Method and device for integrating and processing medical test result mutual recognition data

Abstract

The present application relates to the technical field of medical data management, and discloses a medical test result mutual recognition data integration processing method and device. The present application obtains multi-source clinical diagnosis and treatment data, fuses and standardizes the multi-source clinical diagnosis and treatment data, obtains standard clinical diagnosis and treatment data, constructs a mixed search architecture, searches the standard clinical diagnosis and treatment data based on a user query request, obtains effective clinical diagnosis and treatment data, performs mutual recognition evaluation reasoning based on the effective clinical diagnosis and treatment data, generates a mutual recognition report, and visually displays the mutual recognition report through a man-machine interaction interface. The present application constructs a neural-symbol mixed search architecture for the integrated clinical diagnosis and treatment data, and improves the efficiency of clinical diagnosis and treatment data search and viewing.

Description

Technical Field

[0001] This invention relates to the field of medical data management technology, specifically to a method, apparatus, electronic device, and storage medium for integrating and processing mutually recognized medical test results data. Background Technology

[0002] Patient clinical data includes laboratory data, examination data, clinical records, and follow-up records. This data may span multiple departments and even hospitals, resulting in a wide variety of data types, complex temporal relationships, and even heterogeneity issues. Furthermore, data interoperability rules depend on disease progression and specialty specificity. Therefore, simply merging clinical data can make searching and viewing difficult for doctors and cannot meet actual clinical needs. Summary of the Invention

[0003] This invention provides a method, apparatus, electronic device, and storage medium for integrating and processing medical test results for mutual recognition, in order to solve the technical problem of inconvenience for doctors in searching and viewing clinical diagnosis and treatment data integration in the prior art.

[0004] In a first aspect, the present invention provides a method for integrating and processing medical test result mutual recognition data, comprising: Acquire multi-source clinical diagnosis and treatment data, and fuse and standardize the multi-source clinical diagnosis and treatment data to obtain standard clinical diagnosis and treatment data; A hybrid retrieval architecture is constructed, and based on user query requests, the standard clinical diagnosis and treatment data is retrieved using the hybrid retrieval architecture to obtain effective clinical diagnosis and treatment data; Based on the effective clinical diagnosis and treatment data, mutual recognition assessment reasoning is performed using neural symbolic reasoning to generate a mutual recognition report, which is then visualized through a human-computer interaction interface.

[0005] The above implementation method obtains standard clinical diagnosis and treatment data by preprocessing the integrated clinical diagnosis and treatment data, then constructs a hybrid retrieval architecture to obtain effective clinical diagnosis and treatment data, and outputs mutually recognized reports based on a large language model to realize neural symbol reasoning. This improves the efficiency of clinical diagnosis and treatment data retrieval and viewing.

[0006] In one optional implementation, the process of fusing and standardizing the multi-source clinical diagnosis and treatment data to obtain standard clinical diagnosis and treatment data includes: Define a standardized access protocol and acquire multi-source clinical diagnosis and treatment data based on the standardized access protocol. The multi-source clinical diagnosis and treatment data includes: structured test data, semi-structured medical record text, and time-series labeled data. The multi-source clinical diagnosis and treatment data are standardized according to a preset format to obtain standard format data; A preset time alignment strategy is used to align the standard format data in time to obtain standard clinical diagnosis and treatment data.

[0007] The above implementation method improves mutual recognition compatibility by preprocessing the integrated clinical diagnosis and treatment data to obtain unified standard clinical diagnosis and treatment data.

[0008] In one optional implementation, the preset format includes a preset date format, and the standardization process of the multi-source clinical diagnosis and treatment data according to the preset format includes: Determine whether the patient corresponding to the clinical diagnosis and treatment data has a surgery date. If the patient corresponding to the clinical diagnosis and treatment data has a surgery date, determine the corresponding format of the generation time information based on the time interval between the generation time of the clinical diagnosis and treatment data and the surgery date. If the patient corresponding to the clinical diagnosis and treatment data does not have a surgery date, then the generation time information in the corresponding format is determined according to the time interval between the generation time of the clinical diagnosis and treatment data and the corresponding patient's clinical diagnosis and treatment time. The clinical diagnosis and treatment time includes: the corresponding patient's first visit date, first admission date, first test date, or first diagnosis date.

[0009] The above implementation method allows for flexible anchor point settings and standardized time information. In cases where the surgery date is missing, the clinical diagnosis and treatment time can be used as a supplement, eliminating the problems of inconsistent time formats and time zone misalignment of clinical diagnosis and treatment data across systems, hospitals, and time zones, and ensuring that there is always a reliable anchor point.

[0010] In one optional implementation, the hybrid retrieval architecture includes: vector retrieval, symbolic retrieval, and time-series retrieval. The standard clinical diagnosis and treatment data includes at least one report. The retrieval of the standard clinical diagnosis and treatment data using the hybrid retrieval architecture based on a user query request includes: Based on the vector retrieval, the semantic similarity score between the user query request and each report is calculated; Based on the symbol retrieval, determine whether each report meets the mutual recognition rules to obtain the symbol rule score; Based on the time-series retrieval, calculate the time-series relevance score between each report and the user's query request time; The semantic similarity score, symbol rule score, and temporal relevance score of each report are weighted and calculated to obtain the retrieval score of each report. The reports are then sorted from high to low according to the retrieval scores, and the top preset number of reports are taken as valid clinical diagnosis and treatment data.

[0011] The above implementation method uses three dimensions—semantic similarity score, symbol rule score, and temporal relevance score—to perform query matching, taking into account both semantic understanding and medical logic, thereby enhancing the objectivity of the ranking of each report.

[0012] In one optional implementation, determining whether each report satisfies the mutual recognition rule to obtain a symbol rule score includes: The mutual recognition rules include several mutual recognition sub-rules; The symbol rule score for each report is determined based on the number of times each report meets the mutual recognition sub-rules.

[0013] The above implementation method performs symbol retrieval through multiple mutual recognition sub-rules and uses the number of matches as the basis for judgment. The mutual recognition rules can be used to increase the screening dimensions of clinical diagnosis and treatment data, and prioritize recommending clinical diagnosis and treatment data with practical reference value to query users.

[0014] In one optional implementation, the step of weighting the semantic similarity score, symbol rule score, and temporal relevance score of each report to obtain the retrieval score of each report includes: A first weight is assigned to the semantic similarity score, a second weight is assigned to the symbol rule score, and a third weight is assigned to the temporal relevance score; The first weight, the second weight, and the third weight are dynamically adjusted based on user feedback on the mutual recognition report to optimize the weighted calculation of the search score.

[0015] The above implementation method assigns weights to semantic similarity scores, symbol rule scores, and temporal relevance scores. The dynamic adjustment of these weights can achieve closed-loop adjustment, thereby enhancing the objectivity of the ranking of each report.

[0016] In one optional implementation, based on the effective clinical diagnosis and treatment data, mutual recognition assessment reasoning is performed using neural symbolic reasoning to generate a mutual recognition report, including: Construct prompt words, which include at least valid clinical diagnosis and treatment data, at least a summary of content extracted from mutual recognition rules, and system instructions that define a preset format for mutual recognition reports; The prompt words are input into a large language model to generate a structured mutual recognition report; Based on preset logical consistency constraints, the structured mutual recognition report is verified. If the verification passes, it is used as the final mutual recognition report.

[0017] The above implementation method utilizes a large language model to assist in generating the final structured mutual recognition report. It improves the interactive convenience of the large language model through prompt word generation logic, helping users obtain more accurate and rigorous mutual recognition report content.

[0018] Secondly, the present invention provides a medical test result mutual recognition data integration and processing device, comprising: The data acquisition module is used to acquire multi-source clinical diagnosis and treatment data, and to fuse and standardize the multi-source clinical diagnosis and treatment data to obtain standard clinical diagnosis and treatment data. The query and retrieval module is used to construct a hybrid retrieval architecture and, based on user query requests, use the hybrid retrieval architecture to retrieve the standard clinical diagnosis and treatment data to obtain effective clinical diagnosis and treatment data. The report output module is used to perform mutual recognition assessment reasoning based on the effective clinical diagnosis and treatment data, generate a mutual recognition report, and visualize the mutual recognition report through a human-computer interaction interface.

[0019] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the medical test result mutual recognition data integration processing method of the first aspect or any corresponding embodiment described above.

[0020] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the medical test result mutual recognition data integration processing method of the first aspect or any corresponding embodiment described above. Attached Figure Description

[0021] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the overall technical architecture for the integration and processing of medical test result mutual recognition data according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the first type of data integration and processing method for mutual recognition of medical test results according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the second process of the medical test result mutual recognition data integration and processing method according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the hybrid retrieval architecture according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the mutual recognition sub-rules in an embodiment of the present invention; Figure 6This is a schematic diagram of the prompt word content in an embodiment of the present invention; Figure 7 This is a schematic diagram of the content of the mutual recognition report module A in an embodiment of the present invention; Figure 8 This is a schematic diagram of the contents of the mutual recognition report module B and module C in an embodiment of the present invention; Figure 9 This is a structural block diagram of the medical test result mutual recognition data integration and processing device according to an embodiment of the present invention; Figure 10 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0025] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0026] In related research, based on a real clinical chronic disease dataset of 200 cases from tertiary hospitals in China, including patients with malignant lung tumors and esophageal tumors, and 14 types of laboratory tests, the core pain points of current clinical diagnosis and treatment data processing are analyzed and summarized as follows: First, the integration of multi-dimensional heterogeneous data is challenging. Clinical data is complex, including patient baseline information with 25 dimensions such as age at diagnosis, comorbidities, and smoking history; time-series laboratory data, including multiple follow-up records such as blood gas analysis and blood cell analysis; and data from 14 types of specialized examinations. Furthermore, there is data sparsity across time windows, and some laboratory items have very little postoperative follow-up data, which hinders integrated analysis.

[0027] Second, the temporal relationships are complex. The same test item may be scattered across different test categories, requiring precise alignment by combining patient ID and timestamp; otherwise, data confusion may occur, affecting mutual recognition judgment.

[0028] Third, the mutual recognition rules depend on the evolution of the patient's condition. A single test result cannot be used as a basis for mutual recognition. For example, in blood cell analysis, a patient's white blood cell count may change regularly due to surgical stress after surgery. If only a single result is considered, the risk of postoperative infection may be overlooked. In addition, some patients have data entry anomalies, requiring additional quality control marking.

[0029] Fourth, it has strong specialty specificity. Different diseases, such as lung cancer and esophageal cancer, have significantly different mutual recognition items and risk thresholds. Lung cancer patients need to focus on the dynamic changes of blood gas and tumor markers, while esophageal cancer patients need to focus on electrolyte and nutritional indicators. General mutual recognition standards cannot meet clinical needs.

[0030] Fifth, there is a heterogeneity in the types of test results, with numerical, semi-quantitative, and qualitative data coexisting. The different ways of representing different types of data in the test results vary greatly, and they need to be uniformly processed and cleaned before effective comparison and mutual recognition can be achieved.

[0031] Therefore, for cross-system (such as cross-department, cross-hospital) data integration and processing, especially for clinical diagnosis and treatment data including mutual recognition of medical test and examination results, there is an urgent need for effective management through standardized data interface communication and a unified data storage system.

[0032] like Figure 1 As shown in the figure, this invention proposes a four-layer data integration system, consisting of an intelligent application layer, a neural symbolic reasoning layer, a hybrid retrieval layer, and a multi-source data fusion layer. Each layer communicates through standardized data interfaces, achieving a complete technical chain of upward abstraction and downward dependency. Specifically, the first layer, the multi-source data fusion layer, is responsible for data governance, temporal alignment, and knowledge graph construction. The second layer, the hybrid retrieval layer, is responsible for fusing vector semantic retrieval, logical filtering, and temporal indexing. The third layer, the neural symbolic reasoning layer, is responsible for combining an LLM (Large Language Model) engine, RAG (Retrieval-augmented Generation) technology, and symbolic rule verification. The fourth layer, the intelligent application layer, is responsible for report generation, visualization, and doctor interaction functions.

[0033] According to an embodiment of the present invention, a method for integrating and processing medical test results for mutual recognition is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0034] This embodiment provides a method for integrating and processing mutually recognized medical test results, which can be applied to electronic devices such as computers, servers, and mobile phones. Figure 2 This is a flowchart of a medical test result mutual recognition data integration and processing method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: Obtain multi-source clinical diagnosis and treatment data, and fuse and standardize the multi-source clinical diagnosis and treatment data to obtain standard clinical diagnosis and treatment data.

[0035] Optionally, step S201 can be referred to Figure 1 This is achieved through a multi-source data fusion layer. Clinical diagnosis and treatment data obtained from multiple data sources undergo pre-processing and cleaning to address issues such as missing data, anomalies, and inconsistent time periods. For missing data, processing is tiered based on the proportion of missing data: completely empty or missing data is marked as "not tested," while data with excessive missing data is marked as "to be supplemented." "Completely empty" means that a specific test category for a particular patient has no valid data in the database table; it doesn't mean the entire database or a specific indicator is empty, but rather that the entire database table is empty for that patient. For example, in the pancreatic function data table, most patients have no test records, and all rows for that patient in the table are null values. The system marks this as "not tested" rather than imputation, avoiding the use of unfounded generated data. For anomaly data, marking is done by combining reference ranges and clinical experience to distinguish between numerical anomalies, data entry errors, and semi-quantitative indicator anomalies. For inconsistent time periods, a unified relative time coordinate system is established with the surgery date as the anchor point for all dates, facilitating time-series analysis.

[0036] Step S202: Construct a hybrid retrieval architecture and, based on user query requests, use the hybrid retrieval architecture to retrieve standard clinical diagnosis and treatment data to obtain effective clinical diagnosis and treatment data.

[0037] Optionally, step S202 can be referred to Figure 1 This is achieved through a hybrid retrieval layer. The embodiments of this invention balance semantic understanding and medical logic, making the results of matching query requests more valuable. It can filter out clinical diagnostic data that does not meet the needs of doctors, and input this data into a large language model for reasoning. This minimizes the possibility of the large language model generating false results and ensures that the output conforms to medical rigor.

[0038] Step S203: Based on effective clinical diagnosis and treatment data, use neural symbolic reasoning to perform mutual recognition assessment reasoning, generate a mutual recognition report, and visualize the mutual recognition report through a human-computer interaction interface.

[0039] Optionally, step S203 can be referred to Figure 1 It is implemented by a neural symbolic reasoning layer and an intelligent application layer. Neural symbolic reasoning introduces mutual recognition rules on the basis of large language model reasoning, thereby improving the ability of logical thinking and making the output mutual recognition report conform to the rigor of medicine.

[0040] This embodiment provides a method for integrating and processing mutually recognized medical test results, which can be applied to electronic devices such as computers, servers, and mobile phones. Figure 3 This is a flowchart of a medical test result mutual recognition data integration and processing method according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps: Step S301: Obtain multi-source clinical diagnosis and treatment data, and perform fusion and standardization processing on the multi-source clinical diagnosis and treatment data to obtain standard clinical diagnosis and treatment data.

[0041] Specifically, step S301 includes: Step 3011: Define a standardized access protocol and acquire multi-source clinical diagnosis and treatment data based on the standardized access protocol. Multi-source clinical diagnosis and treatment data includes: structured laboratory test data, semi-structured medical record text, and time-series labeled data.

[0042] The multi-source heterogeneous data can originate from different hospitals. For example, data exported from the LIS (Laboratory Information System) after mutual recognition approval by various hospitals can be transmitted to the clinical laboratory centers in each province and stored in the database. Correspondingly, a standardized access protocol can be defined through the multimodal medical data access module to support the access of structured laboratory test data, semi-structured medical record texts, and time-series labeled data. Structured laboratory test data is data managed in a standard format and can be associated with patient entities through database IDs, registration numbers, names, etc. Semi-structured medical record texts can be used to extract data, with key information including five categories: demographic characteristics (age, gender, height, weight), which determines the reference range selection; clinical diagnosis (e.g., esophageal malignancy), which is the most crucial and directly determines which mutual recognition rule applies; comorbidities (e.g., diabetes, hypertension), affecting the mutual recognition threshold; risk factors (e.g., smoking history, family history), affecting the mutual recognition risk level; and surgical time, serving as the benchmark anchor point for the time-series coordinate system. It can extract key information through Named Entity Recognition (NER). Specifically, it automatically extracts and categorizes machine-understandable keywords from doctors' handwritten medical records. For example, if the chief complaint is "difficulty swallowing for more than 2 months, diagnosed with esophageal squamous cell carcinoma 2 months ago," NER will automatically identify: dysphagia - symptom, esophageal squamous cell carcinoma - disease, etc. In this way, the rule engine can automatically determine which mutual recognition rules apply to the patient based on these structured labels, without the need for manual labeling. Time-series labeled data can be a compilation of multiple follow-up records of the same patient for the same test item.

[0043] Step 3012: Standardize the multi-source clinical diagnosis and treatment data according to the preset format to obtain standard format data.

[0044] Clinical diagnosis and treatment data management uses the registration number as the primary key for cross-table JOIN operations. This is used to definitively link all clinical diagnosis and treatment data, such as test and examination data, for the same patient. The database ID is used only for intra-table location, serving as an auxiliary location key. The name also serves as an auxiliary location key, used for readability and manual verification. There is a clear priority among the registration number, database ID, and name, with the registration number having higher priority than the database ID and name. Accordingly, if the registration numbers are the same, they are considered the same patient, regardless of whether the names are the same. If the registration numbers are different, even if the names are the same, they are considered different patients. If the names are different but the registration numbers are the same, a name difference log is recorded, using the latest or most complete name as the standard, recording all historical name variations and patient entity files. If the database ID conflicts with the registration number or name, the database ID is ignored, and the registration number takes precedence.

[0045] In some alternative implementations, step 3012 above includes: The preset format includes a preset date format. Multi-source clinical diagnosis and treatment data are standardized according to the preset format, including: Step a1: Determine if the patient corresponding to the clinical diagnosis and treatment data has a surgery date. If the patient corresponding to the clinical diagnosis and treatment data has a surgery date, determine the corresponding format of the generation time information based on the time interval between the generation time of the clinical diagnosis and treatment data and the surgery date. Step a2: If the patient corresponding to the clinical diagnosis and treatment data does not have a surgery date, then determine the corresponding format of the generation time information based on the time interval between the generation time of the clinical diagnosis and treatment data and the corresponding patient's clinical diagnosis and treatment time. The clinical diagnosis and treatment time includes: the corresponding patient's first visit date, first admission date, first test date, or first diagnosis date.

[0046] By flexibly setting anchor points, the system eliminates the problems of inconsistent time formats and time zone misalignments in clinical diagnosis and treatment data across systems, hospitals, and time zones, ensuring a consistently reliable anchor point. For patients with prior surgical records, the surgery date is directly used as the anchor point. For patients without prior surgical records, such as those undergoing conservative treatment, chemotherapy / radiotherapy, or those who only had examinations without surgery, the anchor point can be selected based on priority: first visit / admission date, first test date, or first diagnosis date. The system checks if the surgery date is empty upon data access. If it is not empty, the surgery date is used directly as T0; otherwise, the earliest available event time is automatically selected as T0, and the anchor point type is noted in the report to remind the doctor, such as the first visit date. For example, if a laboratory report is generated 8 hours before T0, the corresponding generation time information would be -PT8H; if it is generated 3 days and 8 hours after T0, the corresponding generation time information would be P3DT8H.

[0047] Optionally, for cases involving multiple surgeries in the past, the date of the surgery most relevant to the current mutual recognition decision can be used as the anchor point. If multiple surgeries are for the same disease (e.g., secondary resection), the date of the most recent surgery is taken as T0, as postoperative indicators reflect the latest status. If they are for different diseases, such as first undergoing radical resection for rectal cancer and then undergoing lung resection due to lung metastasis, the surgery date corresponding to the reason for this visit is taken as T0, and the remaining surgery dates are marked as auxiliary events on the timeline. The "General Information and Medical History" notes field in the data can help determine the correspondence between the reason for the visit and the surgery.

[0048] Step 3013: Using a preset time alignment strategy, the standard format data is time-aligned to obtain standard clinical diagnosis and treatment data.

[0049] To address the issue of inconsistent time granularity among various types of clinical diagnosis and treatment data, strategies such as precise alignment, window alignment, and interpolation alignment can be adopted. The choice of strategy can be determined based on the type of data, etc. Clinical diagnosis and treatment data at different time points can be integrated into a unified time series matrix, and data at missing time points can be labeled to ensure the integrity and consistency of time series data.

[0050] The following describes three strategies for precise alignment, window alignment, and interpolation alignment.

[0051] Input: The test set of patient P Each record contains a timestamp. With project portfolio .

[0052] Output: Unified timing matrix , It is a timeline (an ordered set of time points after deduplication and merging). The union of all test items.

[0053] Precise alignment: Multiple tests on the same date are grouped into the same time point. .

[0054] Window alignment: Set mutual recognition time window (Default 72 hours) The weighted average of the tests within the window is combined to form the same time point: , in, It could be the standardized single value of patient i's test item j within the same relative surgical time window W, after exponential time decay weighted fusion. This can represent the k-th test record corresponding to test item j for patient i within the time window W, where λ is the time decay coefficient, which defaults to 0.1. The time point at the center of the window, This refers to the time point of the k-th test record. Aligned time points can be used consistently. And label the window alignment.

[0055] Interpolation alignment: For missing data at time t, the following formula can be used for calculation: , in, The values ​​are the test results after linear interpolation, where t is the missing time point. Let t be the time point before time t. Let t be a time point after time t. for The test value at time. for The test value at any given time.

[0056] Step S302: Construct a hybrid retrieval architecture and, based on user query requests, use the hybrid retrieval architecture to retrieve standard clinical diagnosis and treatment data to obtain effective clinical diagnosis and treatment data.

[0057] like Figure 4 As shown, in order to respond to user queries, it integrates the semantic perception capability of neural networks with the logical reasoning capability of symbol systems to construct a hybrid retrieval architecture that combines vector retrieval, symbol retrieval, and time-series retrieval. At the same time, it obtains the final comprehensive query matching score by analyzing the determined semantic similarity score, symbol rule score, and time-series relevance score, and by weighting and fusing them through hybrid retrieval intersection operations. The comprehensive query matching score can be used for ranking and can even be used as a credibility score.

[0058] Specifically, step S302 includes: The hybrid retrieval architecture includes vector retrieval, symbolic retrieval, and time-series retrieval. Standard clinical diagnostic data includes at least one report. These reports can specifically be laboratory reports, medical records, follow-up reports, etc., and selection can be made from multiple reports.

[0059] Based on user query requests, a hybrid retrieval architecture is used to retrieve the standard clinical diagnosis and treatment data, including: Step b1: Based on vector retrieval, calculate the semantic similarity score between the user's query request and each report.

[0060] Vector retrieval involves encoding patients' clinical data into 768-dimensional vectors using a medical language model. For example, test reports, including test names, diagnostic descriptions, and result interpretations, can be converted into report vectors—essentially encoding text into dense vectors and storing them in the Milvus vector database. When a doctor queries, the user's query request is also converted into a query vector. Cosine similarity analysis between the query vector and the report vector identifies the most semantically similar clinical data. For instance, a query for "post-lung cancer surgery tumor marker follow-up" will match reports for tumor markers such as CEA and CA199. The semantic similarity score can be obtained using the following formula: , in, For semantic similarity scores, For query vector, Here, r is the report vector, and r is the corresponding report. The smaller the cosine angle between two vectors, the higher their semantic similarity. The result range is [-1, 1]. To facilitate subsequent weighted calculations, the result will be... The result range is normalized to [0,1]. Furthermore, vector retrieval also supports matching at multiple granularities, such as single indicator level, single report level, and full patient history level, to meet the retrieval needs of different granularities.

[0061] Step b2: Based on symbol retrieval, determine whether each report meets the mutual recognition rules to obtain the symbol rule score.

[0062] Symbolic retrieval involves writing predefined mutual recognition rules as explicit IF-THEN logic. Examples include rules such as lifelong mutual recognition of blood types, one-year validity for negative infectious disease tests, and non-mutual recognition if postoperative WBC (White Blood Cell) count doubles. By calling these rules, the system checks if the current data meets the conditions and outputs definitive conclusions such as "mutually recognized," "non-mutually recognized," or "warning required." An interpretable mutual recognition rule knowledge base is constructed using a symbolic rule engine, including rule types such as timeliness, disease change thresholds, specialty specificity, and mutual recognition project whitelists. This covers core scenarios for mutual recognition assessment, such as the timeliness of blood cell analysis mutual recognition, dynamic monitoring of tumor markers, and permanent blood type mutual recognition, providing clear evidence for mutual recognition decisions. Preferably, mutual recognition rules are formalized as five-tuples, consisting of attributes such as ID, Condition, Action, Priority, and Source.

[0063] In some alternative implementations, step b2 above includes: Step b201: The mutual recognition rules include several mutual recognition sub-rules; Step b202: Determine the symbol rule score for each report based on the number of compliances of each report with the mutual recognition sub-rules.

[0064] Figure 5 Several mutual recognition sub-rules are described exemplarily: The RULE_T001 rule is used to detect whether a patient's test result is less than or equal to 30 days ago and the patient is stable. The RULE_T002 rule is used to check whether a patient's negative infectious disease test result is valid within one year. The RULE_C001 rule is used to detect whether the postoperative change rate of white blood cells is less than 50%. The RULE_C002 rule is used to detect whether a patient's same laboratory indicator shows a sustained, significant, and linear decline over time. The RULE_W001 rule is used to check whether data conforms to the relevant directory requirements; The RULE_W002 rule checks whether a patient's blood type is a permanently mutually recognized ABO or Rh blood type.

[0065] Accordingly, all mutual recognition sub-rules are iterated through, and each one is compared with a specific report r in the clinical diagnosis and treatment data. If they match, they are counted as 1; if they do not match, they are counted as 0. Finally, the symbol rule score is calculated according to the following formula: , in, Score for symbol rules, The number of iterations required for mutually recognized sub-rules. The judgment value for the j-th mutual recognition sub-rule is calculated as follows: Let 'r' be the corresponding report. For example, if RULE_T001, RULE_T002, RULE_C001, RULE_C002, RULE_W001, and RULE_W002 represent all mutual recognition sub-rules, and each report 'r' is evaluated, and five of the mutual recognition sub-rules are found to be compatible, then the cumulative compatibility count will be 5. That equals 5 / 6.

[0066] Step b3: Based on time-series retrieval, calculate the time-series relevance score between each report and the time of the user's query request.

[0067] Time-series retrieval is based on a time index with the surgery date as an anchor point, which can quickly find all test values, trend directions and cross-indicator correlations of a certain indicator within a specified time range, such as whether blood sodium continues to decrease, whether blood potassium and blood calcium decrease in the same step, and whether the relevant test time is too long.

[0068] In some alternative implementations, step b3 above includes: The time-series relevance score is determined based on the difference between the time of the user's query request and the time of the generation of the clinical diagnosis and treatment data.

[0069] The reference value of a report is determined by measuring the strength of the correlation between the generation time of the corresponding report in clinical diagnosis and treatment data (such as the time of testing or examination) and the current query time. Specifically, the time-series correlation score is calculated according to the following formula: , in, For time-series correlation scores, For the current query time, The time point from which the corresponding report was generated is represented, λ is the decay coefficient (default is 0.1), and r is the corresponding report. Therefore, the closer the corresponding report is to the current time point, the higher the time-series relevance score; the further away it is, the lower the time-series relevance score.

[0070] Furthermore, time-series retrieval also involves trend consistency judgment, which can be achieved through MTI (Multi-Time Index) index primitives for medical time-series data. A dedicated data structure and four types of retrieval operation primitives are designed to quickly realize time-range retrieval, trend detection, cross-project correlation analysis, and mutual recognition filtering, ensuring the efficiency and accuracy of time-series data retrieval. The four types of primitive calculations include: RANGE_QUERY retrieves all values ​​within the time range; TREND_DETECT uses the Pearson correlation coefficient to determine whether there is a continuous increase / decrease; CROSS_PROJECT_CORR uses the Spearman coefficient to determine whether multiple indicators are abnormally correlated; and MUTUAL_RECOG_FILTER performs mutual recognition filtering. The return value of each index operation is a triple: a mutually recognized set, a non-mutually recognized set (with an interpretable rule number), and a warning set. For example, when the correlation coefficient ( A correlation coefficient less than 0.5 indicates a stable trend, and the time-series relevance score remains the value calculated by the above formula. Conversely, a correlation coefficient greater than 0.5 indicates an abnormal trend, and the time-series relevance score is directly set to 0. This ensures that time-series retrieval focuses not only on the newness of the corresponding report but also on the stability of the indicators within that report. New but unstable reports will also have lower time-series relevance scores.

[0071] Step b4: Calculate the weighted scores of semantic similarity, symbol rule, and temporal relevance of each report to obtain the retrieval score of each report. Sort the reports from high to low according to the retrieval scores and select the first preset number of reports as valid clinical diagnosis and treatment data.

[0072] In some alternative implementations, step b4 above includes: Step b401: Assign a first weight to the semantic similarity score, a second weight to the symbol rule score, and a third weight to the temporal relevance score.

[0073] Specifically, the retrieval score can be calculated using the following formula: , in, To obtain a score, For semantic similarity scores, Score for symbol rules, Here, r represents the time-series relevance score, r is the corresponding report, α is the configured first weight, β is the configured second weight, and γ is the configured third weight, with α+β+γ=1. Since the semantic similarity score, symbol rule score, and time-series relevance score are all normalized to [0,1], the overall determined retrieval score will also be between 0 and 1. The higher the retrieval score, the more suitable it is for mutual recognition.

[0074] Step b402: The first weight, the second weight, and the third weight are dynamically adjusted based on user feedback on the mutual recognition report to optimize the weighted calculation of the search score.

[0075] In this model, the first weight can be greater than the second weight, and the second weight can be greater than the third weight. Correspondingly, the first, second, and third weights represent the importance of semantic, rule, and time-related factors, respectively. For example, α=0.4, β=0.35, and γ=0.25. These can be default parameters pre-set by the model or values ​​fine-tuned based on actual matching results. The semantic weight is set highest because semantic similarity matching is the primary goal of retrieval. The rule weight is set second, as while mutual recognition matching has compliance requirements, it is not as direct as semantic similarity matching in terms of query intent. The time weight is set lowest, as the time factor plays a more auxiliary filtering role. These three weights are not fixed and can be automatically adjusted based on doctors' actual adoption feedback through Bayesian optimization to adapt to the mutual recognition preferences of different departments and diseases. Finally, the weighted fusion score of the three factors ensures semantic relevance, compliance with rule logic, and satisfaction of time constraints, meaning that medical testing time must be within the mutual recognition validity period and have a temporal correlation with the current consultation.

[0076] Step S303: Based on effective clinical diagnosis and treatment data, use neural symbolic reasoning to perform mutual recognition assessment reasoning, generate a mutual recognition report, and visualize the mutual recognition report through a human-computer interaction interface.

[0077] Specifically, step S303 includes: Step c1: Construct prompt words. Prompt words should include at least valid clinical diagnosis and treatment data, at least a summary of the content extracted from the mutual recognition rules, and system instructions that define the preset format of the mutual recognition report.

[0078] Figure 6This example describes the generation of prompts using Wang Moumou as the query target. In addition to generating basic patient information and tasks based on the user's query request, the generated prompts can also include valid clinical diagnosis and treatment data and mutual recognition rules to assist the large language model in outputting a mutual recognition report. This ensures the mutual recognition report contains rigorous data and logical analysis and serves as a reference for subsequent output of mutual recognition item suggestion tables and non-mutual recognition item risk explanations. To control the context size and reduce lexical consumption, summary content can be extracted from clinical diagnosis and treatment data and mutual recognition rules to construct prompts. For example, named entities can be used to extract test indicators from laboratory reports and relevant medical record fragments from medical records. Furthermore, the content of mutual recognition rules can be simplified; for instance, only mutual recognition rules that trigger non-mutual recognition or require warnings can be included as summaries in the prompts. Furthermore, the large language model that outputs the report also employs constraint decoding technology, which forces the large language model to output reports that conform to the standard JSON (JavaScript Object Notation) format. That is, system instructions that define the preset format of mutually recognized reports are introduced in the construction prompts. The required fields, field types, nested structures, numerical ranges, array rules and other format information of the output report can be clearly defined through JSON files and other means.

[0079] Step c2: Input the prompt words into the large language model to generate a structured mutual recognition report.

[0080] Large language models can be specific to the medical field. For example, an open-source LLM model fine-tuned from a Chinese medical corpus, such as Medical-Qwen3-14B, can support the understanding of long medical records. Inputs can include structured test data, text medical records, time-series trends, and mutual recognition rules. More accurate and reasonable prompts are then input into the large language model, which ultimately generates a structured mutual recognition report. The specific content of the structured mutual recognition report may include an overview of the patient's time-series test results, a mutual recognition item recommendation table, and a risk explanation for non-mutual recognition items. The mutual recognition item recommendation table and / or the risk explanation for non-mutual recognition items can help users understand the clinical diagnosis and treatment data in the mutual recognition report. Figure 7 , Figure 8 An exemplary description of the mutual recognition report output by a large language model, including... Figure 7 Module A (Patient Time Series Test Overview) in the document. Figure 8The system comprises Module B (Mutual Recognition Item Recommendation Table) and Module C (Non-Mutual Recognition Item Risk Description). In clinical practice, the results can be presented through an intelligent report integration generator and a doctor-interactive review interface. The intelligent report integration generator outputs a structured mutual recognition report, covering an overview of patient time-series tests, a mutual recognition item recommendation table, and a risk description of non-mutual recognition items. In addition, it can output time-series trend graphs, which are used to generate time-series line graphs of key indicators. Descriptive guidelines are automatically annotated in the time-series line graphs: shaded areas indicate upper and lower limits of the reference range; red solid circles mark outliers; orange vertical dashed lines mark surgical or treatment event lines; green background areas mark mutual recognition time windows; and rising / falling / stable arrows mark the trend direction. The doctor-interactive review interface supports human-machine collaborative decision-making. Doctors, as users, can review and modify system suggestions, view the source of original data, rule basis, and similar case references. Feedback records can be used for subsequent reinforcement learning and iterative optimization of the model, such as closed-loop fine-tuning of the first, second, and third weights mentioned above.

[0081] Step c3: Based on the preset logical consistency constraints, verify the structured mutual recognition report. If the verification passes, it will be used as the final mutual recognition report.

[0082] The constraint decoding technology is also reflected in the automatic verification of mutual recognition reports by using at least one of the five formal logical consistency constraints in Table 1, ensuring that mutual recognition recommendations are logically consistent and that the rules are applied accurately, thereby improving the reliability of mutual recognition reports.

[0083] Table 1. List of Logical Consistency Constraints

[0084] This invention eliminates the need for manual data input by users. It automatically retrieves all clinical data of patients directly from the hospital information system. Doctors simply select or open a patient's profile in the system, and the system automatically completes data access, data governance and quality control, time-series alignment, mixed retrieval and rule matching, and report generation. Accordingly, the scope of clinical data can include three categories: Basic patient information, including name, gender, age, registration number, clinical diagnosis, and surgery time, is used to establish patient identity and time-series anchor point T0.

[0085] The full range of test data includes all test records for the patient across all test categories, including blood cell analysis, blood gas analysis, C12 tumor markers, urine analysis, and pre-transfusion examinations. Each record contains the test date, test name, test value, reference range, and result type (commonly numerical, semi-quantitative, and qualitative). These data are used to construct a time-series matrix and trigger rule judgments.

[0086] For example, a post-esophageal cancer patient is transferred from Hospital A to Hospital B for continued treatment. The attending physician at Hospital B opens the patient's file, and the system automatically retrieves all of the patient's clinical treatment data from Hospital A, arranges it chronologically, and marks abnormal trends, such as a sudden drop in blood potassium from 4.0 to 2.1. It also provides mutual recognition suggestions, such as blood type being mutually recognized, but blood gas analysis not being mutually recognized due to unstable condition and requiring retesting, resulting in an integrated, information-based mutual recognition report. The doctor can confirm with one click or mark items as not mutually recognized and have the patient undergo the tests again. Throughout the entire process, the doctor does not need to manually input any data; they only need to make the final decision on the mutual recognition report.

[0087] In some examples, the data integration and processing employs data anonymization techniques, replacing sensitive patient information and storing it encrypted. All mutually recognized decision records are written to an immutable log, enabling end-to-end auditing and traceability, and ensuring data security and privacy.

[0088] In this embodiment of the invention, a temporal anomaly detection algorithm, combined with change point detection and medical rules, is used to accurately identify and classify abnormal changes in time-series data, distinguishing between expected physiological fluctuations, pathological changes, and data quality issues, thus providing evidence to support the mutual recognition of test results. Furthermore, cross-modal data alignment and fusion technology constructs a unified patient representation, integrating four types of information: numerical, textual, temporal, and knowledge graph, further eliminating obstacles to reasoning and analysis of heterogeneous data. Finally, adaptive learning of mutual recognition rules enables continuous iteration of the rule base. In validation based on a 200-case real clinical dataset, the system's core indicators—mutual recognition accuracy, recall, and physician adoption rate—all met expectations. The rule conflict rate was low, and the report generation latency was low, reducing the time for physicians to review patient test reports and complete mutual recognition verification to 20% of the original time, significantly improving diagnostic and treatment efficiency while ensuring the accuracy and logic of mutual recognition judgments. The embodiments of the present invention fully consider the temporal, heterogeneous, sparse, and specialty-specific nature of clinical data, ensuring the accuracy and logical consistency of mutual recognition judgments, while also improving doctors' diagnostic and treatment efficiency. This provides a technical upgrade for the implementation of the mutual recognition system for medical institution examination and test results, and has significant clinical application value.

[0089] According to embodiments of the present invention, a medical test result mutual recognition data integration and processing apparatus is also provided. This apparatus is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0090] This embodiment provides a medical test result mutual recognition data integration and processing device, such as... Figure 9 As shown, it includes: The data acquisition module 901 is used to acquire multi-source clinical diagnosis and treatment data, and to fuse and standardize the multi-source clinical diagnosis and treatment data to obtain standard clinical diagnosis and treatment data. The query and retrieval module 902 is used to construct a hybrid retrieval architecture and, based on user query requests, use the hybrid retrieval architecture to retrieve standard clinical diagnosis and treatment data to obtain effective clinical diagnosis and treatment data. The report output module 903 is used to perform mutual recognition assessment reasoning based on valid clinical diagnosis and treatment data using neural symbolic reasoning, generate a mutual recognition report, and visualize the mutual recognition report through a human-computer interaction interface.

[0091] The medical test result mutual recognition data integration and processing device provided in this embodiment of the invention can execute the data integration and processing method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0092] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0093] The following is a detailed reference. Figure 10 This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from memory 1008 into random access memory (RAM) 1003. The RAM 1003 also stores various programs and data required for the operation of the electronic device. The processor 1001, ROM 1002, and RAM 1003 are interconnected via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0094] Typically, the following devices can be connected to the I / O interface 1005: input devices 1006 including, for example, a touchscreen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 1007 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; memory devices 1008 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows electronic devices to exchange data via wireless or wired communication with other devices. Although Figure 10 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0095] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 1009, or installed from a memory 1008, or installed from a ROM 1002. When the computer program is executed by the processor 1001, it performs the functions defined in the medical test result mutual recognition data integration processing method of the embodiments of the present invention.

[0096] Figure 10 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0097] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the medical test result mutual recognition data integration processing method shown in the above embodiments is implemented.

[0098] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for integrating and processing medical test results for mutual recognition, characterized in that, The method includes: Acquire multi-source clinical diagnosis and treatment data, and fuse and standardize the multi-source clinical diagnosis and treatment data to obtain standard clinical diagnosis and treatment data; A hybrid retrieval architecture is constructed, and based on user query requests, the standard clinical diagnosis and treatment data is retrieved using the hybrid retrieval architecture to obtain effective clinical diagnosis and treatment data; Based on the effective clinical diagnosis and treatment data, mutual recognition assessment reasoning is performed using neural symbolic reasoning to generate a mutual recognition report, which is then visualized through a human-computer interaction interface.

2. The method for integrating and processing medical test result mutual recognition data according to claim 1, characterized in that, The process of fusing and standardizing the multi-source clinical diagnosis and treatment data to obtain standard clinical diagnosis and treatment data includes: Define a standardized access protocol and acquire multi-source clinical diagnosis and treatment data based on the standardized access protocol. The multi-source clinical diagnosis and treatment data includes: structured laboratory test data, semi-structured medical record text, and time-series labeled data. The multi-source clinical diagnosis and treatment data are standardized according to a preset format to obtain standard format data; A preset time alignment strategy is used to align the standard format data in time to obtain standard clinical diagnosis and treatment data.

3. The method for integrating and processing medical test result mutual recognition data according to claim 2, characterized in that, The preset format includes a preset date format, and the standardization process of the multi-source clinical diagnosis and treatment data according to the preset format includes: Determine whether the patient corresponding to the clinical diagnosis and treatment data has a surgery date. If the patient corresponding to the clinical diagnosis and treatment data has a surgery date, determine the corresponding format of the generation time information based on the time interval between the generation time of the clinical diagnosis and treatment data and the surgery date. If the patient corresponding to the clinical diagnosis and treatment data does not have a surgery date, then the generation time information in the corresponding format is determined according to the time interval between the generation time of the clinical diagnosis and treatment data and the corresponding patient's clinical diagnosis and treatment time. The clinical diagnosis and treatment time includes: the corresponding patient's first visit date, first admission date, first test date, or first diagnosis date.

4. The method for integrating and processing medical test result mutual recognition data according to claim 1, characterized in that, The hybrid retrieval architecture includes vector retrieval, symbolic retrieval, and time-series retrieval. The standard clinical diagnosis and treatment data includes at least one report. The retrieval of the standard clinical diagnosis and treatment data using the hybrid retrieval architecture based on user query requests includes: Based on the vector retrieval, the semantic similarity score between the user query request and each report is calculated; Based on the symbol retrieval, determine whether each report meets the mutual recognition rules to obtain the symbol rule score; Based on the time-series retrieval, calculate the time-series relevance score between each report and the user's query request time; The semantic similarity score, symbol rule score, and temporal relevance score of each report are weighted and calculated to obtain the retrieval score of each report. The reports are then sorted from high to low according to the retrieval scores, and the top preset number of reports are taken as valid clinical diagnosis and treatment data.

5. The method for integrating and processing medical test result mutual recognition data according to claim 4, characterized in that, The process of determining whether each report meets the mutual recognition rule to obtain the symbol rule score includes: The mutual recognition rules include several mutual recognition sub-rules; The symbol rule score for each report is determined based on the number of times each report meets the mutual recognition sub-rules.

6. The method for integrating and processing medical test result mutual recognition data according to claim 4, characterized in that, The process of weighting the semantic similarity score, symbol rule score, and temporal relevance score of each report to obtain the retrieval score for each report includes: A first weight is assigned to the semantic similarity score, a second weight is assigned to the symbol rule score, and a third weight is assigned to the temporal relevance score; The first weight, the second weight, and the third weight are dynamically adjusted based on user feedback on the mutual recognition report to optimize the weighted calculation of the search score.

7. The method for integrating and processing medical test result mutual recognition data according to claim 1, characterized in that, Based on the aforementioned effective clinical diagnosis and treatment data, mutual recognition assessment reasoning is performed using neural symbolic reasoning to generate a mutual recognition report, including: Construct prompt words, which include at least valid clinical diagnosis and treatment data, at least a summary of content extracted from mutual recognition rules, and system instructions that define a preset format for mutual recognition reports; The prompt words are input into a large language model to generate a structured mutual recognition report; Based on preset logical consistency constraints, the structured mutual recognition report is verified. If the verification passes, it is used as the final mutual recognition report.

8. A medical test result mutual recognition data integration and processing device, characterized in that, The device includes: The data acquisition module is used to acquire multi-source clinical diagnosis and treatment data, and to fuse and standardize the multi-source clinical diagnosis and treatment data to obtain standard clinical diagnosis and treatment data. The query and retrieval module is used to construct a hybrid retrieval architecture and, based on user query requests, use the hybrid retrieval architecture to retrieve the standard clinical diagnosis and treatment data to obtain effective clinical diagnosis and treatment data. The report output module is used to perform mutual recognition assessment reasoning based on the effective clinical diagnosis and treatment data, generate a mutual recognition report, and visualize the mutual recognition report through a human-computer interaction interface.

9. An electronic device, characterized in that, include: The system includes a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the medical test result mutual recognition data integration processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the medical test result mutual recognition data integration processing method according to any one of claims 1 to 7.

Images