the automatic confirmation of test results and the automatic release of test reports, which refers to the process of automatically performing operations on laboratory results through a series of clear and standard computer algorithms without manual intervention. The main limitation of automatic audit is in the field of automated medical testing, and there are few reports on the automatic audit of clinical immunological tests. In recent years, many medical institutions have implemented automatic audit in ways such as self-coding self-algorithms, middleware, or self-construction. Immunological tests mainly rely on manual operations, with insufficient standardized operations during analysis, and the risks of false positives and false negatives are higher than those of automation, affecting the accuracy of reports. Summary of the Invention

[0003] In view of the above problems, the present invention provides a method for constructing an automatic review rule for clinical immunological tests, which specifically includes: S1. Obtain a clinical immunological test report; S2. Extract a first rule from the immunological report to obtain a first rule label; S3. Perform an automatic review based on the first rule label to obtain the number of missed reviews; S4. Obtain the missed review reports, extract a second rule to obtain a second rule label, perform an automatic review based on the first rule label and the second rule label, and repeat steps S3 - S4 until the number of missed reviews is less than a preset threshold to obtain the final review rule label.

[0004] Step S1 of the method is replaced with: Obtain a clinical immunological test report, the number of reports passed by manual review, and the number of reports intercepted by manual review; Step S3 is replaced with: Perform an automatic review based on the first rule label to obtain the number of reports passed by automatic review and the number of reports intercepted by automatic review; Step S4 is replaced with: Compare the consistency and non-consistency between the number of reports passed by manual review and the number of reports passed by automatic review, and between the number of reports intercepted by manual review and the number of reports intercepted by automatic review; Adjust the first rule label based on the non-consistency to obtain an adjusted rule label, perform an automatic review and non-consistency calculation based on the adjusted rule label, and repeat this step until the non-consistency is less than a preset threshold to obtain the final review rule label; Optionally, S1 further includes obtaining the number of reports passed by manual review and the number of reports intercepted by manual review. In S3, automatic review further includes obtaining the number of reports passed by automatic review and the number of reports intercepted by automatic review, and comparing the consistency and non - consistency between the number of reports passed by manual review and the number of reports passed by automatic review, and between the number of reports intercepted by manual review and the number of reports intercepted by automatic review; based on the non - consistency, perform the first rule label adjustment to obtain the adjusted rule label, and in S4, perform automatic review based on the adjusted rule label and the second rule label.

[0005] The first rule extraction is to extract the first rule label through the first neural network model; Optionally, the first neural network model is trained through medical common sense; Optionally, the second rule extraction is to extract the second rule label through the second neural network model Optionally, the second neural network model is trained through medical experience texts; Optionally, the neural network model adopts one or more of the following: convolutional neural network, bidirectional long - short - term memory neural network, residual network, GRU, Transformer.

[0006] The object of the present invention is to provide a construction system for automatic review rules of clinical immune detection, including: Acquisition module: acquire clinical immune detection reports; First rule module: perform the first rule extraction on the immune report to obtain the first rule label; Review module: perform automatic review based on the first rule label to obtain the number of missed reviews; Optimization module: acquire the missed - review reports to perform the second rule extraction to obtain the second rule label, perform automatic review based on the first rule label and the second rule label, and repeat steps S3 - S4 until the number of missed reviews is less than the preset threshold to obtain the final review rule label.

[0007] The object of the present invention is to provide an intelligent review method for clinical immune detection reports, including: Acquire clinical immune detection reports; The report is automatically reviewed through the final rule label obtained by the above - mentioned construction method of automatic review rules for clinical immune detection to obtain the result of automatic review passing or being intercepted.

[0008] The final rule label includes: hepatitis B group rule label, HIV rule label, hepatitis C label rule, syphilis label rule, other rule labels; Optionally, the hepatitis B group rule label includes one or more of the following: strongly positive hepatitis B surface antigen, HBsAg gray zone, rare pattern, extremely rare pattern; Optionally, the HIV rule label is that when the HIV result is greater than or equal to 1 C.O.I., manual review is required; Optionally, the hepatitis C rule label is that when the HCVAb result is greater than or equal to 0.3 C.O.I., manual review is required; Optionally, the syphilis rule label is that when the TPAb result is greater than or equal to 0.2 C.O.I., manual review is required; Optionally, the other rule label is that the result has an asterisk.

[0009] The rare patterns include one or more of the following: HBeAg alone positive (3), HBsAb positive, HBeAb positive, HBcAb negative (2, 4), HBsAg alone positive (1), HBsAg, HBeAb positive (1, 4), HBsAg, HBeAg, HBeAb, HBcAb positive (1, 3, 4, 5), HBsAg, HBsAb, HBcAb positive (1, 2, 5), HBsAg, HBsAb, HBeAb, HBcAb positive (1, 2, 4, 5), HBeAg and HBcAb positive simultaneously (3, 5), HBeAg, HBeAb, HBcAb positive (3, 4, 5), HBsAb, HBeAg (2, 3), HBsAb, HBeAg, HBcAb (2, 3, 5), HBsAg, HBsAb, HBeAb positive (1, 2, 4), HBsAg, HBeAg positive (1, 3), HBeAb alone positive (4), HBeAg and HBeAb positive simultaneously (3, 4), HBsAg and HBsAb positive simultaneously (1, 2); Optionally, the extremely rare patterns include one or more of the following: HBsAg, HBsAb, HBeAg, HBeAb, HBcAb positive (1, 2, 3, 4, 5), HBsAb, HBeAg, HBeAb positive (2, 3, 4), HBsAb, HBeAg, HBeAb, HBcAb positive (2, 3, 4, 5), HBsAg, HBeAg, HBeAb positive (1, 3, 4), HBsAg, HBsAb, HBeAg, HBeAb positive (1, 2, 3, 4), HBsAg, HBsAb, HBeAg positive (1, 2, 3).

[0010] The final rule label is replaced with: HIV rule label, second syphilis rule label, second hepatitis B group rule label, second hepatitis C rule label, second other rule label; Optionally, for the second syphilis rule label, when the TP result is greater than or equal to 0.7, the TPPA or RPR test is performed; Optionally, for the second hepatitis C rule label, when the HCVAb result is between 1 and 5.25, manual review is required; Optionally, the second hepatitis B group rule label includes one or more of the following: (4, 5) positive, (1, 2) double positive, surface antibody greater than 1000, (1, 3, 5) positive big three yang, (1, 5) positive, e antibody single positive, (2, 5) positive, (2, 4, 5) positive, (1, 4, 5) positive small three yang, (2, 5) double positive, surface antibody greater than 1000, isolated core antibody positive, HBsAg gray zone, second rare pattern, second extremely rare pattern; Optionally, the second rare pattern includes one or more of the following: HBsAb positive, HBeAb positive, HBcAb negative (2, 4), HBeAb isolated positive (4), HBeAg isolated positive (3), HBsAb, HBeAg, HBcAb (2, 3, 5), HBsAb, HBeAg (2, 3), HBeAg and HBeAb simultaneously positive (3, 4), HBeAg, HBeAb, HBcAb positive (3, 4, 5), HBsAg, HBsAb, HBeAb positive (1, 2, 4), HBeAg and HBcAb simultaneously positive (3, 5), HBsAg and HBsAb simultaneously positive (1, 2), HBsAg, HBsAb, HBeAb, HBcAb positive (1, 2, 4, 5), HBsAg, HBsAb, HBcAb positive (1, 2, 5), HBsAg, HBeAg, HBeAb, HBcAb positive (1, 3, 4, 5), HBsAg, HBeAg positive (1, 3), HBsAg, HBeAb positive (1, 4), HBsAg isolated positive (1); Optionally, the second extremely rare pattern includes one or more of the following: HBsAb, HBeAg, HBeAb positive (2, 3, 4), HBsAg, HBsAb, HBeAg, HBeAb, HBcAb positive (1, 2, 3, 4, 5), HBsAb, HBeAg, HBeAb, HBcAb positive (2, 3, 4, 5), HBsAg, HBeAg, HBeAb positive (1, 3, 4), HBsAg, HBsAb, HBeAg, HBeAb positive (1, 2, 3, 4), HBsAg, HBsAb, HBeAg positive (1, 2, 3); Optionally, the second other rule label is that the detection result exceeds the upper limit, the result is "XXX", and the result has an asterisk; Optionally, the syphilis rule label further includes performing a TPPA or RPR test on results where the TP result is greater than or equal to 0.7; Optionally, the hepatitis C rule label further includes that when the HCVAb result is between 1 - 5.25, manual review is required; Optionally, the hepatitis B group rule label also includes one or more of the following: (4, 5) positive, (1, 2) double positive, surface antibody greater than 1000, (1, 3, 5) positive big triple positive, (1, 5) positive, e antibody single positive, (2, 5) positive, (2, 4, 5) positive, (1, 4, 5) positive small triple positive, (2, 5) double positive, surface antibody greater than 1000, single core antibody positive; Optionally, the other rule tags also include the detection result exceeding the upper limit and the result being "XXX".

[0011] The object of the present invention is to provide an intelligent audit system for clinical immunoassay reports, comprising: Report module: obtain clinical immunoassay reports; Automatic review module: The report is automatically reviewed by the final rule label obtained through the above-mentioned method for constructing automatic review rules for clinical immunoassays to obtain a result of automatic review passing or interception.

[0012] The object of the present invention is to provide a computer program product, which includes a computer program or instructions, and the computer program or instructions are executed by a processor to implement the above-mentioned method for constructing automatic review rules for clinical immunoassays or the intelligent review method for clinical immunoassay reports.

[0013] The object of the present invention is to provide a computer device, which includes a memory, a processor and a computer program or instructions stored on the memory, wherein the computer program or instructions are executed by the processor to implement the above-mentioned method for constructing automatic review rules for clinical immunoassays or the intelligent review method for clinical immunoassay reports.

[0014] The object of the present invention is to provide a computer-readable storage medium having a computer program or instructions stored thereon, wherein the computer program or instructions are executed by a processor to implement the above-mentioned method for constructing automatic review rules for clinical immunoassays or the intelligent review method for clinical immunoassay reports.

[0015] Advantages of the present invention: 1. In view of the lack of automatic review rules and processes in clinical immunology testing reports, and the automatic review can improve the efficiency of report review. Based on this, the present invention proposes a method for constructing automatic review rules for clinical immunology testing, which is used to construct review rules for clinical immunology reports, so that the accuracy of automatic review reaches the accuracy of manual review, while saving report review personnel, reducing report review time, and providing test results for other surgical treatments and disease treatments more quickly.

[0016] 2. Regarding the problem that the automatic review rules for the general public are not applicable to the clinical immune reports of pregnant women in maternity hospitals (the clinical immune detection requirements for pregnant women are different from those of the general population, mainly due to the physiological changes during pregnancy, such as hormone levels, immune regulation, blood dilution, and the influence of detection strategies), the clinical immune detection of the present invention mainly constructs and applies automatic review rules based on the clinical immune detection reports of pregnant women, which helps to distinguish pregnant women from the general population, making the review of clinical detection reports more applicable to the detection reports of pregnant women, reducing the omission and misjudgment rates caused by the automatic review system constructed for the general population, improving the detection guarantee for pregnant women, and having good clinical application value.

[0017] 3. The method for constructing immune detection rules based on pregnant women in the present invention constructs 28 and 39 clinical immune detection rules. The 28 - rule and 39 - rule are experimentally compared with 34 clinical immune detection rules for the general population. The 28 - rule and 34 - rule have similar report passing rates and omission rates, but the number of rules is less, and it has higher efficiency in the overall review process. The 39 - rule has a lower report omission rate compared with the 34 - rule. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following - described drawings are only some embodiments of the present invention. For those skilled in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0019] Figure 1 It is a schematic flowchart of the method for constructing automatic review rules for clinical immune detection provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of the system for constructing automatic review rules for clinical immune detection provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of the device for constructing automatic review rules for clinical immune detection provided by an embodiment of the present invention; Figure 4 It is the result of the 28 - rule provided by an embodiment of the present invention; Figure 5 It is the result of the 34 - rule provided by an embodiment of the present invention; Figure 6 It is the result of the 39 - rule provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] In order to enable those skilled in the art to better understand the solution of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention.

[0021] In some of the processes described in the specification, claims and above-mentioned drawings of the present invention, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear herein or may be executed in parallel. The serial numbers of the operations, such as S101, S102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0022] Figure 1 The schematic diagram of the method for constructing the automatic review rules for clinical immunological tests provided by the embodiments of the present invention specifically includes: S1: Obtain a clinical immunological test report; In a specific embodiment, in the antibody tests for hepatitis B, syphilis, acquired immunodeficiency syndrome (HIV), and hepatitis C virus (HCV) in pregnant women, the risks of false positives or false negatives are different from those in the general population, mainly affected by physiological changes during pregnancy (such as hormone levels, immune regulation, blood dilution) and detection strategies. The following is a specific analysis: I. Hepatitis B (HBV) test; False positive risk: The false positive rate of the hepatitis B surface antigen (HBsAg) test is relatively low, but the following situations need to be noted: Blood dilution during pregnancy may lead to weak positive results, and repeated testing or combination with HBV DNA confirmation is required. Cross-reactions may occur in some pregnant women due to immune activation, but they are rare.

[0023] False negative risk: Extremely low. The HBsAg test has high sensitivity and a small risk of missed diagnosis. However, if tested in the very early stage of infection (window period), false negatives may occur.

[0024] II. Syphilis (TP) test; False positive risk: Significantly higher than that in the general population, especially in non-specific tests (such as RPR / TRUST): Physiological anti-phospholipid antibodies, rheumatoid factors, etc. during pregnancy may cause cross-reactions. Confirmation through specific tests (such as TPPA) is required to avoid misdiagnosis.

[0025] False negative risk: Rare. However, if tested in the early stage of infection (window period, about 3 - 4 weeks after infection), the absence of antibody production may lead to false negatives.

[0026] III. Acquired immunodeficiency syndrome (HIV) test; False positive risk: Relatively low, but non-specific cross-reactions may occur in pregnant women due to changes in the immune state, and confirmation through nucleic acid testing or immunoblotting is required.

[0027] Risk of false negatives: Be highly vigilant! If a pregnant woman is tested during the infection window period (before antibodies are produced, about 2 to 6 weeks after infection), the antibody test may be false negative.

[0028] 4. Hepatitis C (HCV) testing; Risk of false positive: Low, but pregnant women may have weak positive results due to blood dilution or immune interference, and HCV RNA testing is required to confirm whether it is an active infection.

[0029] Risk of false negative: If the test is conducted during the infection window period (before antibodies are produced, approximately 8 to 11 weeks after infection), antibodies may not have been produced, resulting in a false negative.

[0030] The main risks of false positive / false negative tests for pregnant women are shown in Table 1 below: Table 1

[0031]

[0032] Recommendations: Measures to reduce false positives / false negatives: 1. Testing time: - First screening in early pregnancy and retesting in late pregnancy (especially for syphilis and AIDS).

[0033] - Increase the frequency of testing for high-risk groups (e.g., those with multiple sexual partners, intravenous drug users).

[0034] 2. Optimization of detection methods: - Syphilis: combined nonspecific (RPR) and specific tests (TPPA).

[0035] - AIDS: After the initial antibody screening is positive, nucleic acid testing or immunoblotting should be performed immediately; high-risk pregnant women should be directly combined with nucleic acid testing.

[0036] - Hepatitis C: Those who are antibody positive must be tested for HCV RNA.

[0037] 3. Comprehensive clinical judgment: - Combined with epidemiological history, symptoms and other laboratory indicators (such as liver function).

[0038] - Infected pregnant women require multidisciplinary management (obstetrics, infectious diseases, and pediatrics).

[0039] The special feature of pregnant women testing is that the health of both mother and child must be protected at the same time. Therefore, it is necessary to be more cautious in eliminating the risk of false positives / false negatives and ensure timely intervention to block mother-to-child transmission.

[0040] The main risks of false positive / false negative in the general population are shown in Table 2 below:

[0041] Table 2

[0042]

[0043] The risks of false positives / false negatives in the detection of the general population are mainly related to the limitations of the detection method, the window period, and the individual immune status. The risk of misdiagnosis needs to be reduced by standardizing operations, reasonably selecting detection strategies, and dynamically rechecking.

[0044] In one embodiment, the clinical immune detection report is the clinical immune detection report of the pregnant population.

[0045] In a specific embodiment, specimens for the preoperative eight-item detection during the period from 20231101 to 20240223 are selected and randomly divided into two groups for detection, and a total of 17,868 groups of detection data are obtained.

[0046] S2: Extract the first rule label from the immune report. In one embodiment, the first rule extraction is performed by a first neural network model to obtain the first rule label.

[0047] The first neural network model is trained with medical common sense.

[0048] The training process of the first neural network model is as follows: Obtain the medical common sense dataset and labels of clinical immunity. Input the medical common sense dataset and labels of clinical immunity into the first neural network model to be trained for feature training until the loss function of the model remains unchanged, and the first neural network model is obtained.

[0049] In one embodiment, the neural network model adopts one or more of the following: convolutional neural network, bidirectional long short-term memory neural network, residual network, GRU, Transformer.

[0050] S3: Perform automatic review based on the first rule label to obtain the number of missed reviews. In one embodiment, the first rule label is input into the automatic review system, and the automatic review system reviews the clinical immune report based on the first rule label to obtain the number of reviews, the number of missed reviews; it also includes the number of reviews passed.

[0051] In a specific embodiment, the automatic review system adopts the Laboman system and the supporting detection reagents that are supporting the HISCL-5000 chemiluminescence immunoassay analyzer of Sysmex Corporation.

[0052] S4: Obtain the missed review report, perform second rule extraction to obtain second rule tags, and conduct automatic review based on the first rule tags and the second rule tags. Repeat steps S3 - S4 until the number of missed reviews is less than the preset threshold to obtain the final review rule tags.

[0053] In one embodiment, the second rule extraction is performed by a second neural network model to obtain second rule tags.

[0054] The second neural network model is trained with medical experience texts.

[0055] The training process of the second neural network model is as follows: Obtain the medical experience text dataset and labels of clinical immunology; Input the medical experience text dataset and labels of clinical immunology into the second neural network model to be trained for feature training until the loss function of the model remains unchanged, obtaining the second neural network model.

[0056] Among them, the medical experience texts of clinical immunology are texts statistically analyzed and labeled by doctors who have been engaged in the review of clinical immunology tests for more than ten years.

[0057] In one embodiment, method S1 is replaced with: Obtain the clinical immunology test report, the number of reports passed by manual review, and the number of reports intercepted by manual review; S3 is replaced with: Conduct automatic review based on the first rule tags to obtain the number of reports passed by automatic review and the number of reports intercepted by automatic review; S4 is replaced with: Compare the consistency and non - consistency between the number of reports passed by manual review and the number of reports passed by automatic review, and between the number of reports intercepted by manual review and the number of reports intercepted by automatic review; Adjust the first rule tags based on the non - consistency to obtain the adjusted rule tags, and conduct automatic review and non - consistency calculation based on the adjusted rule tags. Repeat this step until the non - consistency is less than the preset threshold to obtain the final review rule tags.

[0058] In one embodiment, S1 further includes obtaining the number of reports passed by manual review and the number of reports intercepted by manual review. In S3, the automatic review also includes obtaining the number of reports passed by automatic review and the number of reports intercepted by automatic review, and comparing the consistency and non - consistency between the number of reports passed by manual review and the number of reports passed by automatic review, and between the number of reports intercepted by manual review and the number of reports intercepted by automatic review; Adjust the first rule tags based on the non - consistency to obtain the adjusted rule tags. In S4, conduct automatic review based on the adjusted rule tags and the second rule tags.

[0059] In a specific embodiment, 17,868 sets of data are randomly and unequally allocated into two groups. The first group is the research group, with a total of 8,865 sets of data, and the second group is the verification group, with a total of 9,003 sets of data. The 8,865 sets of data in the research group are manually reviewed, and the number of manually reviewed passes and the number of manually reviewed interceptions are counted. Summarize the interception rules according to medical common sense, complete the interception rule tags, enter the interception rule tags into the Laboman system, and determine the automatic review program. Then, perform automatic review by this system to obtain the number of automatic review passes and the number of automatic review interceptions. Taking the manual review as the standard, compare the above-mentioned number of manually reviewed passes and the number of manually reviewed interceptions, the number of automatic review passes and the number of automatic review interceptions. Obtain the number of detection consistencies, the number of inconsistencies (invalid interceptions), and the number of missed reviews. Analyze the missed review results again, summarize the interception rules according to daily work experience, and complete the new interception rule tags. Repeat the above operations to obtain the new number of detection consistencies, the number of inconsistencies (invalid interceptions), and the number of missed reviews.

[0060] Verification: Enter all the summarized rules into the database of the verification group to obtain the number of automatic review passes, the number of automatic review interceptions, the number of consistencies, the number of inconsistencies (invalid interceptions), and the number of missed reviews of the verification group. Compare the number of automatic review passes, the number of automatic review interceptions, the number of consistencies, the number of inconsistencies (invalid interceptions), and the number of missed reviews of the research group with the number of automatic review passes, the number of automatic review interceptions, the number of consistencies, the number of inconsistencies (invalid interceptions), and the number of missed reviews of the verification group to obtain the verification result.

[0061] The disclosed embodiment of the present invention also provides a computer program product or system, including a computer program, which when executed by a processor implements the steps of the above-mentioned method for constructing an automatic review rule for clinical immunological tests.

[0062] Figure 2 The schematic diagram of the system for constructing an automatic review rule for clinical immunological tests provided by the embodiment of the present invention specifically includes: Acquisition module: Acquire a clinical immunological test report; First rule module: Extract the first rule from the immunological report to obtain the first rule tag; Review module: Perform automatic review based on the first rule tag to obtain the number of missed reviews; Optimization module: Acquire the missed review report, extract the second rule to obtain the second rule tag, perform automatic review based on the first rule tag and the second rule tag, and repeat steps S3 - S4 until the number of missed reviews is less than a preset threshold to obtain the final review rule tag.

[0063] The embodiment of the present invention provides an intelligent review method for a clinical immunological test report, including: Acquire a clinical immunological test report; The above-mentioned report is automatically audited by using the final rule tags obtained through the above-mentioned method for constructing the automatic audit rules for clinical immunological tests to obtain the result of automatic audit passing or interception.

[0064] In one embodiment, the final rule tags include: hepatitis B group rule tags, HIV rule tags, hepatitis C rule tags, syphilis rule tags, and other rule tags; Optionally, the hepatitis B group rule tags include one or more of the following: strongly positive hepatitis B surface antigen, HBsAg gray zone, rare pattern, and extremely rare pattern; Optionally, the HIV rule tag is that when the HIV result is greater than or equal to 1 C.O.I., manual review is required; Optionally, the hepatitis C rule tag is that when the HCVAb result is greater than or equal to 0.3 C.O.I., manual review is required; Optionally, the syphilis rule tag is that when the TPAb result is greater than or equal to 0.2 C.O.I., manual review is required; Optionally, the other rule tag is that the result has an asterisk.

[0065] The rare pattern includes one or more of the following: HBeAg alone is positive (3), HBsAb is positive, HBeAb is positive, HBcAb is negative (2, 4), HBsAg alone is positive (1), HBsAg, HBeAb are positive (1, 4), HBsAg, HBeAg, HBeAb, HBcAb are positive (1, 3, 4, 5), HBsAg, HBsAb, HBcAb are positive (1, 2, 5), HBsAg, HBsAb, HBeAb, HBcAb are positive (1, 2, 4, 5), HBeAg and HBcAb are simultaneously positive (3, 5), HBeAg, HBeAb, HBcAb are positive (3, 4, 5), HBsAb, HBeAg (2, 3), HBsAb, HBeAg, HBcAb (2, 3, 5), HBsAg, HBsAb, HBeAb are positive (1, 2, 4), HBsAg, HBeAg are positive (1, 3), HBeAb alone is positive (4), HBeAg and HBeAb are simultaneously positive (3, 4), HBsAg and HBsAb are simultaneously positive (1, 2); Optionally, the rare patterns include one or more of the following: HBsAg, HBsAb, HBeAg, HBeAb, HBcAb positive (1, 2, 3, 4, 5), HBsAb, HBeAg, HBeAb positive (2, 3, 4), HBsAb, HBeAg, HBeAb, HBcAb positive (2, 3, 4, 5), HBsAg, HBeAg, HBeAb positive (1, 3, 4), HBsAg, HBsAb, HBeAg, HBeAb positive (1, 2, 3, 4), HBsAg, HBsAb, HBeAg, HBeAg positive (1, 2, 3).

[0066] In another embodiment, the syphilis rule tag further includes a TP result greater than or equal to 0.7 for a TPPA or RPR test; Optionally, the hepatitis C rule label also includes that the HCVAb result is between 1-5.25 and requires manual review; Optionally, the hepatitis B group rule label also includes one or more of the following: (4, 5) positive, (1, 2) double positive, surface antibody greater than 1000, (1, 3, 5) positive big triple positive, (1, 5) positive, e antibody single positive, (2, 5) positive, (2, 4, 5) positive, (1, 4, 5) positive small triple positive, (2, 5) double positive, surface antibody greater than 1000, single core antibody positive; Optionally, the other rule tags also include the detection result exceeding the upper limit and the result being "XXX".

[0067] In another embodiment, the last rule tag is replaced by: HIV rule tag, second syphilis rule tag, second hepatitis B group rule tag, second hepatitis C rule tag, second other rule tag; Optionally, the second syphilis rule label is a TPPA or RPR test for a TP result greater than or equal to 0.7; Optionally, the second hepatitis C rule label is that the HCVAb result is between 1-5.25 and requires manual review; Optionally, the second hepatitis B group rule label includes one or more of the following: (4, 5) positive, (1, 2) double positive, surface antibody greater than 1000, (1, 3, 5) positive big triple positive, (1, 5) positive, e antibody single positive, (2, 5) positive, (2, 4, 5) positive, (1, 4, 5) positive small triple positive, (2, 5) double positive, surface antibody greater than 1000, single core antibody positive, HBsAg gray area, second rare pattern, second rare pattern; Optionally, the second uncommon pattern includes one or more of the following: HBsAb positive, HBeAb positive, HBcAb negative (2, 4), HBeAb positive alone (4), HBeAg positive alone (3), HBsAb, HBeAg, HBcAb (2, 3, 5), HBsAb, HBeAg (2, 3), HBeAg and HBeAb positive at the same time (3, 4), HBeAg, HBeAb, HBcAb positive (3, 4, 5), HBsAg, HBsAb, HBeAb positive (1, 2, 4), HBeAg and HBcAb are positive at the same time (3, 5), HBsAg and HBsAb are positive at the same time (1, 2), HBsAg, HBsAb, HBeAb, HBcAb are positive (1, 2, 4, 5), HBsAg, HBsAb, HBcAb are positive (1, 2, 5), HBsAg, HBeAg, HBeAb, HBcAb are positive (1, 3, 4, 5), HBsAg, HBeAg, HBeAb, HBcAb are positive (1, 3), HBsAg, HBeAb are positive (1, 4), HBsAg is positive alone (1); Optionally, the second rare pattern includes one or more of the following: HBsAb, HBeAg, HBeAb positive (2, 3, 4), HBsAg, HBsAb, HBeAg, HBeAb, HBcAb positive (1, 2, 3, 4, 5), HBsAb, HBeAg, HBeAb, HBcAb positive (2, 3, 4, 5), HBsAg, HBeAg, HBeAb positive (1, 3, 4), HBsAg, HBsAb, HBeAg, HBeAb positive (1, 2, 3, 4), HBsAg, HBsAb, HBeAg, HBeAg positive (1, 2, 3); Optionally, the second other rule label is that the detection result exceeds the upper limit, the result is "XXX", and the result is marked with an asterisk.

[0068] In a specific embodiment, the present invention constructs 28 rules and 39 rules through a construction method, as shown in Table 3 and Table 4. High-sensitivity chemiluminescence method detects five items of hepatitis B, syphilis, hepatitis C, and AIDS, a total of 8 items of preoperative detection, and the instrument uses Sysmex high-sensitivity chemiluminometer HISCL5000, supporting reagents, quality control products and calibrators. The Laboman system of Sysmex HISCL5000 has 34 audit rules, which is suitable for the general population. The 34 rules are replaced with 28 rules or 39 rules for automatic audit of clinical immune test reports of pregnant women.

[0069] Table 3 28 rules for clinical immunization audit

[0070] Twenty-eight rules are automatically audited through Machine 1 and Machine 2, and the results are as Figure 4 shown. The non-audit rates are 13.09% and 12.9%, which are almost equal to the non-audit rate audit results of 34 rules (as Figure 5 shown). Later, the strongly positive hepatitis B surface antigen was deleted, and the HCVAb result ≥ 0.3 C.O.I. requires manual review; the TPAb result ≥ 0.2 C.O.I. requires manual review. Retest rules were added for TP samples and HCV samples, and audit rules were added for TP samples and HCV samples. Finally, the non-audit rates of 39 audit rules are 2.54% and 2.53%, as Figure 6 shown, which greatly reduces the non-audit rate and significantly improves the consistency rate with manual audits.

[0071] Table 4 Thirty-nine rules for clinical immunology audits

[0072] An embodiment of the present invention provides an intelligent audit system for clinical immunology test reports, including: Report module: Obtain clinical immunology test reports; Automatic audit module: The report is automatically audited through the final rule tags obtained by the method for constructing the automatic audit rules for clinical immunology tests described in claims 1-3 to obtain the results of automatic audit passing or interception.

[0073] Figure 3 A schematic diagram of a computer device provided by an embodiment of the present invention specifically includes: A memory and a processor; the memory is used to store program instructions; the processor is used to call program instructions, and when the program instructions are executed, any one of the above-mentioned methods for constructing automatic audit rules for clinical immunology tests or the intelligent audit method for clinical immunology test reports.

[0074] An embodiment of the present invention also provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, any one of the above-mentioned methods for constructing automatic audit rules for clinical immunology tests or the intelligent audit method for clinical immunology test reports.

[0075] The verification results of this verification embodiment show that allocating fixed weights for indications can improve the performance of this method compared to the default settings. Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be elaborated here. In several embodiments provided by this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of devices or units can be in electrical, mechanical, or other forms. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in each embodiment of the present invention, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units. Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program, and this program can be stored in a computer-readable storage medium. The storage medium can include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), magnetic disk, or optical disc, etc.

[0076] Those of ordinary skill in the art can understand that all or part of the steps in implementing the methods of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. The above-mentioned medium storage can be read-only memory, magnetic disk, or optical disc, etc.

[0077] The above has introduced in detail a computer device provided by the present invention. For those of ordinary skill in the art, according to the idea of the embodiments of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for constructing automatic review rules for clinical immunoassays, characterized in that: include: S1. Obtain clinical immunoassay report; S2. Performing a first rule extraction on the immunization report to obtain a first rule label; S3, automatically review based on the first rule label to obtain the number of missed reviews; S4. Obtain the missed review report and perform second rule extraction to obtain the second rule label. Perform automatic review based on the first rule label and the second rule label. Repeat steps S3-S4 until the number of missed reviews is less than the preset threshold, and obtain the final review rule label.

2. The method for constructing automatic audit rules for clinical immunoassay according to claim 1, characterized in that: The method S1 is replaced by: obtaining a clinical immunoassay report, a number of reports that have been manually reviewed and a number of reports that have been intercepted by manual review; S3 is replaced by: performing automatic review based on a first rule label to obtain a number of automatic review passes and a number of automatic review intercepts; S4 is replaced by: comparing the consistency and inconsistency of the number of manual review passes and the number of automatic review passes, the number of manual review intercepts and the number of automatic review intercepts; adjusting the first rule label based on the inconsistency to obtain an adjusted rule label, performing automatic review and inconsistency calculation based on the adjusted rule label, and repeating this step until the inconsistency is less than a preset threshold, thereby obtaining a final review rule label; Optionally, S1 also includes obtaining the number of reports that have passed manual review and the number of reports that have been intercepted for manual review, and automatic review in S3 also includes obtaining the number of reports that have passed automatic review and the number of reports that have been intercepted for automatic review, and comparing the consistency and inconsistency between the number of reports that have passed manual review and the number of reports that have passed automatic review, and between the number of reports that have been intercepted for manual review and the number of reports that have been intercepted for automatic review; adjusting the first rule label based on the inconsistency to obtain an adjusted rule label, and performing automatic review based on the adjusted rule label and the second rule label in S4.

3. The method for constructing automatic audit rules for clinical immunoassay according to claim 1, characterized in that: The first rule extraction is performed by extracting a first neural network model to obtain a first rule label; Optionally, the first neural network model is obtained by training through medical common sense; Optionally, the second rule extraction is performed by extracting a second rule label through a second neural network model. Optionally, the second neural network model is obtained by training with medical experience text; Optionally, the neural network model adopts one or more of the following: convolutional neural network, bidirectional long short-term memory neural network, residual network, GRU, Transformer.

4. An intelligent audit method for clinical immunoassay reports, characterized in that: include: Obtain clinical immunoassay reports; The report is automatically audited by the final rule label obtained by the method for constructing automatic audit rules for clinical immunoassays described in claims 1-3 to obtain an automatic audit pass or intercept result.

5. The intelligent audit method for clinical immunoassay report according to claim 4, characterized in that: The final rule labels include: hepatitis B group rule labels, HIV rule labels, hepatitis C label rules, syphilis label rules, and other rule labels; Optionally, the hepatitis B group rule labels include one or more of the following: strong positive hepatitis B surface antigen, HBsAg gray area, uncommon pattern, rare pattern; Optionally, the HIV rule label is HIV result greater than or equal to 1 COI requires manual review; Optionally, the hepatitis C rule label is that HCVAb results greater than or equal to 0.3 COI require manual review; Optionally, the syphilis rule label is that the TPAb result is greater than or equal to 0.2 COI and requires manual review; Optionally, the other rule label is a result with an asterisk.

6. The intelligent audit method for clinical immunoassay report according to claim 5, characterized in that: The rare patterns include one or more of the following: HBeAg alone positive (3), HBsAb positive, HBeAb positive, HBcAb negative (2, 4), HBsAg positive alone (1), HBsAg and HBeAb positive (1, 4), HBsAg, HBeAg, HBeAb, and HBcAb positive (1, 3, 4, 5), HBsAg, HBsAb, and HBcAb positive (1, 2, 5), HBsAg, HBsAb, HBeAb, and HBcAb positive (1, 2, 4, 5), HBeAg and HBcAb positive at the same time (3, 5), HBeAg, HBeAb, and HBcAb positive (3, 4, 5), HBsAb and HBeAg (2, 3), HBsAb, HBeAg, HBcAb (2, 3, 5), HBsAg, HBsAb, and HBeAb positive (1, 2, 4), HBsAg and HBeAg positive (1, 3), HBeAb positive alone (4), HBeAg and HBeAb positive at the same time (3, 4), HBsAg and HBsAb positive at the same time (1, 2); Optionally, the rare patterns include one or more of the following: HBsAg, HBsAb, HBeAg, HBeAb, HBcAb positive (1, 2, 3, 4, 5), HBsAb, HBeAg, HBeAb positive (2, 3, 4), HBsAb, HBeAg, HBeAb, HBcAb positive (2, 3, 4, 5), HBsAg, HBeAg, HBeAb positive (1, 3, 4), HBsAg, HBsAb, HBeAg, HBeAb positive (1, 2, 3, 4), HBsAg, HBsAb, HBeAg, HBeAg positive (1, 2, 3).

7. The intelligent audit method for clinical immunoassay report according to claim 4, characterized in that: The last rule tag is replaced by: HIV rule tag, second syphilis rule tag, second hepatitis B group rule tag, second hepatitis C rule tag, second other rule tag; Optionally, the second syphilis rule label is a TPPA or RPR test for a TP result greater than or equal to 0.7; Optionally, the second hepatitis C rule label is that the HCVAb result is between 1-5.25 and requires manual review; Optionally, the second hepatitis B group rule label includes one or more of the following: (4, 5) positive, (1, 2) double positive, surface antibody greater than 1000, (1, 3, 5) positive big triple positive, (1, 5) positive, e antibody single positive, (2, 5) positive, (2, 4, 5) positive, (1, 4, 5) positive small triple positive, (2, 5) double positive, surface antibody greater than 1000, single core antibody positive, HBsAg gray area, second rare pattern, second rare pattern; Optionally, the second uncommon pattern includes one or more of the following: HBsAb positive, HBeAb positive, HBcAb negative (2, 4), HBeAb positive alone (4), HBeAg positive alone (3), HBsAb, HBeAg, HBcAb (2, 3, 5), HBsAb, HBeAg (2, 3), HBeAg and HBeAb positive at the same time (3, 4), HBeAg, HBeAb, HBcAb positive (3, 4, 5), HBsAg, HBsAb, HBeAb positive (1, 2, 4), HBeAg and HBcAb positive at the same time (3 , 5), HBsAg and HBsAb are simultaneously positive (1, 2), HBsAg, HBsAb, HBeAb, HBcAb are positive (1, 2, 4, 5), HBsAg, HBsAb, HBcAb are positive (1, 2, 5), HBsAg, HBeAg, HBeAb, HBcAb are positive (1, 3, 4, 5), HBsAg, HBeAg are positive (1, 3), HBsAg, HBeAb are positive (1, 4), HBsAg is positive alone (1); Optionally, the second rare pattern includes one or more of the following: HBsAb, HBeAg, HBeAb positive (2, 3, 4), HBsAg, HBsAb, HBeAg, HBeAb, HBcAb positive (1, 2, 3, 4, 5), HBsAb, HBeAg, HBeAb, HBcAb positive (2, 3, 4, 5), HBsAg, HBeAg, HBeAb positive (1, 3, 4), HBsAg, HBsAb, HBeAg, HBeAb positive (1, 2, 3, 4), HBsAg, HBsAb, HBeAg, HBeAg positive (1, 2, 3); Optionally, the second other rule label is that the detection result exceeds the upper limit, the result is "XXX", and the result is asterisked; Optionally, the syphilis rule label also includes a TP result greater than or equal to 0.7 for a TPPA or RPR test; Optionally, the hepatitis C rule label also includes that the HCVAb result is between 1-5.25 and requires manual review; Optionally, the hepatitis B group rule label also includes one or more of the following: (4, 5) positive, (1, 2) double positive, surface antibody greater than 1000, (1, 3, 5) positive big triple positive, (1, 5) positive, e antibody single positive, (2, 5) positive, (2, 4, 5) positive, (1, 4, 5) positive small triple positive, (2, 5) double positive, surface antibody greater than 1000, single core antibody positive; Optionally, the other rule tags also include the detection result exceeding the upper limit and the result is "XXX".

8. A computer program product comprising a computer program or instructions, characterized in that: The computer program or instructions are executed by a processor to implement the method for constructing automatic review rules for clinical immunoassays as described in any one of claims 1-3, or the intelligent review method for clinical immunoassay reports as described in any one of claims 4-7.

9. A computer device comprising a memory, a processor and a computer program or instruction stored in the memory, characterized in that: The computer program or instructions are executed by a processor to implement the method for constructing automatic review rules for clinical immunoassays as described in any one of claims 1-3, or the intelligent review method for clinical immunoassay reports as described in any one of claims 4-7.

10. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: The computer program or instructions are executed by a processor to implement the method for constructing automatic review rules for clinical immunoassays as described in any one of claims 1-3, or the intelligent review method for clinical immunoassay reports as described in any one of claims 4-7.

Images