Medical Test Item Reference Interval Appropriateness Evaluation System

Abstract

This application discloses a system for evaluating the suitability of reference intervals for medical test items. The system for evaluating the suitability of reference intervals for medical test items evaluates the abnormality rate of the reference interval and the distribution of abnormality levels of test results through a preset reference interval and a reference interval obtained by calculating based on a set of physical examination data, and also evaluates the consistency between the preset reference interval and the literature reference interval of the test item, and then generates an evaluation result on the suitability of the preset reference interval for each test item. In this way, the system for evaluating the suitability of reference intervals for medical test items can evaluate the suitability of the preset reference interval, which is of great significance for value-based medicine. The data extracted by the system for evaluating the suitability of reference intervals for medical test items comes from real-world test data, without the need to recruit additional subjects for testing, and can greatly reduce the cost of evaluating the suitability of reference intervals.

Description

Technical Field

[0001] The present application relates to the field of reference interval evaluation, and particularly to a system for evaluating the suitability of reference intervals for medical test items. Background Art

[0002] The reference intervals of various test items used in medical laboratories mainly come from reagent specifications, health industry standards, authoritative literature or textbooks, etc. Due to differences in social and economic levels, regional living environments and eating habits, etc., the existing reference intervals of various detection items may not be applicable to the population served by the laboratory. When the reference interval is too wide, it may lead to delays in the treatment of some patients; when the reference interval is too narrow, it will cause excessive examinations of patients and waste of medical resources. Therefore, regular evaluation of the suitability of reference intervals is of great significance for value-based medicine.

[0003] Currently, for the suitability evaluation of reference intervals: at least 20 or 60 qualified reference individuals are screened, and tested according to the standard operating procedure to obtain preliminary test results; the outliers in the preliminary test results are checked and removed; when the proportion of the number of reference individuals whose preliminary test results after removing outliers are within the preset reference interval ≥ 90%, it is regarded as passing the verification of the preset reference interval; when the verification of the preset reference interval passes, it is considered that the preset reference interval is suitable.

[0004] However, the above method for evaluating the suitability of preset reference intervals has high costs, cumbersome processes, less data volume, deviation in the selection of reference individuals, and does not combine the large test data in the real world or the latest research literature progress. The reliability of the evaluation results is poor, and there is an urgent need for a new evaluation scheme for quickly and accurately evaluating the suitability of reference intervals. Summary of the Invention

[0005] One advantage of the present application is to provide a system for evaluating the suitability of reference intervals for medical test items. Among them, the system for evaluating the suitability of reference intervals for medical test items can evaluate the suitability of the preset reference interval, which is of great significance for value-based medicine.

[0006] One advantage of the present application is to provide a system for evaluating the suitability of reference intervals for medical test items. Among them, the data selected by the system for evaluating the suitability of reference intervals for medical test items comes from the test data in the real world, and there is no need to recruit additional subjects for testing, which can greatly reduce the cost of evaluating the suitability of reference intervals.

[0007] One advantage of the present application is to provide a system for evaluating the suitability of reference intervals for medical test items. Among them, the system for evaluating the suitability of reference intervals for medical test items can also evaluate the suitability of the reference intervals in the literature.

[0008] According to one aspect of the present application, a system for evaluating the suitability of reference intervals for medical test items is provided, which includes:

[0009] A preset reference interval extraction module for extracting the preset reference intervals of at least one test item from a test item reference interval database;

[0010] A physical examination data collection and analysis module for obtaining a set of physical examination data and determining the reconstructed reference intervals of the test items by using the EP28 non-parametric method;

[0011] A test data collection module for extracting a test data set of the test items within a predetermined time period from a laboratory information system;

[0012] A data grouping module for splitting the test data set of the test items according to three types of medical visits, namely physical examination, outpatient service, and inpatient service, to obtain a first test data subset, a second test data subset, and a third test data subset;

[0013] A reference interval abnormality rate evaluation module for calculating the upper limit abnormality rate and the lower limit abnormality rate of the first test data subset, the second test data subset, and the third test data subset with respect to the preset reference interval and the reconstructed reference interval respectively to obtain a reference interval abnormality rate evaluation result;

[0014] A suitability management module for generating a suitability evaluation result of the test items based on the reference interval abnormality rate evaluation result.

[0015] In the system for evaluating the suitability of reference intervals for medical test items according to the present application, the reference interval abnormality rate evaluation module includes:

[0016] A preset reference interval abnormality rate evaluation unit for calculating the upper limit abnormality rate and the lower limit abnormality rate of the preset reference intervals corresponding to the first test data subset, the second test data subset, and the third test data subset respectively to obtain a preset reference interval abnormality rate evaluation result;

[0017] A reconstructed reference interval abnormality rate evaluation unit for calculating the upper limit abnormality rate and the lower limit abnormality rate of the reconstructed reference intervals corresponding to the first test data subset, the second test data subset, and the third test data subset respectively to obtain a reconstructed reference interval abnormality rate evaluation result.

[0018] In the system for evaluating the suitability of reference intervals for medical test items according to the present application, the suitability management module includes:

[0019] A consistency evaluation unit for performing consistency evaluation and analysis on the preset reference interval abnormality rate evaluation result and the reconstructed reference interval abnormality rate evaluation result to obtain a consistency evaluation result;

[0020] An appropriateness evaluation result determination unit for generating an appropriateness evaluation result of the test item based on the consistency evaluation result, the preset and reconstructed reference interval abnormality rate evaluation results.

[0021] In the medical test item reference interval appropriateness evaluation system according to the present application, the consistency evaluation unit is further configured to:

[0022] Calculate the difference degree between the upper limit abnormality rate of the reconstructed reference interval and the upper limit abnormality rate of the preset reference interval and calculate the difference degree between the lower limit abnormality rate of the reconstructed reference interval and the lower limit abnormality rate of the preset reference interval to obtain a first difference degree and a second difference degree;

[0023] Generate the consistency evaluation result based on the comparison between the first difference degree and the second difference degree and a preset threshold.

[0024] In the medical test item reference interval appropriateness evaluation system according to the present application, the consistency evaluation unit is further configured to: calculate the ratio between the upper limit abnormality rate and the lower limit abnormality rate of the preset reference interval of any two of the first test data subset, the second test data subset, and the third test data subset to obtain a plurality of abnormality rate ratios; generate the consistency evaluation result based on the comparison between each abnormality rate ratio in the plurality of abnormality rate ratios and a preset threshold.

[0025] In the medical test item reference interval appropriateness evaluation system according to the present application, it further includes: an abnormal level judgment criterion determination module for determining an abnormal level judgment criterion based on the set of physical examination data; an abnormal level calculation module for analyzing the second test data subset and the third test data subset based on the abnormal level criterion to obtain the abnormal level distribution of the outpatient test results and the abnormal level distribution of the inpatient test results.

[0026] In the medical test item reference interval appropriateness evaluation system according to the present application, the abnormal level judgment criterion determination module is further configured to:

[0027] Determine the upper and lower limits of the normal level based on the grade division condition that the set distribution of the physical examination data is greater than 2.5% or less than 97.5% bilaterally or the grade division condition that the set distribution of the physical examination data is less than 95% unilaterally;

[0028] Determine the upper and lower limits of the mild level based on the grading conditions where the set distribution of the physical examination data is greater than 1.5% and less than or equal to 2.5% or greater than or equal to 97.5% and less than 98.5% bilaterally, or based on the grading conditions where the set distribution of the physical examination data is greater than or equal to 95% and less than 97% unilaterally;

[0029] Determine the upper and lower limits of the moderate level based on the grading conditions where the set distribution of the physical examination data is greater than 1.0% and less than or equal to 1.5% or greater than or equal to 98.5% and less than 99.0% bilaterally, or based on the grading conditions where the set distribution of the physical examination data is greater than or equal to 97.0% and less than 98.0% unilaterally;

[0030] Determine the upper and lower limits of the severe level based on the grading conditions where the set distribution of the physical examination data is greater than 0.5% and less than or equal to 1.0% or greater than or equal to 99.0% and less than 99.5% bilaterally, or based on the grading conditions where the set distribution of the physical examination data is greater than or equal to 98.0% and less than 99.0% unilaterally;

[0031] Determine the upper and lower limits of the extreme level based on the grading conditions where the set distribution of the physical examination data is less than or equal to 0.5% or greater than or equal to 99.5% bilaterally, or based on the grading conditions where the set distribution of the physical examination data is greater than or equal to 99.0% unilaterally.

[0032] In the medical test item reference interval suitability evaluation system according to the present application, it further includes: a reference collection module for searching for references; a reference level evaluation module for evaluating the quality of the references to obtain a reference quality level evaluation result; and a reference reference interval management module for extracting and managing relevant data on the recommended reference interval for the test item from the references based on a large language model.

[0033] In the medical test item reference interval suitability evaluation system according to the present application, the reference reference interval management module is further configured to:

[0034] Calculate the bias ratio between the recommended reference interval of the test item and the reconstructed reference interval;

[0035] When the bias ratio is greater than a preset threshold, it is determined that the consistency between the recommended reference interval of the test item and the reconstructed reference interval of the test item does not meet the preset requirements;

[0036] When the bias ratio is less than or equal to the preset threshold, it is determined that the consistency between the recommended reference interval of the test item and the reconstructed reference interval of the test item meets the preset requirements.

[0037] Through the understanding of the subsequent description and drawings, the further objectives and advantages of the present application will be fully reflected.

[0038] These and other objects, features, and advantages of the present application will be fully embodied in the following detailed description, the accompanying drawings, and the claims. Description of the Drawings

[0039] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other objects, features, and advantages of the present application will become more obvious. The accompanying drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation to the present application. In the accompanying drawings, the same reference numerals generally represent the same components.

[0040] Figure 1 It is a block diagram of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0041] Figure 2 It is a schematic diagram of an interface for calculating the EP28 non-parametric method of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0042] Figure 3 It is a schematic diagram of an interface showing the preset reference interval of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0043] Figure 4 It is another schematic diagram of an interface showing the preset reference interval of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0044] Figure 5 It is a block diagram of an adaptability management module in a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0045] Figure 6 It is a schematic diagram of an interface for calculating the abnormal rate level of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0046] Figure 7 It is another schematic diagram of an interface for calculating the abnormal rate level of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0047] Figure 8 It is a schematic diagram of an interface for determining the abnormal level standard of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0048] Figure 9 It is another schematic diagram of an interface for determining the abnormal level standard of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0049] Figure 10 It shows a schematic diagram of an interface of a reference interval literature list of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0050] Figure 11 It shows another schematic diagram of an interface of reference interval literature content management of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0051] Figure 12 It shows a schematic diagram of an interface of reference interval literature comparison of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application.

[0052] Figure 13 It shows another schematic diagram of an interface of reference interval literature comparison of a reference interval suitability evaluation system for medical test items according to an embodiment of the present application. Detailed implementation manners

[0053] Next, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0054] It can be understood that the term "one" should be understood as "at least one" or "one or more". That is, in one embodiment, the number of one element can be one, while in other embodiments, the number of the element can be multiple. The term "one" cannot be understood as a limitation on the number. "Multiple" means greater than or equal to two.

[0055] Although ordinal numbers such as "first", "second", etc. will be used to describe various components, those components are not limited herein. The term is only used to distinguish one component from another. For example, the first component can be called the second component, and similarly, the second component can also be called the first component without departing from the teachings of the concept of the present application. The term "and / or" used herein includes any and all combinations of one or more of the associated listed items.

[0056] The terms used herein are only for the purpose of describing various embodiments and are not intended to be limiting. As used herein, the singular form also includes the plural form unless the context clearly indicates otherwise. Additionally, it will be understood that the terms "comprise" and / or "have" when used in this specification specify the presence of the described features, numbers, operations, components, elements, or combinations thereof, without excluding the presence or addition of one or more other features, numbers, operations, components, elements, or combinations thereof.

[0057] As Figures 1 to 13As shown, a medical test item reference interval suitability evaluation system according to an embodiment of the present application is illustrated. The medical test item reference interval suitability evaluation system 1000 can evaluate the suitability of a preset reference interval, which is of great significance to value-based medicine. The data extracted by the medical test item reference interval suitability evaluation system 1000 comes from real-world test data, and there is no need to recruit additional subjects for testing, which can greatly reduce the cost of evaluating the suitability of the reference interval. The medical test item reference interval suitability evaluation system 1000 can also evaluate the suitability of the reference interval in the literature.

[0058] Specifically, as Figure 1 shown, the medical test item reference interval suitability evaluation system 1000 includes a preset reference interval extraction module 10, a physical examination data collection and analysis module 20, a test data collection module 30, a data grouping module 40, a reference interval abnormality rate evaluation module 50, and a suitability management module 60. Among them, the preset reference interval extraction module 10 is used to extract the preset reference interval of at least one test item from the test item reference interval database; the physical examination data collection and analysis module 20 is used to obtain a set of physical examination data and determine the reconstructed reference interval of the test item by using the EP28 non-parametric method; the test data collection module 30 is used to extract the test data set of the test item within a predetermined time period from the laboratory information system; the data grouping module 40 is used to perform data segmentation on the test data set of the test item according to three visit types of physical examination, outpatient, and inpatient to obtain a first test data subset, a second test data subset, and a third test data subset; the reference interval abnormality rate evaluation module 50 is used to calculate the upper limit abnormality rate and the lower limit abnormality rate of the first test data subset, the second test data subset, and the third test data subset relative to the preset reference interval and the reconstructed reference interval respectively to obtain a reference interval abnormality rate evaluation result; the suitability management module 60 is used to generate a suitability evaluation result of the test item based on the reference interval abnormality rate evaluation result.

[0059] Specifically, during the operation of the preset reference interval extraction module, that is, the preset reference interval of at least one test item is extracted from the test item reference interval database through the preset reference interval extraction module 10. The preset reference interval of the at least one test item includes, but is not limited to, the preset reference interval of platelet count, the preset reference interval of total protein, the preset reference interval of albumin, the preset reference interval of alanine aminotransferase, the preset reference interval of aspartate aminotransferase, etc.

[0060] It is worth mentioning that the preset reference intervals for each of the item test items can also be divided into multiple groups. For example, the preset reference interval for the platelet count test item includes the preset reference interval for the platelet count test item in the first group, the preset reference interval for the platelet count test item in the second group, the preset reference interval for the platelet count test item in the third group, the preset reference interval for the platelet count test item in the fourth group, the preset reference interval for the platelet count test item in the fifth group, the preset reference interval for the platelet count test item in the sixth group, and the preset reference intervals for the platelet count test items in more groups. In a specific example, the preset reference interval for each group corresponds to a group of test subjects, and the test subjects are grouped according to age and gender; for example, the preset reference interval for the platelet count test item in the first group is the preset reference interval for the platelet count test item in the group of female 6 months - <1 year old; the preset reference interval for the platelet count test item in the second group is the preset reference interval for the platelet count test item in the group of female 1 year - <2 years old; the preset reference interval for the platelet count test item in the third group is the preset reference interval for the platelet count test item in the group of female 2 years - <6 years old; the preset reference interval for the platelet count test item in the fourth group is the preset reference interval for the platelet count test item in the group of female 6 years - <12 years old; the preset reference interval for the platelet count test item in the fifth group is the preset reference interval for the platelet count test item in the group of female 12 years - <18 years old; the preset reference interval for the platelet count test item in the sixth group is the preset reference interval for the platelet count test item in the group of female 18 years - <53 years old; the preset reference intervals for the platelet count test items in other groups are for the test subjects corresponding to other groups. It is also worth mentioning that in other examples, the data in the test item reference interval database can be first grouped according to the test subjects, and then grouped according to the test item types.

[0061] During the operation of the physical examination data collection and analysis module 20, it first obtains a set of physical examination data and processes the set of physical examination data using the EP28 non-parametric method to determine the reconstructed reference intervals for at least one test item, such as Figure 2As shown. In addition to the EP28 non-parametric method, other suitable methods for establishing reference intervals such as the EP28 parametric method, refineR, TMC, Kosmic, etc. may also be used. Those skilled in the art should be aware that the EP28 non-parametric method includes the exclusion of abnormal physical examination data. Specifically, the abnormal physical examination data exclusion process includes: deleting incomplete data, selecting only the most recent result from multiple test results of the same individual, deleting data with positive results (chest X-ray, ultrasound, CT, etc.), and using the LAVE method to exclude data that exceeds the reference interval (such as WBC, Hb, MCV, MCHC, ALT, GGT, BUN, Cr, CEA, AFP, CA19-9, etc.). After using the Tukey method to exclude out-of-threshold values, the remaining set of physical examination data includes n test data. The n test data can be arranged in ascending order, and the n test data after ascending order can be used separately , , …, express, , To test the minimum value in the project data, is the maximum value in the test item data. Divide the n test data into 100 equal parts, and the number corresponding to the r% rank is the rth percentile, represented by the symbol Pr. The rank of the reference lower limit value and the reference upper limit value of the reconstructed reference interval can be represented by P 2.5 and P 97.5 Indicates that, where is calculated by the following formula: and If the calculated values ​​are not integers, they can be rounded to the nearest integer. The confidence interval for the lower limit of the reconstructed reference interval is P 1 -P 5 , the confidence interval of the upper limit of the reconstructed reference interval is P 95 -P 99 .

[0062] Accordingly, the EP28 non-parametric method can be used to obtain the reconstructed reference intervals for each test item in the physical examination category. For example, the reconstructed reference interval for platelet count in the physical examination category, the reconstructed reference interval for total protein in the physical examination category, the reconstructed reference interval for albumin test item in the physical examination category, the reconstructed reference interval for alanine aminotransferase test item in the physical examination category, the reconstructed reference interval for aspartate aminotransferase test item in the physical examination category, etc. Accordingly, the reconstructed reference interval for each test item in the physical examination category can be divided into multiple grouped reconstructed reference intervals for the test item. For example, the reconstructed reference interval for platelet count in the physical examination category includes the reconstructed reference interval for platelet count in the first group in the physical examination category, the reconstructed reference interval for platelet count in the second group in the physical examination category, the reconstructed reference interval for platelet count in the third group in the physical examination category, the reconstructed reference interval for platelet count in the fourth group in the physical examination category, the reconstructed reference interval for platelet count in the fifth group in the physical examination category, the reconstructed reference interval for platelet count in the sixth group in the physical examination category, and the reconstructed reference intervals for platelet count in more groups in the physical examination category. For example, the reconstructed reference interval for each group in the physical examination category corresponds to a group of test subjects, and the test subjects are grouped according to age and gender. For example, the reference interval for platelet count in the first group in the physical examination category is for test subjects who are females aged 6 months - <1 year; the reference interval for platelet count in the second group in the physical examination category is for test subjects who are females aged 1 year - <2 years; the reference interval for platelet count in the third group in the physical examination category is for test subjects who are females aged 2 years - <6 years; the reference interval for platelet count in the fourth group in the physical examination category is for test subjects who are females aged 6 years - <12 years; the reference interval for platelet count in the fifth group in the physical examination category is for test subjects who are females aged 12 years - <18 years; the reference interval for platelet count in the sixth group in the physical examination category is for test subjects who are females aged 18 years - <53 years; the reference intervals for platelet count in other groups in the physical examination category correspond to test subjects in other groups.

[0063] During the operation of the test data collection module, that is, at least one test data set of test items within a predetermined time period is extracted from the laboratory information system through the test data collection module 30. In particular, compared with the existing reference interval adaptability evaluation process, test data from the real world can be extracted from the laboratory information system, and there is no need to recruit additional subjects for testing, thereby greatly reducing the cost of evaluating the suitability of the reference interval.

[0064] In a specific implementation, the test data collection module 30 first extracts the test data set within a predetermined time period from the laboratory information system, and then groups the test data set within the predetermined time period according to test subjects (in a specific example, the test subjects are grouped according to age and gender), such as Figure 3 and Figure 4As shown, the data is then grouped according to the type of medical visit, and then further grouped according to the type of test item.

[0065] Similarly, the test data set within the predetermined time period can be sliced into multiple grouped test data subsets. For example, the first grouped test data subset, the second grouped test data subset, the third grouped test data subset, the fourth grouped test data subset, the fifth grouped test data subset, the sixth grouped test data subset, and more grouped test data subsets. The first grouped test data subset has a test object of female, aged 6 months - <1 year; the second grouped test data subset has a test object of female, aged 1 year - <2 years; the third grouped test data subset has a test object of female, aged 2 years - <6 years; the fourth grouped test data subset has a test object of female, aged 6 years - <12 years; the fifth grouped test data subset has a test object of female, aged 12 years - <18 years; the sixth grouped test data subset has a test object of female, aged 18 years - <53; the test data subsets of other groups have test objects of other grouped test data subsets.

[0066] Each grouped test data subset includes a first type of test item subset with a medical visit type of physical examination, a second type of test item subset with a medical visit type of outpatient, and a third type of test item subset with a medical visit type of inpatient. For example, the first grouped test data subset of the group with a test object of female, aged 12 years - <18 years includes a first type of test item subset, a second type of test item subset, and a third type of test item subset; the test data subset of the group with a test object of female, aged 18 years - <53 years includes another first type of test item subset, another second type of test item subset, and another third type of test item subset.

[0067] Each type of test item subset of each grouped test data subset with each medical visit type includes test data of at least one test item. Correspondingly, the first type of test item subset of each grouped test data subset includes at least one single test item data subset of the physical examination category; the second type of test item subset of each grouped test data subset includes at least one single test item data subset of the outpatient category; the third type of test item subset of each grouped test data subset includes at least one single test item data subset of the inpatient category; for example, the first type of test item subset with a medical visit type of physical examination in the first grouped test data subset of the group with a test object of female, aged 18 years - <53 years includes test data of the platelet count test item under the physical examination category for the group with a test object of female, aged 18 years - <53 years. It should be understood that the test data set extracted from the laboratory information system within the predetermined time period can also be first sliced according to the grouping of the test objects (the test objects are grouped according to age and gender), then sliced according to the test item data, and then sliced according to the medical visit type.

[0068] Accordingly, in the process of the data grouping module 40 splitting the test data set of each test item according to three types of medical visits, namely physical examination, outpatient service, and inpatient service, to obtain a first test data subset, a second test data subset, and a third test data subset, the first test data subset in the test data set of each test item is the test data subset of the test item with the type of physical examination; the second test data subset in the test data set of each test item is the test data subset of the test item with the type of outpatient service; and the third test data subset in the test data set of each test item is the test data subset of the test item with the type of inpatient service.

[0069] Specifically, during the operation of the reference interval abnormality rate evaluation module 50, that is, calculating the upper limit abnormality rate and the lower limit abnormality rate of the reference interval of the first test data subset, the second test data subset, and the third test data subset with respect to the preset reference interval and the reconstructed reference interval respectively to obtain the reference interval abnormality rate evaluation result. For the sake of illustration, a specific test item is taken as an example for exemplary explanation, for example, platelet count.

[0070] In a specific example, the reference interval abnormality rate evaluation module includes: a preset reference interval abnormality rate evaluation unit and a reconstructed reference interval abnormality rate evaluation unit. Among them, the preset reference interval abnormality rate evaluation unit is used to calculate the upper limit abnormality rate and the lower limit abnormality rate of the reference interval corresponding to the first test data subset, the second test data subset, and the third test data subset respectively to obtain the preset reference interval abnormality rate evaluation result; the reconstructed reference interval abnormality rate evaluation unit is used to calculate the upper limit abnormality rate and the lower limit abnormality rate of the reconstructed reference interval corresponding to the first test data subset, the second test data subset, and the third test data subset respectively to obtain the reconstructed reference interval abnormality rate evaluation result.

[0071] Specifically, the preset reference interval abnormality rate evaluation unit is further configured to calculate, respectively, the proportions of the numbers of the test data exceeding the upper limit value of the preset reference interval in the first test data subset, the second test data subset, and the third test data subset of the test item to obtain the upper limit abnormality rates of the preset reference intervals corresponding to the first test data subset, the second test data subset, and the third test data subset. Here, the proportion is the proportion of the number of the test data exceeding the upper limit value of the preset reference interval divided by the cardinality of each test data subset. Similarly, the preset reference interval abnormality rate evaluation unit is further configured to calculate, respectively, the proportions of the numbers of the test data lower than the lower limit value of the preset reference interval in the first test data subset, the second test data subset, and the third test data subset of the test item to obtain the lower limit abnormality rates of the preset reference intervals corresponding to the first test data subset, the second test data subset, and the third test data subset.

[0072] Correspondingly, the reconstructed reference interval abnormality rate evaluation unit can calculate the upper limit abnormality rate and the lower limit abnormality rate of the reconstructed reference interval corresponding to the first test data subset, the second test data subset, and the third test data subset in a running mode similar to that of the preset reference interval abnormality rate evaluation unit. For the sake of brevity of the text, it will not be elaborated here.

[0073] As Figure 5 shown, in the embodiment of the present application, the suitability management module 60 includes a consistency evaluation unit 61 and a suitability evaluation result determination unit 62. Among them, the consistency evaluation unit 61 is configured to perform a consistency evaluation analysis on the preset reference interval abnormality rate evaluation result and the reconstructed reference interval abnormality rate evaluation result to obtain a consistency evaluation result; and the suitability evaluation result determination unit 62 is configured to generate a suitability evaluation result of the test item based on the consistency evaluation result, the preset and reconstructed reference interval abnormality rate evaluation results.

[0074] In the embodiment of the present application, the running logic of the consistency evaluation unit 61 is as follows: calculate the difference degree between the upper limit abnormality rate of the reconstructed reference interval and the upper limit abnormality rate of the preset reference interval and calculate the difference degree between the lower limit abnormality rate of the reconstructed reference interval and the lower limit abnormality rate of the preset reference interval to obtain a first difference degree and a second difference degree, and generate the consistency evaluation result based on the comparison between the first difference degree and the second difference degree and a preset threshold.

[0075] In a specific example, calculate the difference degree between the upper limit abnormality rate of the reconstructed reference interval and the upper limit abnormality rate of the preset reference interval. That is, compare the evaluated reference interval abnormality rate with that obtained by the EP28 non-parametric method. If the difference is greater than 50%, it is not suitable; if it is less than 50%, it may be suitable.

[0076] In another embodiment of the present application, the operation logic of the consistency evaluation unit 61 is as follows: First, calculate the ratio between the upper limit abnormality rate and the lower limit abnormality rate of the preset reference interval of any two of the first test data subset, the second test data subset, and the third test data subset to obtain a plurality of abnormality rate ratios; then, generate the consistency evaluation result based on the comparison between each abnormality rate ratio in the plurality of abnormality rate ratios and a preset threshold. Specifically, in a specific example, when the abnormality rate of outpatient or inpatient is too different (too large or too small) from the abnormality rate of physical examination, the reference interval may not be suitable. For example, when the outpatient abnormality rate is higher than 5 times that of physical examination or the inpatient abnormality rate is higher than 20 times that of physical examination, the reference interval may not be suitable.

[0077] It is worth mentioning that in specific implementation, the judgment range can be set according to the characteristics of different items, population distribution, and hospital specialty features. However, the present application is not limited thereto.

[0078] In a specific example, set conditions according to the age, gender, etc. of the reference interval, query the data in the laboratory information system from October 22, 2024 to November 22, 2024, and calculate the abnormality rate of the upper or lower limit of the PLT reference interval (male: 83 - 303×10 9 / L, female: 101 - 320×10 9 / L) respectively according to three different visit types of physical examination, outpatient, and inpatient. Then, calculate the abnormality rate of the upper or lower limit of the reference interval (male: 137 - 354×10 9 / L, female 144 - 384×10 9 / L) recalculated by the non-parametric method of EP28 using the physical examination data.

[0079] For the suitability evaluation result, the preset reference interval of PLT is significantly lower than the reference intervals in the EP28 non-parametric method and the reference literature. Figure 7 The upper and lower limit abnormality rates of the preset reference interval for physical examination, outpatient, and inpatient are 15.75%, 13.82%, 10.56% and 0.16%, 2.71%, 18.27% respectively. The extremely high rate of inpatients in the abnormal grade distribution accounts for 20.51%. The suitability of the PLT reference interval needs further study.

[0080] Such as Figure 1As shown, in a preferred embodiment of the present application, the medical test item reference range suitability evaluation system 1000 further includes an abnormal level judgment criterion determination module 70 and an abnormal level calculation module 80. The abnormal level judgment criterion determination module 70 is used to determine the abnormal level judgment criterion based on the set of physical examination data. The abnormal level calculation module 80 is used to analyze the second test data subset and the third test data subset based on the abnormal level judgment criterion to obtain the abnormal level distribution of outpatient test results and the abnormal level distribution of inpatient test results.

[0081] In a specific example, the abnormal level judgment criterion determination module 70 determines, based on the set of physical examination data, the upper and lower limits of five levels (as Figure 6 , Figure 7 , Figure 8 and Figure 9 shown), namely, the five levels of normal, mild abnormality, moderate abnormality, severe abnormality, and extreme abnormality. Among them, the physical examination data that does not belong to the normal level is considered abnormal physical examination data. That is, the abnormal physical examination data includes the physical examination data of the mild abnormality level, the moderate abnormality level, the severe abnormality level, and the extreme abnormality level.

[0082] In this specific example, the abnormal level judgment criterion determination module is further used to: determine the upper and lower limits of the normal level based on the level division condition that the set distribution of the physical examination data is bilaterally greater than 2.5% or less than 97.5% or the level division condition that the set distribution of the physical examination data is unilaterally less than 95%; determine the upper and lower limits of the mild level based on the level division condition that the set distribution of the physical examination data is bilaterally greater than 1.5% and less than or equal to 2.5% or greater than or equal to 97.5% and less than 98.5% or the level division condition that the set distribution of the physical examination data is unilaterally greater than or equal to 95% and less than 97%; determine the upper and lower limits of the moderate level based on the level division condition that the set distribution of the physical examination data is bilaterally greater than 1.0% and less than or equal to 1.5% or greater than or equal to 98.5% and less than 99.0% or the level division condition that the set distribution of the physical examination data is unilaterally greater than or equal to 97.0% and less than 98.0%; determine the upper and lower limits of the severe level based on the level division condition that the set distribution of the physical examination data is bilaterally greater than 0.5% and less than or equal to 1.0% or greater than or equal to 99.0% and less than 99.5% or the level division condition that the set distribution of the physical examination data is unilaterally greater than or equal to 98.0% and less than 99.0%; and determine the upper and lower limits of the extreme level based on the level division condition that the set distribution of the physical examination data is bilaterally less than or equal to 0.5% or greater than or equal to 99.5% or the level division condition that the set distribution of the physical examination data is unilaterally greater than or equal to 99.0%.

[0083] In this specific example, the operating logic of the abnormal level calculation module 80 is as follows: Based on the abnormal level standard, count the proportions of the test data belonging to normal, mild abnormality, moderate abnormality, severe abnormality, and extreme abnormality in the second test data subset and the third test data subset respectively, and in this way, determine the abnormal levels of the test results of outpatient patients and inpatients.

[0084] In a specific example, for the data of a group of test items in a laboratory during a certain period, filter out the data with the visit type of physical examination and calculate the abnormal levels of the test results of outpatient and inpatient patients based on the big data distribution of physical examination. Furthermore, use the big data distribution of physical examination as the basis for dividing the abnormal levels. Normal: For the bilateral range of the physical examination data distribution, it is >2.5% or <97.5%, and for the one-way range, it is <95%. Mild: For the bilateral range of the physical examination data distribution, it is 1.5 - 2.5% or 97.5 - 98.5%, and for the one-way range, it is 95 - 97%. Moderate: For the bilateral range of the physical examination data distribution, it is 1.0 - 1.5% or 98.5 - 99.0%, and for the one-way range, it is 97.0 - 98.0%. Severe: For the bilateral range of the physical examination data distribution, it is 0.5 - 1.0% or 99.0 - 99.5%, and for the one-way range, it is 98.0 - 99.0%. Extreme: For the bilateral range of the physical examination data distribution, it is ≤0.5 or ≥99.5%, and for the one-way range, it is ≥99.0%. Then, use the loop algorithm to update the number of test results of the previous day at a fixed time every day. The calculation of the cumulative percentage value is by counting the proportion of the number lower than this result value in the total number. The population percentage is classified and counted according to 0.5, 1, 1.5, 2, 2.5, 3, 4, 5, 6......95, 96, 97, 97.5, 98, 98.5, 99.5, 100. After obtaining the population percentage of each physical examination item, the result value corresponding to the percentage used to divide the abnormal level can be obtained from it.

[0085] The actual example is as follows: On a certain day, the test result of platelet count in physical examination is 109×10 9 / L for a total of 5. If the value 109×10 9 / L already exists in the previously counted cumulative percentage, then update the item quantity, add the cumulative quantity of 27, and the total should be 32. There are 104 with the result of 109×10 9 / L and below, and the cumulative percentage is 0.501%. If the value 109×10 9 / L does not exist, then add this result value with a quantity of 5, and then calculate the cumulative percentage.

[0086] It is worth mentioning that when managing the suitability of the preset reference intervals for multiple test items, multi-threading can be used to evaluate the suitability of the preset reference intervals for multiple test items respectively through the above-mentioned reference interval suitability evaluation system 1000 for medical test items. It is also worth mentioning that during the process of evaluating the suitability of the reference interval, the abnormal level evaluation result can also be used as an inclusion index to more comprehensively evaluate the actual suitability of the reference interval, but this is not the case for this application.

[0087] As Figure 1 shown, in this preferred embodiment of the present application, the reference interval suitability evaluation system 1000 for medical test items also manages the suitability of the reference interval in combination with reference documents. It is worth mentioning that managing the suitability of the reference interval in combination with reference documents is an optional item.

[0088] Specifically, as Figure 5 shown, the reference interval suitability evaluation system 1000 for medical test items further includes a reference document collection module 90, a reference document level evaluation module 100, and a reference document reference interval management module 110. The reference document collection module 90 is used to search for reference documents. The reference document level evaluation module 100 is used to evaluate the quality of the reference documents to obtain a reference document quality level evaluation result. The reference document reference interval management module 110 is used to extract and manage the relevant data of the recommended reference intervals for at least one test item in the reference documents based on a large language model.

[0089] Correspondingly, the reference interval suitability evaluation system 1000 for medical test items also searches for reference documents through the reference document collection module 90, evaluates the quality of the reference documents through the reference document level evaluation module 100 to obtain a reference document quality level evaluation result, and extracts and manages the relevant data of the recommended reference intervals for at least one test item in the reference documents based on a large language model through the reference document reference interval management module 110.

[0090] During the process of searching for reference documents through the reference document collection module 90, it is possible to link to online databases and library resources through links to search for reference documents from major literature databases, such as CNKI Literature Database, Wanfang Database, Wiley Database, etc.; it is also possible to add new reference documents (such as Figure 11As shown). Convert the keywords (such as inspection item names, detection systems, reference ranges, etc.) into high-dimensional feature vectors, capture the semantic and syntactic information in the keywords through these vectors, then calculate the similarity between the high-dimensional feature vectors and the high-dimensional feature vectors corresponding to each corpus knowledge text block, retrieve the most relevant corpus knowledge text library, and sort the retrieval results according to the similarity to obtain a number of preliminarily screened documents; further, they can be screened according to the partition, impact factor, citation times, publication time, and document type to obtain a number of target documents. Support importing external documents into the document library by uploading files.

[0091] That is, in a specific example of the present application, the process of searching for reference documents includes the steps:

[0092] S1: Input the reference document search keywords;

[0093] S2: Perform semantic embedding encoding on the reference document search keywords to obtain a reference document search keyword semantic embedding encoding vector;

[0094] S3: Extract a set of corpus knowledge text blocks from the corpus knowledge base;

[0095] S4: Perform semantic encoding on each corpus knowledge text block in the set of corpus knowledge text blocks to obtain a set of corpus knowledge text block semantic embedding encoding vectors;

[0096] S5: Calculate the similarity between the reference document search keyword semantic embedding encoding vector and each corpus knowledge text block semantic embedding encoding vector in the set of corpus knowledge text block semantic embedding encoding vectors to obtain a set of similarities;

[0097] S6: Sort the set of similarities to obtain a number of preliminarily screened documents.

[0098] In a preferred embodiment, in order to improve the calculation accuracy of the similarity, the process of calculating the similarity between the reference document search keyword semantic embedding encoding vector and each corpus knowledge text block semantic embedding encoding vector in the set of corpus knowledge text block semantic embedding encoding vectors includes the following steps:

[0099] S51: Perform semantic fine-grained interaction encoding on the reference document search keyword semantic embedding encoding vector and the corpus knowledge text block semantic embedding encoding vector to obtain a search keyword-corpus knowledge text block bidirectional fine-grained semantic association encoding vector;

[0100] S52: Input the search keyword-corpus knowledge text block bidirectional fine-grained semantic association encoding vector into a similarity estimation module based on a decoder to obtain the similarity.

[0101] In this specific example, performing semantic fine-grained interaction encoding on the semantic embedding encoding vector of the reference search keywords and the semantic embedding encoding vector of the corpus knowledge text block includes:

[0102] S511: Performing principal component analysis on the semantic embedding encoding vector of the corpus knowledge text block to obtain a set of principal component feature encoding vectors of the corpus knowledge text block; this process is represented by the formula:

[0103]

[0104] Where, represents the semantic embedding encoding vector of the corpus knowledge text block, represents the transpose of the vector, represents the length of the semantic embedding encoding vector of the corpus knowledge text block, represents the covariance matrix, represents the set of sets of principal component feature encoding vectors of the corpus knowledge text block, 、 、 respectively represent the first, second, and th principal component feature encoding vectors of the corpus knowledge text block, represents the total number of principal component feature encoding vectors of the corpus knowledge text block, represents the diagonal matrix, represents the diagonal elements, 、 、 represent the first, second, and th eigenvalues on the diagonal of the diagonal matrix.

[0105] It should be understood that through principal component analysis, the semantic embedding encoding vectors of high-dimensional corpus knowledge text blocks are first converted into a low-dimensional but more compact and information-rich representation - namely, a set of principal component feature encoding vectors of corpus knowledge text blocks. This process is not just a simple dimensionality reduction operation; it is actually a deep information purification of the original data. In this process, the information originally scattered in the high-dimensional space is reorganized and concentrated on a set of ordered principal component axes. This means that a small number of head principal components can capture most of the variation information in the data, while the redundant or noisy information is naturally filtered out. For example, when constructing a literature retrieval system, by performing principal component analysis on each text block in the corpus, the potential multicollinearity problem within the text can be effectively eliminated, making the differences between different text blocks more obvious, thereby improving the relevance and accuracy of the retrieval results. In addition, due to the reduction of feature dimensions, the time cost of model training will be greatly reduced, and at the same time, the risk of overfitting can be reduced because fewer features mean fewer model parameters, thus reducing the model complexity.

[0106] S512: Perform a linear transformation on each principal component feature encoding vector of the set of principal component feature encoding vectors of the corpus knowledge text blocks to obtain a set of principal component linear transformation feature encoding vectors of the corpus knowledge text blocks, where each principal component linear transformation feature encoding vector of the set of principal component linear transformation feature encoding vectors of the corpus knowledge text blocks has the same feature scale as the semantic embedding encoding vector of the reference document search keyword; this process is represented by the formula:

[0107]

[0108] where, represents performing a linear transformation on , represents the set of principal component linear transformation feature encoding vectors of the corpus knowledge text blocks, , , represent the first, second, and the th principal component feature encoding vectors of the corpus knowledge text blocks.

[0109] It should be understood that through linear transformation, the scale of each principal component feature encoding vector of the corpus knowledge text blocks can be adjusted so that they can be compared and interacted under the same reference framework. This scale adjustment is not just to make the data of different modalities numerically closer; more importantly, it is to ensure that in the process of information fusion, the data of each modality can contribute its unique information fairly, rather than some modality's information being over-amplified or ignored due to scale differences.

[0110] S513: Perform inter-modal independence encoding on each principal component feature encoding vector of the corpus knowledge text block in the set of the semantic embedding encoding vector of the reference search keyword and the principal component feature encoding vector of the corpus knowledge text block to obtain a set of inter-modal independence encoding matrices of the search keyword - corpus knowledge text block principal component; this process is represented by the formula:

[0111]

[0112] where represents the -th principal component feature encoding vector of the corpus knowledge text block, and are feature mapping functions, such as linear mapping or non-linear kernel functions, represents the semantic embedding encoding vector of the reference search keyword, represents the length of the semantic embedding encoding vector of the reference search keyword, represents the -th inter-modal independence encoding matrix of the search keyword - corpus knowledge text block principal component.

[0113] It should be understood that through inter-modal independence coding, an optimized space can be created in which the search keywords and each text block in the corpus are represented in a new and more refined way, that is, a set of principal component inter-modal independence coding matrices of search keyword-corpus knowledge text blocks is formed. This coding method not only combines two types of vectors formally, but also is based on a profound understanding of the relationship between them, emphasizing the differences and unique contributions of the two in terms of the expressed content. Specifically, when comparing search keywords with a large number of corpus knowledge text blocks, directly fusing these vectors may lead to some problems. For example, some high-dimensional features may have a high correlation between different modalities, which will lead to information duplication and increase the complexity of the model. Through inter-modal independence coding, we can identify those truly independent features, so as to more effectively use these features to enhance the performance of the model. This method encourages the model to focus on those parts that provide unique information between different modalities, rather than simply stacking all available information. Through inter-modal independence coding, the system can better identify which documents not only contain terms directly related to the query word, but also contain unique information that can supplement and expand the meaning of the query word. This method helps to improve the relevance of the retrieval results, so that the literature finally presented to the user is not only superficially matched, but also a high-quality resource closely related to the query intention in a deep sense. In addition, this coding strategy also helps to balance the importance of different modalities, ensuring that no modality dominates the final decision due to its inherent characteristics (such as scale size). In this way, both the search keywords and the text blocks in the corpus can play their maximum role within their due scope, jointly providing more accurate and valuable retrieval results for users.

[0114] S514: Based on the set of principal component inter-modal independence coding matrices of search keyword-corpus knowledge text blocks, calculate the set of principal component inter-modal independence soft constraint factors of search keyword-corpus knowledge text blocks; this process is expressed by the formula:

[0115]

[0116] where represents the square of the Frobenius norm of the matrix, represents the th principal component inter-modal independence soft constraint factor of search keyword-corpus knowledge text blocks.

[0117] It should be understood that the inter-modal independence soft constraint factor between the computed search keywords and each corpus knowledge text block is calculated. This process essentially creates a dynamic adjustment mechanism that allows the model to adaptively weigh the importance of each modality according to the specific situation of the input data. Specifically, the soft constraint factor provides a flexible way for the model to emphasize the independence between modalities while also taking into account the inevitable correlations between them. This flexibility is crucial for dealing with complex data in the real world, as actual data often contains information that is both complementary and overlapping to some extent. In addition, the dynamic nature of the soft constraint factor means that it can adjust its behavior as the input changes. If a particular query term is particularly relevant to certain types of literature, the soft constraint factor can automatically adapt to this situation to ensure that the most relevant literature is given priority. In this way, the system can maintain a high level of accuracy and relevance whether dealing with highly specialized queries or broad topic searches.

[0118] S515: Input each principal component feature encoding vector of the corpus knowledge text blocks in the set of the semantic embedding encoding vectors of the reference search keywords and the principal component feature encoding vectors of the corpus knowledge text blocks into the feature interaction response unit to obtain a set of fine-grained response interaction encoding vectors between the search keywords and the principal components of the corpus knowledge text blocks; this process is represented by the formula:

[0119]

[0120] where, represents element-wise multiplication, represents element-wise vector, represents element-wise division, represents the concatenation function, represents the th weight matrix, represents the th bias vector, represents the th fine-grained response interaction encoding vector between the search keywords and the principal components of the corpus knowledge text blocks.

[0121] It should be understood that when these vectors are input into the feature interaction response unit, the system begins to perform fine-grained feature interactions. This is not just a simple vector addition or dot product operation, but a deep exploration of the complex relationships and potential patterns between different modal features. Specifically, the feature interaction response unit explores the multi-level associations between the search keywords and each text block, including but not limited to semantic similarities, thematic overlaps, and subtle connections in the context. This fine-grained interaction does not just stay at the surface level, it also goes deep into the feature dimension level to explore nonlinear relationships and hierarchical structures. This means that the system can capture deeper information in the document. Ultimately, the set of fine-grained response interaction encoding vectors between the search keyword-corpus knowledge text block principal component modalities generated through this process provides a more comprehensive and sophisticated representation. These encoding vectors not only reflect the direct match between the query terms and the documents, but also reveal a more complex interaction and complementary relationship between the two.

[0122] S516: Based on the set of the soft constraint factors of independence between the search keyword and the principal component modalities of the corpus knowledge text block, the set of fine-grained response interaction encoding vectors between the search keyword and the principal component modalities of the corpus knowledge text block is dynamically and adaptively aggregated to obtain the interaction response encoding vector; this process is expressed by the formula:

[0123]

[0124] in, represents the normalized exponential function, Represents the interaction response encoding vector.

[0125] It should be understood that, by using the set of soft constraint factors for the independence between the principal component modes of the search keyword-corpus knowledge text chunks calculated previously, the system begins to perform dynamic adaptive aggregation on the fine-grained response interaction coding vectors between the principal component modes of the search keyword-corpus knowledge text chunks. The core of this process lies in dynamically adjusting the contribution weights of each fine-grained response according to the soft constraint factors, so as to selectively fuse the most relevant feature information. This dynamic adjustment mechanism enables the system to flexibly adapt to the characteristics of different input data. For example, when a certain query word is particularly relevant to a specific type of literature, the system will automatically increase the weight of this type of literature in the final aggregation result. On the contrary, if some literatures contain partial information related to the query word but deviate significantly from the query intention as a whole, their weights will be correspondingly reduced. This ensures that the finally generated interaction response coding vector not only contains the direct matching information between the query word and the literature, but also reflects the more complex complementary relationship between the two. In addition, the dynamic adaptive aggregation process can also effectively reduce the influence of redundant information. Since the soft constraint factors emphasize the independence and complementarity between the modes, the system tends to retain those parts that provide unique perspectives and information, while suppressing repetitive or irrelevant information. In this way, the finally obtained interaction response coding vector is not only more compact and refined, but also has higher discrimination and expressiveness.

[0126] In the process of evaluating the quality of the reference documents through the reference document level evaluation module 100 to obtain the reference document quality level evaluation result, techniques for extracting and / or translating the target content of the documents can be adopted. The PDF document can be scanned through optical character recognition (OCR) technology and converted into editable text. For foreign language documents, translation can be carried out by using an encoder-decoder architecture with the aid of a deep learning model. The encoder converts the source language text into an intermediate text of a vector sequence, and the decoder then translates this intermediate text to generate the target text. At the same time, the model can understand the semantics more accurately through the attention mechanism and then generate a translation result that is more in line with the expression habits of the target language. The large language model is trained through a corpus of a large number of medical documents, and the professional optimization for the medical field enhances the understanding and translation ability of medical terms. Using the large language model to automatically retrieve documents and extract the relevant data in the reference intervals therein avoids manual searching for documents and manual searching for the content in the documents, saving time and effort for researchers.

[0127] Further, a page layout detection model can be used to determine the reading order of the text, and cleaning and formatting algorithms can be used to process text blocks. For example, functions of abstract sentences are classified, functions of chapters are identified, functions of citations are identified, etc., that is, attributes or relationships of one or more types of knowledge units are discriminated. Then, combination and post - processing algorithms are adopted to improve the text quality. Finally, structured information in the literature, such as detection year, region, population race, detection instrument, detection method, method for establishing reference intervals, reference intervals, and confidence intervals, etc., is extracted through natural language processing (NLP) techniques and machine learning in the PDF parsing library. After extracting these fields, they are automatically entered into the system to form a literature database for management. In addition, since medical literature contains some unstructured information, including inclusion criteria, exclusion criteria, etc., these information need to be extracted and organized into structured information for comparative analysis.

[0128] Through the classification model of machine learning, the literature on reference intervals of the same type of test items is grouped into one category. The citation network and academic social network of the literature are analyzed using graph neural networks to strengthen the understanding of the relationships and influence among the literature. A pre - trained language model is used for deeper text understanding and analysis to improve the accuracy of literature quality assessment, as shown in Table 1.

[0129] To more accurately evaluate the quality of the literature, different literatures can be rated as four grades A, B, C, and D according to the literature quality evaluation standard checklist. Generally, grade A means that all quality evaluation criteria of a literature reach grade A; grade B means that the lowest grade reached in the quality evaluation criteria is grade B; grade C means that the lowest grade reached in the quality evaluation criteria is grade C; grade D means that the lowest grade reached in the quality evaluation criteria is grade D. Among them, grade D indicates that the reference intervals evaluated in this literature should not be applied to clinical practice.

[0130] Table 1 Checklist for evaluating the quality of reference interval literature

[0131]

[0132]

[0133] Subsequent data processing can be carried out for the literature with a relatively high quality level of references. In the process of extracting and managing the relevant data on the recommended reference intervals for at least one test item in the reference literature based on the large - language model through the reference interval management module 110 of the reference literature, in the literature database, the literature name, literature quality, impact factor, publication date, journal, author, full text of the literature, abstract, citations, etc. can be browsed; the project name, unit, ethnicity, instrument (such as Figure 11Manage the inspection items (as shown), reagents, inspection methods, regions, reference interval establishment methods, out-of-threshold values, inclusion and exclusion criteria, and reference intervals grouped by gender and age; in addition, a literature database can be associated, and there is a one-to-many relationship between the literature and the project, which are grouped into one category according to the same project name; the recommended reference intervals for the same inspection item in different literatures can be displayed in the form of generated pictures. The abnormal rate and abnormal level of the reference interval in the real-world data of different visit types (physical examination, outpatient, inpatient) can be analyzed by importing data from the laboratory information system, that is, analyze the abnormal rate and abnormal level of the reconstructed reference interval corresponding to the first inspection data subset, the second inspection data subset, and the third inspection data subset, and display them in pictures.

[0134] In a specific example, the process of extracting and managing the relevant data of the recommended reference intervals for at least one inspection item in the reference literature by the reference literature reference interval management module 110 based on the large language model includes: evaluating the consistency between the inspection item and the recommended reference period based on the reference literature. Its operation logic is as follows:

[0135] Automatically evaluate the quality of the reference interval by calculating the bias ratio between the reference interval of the reference literature and the reference interval of the inspection item, so as to judge whether the reference interval is appropriate and perform visual display.

[0136]

[0137]

[0138]

[0139]

[0140] Among them, LL 0 、UL 0 and Me 0 represent the lower limit, upper limit and median of the reference interval of the inspection item, while LL, UL and Me represent the lower limit, upper limit and median of the reference interval in the literature. When BR LL or BR UL is greater than 0.375, it indicates that the bias ratio is relatively large and the consistency between the reference interval of the inspection item and the reference interval in the literature is poor; on the contrary, it indicates that the bias ratio is relatively small and the consistency between the reference interval of the inspection item and the reference interval in the literature is good. Visualize the comparison results of the reference literature in a graphical manner, as shown in Figure 12 、 Figure 13 shown. The reference literature list is displayed on the left side of the graph, including the author (English), year, gender, age, and pregnancy cycle, etc. The reference intervals with a bias ratio greater than 0.375 are displayed in red. BR MeFor reference only, no consistency judgment criteria are set.

[0141] Based on the data such as the regions and races studied in the literature reference interval, detection instruments and methods, methods for establishing reference intervals, population grouping and sample size, etc., combined with the current status of laboratory test items, the upper and lower limits of the recommended reference intervals that can be selected or estimated (median or mode) for a certain test item.

[0142] That is, its technical logic is abstracted and summarized as follows: Calculate the bias ratio between the upper limit value of the preset reference interval for each group corresponding to each subset of test data for each test item and the upper limit value of the recommended reference interval for the corresponding group of the subset of test data for this test item according to the following formula:

[0143]

[0144]

[0145] Among them, SD RI Represents the reference value; UL 0 Represents the upper limit value of the recommended reference interval for a group corresponding to a subset of test data of a test item; LL 0 Represents the lower limit value of the recommended reference interval for a group corresponding to a subset of test data of a test item; BR UL Represents the upper limit bias ratio of the preset reference interval for a group corresponding to a subset of test data of a test item relative to the upper limit of the recommended reference interval for the corresponding group of the subset of test data of this test item; UL Represents the upper limit value of the preset reference interval for a group corresponding to a subset of test data of a test item;

[0146] Calculate the bias ratio between the lower limit value of the preset reference interval for each test item and the lower limit value of the recommended reference interval for this test item according to the following formula:

[0147]

[0148]

[0149] Among them, BR LL Represents the lower limit bias ratio of the preset reference interval for a group corresponding to a subset of test data of a test item relative to the lower limit of the recommended reference interval for the corresponding group of the subset of test data of this test item; LL Represents the lower limit value of the preset reference interval for a group corresponding to a subset of test data of a test item;BR Me It represents the bias ratio of the median of the preset reference interval corresponding to a group of a subset of test data of a test item to the median of the recommended reference interval corresponding to the group of the subset of test data of this type of test item for this test item. Me It represents the median of the preset reference interval corresponding to a group of a subset of test data of a test item.

[0150] In a specific example, the preset threshold is equal to 0.375. When or is greater than 0.375, it is determined that the consistency between the recommended reference interval and the reconstructed reference interval of this test item does not meet the preset requirements; when and is less than or equal to 0.375, it is determined that the consistency between the recommended reference interval and the reconstructed reference interval of this test item meets the preset requirements. BR Me For reference only, no consistency judgment criteria are set.

[0151] In summary, the medical test item reference interval suitability evaluation system 1000 according to the embodiments of the present application is clarified. The medical test item reference interval suitability evaluation system 1000 groups the test data according to gender and age, and establishes reference intervals corresponding to each age group of different genders, which can improve the accuracy of the reference interval to a certain extent. The set of test data of the subject object in the medical test item reference interval suitability evaluation system 1000 comes from the real medical record data of real patients. In this way, on the one hand, the data source is reliable, and on the other hand, there is no need to specifically recruit subject objects for testing for the parameter interval experiment, which can improve the feasibility to a certain extent and reduce the cost.

[0152] The above describes the present application and its implementation manners. Such a description is not restrictive. What is shown in the drawings is only one of the implementation manners of the present application, and the actual structure is not limited thereto. Generally speaking, if those of ordinary skill in the art are inspired by it and design similar structural manners and embodiments without creative work without departing from the creative purpose of the present application, they shall fall within the protection scope of the present application.

Claims

1. A reference interval suitability assessment system for medical test items, characterized in that: include: A preset reference interval extraction module, used to extract a preset reference interval of at least one test item from a test item reference interval database; A physical examination data collection and analysis module, used to obtain a set of physical examination data, and process the set of physical examination data using the EP28 non-parametric method to determine the reconstructed reference interval of the test item; A test data collection module, used to extract test data sets of the test items within a predetermined time period from a laboratory information system; A data grouping module, used for segmenting the test data set of the test items according to the three types of medical examination, outpatient and hospitalization to obtain a first test data subset, a second test data subset and a third test data subset; A reference interval abnormality rate assessment module, used to calculate the upper and lower abnormality rates of the first test data subset, the second test data subset, and the third test data subset relative to a preset reference interval, so as to obtain a preset reference interval abnormality rate assessment result; Calculating the upper limit abnormality rate and the lower limit abnormality rate of the first test data subset, the second test data subset and the third test data subset relative to the reconstructed reference interval to obtain a reconstructed reference interval abnormality rate evaluation result; The suitability management module is used to perform consistency assessment analysis on the preset reference interval abnormality rate assessment results and the reconstructed reference interval abnormality rate assessment results to obtain consistency assessment results.

2. The medical test item reference interval suitability evaluation system according to claim 1, characterized in that: The reference interval abnormality rate assessment module includes: A preset reference interval abnormality rate evaluation unit, used to respectively calculate the upper limit abnormality rate of the preset reference interval and the lower limit abnormality rate of the preset reference interval corresponding to the first test data subset, the second test data subset and the third test data subset to obtain a preset reference interval abnormality rate evaluation result; The reconstructed reference interval abnormality rate assessment unit is used to respectively calculate the upper limit abnormality rate and the lower limit abnormality rate of the reconstructed reference interval corresponding to the first test data subset, the second test data subset and the third test data subset to obtain the reconstructed reference interval abnormality rate assessment result.

3. The medical test item reference interval suitability evaluation system according to claim 2, characterized in that: The suitability management module comprises: A consistency evaluation unit, configured to perform consistency evaluation analysis on the preset reference interval abnormality rate evaluation result and the reconstructed reference interval abnormality rate evaluation result to obtain a consistency evaluation result; A suitability assessment result determination unit is used to generate a suitability assessment result of the inspection item based on the consistency assessment result, the preset reference interval abnormality rate assessment result and the reconstructed reference interval abnormality rate assessment result.

4. The medical test item reference interval suitability evaluation system according to claim 3, characterized in that: The consistency assessment unit is further used for: Calculating the difference between the upper limit abnormality rate of the reconstructed reference interval and the upper limit abnormality rate of the preset reference interval and calculating the difference between the lower limit abnormality rate of the reconstructed reference interval and the lower limit abnormality rate of the preset reference interval to obtain a first difference and a second difference; The consistency evaluation result is generated based on a comparison between the first difference degree and the second difference degree and a preset threshold.

5. The medical test item reference interval suitability evaluation system according to claim 3, characterized in that: The consistency assessment unit is further used for: Calculating the ratio between the upper limit abnormality rate of the preset reference interval or the lower limit abnormality rate of the preset reference interval of any two test data subsets among the first test data subset, the second test data subset and the third test data subset to obtain a plurality of abnormality rate ratios; The consistency evaluation result is generated based on a comparison between each abnormality rate ratio of the multiple abnormality rate ratios and a preset threshold.

6. The medical test item reference interval suitability evaluation system according to claim 1, characterized in that: Also includes: An abnormality level judgment standard determination module, used to determine the abnormality level judgment standard based on the set of physical examination data; The abnormality level calculation module is used to analyze the second test data subset and the third test data subset based on the abnormality level standard to obtain the abnormality level distribution of the outpatient test results and the abnormality level distribution of the inpatient test results.

7. The medical test item reference interval suitability evaluation system according to claim 6, characterized in that: The abnormal level judgment standard determination module is further used to: Determine the upper and lower limits of the normal level based on the level division condition that the set distribution of the physical examination data is greater than 2.5% or less than 97.5% on both sides or the level division condition that the set distribution of the physical examination data is less than 95% on one side; Determine the upper and lower limits of the mild level based on the level division condition that the set distribution of the physical examination data is greater than 1.5% and less than 2.5% on both sides or greater than or equal to 97.5% and less than 98.5% on one side, or the level division condition that the set distribution of the physical examination data is greater than or equal to 95% and less than 97% on one side; Determine the upper and lower limits of the moderate level based on the level division condition that the set distribution of the physical examination data is greater than 1.0% and less than or equal to 1.5% on both sides or greater than or equal to 98.5% and less than 99.0% on one side, or the level division condition that the set distribution of the physical examination data is greater than or equal to 97.0% and less than 98.0% on one side; Determine the upper and lower limits of the severe level based on the level classification condition that the set distribution of the physical examination data is greater than 0.5% and less than or equal to 1.0% on both sides or greater than or equal to 99.0% and less than 99.5% on one side, or the level classification condition that the set distribution of the physical examination data is greater than or equal to 98.0% and less than 99.0% on one side; The upper and lower limits of the extreme level are determined based on the level division condition that the set distribution of the physical examination data is less than or equal to 0.5% or greater than or equal to 99.5% on both sides or the level division condition that the set distribution of the physical examination data is greater than or equal to 99.0% on one side.

8. The medical test item reference interval suitability evaluation system according to claim 1, characterized in that: Also includes: Reference collection module, used to search for references; A reference document quality assessment module, used for performing quality assessment on the reference document to obtain a reference document quality assessment result; The reference reference interval management module is used to extract and manage relevant data on the recommended reference interval of the test item from the reference based on the large language model.

9. The medical test item reference interval suitability evaluation system according to claim 8, characterized in that: The reference document reference interval management module is further used to: Calculating the bias ratio between the recommended reference interval of the test item and the reconstructed reference interval; When the bias ratio is greater than a preset threshold, it is determined that the consistency between the recommended reference interval of the test item and the reconstructed reference interval of the test item does not meet the preset requirement; When the bias ratio is less than or equal to a preset threshold, it is determined that the consistency between the recommended reference interval of the test item and the reconstructed reference interval of the test item meets the preset requirement.

Images