A method and device for evaluating the application performance of HIV rapid detection reagent

Abstract

The present invention provides an application performance evaluation method and device for HIV rapid detection reagents, which relates to the technical field of performance evaluation, including: obtaining a test sample set, including a test urine sample set and a homologous blood sample set; using a standard blood HIV antibody detection reagent for detection, obtaining a standard detection result set and a standard detection time set; using a target HIV rapid detection reagent for detection, obtaining a target detection result set and a target detection time set; using a twin neural network to construct a detection consistency verification model, perform detection consistency verification, and output a first performance evaluation result; perform detection time evaluation, output a second performance evaluation result; and output an application performance evaluation result. The present invention solves the technical problem that the prior art usually relies on simple indicators to perform performance evaluation of HIV rapid detection reagents, lacks a unified and standardized process to evaluate overall performance, and results in poor reliability of the detection results.

Description

Technical Field

[0001] The present invention relates to the technical field of performance evaluation, and in particular to a method and a device for evaluating the application performance of a rapid HIV detection reagent. Background Art

[0002] HIV self-testing is a process in which an individual collects samples, tests, and reads the results in private, alone or in the company of someone they trust. Compared with traditional HIV testing, HIV self-testing can increase the frequency of HIV testing in key populations. However, the performance evaluation of rapid test kits used for HIV self-testing mostly relies on simple indicators such as positive rate, negative rate, sensitivity, and specificity. There is a lack of a unified and standardized process to evaluate the overall performance of the test kits, especially compared with traditional standard blood tests. There may be errors. The test results of non-blood samples such as urine may be affected by sample quality, environmental factors, etc., resulting in false positive or false negative results, which in turn affects the reliability and accuracy of the test. Summary of the invention

[0003] This application provides a method and device for evaluating the application performance of HIV rapid detection reagents, aiming to solve the technical problem that the existing technology usually relies on simple indicators to evaluate the performance of HIV rapid detection reagents, lacks a unified and standardized process to evaluate the overall performance, and leads to poor reliability of the test results.

[0004] The first aspect disclosed in the present application provides an application performance evaluation method for an HIV rapid detection reagent, the method comprising: obtaining a test sample set, wherein the test sample set comprises a test urine sample set and a homologous blood sample set collected synchronously; in a preset test scenario, using a standard blood HIV antibody detection reagent to detect the homologous blood sample set to obtain a standard test result set and a standard test time set; in the preset test scenario, using a target HIV rapid detection reagent to detect the test urine sample set to obtain a target test result set and a target detection time set; using the standard test result set as a comparison benchmark, using a twin neural network to construct a detection consistency verification model, performing detection consistency verification on the target detection result set based on the detection consistency verification model, and outputting a first performance evaluation result according to the detection consistency verification result; performing detection time evaluation based on the standard detection time set and the target detection time set, and outputting a second performance evaluation result according to the detection time evaluation result; outputting an application performance evaluation result of the target HIV rapid detection reagent based on the first performance evaluation result and the second performance evaluation result.

[0005] The second aspect disclosed in the present application provides an application performance evaluation device for HIV rapid detection reagents, the device being used for the above-mentioned application performance evaluation method for HIV rapid detection reagents, the device comprising: a test sample acquisition module, the test sample acquisition module being used to acquire a test sample set, wherein the test sample set comprises a test urine sample set and a homologous blood sample set collected synchronously; a blood sample detection module, the blood sample detection module being used to detect the homologous blood sample set using a standard blood HIV antibody detection reagent under a preset test scenario to obtain a standard test result set and a standard test time set; a urine sample detection module, the urine sample detection module being used to detect the test urine sample set using a target HIV rapid detection reagent under the preset test scenario Test, obtain a target detection result set and a target detection time set; a consistency verification module, the consistency verification module is used to use the standard detection result set as a comparison benchmark, adopt a twin neural network to build a detection consistency verification model, perform detection consistency verification on the target detection result set based on the detection consistency verification model, and output a first performance evaluation result according to the detection consistency verification result; a detection time evaluation module, the detection time evaluation module is used to perform detection time evaluation based on the standard detection time set and the target detection time set, and output a second performance evaluation result according to the detection time evaluation result; an evaluation result output module, the evaluation result output module is used to output the application performance evaluation result of the target HIV rapid detection reagent based on the first performance evaluation result and the second performance evaluation result.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] By collecting homologous blood samples and urine samples, and using standard HIV antibody detection reagents and target HIV rapid detection reagents to test these two types of samples respectively, the comparison basis of the test is ensured, and the twin neural network is used for consistency verification to accurately analyze the consistency between the test results of blood samples and urine samples, thereby improving the accuracy of the consistency evaluation of the test results and avoiding the deviation that may be caused by manual analysis. In terms of detection time, the standard detection time set and the target detection time set are used for comparison and analysis to accurately quantify the detection efficiency of the target detection reagent. Combined with the time deviation feature evaluation, it can be judged whether the target detection reagent meets the efficiency requirements, that is, whether it can give reliable test results within a suitable time range. Combined with the first performance evaluation result based on the consistency evaluation of the test results and the second performance evaluation result based on the detection time evaluation, it is possible to make an accurate and comprehensive judgment on the overall performance of the target HIV rapid detection reagent, improve the automation level and accuracy of the evaluation process, and be able to complete the performance evaluation of a large number of samples in a relatively short time, providing an efficient evaluation tool for the development and practical application of reagents.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A schematic diagram of a method for evaluating the application performance of an HIV rapid test kit provided in an embodiment of the present application;

[0010] Figure 2 A schematic diagram of the structure of an application performance evaluation device for an HIV rapid detection reagent provided in an embodiment of the present application.

[0011] Description of the reference numerals: test sample acquisition module 10 , blood sample detection module 20 , urine sample detection module 30 , consistency verification module 40 , detection time evaluation module 50 , evaluation result output module 60 . DETAILED DESCRIPTION

[0012] The embodiments of the present application provide a method and device for evaluating the application performance of HIV rapid detection reagents, thereby solving the technical problem that the prior art generally relies on simple indicators to evaluate the performance of HIV rapid detection reagents, lacks a unified and standardized process to evaluate the overall performance, and leads to poor reliability of the detection results.

[0013] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically introduced below in conjunction with the accompanying drawings of the specification. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0014] Embodiment 1, as Figure 1 As shown, the present application embodiment provides a method for evaluating the application performance of HIV rapid detection reagents, the method comprising:

[0015] A test sample set is obtained, wherein the test sample set includes a test urine sample set and a homologous blood sample set collected synchronously.

[0016] Collect test sample sets, such as recruiting volunteers for sample collection through voluntary consultation and testing points. The gender and age of the subjects are not limited. The test urine sample set obtained refers to the sample set formed by collecting human urine, which is used for subsequent rapid HIV testing. The homologous blood sample set refers to the blood sample collected from the same subject. Normally, blood is a standard sample for detecting HIV antibodies. It is used here to provide standard test results as a comparison benchmark for urine sample test results. The collection process involves synchronous collection of urine samples and blood samples, which means that urine samples and blood samples are collected from the same subject in the same time period. This synchronization is to avoid interference with the test results due to time differences or physiological changes. For example, HIV antibody levels may fluctuate over time and with changes in individual conditions, so samples collected at the same time can better reflect the biological status at the same time.

[0017] In a preset test scenario, a standard blood HIV antibody detection reagent is used to detect the homologous blood sample set to obtain a standard test result set and a standard test time set.

[0018] The preset test scenario refers to testing under standardized and controllable experimental conditions, usually in an environment that meets laboratory specifications, such as specified temperature, humidity, time, etc., to ensure that the test results are consistent and repeatable. Standard blood HIV antibody test reagents are HIV antibody test reagents that have been widely used and verified for their accuracy and reliability, or other highly sensitive laboratory blood test methods. This standard reagent is selected to provide a reliable reference result for subsequent urine testing. Use a standard HIV antibody test kit and operate according to its instructions, which generally includes sample addition, reagent reaction, color development detection and other processes. After the test is completed, record the HIV antibody results of each blood sample to form a standard test result set, which includes positive and negative results. At the same time, record the time spent on each test to form a standard test time set for subsequent evaluation of test efficiency.

[0019] In the preset test scenario, the test urine sample set is tested using a target HIV rapid test reagent to obtain a target test result set and a target test time set.

[0020] The comparative test was carried out under the same test environment as the standard blood HIV antibody test reagent. The test operation of the same product was completed by the same tester, and the operation method was carried out in accordance with the requirements of the respective product instructions. This ensured consistent experimental conditions and avoided the influence of external factors on the test results.

[0021] The test was performed according to the instructions for use of the target HIV rapid test reagent. The target HIV rapid test reagent uses immunochromatography technology and the principle of indirect method to detect HIV-1 and HIV-2 antibodies in human urine samples. The test line (T line) of the nitrocellulose membrane is coated with HIV recombinant antigens. If the sample contains HIV1 / 2 antibodies, the test line (T line) agglutinates to form a red line, which is a positive result; if there is no HIV1 / 2 antibody in the sample, no red line will be formed on the test line (T line), which is a negative result. After the test is completed, the HIV antibody test results of each urine sample are recorded to form a target test result set. The time spent on each test is also recorded to form a target test time set.

[0022] Taking the standard detection result set as a comparison benchmark, a twin neural network is used to build a detection consistency verification model, and the target detection result set is verified for detection consistency based on the detection consistency verification model, and a first performance evaluation result is output according to the detection consistency verification result.

[0023] The standard test result set is a reliable result obtained using a standard blood HIV antibody test. The urine test results of the target HIV rapid test kit will be compared with this to verify its accuracy. The twin neural network is a neural network architecture specifically used to compare the similarity between two inputs. It consists of two sub-networks that share weights. The two input samples (standard test results and target test results) pass through the sub-network to obtain a feature vector. Then, by calculating the similarity between the two feature vectors, such as cosine similarity, Euclidean distance, etc., it is determined whether they are consistent. A detection consistency verification model is constructed based on the twin neural network. The consistency is evaluated by comparing the test results of urine samples with the test results of blood samples, and the consistency evaluation coefficient is output. If the consistency evaluation coefficient reaches the preset threshold, the detection consistency is determined to be good, and the first performance evaluation result is output as passed.

[0024] A detection time evaluation is performed based on the standard detection time set and the target detection time set, and a second performance evaluation result is output according to the detection time evaluation result.

[0025] The data in the standard detection time set and the target detection time set are sorted to generate a standard detection time data sequence and a target detection time data sequence. For the convenience of comparison, the two time series are standardized to generate a standard detection time feature sequence and a target detection time feature sequence. The standardization process can eliminate the difference in time distribution between different samples so that the time data can be compared on the same scale. The deviation of the target detection time feature sequence is calculated based on the standard detection time feature sequence to generate a detection time deviation feature sequence for quantifying the difference between the urine detection time and the blood detection time. The detection time feature is evaluated based on the deviation sequence, and the detection time feature evaluation coefficient is calculated. If the coefficient meets the preset threshold, it means that the time performance of the target detection reagent is good, and the second performance evaluation result is output as passed.

[0026] The application performance evaluation result of the target HIV rapid detection reagent is output based on the first performance evaluation result and the second performance evaluation result.

[0027] Integrate the first performance evaluation results and the second performance evaluation results. If both the first performance evaluation results and the second performance evaluation results are passed, it means that the performance of the target HIV rapid detection reagent meets expectations and can be effectively used for HIV rapid detection of urine samples; conversely, if any evaluation result fails, further analysis and optimization of the design and use of the target detection reagent are needed.

[0028] Furthermore, the method of using the standard test result set as a comparison benchmark and using a twin neural network to construct a test consistency verification model includes:

[0029] Acquire the test sample set, wherein the test sample set includes a case group, a control group, and an interference group in preset proportions; based on the case group, the control group, and the interference group, use a twin neural network to respectively construct a case group detection comparison channel, a control group detection comparison channel, and an interference group detection comparison channel; connect the case group detection comparison channel, the control group detection comparison channel, and the interference group detection comparison channel in parallel to obtain the detection comparison layer; construct a detection consistency verification layer, and cascade the detection comparison layer and the detection consistency verification layer to obtain the detection consistency verification model.

[0030] The test sample set is divided into a case group, a control group, and an interference group, wherein the case group includes HIV1 / 2 positive samples, the control group includes HIV1 / 2 negative samples, and the interference group includes rheumatoid factor (RF+), HCV, SYP positive or pregnant women samples, etc. The preset ratio is set according to the specific situation. For example, the case group samples are 200 cases, the control group samples are 200 cases, and the interference group is 50 cases. Blood samples and urine samples are collected simultaneously. The case group samples contain clear HIV genotypes, including B / B', BC recombinant type, and AE recombinant type, with no less than 10 cases of each.

[0031] The core structure of the twin neural network consists of two weight-sharing sub-networks, which are used to process different inputs, namely standard test results and target detection results, and generate feature vectors for comparing the similarities between them. Among them, the input data of the case group detection comparison channel are the standard test results and target detection results of the case group (positive samples), and the output is to compare the consistency of the standard results and urine test results in positive cases, which is used to evaluate the sensitivity of the target detection reagent; the input of the control group detection comparison channel is the standard test results and target detection results of the control group (negative samples), and the output is to compare the consistency of the standard results and urine test results in the negative control group, which is used to evaluate the specificity of the reagent; the input of the interference group detection comparison channel is the standard test results and target detection results of the interference group, and the output is to evaluate whether the target detection reagent can avoid interference and maintain the accuracy of the results in samples with interference factors.

[0032] The comparison channels of the case group, control group, and interference group are run in parallel, so that the sample tests of three different groups can be processed at the same time, ensuring the consistency verification of the test results of all samples in the same model. In each channel, the twin neural network processes the corresponding sample group data respectively and outputs the consistency evaluation results of each group. Through the parallel channels, the test comparison channels of all groups together constitute the test comparison layer, which is responsible for summarizing the test consistency evaluation results of each group.

[0033] After the detection comparison layer, a detection consistency verification layer is added to process the detection results of different groups from the comparison layer and calculate the overall consistency evaluation coefficient. The verification layer calculates the consistency coefficient of the detection results based on the detection consistency evaluation function. For example, the verification layer calculates the observed consistency coefficient and the expected consistency coefficient, and finally outputs the detection consistency evaluation coefficient to evaluate the accuracy of the detection results. The two are cascaded to finally obtain a complete detection consistency verification model. The first layer of the model outputs the consistency data of the detection results of each group, and the second layer performs further verification and calculation based on these data, and finally outputs the overall consistency evaluation.

[0034] Furthermore, the detection consistency verification layer embeds the detection consistency evaluation function as follows:

[0035]

[0036] Among them, k is the detection consistency evaluation coefficient, P0 is the observation consistency coefficient, P e is the expected consistency coefficient.

[0037] Specifically, the detection consistency evaluation function is constructed based on the Kappa coefficient formula. The Kappa coefficient is a method for evaluating the consistency of classification results, especially for evaluating whether two detection tools are consistent, excluding the influence of random consistency. The observed consistency coefficient is the observed accuracy, that is, the probability of actual consistency, and the expected consistency coefficient is the probability of expected consistency, that is, the probability of accidental consistency between the two according to the assumption of random consistency. The calculated detection consistency evaluation coefficient k value ranges from -1 to 1, where 1 represents complete consistency, 0 represents the same as random consistency, that is, no consistency, and -1 represents negative consistency, which is worse than random.

[0038] Further, performing detection consistency verification on the target detection result set based on the detection consistency verification model, and outputting a first performance evaluation result according to the detection consistency verification result, includes:

[0039] A detection four-grid table is constructed based on the output results of the case group detection comparison channel, the control group detection comparison channel, and the interference group detection comparison channel, wherein the data points of the detection four-grid table include true positive, false positive, true negative, and false negative; based on the detection four-grid table, the random consistency probability of the target detection result and the standard detection result is calculated to obtain the observed consistency coefficient and the expected consistency coefficient; based on the detection consistency evaluation function, the observed consistency coefficient and the expected consistency coefficient are calculated to obtain the detection consistency evaluation coefficient as the detection consistency verification result; when the detection consistency evaluation coefficient meets the preset detection consistency evaluation coefficient threshold, the first performance evaluation result is output as the evaluation passed.

[0040] Data were extracted from the output results of the detection comparison channels of the case group, control group and interference group, and filled in the four-cell detection table one by one. This table shows the accuracy of the target reagent in the test results, which is used to further calculate the consistency index. Among them, a true positive means that both the standard test and the target test results are positive, a false positive means that the standard test is negative, but the target test result is positive, a true negative means that the standard test is positive, but the target test result is negative, and a false negative means that both the standard test and the target test results are negative.

[0041] Observed consistency refers to the probability that the target test results are consistent with the standard test results in all tests. This can be calculated by the true positive and true negative data in the four-cell table, that is, when the test results are positive and negative, the number of samples that are consistent with the two is divided by the total number of samples. Expected consistency refers to the probability that the target test results are consistent with the standard test results under random circumstances. This can be calculated by the sum of the test results of each category in the four-cell table, and is used to estimate the probability that the target test results and the standard results are consistent under random circumstances.

[0042] The test consistency evaluation function is constructed based on the Kappa coefficient formula. The Kappa coefficient is a method for evaluating the consistency of classification results, especially for evaluating whether two test tools are consistent, excluding the influence of random consistency. When the Kappa coefficient is 1, it means complete consistency; when the Kappa coefficient is 0, it means that the consistency is completely random; negative values ​​indicate worse than random consistency. By calculating the test consistency evaluation coefficient, the consistency level of the target HIV rapid test reagent and the standard test results is obtained as the final consistency verification result.

[0043] The test consistency assessment coefficient threshold is a preset threshold set by industry standards or laboratory requirements, such as 0.85. When the test consistency assessment coefficient meets or exceeds this threshold, it means that the test results of the target HIV rapid test reagent are sufficiently consistent with the standard blood test results, and the verification is passed; if it is lower than the set threshold, the consistency verification fails, indicating that the test results of the target detection reagent are inconsistent with the standard results, and further adjustment of the reagent or improvement of the detection method may be required.

[0044] Furthermore, the method further comprises:

[0045] When the test consistency evaluation coefficient is not 1, a plurality of inconsistent test samples are extracted, and HIV antibody retests are performed on the plurality of inconsistent test samples to obtain a plurality of retest results; if the plurality of retest results have not changed, the test consistency evaluation coefficient is output as the test consistency verification result; if the plurality of retest results have changed, the test consistency evaluation coefficient is corrected based on the changed retest results, and the corrected test consistency evaluation coefficient is output as the test consistency verification result.

[0046] The test consistency assessment coefficient is used to measure the consistency between the test results of the target test reagent and the standard test results. The assessment coefficient is 1, which means that all test results are completely consistent. However, in reality, it is often impossible to achieve complete consistency, so samples with inconsistent test results may appear. At this time, samples with inconsistent test results during the initial test need to be retested to eliminate accidental errors or operational errors in the initial test. Specifically, samples whose target test results are inconsistent with the standard test results are extracted from all samples. These samples may show false positives, false negatives or other forms of abnormal results. The extracted inconsistent samples are retested. The purpose of the retest is to confirm whether the inconsistent results are caused by some operational errors or reagent problems, or reflect some special biological phenomena.

[0047] If the re-test results are consistent with the initial test results, it indicates that the initial test results are reliable. In this case, the consistency evaluation coefficient calculated initially is kept unchanged as the final test consistency verification result, which means that the existing data is sufficient to represent the actual performance of the target detection reagent and there is no need to further adjust the evaluation coefficient.

[0048] If the retest results differ from the initial test results, it means that there may be some errors or operational problems in the initial test, resulting in inaccurate results. In this case, the four-cell table calculated initially is corrected based on the retest results, and the changed results in the retest replace the initial test results. The data of true positive, false positive, true negative and false negative are updated, and the evaluation coefficient is recalculated based on the updated four-cell table data to improve the accuracy of the test consistency evaluation. The corrected evaluation coefficient is used as the final consistency verification result, reflecting the performance of the target reagent after the retest.

[0049] Furthermore, the method of performing detection time evaluation based on the standard detection time set and the target detection time set, and outputting a second performance evaluation result according to the detection time evaluation result, includes:

[0050] The standard detection time set and the target detection time set are sorted based on the sample sequence number of the test sample set to generate a standard detection time data sequence and a target detection time data sequence; the standard detection time data sequence and the target detection time data sequence are standardized to generate a standard detection time feature sequence and a target detection time feature sequence; based on the standard detection time feature sequence, deviation calculation is performed on the target detection time feature sequence to obtain a detection time deviation feature sequence; based on the detection time deviation feature sequence, detection time feature evaluation is performed to calculate a detection time feature evaluation coefficient as the detection time evaluation result; when the detection time feature evaluation coefficient meets a preset detection time feature evaluation coefficient threshold, the second performance evaluation result is output as evaluation passed.

[0051] According to the number or serial number of each sample, the data in the two time sets are sorted to ensure that the standard detection time and target detection time of the same sample are in corresponding positions. The sorted data constitute the standard detection time data sequence and the target detection time data sequence. Each point in these sequences is matched one by one with the corresponding sample serial number to ensure the comparability of the data.

[0052] Standardize the time data, for example, use the maximum and minimum value standardization formula to process it so that it is in a unified interval, usually between 0 and 1 or other ranges. After processing, generate a standard detection time feature sequence, which represents the normalized characteristics of the standard blood detection time, and generate a target detection time feature sequence, which represents the normalized characteristics of the target urine detection time. Standardization eliminates the scale differences between sample detection times, allowing the detection times of different samples to be compared at the same level, avoiding analytical deviations caused by abnormally large or small individual sample times.

[0053] For each sample, the difference between the target detection time characteristic sequence and the standard detection time characteristic sequence is calculated to obtain the time deviation of each sample. A positive deviation indicates that the detection time of the target detection reagent is longer than the standard time, and a negative deviation indicates that the detection time is shorter. The time deviation of each sample constitutes the detection time deviation characteristic sequence, which reflects the time deviation of the target HIV rapid test reagent relative to the standard blood test. This sequence is used to evaluate the time performance of the target detection reagent.

[0054] Calculate the average value of the deviation sequence to evaluate whether the target detection reagent is faster or slower than the standard time overall; calculate the standard deviation of the deviation sequence to evaluate the fluctuation of the time deviation and reflect whether the time performance of the target reagent is stable. Consider the average deviation and standard deviation comprehensively to generate a comprehensive detection time characteristic evaluation coefficient to reflect the overall time performance of the target detection reagent. If the coefficient is small, it means that the time efficiency of the target reagent is good and relatively stable; if the coefficient is large, it may indicate low time efficiency or large fluctuations.

[0055] According to the experimental design or industry standards, the detection time characteristic evaluation coefficient threshold is set to determine whether the time efficiency of the target detection reagent is qualified. The threshold can represent the allowable time deviation range or the maximum fluctuation of the detection time. If the detection time characteristic evaluation coefficient is lower than or equal to the preset threshold, it means that the time efficiency of the target detection reagent meets the standard and the time evaluation passes; if the detection time characteristic evaluation coefficient is higher than the preset threshold, the time evaluation fails, indicating that the target reagent has deficiencies in detection speed or stability.

[0056] Furthermore, the method of performing detection time feature evaluation based on the detection time deviation feature sequence and calculating a detection time feature evaluation coefficient includes:

[0057] The detection time deviation feature sequence is mean-calculated to obtain the detection time deviation central coefficient; the detection time deviation feature sequence is standard-calculated to obtain the detection time deviation dispersion coefficient; based on the detection time feature evaluation function, the detection time deviation central coefficient and the detection time deviation dispersion coefficient are calculated to generate the detection time feature evaluation coefficient; wherein the detection time feature evaluation function is as follows:

[0058]

[0059] Among them, T is the detection time characteristic evaluation coefficient, μ t is the detection time deviation center coefficient, σ t is the detection time deviation dispersion coefficient, ω1 and ω2 are the weights of the detection time deviation central coefficient and the detection time deviation dispersion coefficient respectively, and ω1+ω2=1.

[0060] A detection time deviation characteristic sequence is obtained, where each element in the sequence represents the deviation between the target detection time and the standard detection time of the target HIV rapid detection reagent. The detection time deviation central coefficient is obtained by calculating the mean of the sequence, which reflects the average deviation degree of the target detection time relative to the standard detection time.

[0061] Continue to calculate the standard deviation of the detection time deviation feature sequence to obtain the detection time deviation dispersion coefficient, which is used to measure the dispersion of the detection time deviation, that is, the dispersion of the target detection time relative to the standard detection time. If the dispersion coefficient is large, it means that the detection time fluctuates greatly in different samples; if the dispersion coefficient is small, it indicates that the detection time is relatively concentrated.

[0062] The predefined detection time characteristic evaluation function is used to comprehensively calculate the detection time deviation center coefficient and the detection time deviation dispersion coefficient to generate the final detection time characteristic evaluation coefficient, which is used to comprehensively evaluate the detection time characteristic performance of the target HIV rapid test reagent. Among them, the detection time characteristic evaluation function is as follows:

[0063]

[0064] This function balances the degree of deviation and volatility of detection time in a weighted manner, and the generated evaluation coefficient T can provide an overall evaluation of the detection time characteristics.

[0065] In summary, the application performance evaluation method of a rapid HIV detection reagent provided in the embodiment of the present application has the following technical effects:

[0066] By collecting homologous blood samples and urine samples, and using standard HIV antibody detection reagents and target HIV rapid detection reagents to test these two types of samples respectively, the comparison basis of the test is ensured, and the twin neural network is used for consistency verification to accurately analyze the consistency between the test results of blood samples and urine samples, thereby improving the accuracy of the consistency evaluation of the test results and avoiding the deviation that may be caused by manual analysis. In terms of detection time, the standard detection time set and the target detection time set are used for comparison and analysis to accurately quantify the detection efficiency of the target detection reagent. Combined with the time deviation feature evaluation, it can be judged whether the target detection reagent meets the efficiency requirements, that is, whether it can give reliable test results within a suitable time range. Combined with the first performance evaluation result based on the consistency evaluation of the test results and the second performance evaluation result based on the detection time evaluation, it is possible to make an accurate and comprehensive judgment on the overall performance of the target HIV rapid detection reagent, improve the automation level and accuracy of the evaluation process, and be able to complete the performance evaluation of a large number of samples in a relatively short time, providing an efficient evaluation tool for the development and practical application of reagents.

[0067] Example 2, based on the same inventive concept as the method for evaluating the application performance of a rapid HIV detection reagent in the previous example, Figure 2 As shown, the embodiment of the present application provides an application performance evaluation device for HIV rapid detection reagents, the device comprising:

[0068] A test sample acquisition module 10, the test sample acquisition module 10 is used to acquire a test sample set, wherein the test sample set includes a test urine sample set and a homologous blood sample set collected synchronously; a blood sample detection module 20, the blood sample detection module 20 is used to detect the homologous blood sample set using a standard blood HIV antibody detection reagent under a preset test scenario to obtain a standard test result set and a standard test time set; a urine sample detection module 30, the urine sample detection module 30 is used to detect the test urine sample set using a target HIV rapid detection reagent under the preset test scenario to obtain a target test result set and a target detection time set; a consistency verification module 40, the consistency The performance verification module 40 is used to use the standard detection result set as a comparison benchmark, adopt a twin neural network to build a detection consistency verification model, perform detection consistency verification on the target detection result set based on the detection consistency verification model, and output a first performance evaluation result according to the detection consistency verification result; the detection time evaluation module 50, the detection time evaluation module 50 is used to perform detection time evaluation based on the standard detection time set and the target detection time set, and output a second performance evaluation result according to the detection time evaluation result; the evaluation result output module 60, the evaluation result output module 60 is used to output the application performance evaluation result of the target HIV rapid detection reagent based on the first performance evaluation result and the second performance evaluation result.

[0069] Furthermore, the device also includes a detection consistency verification model building module to perform the following operation steps:

[0070] Acquire the test sample set, wherein the test sample set includes a case group, a control group, and an interference group in preset proportions; based on the case group, the control group, and the interference group, use a twin neural network to respectively construct a case group detection comparison channel, a control group detection comparison channel, and an interference group detection comparison channel; connect the case group detection comparison channel, the control group detection comparison channel, and the interference group detection comparison channel in parallel to obtain the detection comparison layer; construct a detection consistency verification layer, and cascade the detection comparison layer and the detection consistency verification layer to obtain the detection consistency verification model.

[0071] Furthermore, the detection consistency verification layer embeds the detection consistency evaluation function as follows:

[0072]

[0073] Among them, k is the detection consistency evaluation coefficient, P0 is the observation consistency coefficient, P e is the expected consistency coefficient.

[0074] Furthermore, the device further includes a first performance evaluation result output module to perform the following operation steps:

[0075] A detection four-grid table is constructed based on the output results of the case group detection comparison channel, the control group detection comparison channel, and the interference group detection comparison channel, wherein the data points of the detection four-grid table include true positive, false positive, true negative, and false negative; based on the detection four-grid table, the random consistency probability of the target detection result and the standard detection result is calculated to obtain the observed consistency coefficient and the expected consistency coefficient; based on the detection consistency evaluation function, the observed consistency coefficient and the expected consistency coefficient are calculated to obtain the detection consistency evaluation coefficient as the detection consistency verification result; when the detection consistency evaluation coefficient meets the preset detection consistency evaluation coefficient threshold, the first performance evaluation result is output as the evaluation passed.

[0076] Furthermore, the device also includes a detection consistency verification result output module to perform the following operation steps:

[0077] When the test consistency evaluation coefficient is not 1, a plurality of inconsistent test samples are extracted, and HIV antibody retests are performed on the plurality of inconsistent test samples to obtain a plurality of retest results; if the plurality of retest results have not changed, the test consistency evaluation coefficient is output as the test consistency verification result; if the plurality of retest results have changed, the test consistency evaluation coefficient is corrected based on the changed retest results, and the corrected test consistency evaluation coefficient is output as the test consistency verification result.

[0078] Furthermore, the device further includes a second performance evaluation result output module to perform the following operation steps:

[0079] The standard detection time set and the target detection time set are sorted based on the sample sequence number of the test sample set to generate a standard detection time data sequence and a target detection time data sequence; the standard detection time data sequence and the target detection time data sequence are standardized to generate a standard detection time feature sequence and a target detection time feature sequence; based on the standard detection time feature sequence, deviation calculation is performed on the target detection time feature sequence to obtain a detection time deviation feature sequence; based on the detection time deviation feature sequence, detection time feature evaluation is performed to calculate a detection time feature evaluation coefficient as the detection time evaluation result; when the detection time feature evaluation coefficient meets a preset detection time feature evaluation coefficient threshold, the second performance evaluation result is output as evaluation passed.

[0080] Furthermore, the device further includes a detection time feature evaluation coefficient acquisition module to perform the following operation steps:

[0081] The detection time deviation feature sequence is mean-calculated to obtain the detection time deviation central coefficient; the detection time deviation feature sequence is standard-calculated to obtain the detection time deviation dispersion coefficient; based on the detection time feature evaluation function, the detection time deviation central coefficient and the detection time deviation dispersion coefficient are calculated to generate the detection time feature evaluation coefficient; wherein the detection time feature evaluation function is as follows:

[0082]

[0083] Among them, T is the detection time characteristic evaluation coefficient, μ t is the detection time deviation center coefficient, σ t is the detection time deviation dispersion coefficient, ω1 and ω2 are the weights of the detection time deviation central coefficient and the detection time deviation dispersion coefficient respectively, and ω1+ω2=1.

[0084] Through the above-mentioned detailed description of the application performance evaluation method of an HIV rapid detection reagent in this specification, those skilled in the art can clearly understand the application performance evaluation device of an HIV rapid detection reagent in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0085] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for evaluating the application performance of a rapid HIV detection reagent, characterized in that: The method comprises: Acquire a test sample set, wherein the test sample set includes a test urine sample set and a homologous blood sample set collected synchronously; In a preset test scenario, a standard blood HIV antibody detection reagent is used to detect the homologous blood sample set to obtain a standard test result set and a standard test time set; In the preset test scenario, the test urine sample set is tested using a target HIV rapid test reagent to obtain a target test result set and a target test time set; Taking the standard detection result set as a comparison benchmark, a twin neural network is used to build a detection consistency verification model, based on the detection consistency verification model, the target detection result set is subjected to detection consistency verification, and a first performance evaluation result is output according to the detection consistency verification result; Performing a detection time evaluation based on the standard detection time set and the target detection time set, and outputting a second performance evaluation result according to the detection time evaluation result, the method comprising: Sort the standard detection time set and the target detection time set based on the sample sequence number of the test sample set to generate a standard detection time data sequence and a target detection time data sequence; Standardizing the standard detection time data sequence and the target detection time data sequence to generate a standard detection time feature sequence and a target detection time feature sequence; Taking the standard detection time feature sequence as a benchmark, performing deviation calculation on the target detection time feature sequence to obtain a detection time deviation feature sequence; Performing a detection time feature evaluation based on the detection time deviation feature sequence, and calculating a detection time feature evaluation coefficient as the detection time evaluation result; When the detection time characteristic evaluation coefficient meets a preset detection time characteristic evaluation coefficient threshold, outputting the second performance evaluation result as evaluation passed; The application performance evaluation result of the target HIV rapid detection reagent is output based on the first performance evaluation result and the second performance evaluation result.

2. The method for evaluating the application performance of a rapid HIV detection reagent according to claim 1, characterized in that: The method of using the standard test result set as a comparison benchmark and using a twin neural network to build a test consistency verification model includes: Acquire the test sample set, wherein the test sample set includes a case group, a control group, and an interference group in preset proportions; Based on the case group, control group, and interference group, a twin neural network is used to construct a case group detection comparison channel, a control group detection comparison channel, and an interference group detection comparison channel respectively; The case group detection comparison channel, the control group detection comparison channel, and the interference group detection comparison channel are connected in parallel to obtain a detection comparison layer; A detection consistency verification layer is constructed, and the detection comparison layer and the detection consistency verification layer are cascaded to obtain the detection consistency verification model.

3. The method for evaluating the application performance of a rapid HIV detection reagent according to claim 2, characterized in that: The detection consistency verification layer embeds the detection consistency evaluation function as follows: Among them, k is the detection consistency evaluation coefficient, P0 is the observation consistency coefficient, P e is the expected consistency coefficient.

4. The method for evaluating the application performance of a rapid HIV detection reagent according to claim 3, characterized in that: The performing detection consistency verification on the target detection result set based on the detection consistency verification model, and outputting a first performance evaluation result according to the detection consistency verification result, comprises: Constructing a detection four-grid table based on the output results of the case group detection comparison channel, the control group detection comparison channel, and the interference group detection comparison channel, wherein the data points of the detection four-grid table include true positive, false positive, true negative, and false negative; Based on the detection four-grid table, the random consistency probability of the target detection result and the standard detection result is calculated to obtain the observed consistency coefficient and the expected consistency coefficient; Calculating the observed consistency coefficient and the expected consistency coefficient based on the detection consistency evaluation function to obtain a detection consistency evaluation coefficient as a detection consistency verification result; When the detection consistency evaluation coefficient meets a preset detection consistency evaluation coefficient threshold, the first performance evaluation result is output as evaluation passed.

5. The method for evaluating the application performance of a rapid HIV detection reagent according to claim 4, characterized in that: The method further comprises: When the test consistency evaluation coefficient is not 1, extracting multiple inconsistent test samples, retesting the multiple inconsistent test samples for HIV antibodies, and obtaining multiple retest results; If the multiple re-test results have not changed, outputting the detection consistency assessment coefficient as the detection consistency verification result; If there are changes in the multiple re-inspection results, the detection consistency evaluation coefficient is corrected based on the changed re-inspection results, and the corrected detection consistency evaluation coefficient is output as the detection consistency verification result.

6. The method for evaluating the application performance of a rapid HIV detection reagent according to claim 1, characterized in that: The method of performing detection time feature evaluation based on the detection time deviation feature sequence and calculating a detection time feature evaluation coefficient comprises: Calculating the mean of the detection time deviation feature sequence to obtain a detection time deviation center coefficient; Calculating the standard deviation of the detection time deviation characteristic sequence to obtain a detection time deviation dispersion coefficient; Based on the detection time feature evaluation function, the detection time deviation central coefficient and the detection time deviation discrete coefficient are calculated to generate the detection time feature evaluation coefficient; Wherein, the detection time feature evaluation function is as follows: Among them, T is the detection time characteristic evaluation coefficient, μ t is the detection time deviation center coefficient, σ t is the detection time deviation dispersion coefficient, ω1 and ω2 are the weights of the detection time deviation central coefficient and the detection time deviation dispersion coefficient respectively, and ω1+ω2=1.

7. An application performance evaluation device for HIV rapid detection reagents, characterized in that: A method for evaluating the application performance of an HIV rapid detection reagent according to any one of claims 1 to 6, the device comprising: A test sample acquisition module, the test sample acquisition module is used to acquire a test sample set, wherein the test sample set includes a test urine sample set and a homologous blood sample set collected synchronously; A blood sample detection module, wherein the blood sample detection module is used to detect the homologous blood sample set using a standard blood HIV antibody detection reagent under a preset test scenario to obtain a standard test result set and a standard test time set; A urine sample detection module, wherein the urine sample detection module is used to detect the test urine sample set using a target HIV rapid detection reagent under the preset test scenario to obtain a target detection result set and a target detection time set; A consistency verification module, wherein the consistency verification module is used to use the standard detection result set as a comparison benchmark, use a twin neural network to build a detection consistency verification model, perform detection consistency verification on the target detection result set based on the detection consistency verification model, and output a first performance evaluation result according to the detection consistency verification result; A detection time evaluation module, the detection time evaluation module is used to perform detection time evaluation based on the standard detection time set and the target detection time set, and output a second performance evaluation result according to the detection time evaluation result; An evaluation result output module is used to output an application performance evaluation result of the target HIV rapid detection reagent based on the first performance evaluation result and the second performance evaluation result.

Images