Cardiovascular disease assessment management method and system based on big data

Abstract

The invention discloses a cardiovascular disease assessment management method and system based on big data, and relates to the field of big data, and the method comprises the steps: obtaining original medical record data of a target patient, carrying out the standardization processing, obtaining to-be-assessed medical record data, carrying out the comparison analysis of the to-be-assessed medical record data and desensitization archived medical record data, and obtaining an assessment result; and then corresponding intervention strategies are selected from a preset database, so that scientific assessment and effective management of cardiovascular diseases based on big data are realized.

Description

Technical Field

[0001] The present invention relates to the field of big data, and in particular to a cardiovascular disease assessment and management method and system based on big data. Background Art

[0002] Cardiovascular disease is one of the major diseases that seriously threaten human health, and accurate assessment and effective management are crucial. Traditional assessments rely on doctors' experience and limited case analysis, which are subjective and limited. With the development of medical big data, using it to assess and manage cardiovascular disease has become a trend. However, existing related technologies still need to be improved in terms of data processing, assessment accuracy, and targeted intervention strategies. Summary of the invention

[0003] The purpose of the present invention is to provide a cardiovascular disease assessment and management method and system based on big data.

[0004] In a first aspect, an embodiment of the present invention provides a cardiovascular disease assessment and management method based on big data, comprising: Obtain original medical records of target patients; Performing data standardization processing on the original medical record data to obtain medical record data to be evaluated; Compare and analyze the medical record data to be evaluated with at least two desensitized archived medical record data to obtain the cardiovascular disease evaluation results corresponding to the medical record data to be evaluated; According to the cardiovascular disease assessment results, a corresponding cardiovascular disease intervention strategy is selected from a preset cardiovascular disease intervention database.

[0005] In a second aspect, an embodiment of the present invention provides a server system, including a server, wherein the server is used to execute the method described in the first aspect.

[0006] Compared with the prior art, the beneficial effects provided by the present invention include: adopting a cardiovascular disease assessment and management method and system based on big data disclosed by the present invention, obtaining the original medical record data of the target patient and standardizing it to obtain the medical record data to be assessed, comparing and analyzing it with the desensitized archived medical record data to obtain the assessment results, and then selecting the corresponding intervention strategy from the preset database, thereby realizing scientific assessment and effective management of cardiovascular diseases based on big data. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can also be obtained based on these drawings without creative work.

[0008] Figure 1 A schematic diagram of the steps of a cardiovascular disease assessment and management method based on big data provided by an embodiment of the present invention; Figure 2 A schematic block diagram of the structure of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0009] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0010] The specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings.

[0011] In order to solve the technical problems in the aforementioned background technology, Figure 1 This is a flow chart of a cardiovascular disease assessment and management method based on big data provided in an embodiment of the present disclosure. The cardiovascular disease assessment and management method based on big data is introduced in detail below.

[0012] Step S201, obtaining the original medical record data of the target patient; Step S202, performing data standardization processing on the original medical record data to obtain the medical record data to be evaluated; Step S203, comparing and analyzing the medical record data to be evaluated with at least two desensitized archived medical record data to obtain a cardiovascular disease evaluation result corresponding to the medical record data to be evaluated; Step S204: selecting a corresponding cardiovascular disease intervention strategy from a preset cardiovascular disease intervention database according to the cardiovascular disease assessment result.

[0013] In an embodiment of the present invention, for example, a patient named Mr. Li comes to see a doctor because of palpitations and chest tightness. After the doctor conducts a preliminary examination on Mr. Li, he enters the relevant information into the electronic medical record system of the hospital. At this time, the server will obtain Mr. Li's original medical record data from the electronic medical record system. These data include Mr. Li's basic personal information (such as age, gender, past medical history, etc.), description of the symptoms of this visit (frequency of palpitations, degree of chest tightness, etc.), various preliminary examination results (such as blood pressure, heart rate, etc.) and some related examination reports done before (such as text descriptions corresponding to electrocardiograms, cardiac color ultrasound and other imaging materials). After the server obtains Mr. Li's original medical record data, it begins to perform data standardization processing. Since different departments and different doctors may have differences in format, expression, etc. when entering medical record data, the server needs to standardize these data. For example, for Mr. Li's blood pressure data, some doctors may record it as "120 / 80mmHg", while others may write it as "12080". The server will standardize it as "120 / 80mmHg" according to the unified format standard. For another example, for the symptom description part, some may be more colloquial, and the server will convert it into a standardized expression of medical terminology. After such a series of processing, Mr. Li's medical record data becomes the medical record data to be evaluated that conforms to the standard format and standardized expression, so as to carry out accurate analysis and comparison later. The server first extracts cardiovascular disease-related images from Mr. Li's medical record data to be evaluated, such as the electrocardiogram and cardiac color Doppler ultrasound images taken previously. Assume that the server divides the electrocardiogram images into one data set and the cardiac color Doppler ultrasound images into another data set. Then, the target cardiovascular disease-related images are obtained from these data sets, such as selecting the electrocardiogram image at the time of the current symptom onset as the target cardiovascular disease-related image, extracting image features, and obtaining the first medical record image features, such as the waveform features on the electrocardiogram, the heart rate change features, etc. At the same time, the server obtains the first medical record document data corresponding to Mr. Li's medical record data to be evaluated, that is, the doctor's text description of his condition. Semantic description features are extracted from these document data to obtain the first semantic description feature, for example, semantic features about the severity of symptoms, frequency of onset, etc. Next, the server obtains the first multidimensional feature based on the first medical record image feature and the first semantic description feature, which includes the target image feature and the first semantic description feature. Then, the server obtains the pre-established image feature matching pool and semantic description feature matching pool. The image feature matching pool records the mutual correlation between each pending medical record data and its corresponding image feature, and the same is true for the semantic description feature matching pool. The server matches these two matching pools according to the target image feature, and also matches these two matching pools according to the first semantic description feature, thereby obtaining at least two target pending medical record data that match Mr. Li's medical record data to be evaluated.For example, the medical records of several patients who also had symptoms of palpitation and chest tightness and whose related image features and semantic description features were similar were matched as target pending medical records. For at least two target pending medical records matched by each matching method, the server calculates the matching coefficient of Mr. Li's medical records to be evaluated and each target pending medical records. For example, the matching coefficient is calculated by comparing the similarity of image features, the fit of semantic description features, etc. Then, for each target pending medical records, the server calculates the arithmetic mean of its matching coefficient with the medical records to be evaluated in each matching method. According to this arithmetic mean, the desensitized archived medical records are determined from at least two target pending medical records. Assume that the desensitized archived medical records of three patients are finally determined, and these three patients have undergone a clear diagnosis of cardiovascular disease and their medical records are complete. The server constructs the disease association information according to the second multidimensional features corresponding to each desensitized archived medical records, the cardiovascular disease target value configured for each desensitized archived medical records, and the first multidimensional features of Mr. Li's medical records to be evaluated. For example, Mr. Li's electrocardiogram image features, symptom description semantic features, etc. are comprehensively correlated and analyzed with the corresponding features in the three determined desensitized archived medical records to construct the disease association information. Then, based on this disease association information, the cardiovascular disease target value of Mr. Li's medical record data to be evaluated is identified to obtain the cardiovascular disease evaluation result. If most of the patients in the three desensitized archived medical record data are diagnosed with coronary heart disease and the degree of the disease is graded differently, through comparative analysis, the server may obtain the evaluation result that Mr. Li suffers from coronary heart disease and his condition is at a certain level. After the server obtains the evaluation result that Mr. Li suffers from coronary heart disease and his condition is at a certain level, it will select the corresponding intervention strategy in the preset cardiovascular disease intervention database. For example, for patients with coronary heart disease of Mr. Li's disease level, the intervention database may recommend drug treatment first, including taking certain antiplatelet drugs, blood lipid regulating drugs, etc., and cooperate with regular review (such as reviewing electrocardiograms and cardiac ultrasound every other month), and at the same time, it is recommended that Mr. Li adjust his lifestyle, such as reducing the intake of greasy food and increasing moderate exercise. The server pushes these intervention strategies to the doctor so that the doctor can develop a more appropriate treatment plan for Mr. Li.

[0014] In an embodiment of the present invention, the medical record data to be evaluated is compared and analyzed with at least two desensitized archived medical record data to obtain a cardiovascular disease evaluation result corresponding to the medical record data to be evaluated, which can be implemented through the following examples.

[0015] Acquire medical record data to be evaluated, and perform multidimensional feature extraction on the medical record data to be evaluated to obtain a first multidimensional feature, where the first multidimensional feature includes at least two data dimension features corresponding to at least two data dimensions respectively; At least two desensitized archived medical record data matching the medical record data to be evaluated according to the first multidimensional feature matching, each of the desensitized archived medical record data being configured with a cardiovascular disease target value; Constructing disease association information according to the second multidimensional features corresponding to each of the desensitized archived medical record data, the cardiovascular disease target value configured for each of the desensitized archived medical record data, and the first multidimensional features of the medical record data to be evaluated; According to the disease association information, the cardiovascular disease target value of the medical record data to be evaluated is identified to obtain a cardiovascular disease evaluation result.

[0016] In an embodiment of the present invention, illustratively, the server has completed the standardization of Mr. Li's original medical record data and obtained the medical record data to be evaluated. Now, the server will perform multi-dimensional feature extraction on this medical record data to be evaluated. From the perspective of data dimensions, on the one hand, there is the basic information dimension of Mr. Li, from which the server extracts features such as age and gender. For example, Mr. Li is 55 years old and male. These features have a certain reference value for the evaluation of cardiovascular disease, because the risks and manifestations of cardiovascular disease in people of different ages and genders may be different. On the other hand, there is the symptom description dimension. The server carefully analyzes the specific conditions of Mr. Li's palpitations and chest tightness recorded by the doctor, including the frequency of attacks, the duration of each attack, and whether there are other symptoms (such as chest pain, dyspnea, etc.). For example, Mr. Li's palpitation symptoms occur 2-3 times a week on average, each lasting about 10-15 minutes, occasionally accompanied by mild chest pain. There is also the examination result dimension, and the server focuses on Mr. Li's various cardiovascular-related examination data. Taking blood pressure as an example, Mr. Li's blood pressure is 130 / 85mmHg, which reflects the stress state of his cardiovascular system. Looking at the results of the electrocardiogram, the server extracts the waveform features on the electrocardiogram, such as whether there are ST segment changes, T wave inversion, etc., which are important clues for judging cardiovascular disease. Through the analysis and extraction of these different data dimensions, the server obtains the first multidimensional feature, which covers multiple aspects of data dimension features, laying the foundation for the subsequent matching and analysis with desensitized archived medical records. After the server has the first multidimensional feature of Mr. Li's medical record data to be evaluated, it begins to match and search in the huge desensitized archived medical record database. This database stores a large number of medical record data of previous patients after desensitization, and each medical record data is configured with a corresponding cardiovascular disease target value, that is, information such as the disease condition and degree after a clear diagnosis. The server first compares each data dimension feature in Mr. Li's first multidimensional feature with the medical record data in the database one by one. For example, in the age dimension, the server will filter out the medical records of patients with similar ages, because people of similar ages may have more similarities in cardiovascular health status. For Mr. Li, who is 55 years old, the server may first lock the medical records of patients aged 45-65 as the initial matching range. In the dimension of symptom description, the server will look for medical records with the same symptoms of palpitations and chest tightness, and similar frequency, duration, and accompanying symptoms. Suppose there is a medical record of Ms. Wang in the database. She also suffers from frequent palpitations, with a similar number of attacks per week as Mr. Li, and the duration of each attack is also similar, and occasionally accompanied by mild chest pain. Then Ms. Wang's medical record data will be initially matched in this dimension. In the dimension of examination results, the server compares blood pressure, electrocardiogram and other examination data.If the blood pressure value of another Mr. Zhang in his medical record is close to that of Mr. Li, and there are similar ST segment changes on the electrocardiogram, then Mr. Zhang's medical record data will also be included in the scope of further matching. Through comprehensive comparison and screening of multiple data dimension features, the server finally determined at least two desensitized archived medical record data with a high degree of matching with Mr. Li's medical record data to be evaluated. For example, in addition to the medical records of Ms. Wang and Mr. Zhang mentioned above, a medical record of Mr. Zhao was also matched. Their medical record data have certain similarities with Mr. Li's situation in various dimensions, and these desensitized archived medical record data are clearly marked with their corresponding cardiovascular disease target values, such as Ms. Wang was diagnosed with mild coronary heart disease, Mr. Zhang was diagnosed with moderate coronary heart disease, and Mr. Zhao was diagnosed with early myocardial infarction. After the server has determined the desensitized archived medical record data of Ms. Wang, Mr. Zhang, and Mr. Zhao that match Mr. Li's medical record data to be evaluated, it will then construct the disease association information. For Ms. Wang's desensitized archived medical record data, the server extracts its corresponding second multidimensional features. This includes the basic information dimension features of Ms. Wang, such as her age is 52 years old and her gender is female; the symptom description dimension features, such as her specific feelings when the symptoms of palpitation occur, the frequency of the attacks, etc.; the examination result dimension features, such as her blood pressure value, electrocardiogram characteristics, etc. At the same time, combined with the cardiovascular disease target value of Ms. Wang diagnosed with mild coronary heart disease, and the first multidimensional features of Mr. Li's medical record data to be evaluated (such as Mr. Li's age, symptoms, examination results, etc.), the server begins to build related information. The server will analyze how this age difference may affect the development of the disease, although there are differences between Mr. Li and Ms. Wang in age, but they are within a certain range. In terms of symptoms, compare the similarities and differences between the two people's symptoms such as palpitations and chest tightness, and the relationship between these symptoms and the diagnosis of coronary heart disease. For the examination results, compare the similarities and differences between the two people's blood pressure, electrocardiogram and other data, and how these data reflect the status of cardiovascular disease. Through such comprehensive analysis and synthesis, the server constructs a part of the disease-related information for the medical record data of Mr. Li and Ms. Wang. Similarly, for the desensitized archived medical records of Mr. Zhang and Mr. Zhao, the server also extracts their second multidimensional features respectively, and combines their respective cardiovascular disease target values ​​and Mr. Li’s first multidimensional features to conduct similar comprehensive analysis and synthesis. For example, for Mr. Zhang’s moderate coronary heart disease, the server will focus on analyzing the correlation between his condition and Mr. Li’s current condition, as well as the impact of differences in examination results on the judgment of the condition. Finally, the server will integrate the condition-related information constructed for the medical records of different patients such as Ms. Wang, Mr. Zhang, and Mr. Zhao to form a complete set of condition-related information. This information comprehensively covers the correlation between Mr. Li and these matching desensitized archived medical records in various dimensions and the corresponding cardiovascular disease target values.After the server has constructed complete disease-related information, it uses this as a basis to identify the cardiovascular disease target value of Mr. Li's medical record data to be evaluated, thereby obtaining the cardiovascular disease evaluation result. The server will carefully analyze the comparison of Mr. Li with Ms. Wang, Mr. Zhang, Mr. Zhao and other patients in various dimensions in the disease-related information and the corresponding cardiovascular disease target values. For example, from the perspective of age, Mr. Li is slightly older than Ms. Wang, which may mean that he has a certain potential increased risk of developing cardiovascular disease; from the perspective of symptoms, the frequency and duration of Mr. Li's palpitations and chest tightness symptoms may be in a certain range compared with Ms. Wang, Mr. Zhang, etc., which has a certain reference value for judging the degree of the disease; from the perspective of examination results, Mr. Li's blood pressure values, electrocardiogram characteristics, etc., and other similarities and differences with other patients can also reflect his cardiovascular disease status. Taking these factors into consideration, the server combines the cardiovascular disease target values ​​such as Ms. Wang being diagnosed with mild coronary heart disease, Mr. Zhang being diagnosed with moderate coronary heart disease, and Mr. Zhao being diagnosed with early myocardial infarction, as well as their various associations with Mr. Li in the disease-related information, and finally obtains Mr. Li's cardiovascular disease evaluation results after complex data analysis and judgment. Assuming that after analysis, the server determines that Mr. Li is more likely to suffer from moderate coronary heart disease, then this "suffering from moderate coronary heart disease" is the cardiovascular disease assessment result corresponding to Mr. Li's medical record data to be assessed. In this way, through a series of meticulous comparisons, analyses, and comprehensive judgments, the server completes the transformation process from the medical record data to be assessed to the cardiovascular disease assessment results in a scientific and accurate manner, providing an important basis for subsequent treatment and intervention.

[0017] In an embodiment of the present invention, the first multidimensional feature includes a first medical record image feature, and the second multidimensional feature includes a second medical record image feature; the condition association information is constructed based on the second multidimensional feature corresponding to each of the desensitized archived medical record data and the cardiovascular disease target value configured for each of the desensitized archived medical record data, and the first multidimensional feature of the medical record data to be evaluated, which can be implemented through the following examples.

[0018] Performing feature mapping processing on the first medical record image feature to obtain first condition semantic description information, and performing feature mapping processing on the second medical record image feature corresponding to each desensitized archived medical record data to obtain second condition semantic description information corresponding to each desensitized archived medical record data; Obtaining first medical record document data corresponding to the medical record data to be evaluated and second medical record document data corresponding to each of the desensitized archived medical record data; The condition association information is constructed based on the second condition semantic description information, the second medical record document data and the cardiovascular disease target value corresponding to each desensitized archived medical record data, as well as the first condition semantic description information and the first medical record document data.

[0019] In an embodiment of the present invention, for example, for Mr. Li's medical record data to be evaluated, which includes cardiovascular disease-related images such as electrocardiograms and cardiac color ultrasound, the first medical record image features extracted by the server are very critical. Taking Mr. Li's electrocardiogram image as an example, the server must first perform feature mapping processing on these first medical record image features. It converts various waveform features, time interval features, etc. on the electrocardiogram through built-in professional image analysis algorithms and models. For example, the morphology, amplitude, and time interval between the waveforms of the P wave, QRS complex, and T wave of a normal electrocardiogram will be converted into description information with semantic meaning after feature mapping processing. Assume that after processing, the server maps the image feature of the slightly wider P wave morphology in Mr. Li's electrocardiogram to the first condition semantic description information such as "there may be slight abnormalities in the atrial depolarization process", and maps the feature of the QRS complex duration being normal but the voltage being slightly lower to similar semantic descriptions such as "the overall ventricular depolarization is normal but the intensity of myocardial electrical activity may be slightly weak". Similarly, for the desensitized archived medical records of patients such as Ms. Wang, Mr. Zhang, and Mr. Zhao, the server will also perform similar feature mapping processing on the corresponding second medical record image features. Taking Ms. Wang's cardiac color Doppler ultrasound image as an example, the server analyzes the second medical record image features such as the size of the heart chamber, the thickness of the ventricular wall, and the blood flow in the image. For example, if Ms. Wang's cardiac color Doppler ultrasound shows that the thickness of the left ventricular wall is slightly thicker than the normal range, after feature mapping processing, the server will obtain the second condition semantic description information such as "the left ventricular myocardium may have compensatory thickening, indicating that there may be a potential increase in cardiovascular burden". For Mr. Zhang's electrocardiogram image, if there is a feature of mild ST segment depression, it will be mapped to the second condition semantic description information of "the myocardium may have local ischemia" after processing. Through such feature mapping processing, the server obtains the corresponding condition semantic description information for each medical record data, which can more intuitively reflect the cardiovascular disease-related conditions implied by the image features. After completing the above-mentioned image feature processing, the server will then obtain the relevant medical record document data. For Mr. Li's medical record data to be evaluated, the server accurately extracts the corresponding first medical record document data from the electronic medical record system. These document data are the text content of the doctor's detailed record of Mr. Li's condition during the consultation. These include Mr. Li's chief complaint, that is, the specific feelings, frequency of attacks, duration, etc. of his own description of symptoms such as palpitations and chest tightness; the doctor's physical examination results, such as whether there are murmurs when auscultating the heart, whether the heart rate is regular, etc.; and the doctor's preliminary diagnostic impression based on preliminary examinations and clinical experience. Similarly, the server will also obtain the second medical record document data corresponding to the desensitized archived medical record data of patients such as Ms. Wang, Mr. Zhang, and Mr. Zhao from the storage system.Taking Ms. Wang as an example, her second medical record document data records similar information when she first visited the doctor, such as her description of palpitations accompanied by a slight dizziness, the frequency of attacks is slightly lower than that of Mr. Li, the doctor found slight murmurs in her heart during the physical examination, and the detailed diagnosis process and basis for the final diagnosis of mild coronary heart disease. By obtaining these detailed medical record document data, the server is fully prepared for the subsequent construction of comprehensive and accurate disease-related information. After the server has all the necessary data mentioned above, it begins to build disease-related information. Taking Ms. Wang's desensitized archived medical record data as an example, the server first combines her second disease semantic description information "compensatory thickening of the left ventricular myocardium may exist, indicating that there may be a potential increase in cardiovascular burden", the second medical record document data (including symptom description, physical examination results, diagnosis process, etc.) and her cardiovascular disease target value (mild coronary heart disease), and then conducts a comprehensive comparative analysis with Mr. Li's first disease semantic description information (such as the semantic description after electrocardiogram image feature mapping mentioned above) and the first medical record document data (Mr. Li's symptoms, physical examination results, etc.). The server will analyze the similarities and differences between Ms. Wang and Mr. Li in terms of symptoms. For example, Ms. Wang's palpitations are accompanied by slight dizziness, while Mr. Li's palpitations are occasionally accompanied by slight chest pain. These two different accompanying symptoms may imply different directions of cardiovascular disease development. In terms of physical examination results, Ms. Wang's heart auscultation has slight murmurs, and Mr. Li's heart rate may be slightly irregular. These differences and similarities are also important for judging the condition of cardiovascular disease. At the same time, by comparing the semantic description information of the two people's conditions, Ms. Wang's compensatory thickening of the left ventricular myocardium and the intensity of myocardial electrical activity implied by Mr. Li's electrocardiogram image characteristics reflect the state of the cardiovascular system from different angles. By analyzing these associations, we can have a deeper understanding of the similarities and differences between the two people's conditions. For the desensitized archived medical record data of Mr. Zhang and Mr. Zhao, the server will also conduct a comprehensive comparative analysis of their respective second condition semantic description information, second medical record document data, and cardiovascular disease target values ​​with Mr. Li's first condition semantic description information and first medical record document data in the same way. For example, for Mr. Zhang, the second condition semantic description information "possible myocardial ischemia", the second medical record document data (such as symptom onset, physical examination results, etc.) and the cardiovascular disease target value (moderate coronary heart disease) are combined to compare with Mr. Li's relevant data to analyze the difference between the two in myocardial ischemia and the impact of this difference on the judgment of the condition. Finally, the server will integrate all the related information obtained through a comprehensive comparison and analysis of the medical records of different patients such as Ms. Wang, Mr. Zhang and Mr. Zhao with the medical records of Mr. Li, and construct a complete set of condition-related information.This disease-related information covers the similarities and differences between patients in terms of symptoms, physical examination results, semantic descriptions of the disease implied by imaging features, and corresponding cardiovascular disease target values, etc., providing an important basis for the subsequent accurate identification of cardiovascular disease target values ​​in the medical record data to be evaluated.

[0020] In the embodiment of the present invention, the number of the first medical record image features is at least two; and feature mapping processing is performed on the first medical record image features to obtain first condition semantic description information, which can be implemented through the following example.

[0021] Performing a feature integration operation on at least two of the first medical record image features to obtain a target image feature; The target image feature is converted into a semantic feature domain through a preset feature conversion component to obtain the first semantic description information of the disease condition.

[0022] In the embodiment of the present invention, for example, multiple image features such as P wave morphology, QRS wave group duration and voltage, T wave morphology, etc. are extracted from Mr. Li's electrocardiogram image, and features such as cardiac cavity size and ventricular wall thickness are obtained from the cardiac color Doppler ultrasound image. The server will perform feature integration operations on these different first medical record image features. It uses a specific algorithm to comprehensively process these scattered image features and integrate the information about the state of the cardiovascular system contained in each of them. Just like combining the features of each wave of the electrocardiogram with the cardiac structure features displayed by cardiac color Doppler ultrasound, after complex calculations and integration, a target image feature that can comprehensively reflect Mr. Li's cardiovascular condition is obtained. This target image feature is no longer an isolated image feature, but an overall feature description that integrates multiple image information. After obtaining the target image feature, the server uses a preset feature conversion component to perform the next step of processing. This feature conversion component is like a translator, which can convert the target image feature, which exists in the form of data and graphics, into a semantic feature domain, that is, into a form that can accurately describe the cardiovascular condition in text language. For example, after being processed by the feature conversion component, the originally complex target image features may be converted into specific and intuitively understandable first semantic description information of the disease, such as "the electrical activity and structure of the heart as a whole present a certain specific state, and the myocardium may have a certain degree of functional abnormality", thereby providing a clearer and easier-to-understand basis for subsequent operations such as constructing disease-related information.

[0023] In an embodiment of the present invention, the condition association information is constructed based on the second condition semantic description information, the second medical record document data and the cardiovascular disease target value corresponding to each desensitized archived medical record data, as well as the first condition semantic description information and the first medical record document data, which can be implemented through the following examples.

[0024] Constructing context deduction information according to a first identifier for distinguishing different contents included in the same medical record data, and the second condition semantic description information, the second medical record document data, and the cardiovascular disease target value corresponding to the same desensitized archived medical record data, so as to obtain at least two context deduction information corresponding to at least two desensitized archived medical record data; Integrating the at least two pieces of context deduction information to obtain context integration information, wherein at least two pieces of context deduction information in the context integration information are divided by a second identifier for distinguishing contents of different medical record data; Constructing association information according to the first identifier, the first condition semantic description information and the first medical record document data; The disease condition related information is generated according to the context integration information and the related information.

[0025] In an embodiment of the present invention, exemplarily, in the hospital's information management system, the server is responsible for processing a large amount of medical record data to perform accurate cardiovascular disease assessment. Take the previously mentioned Mr. Li's to-be-assessed medical record data and the matching desensitized archived medical record data of Ms. Wang, Mr. Zhang, and Mr. Zhao as examples to illustrate this step. For each desensitized archived medical record data, there are different types of information. In order to better sort out and use this information to construct disease-related information, the server will use a pre-set first identifier. This first identifier is like a label attached to different contents in the medical record data to clearly distinguish them. Taking Ms. Wang's desensitized archived medical record data as an example, the server first extracts the second disease semantic description information therein, such as "there may be compensatory thickening of the left ventricular myocardium, indicating that there may be a potential increase in cardiovascular burden", which is obtained by mapping its cardiac color ultrasound image features. Then there is the second medical record document data, which records in detail the description of Ms. Wang's symptoms when she visited the doctor (palpitation accompanied by slight dizziness, frequency of attacks, etc.), physical examination results (slight murmurs in heart auscultation, etc.) and the detailed process of the final diagnosis of mild coronary heart disease. And its cardiovascular disease target value is mild coronary heart disease. The server uses the first identifier to distinguish these different contents. For example, the content of the symptom description part is affixed with a "symptom identifier", the physical examination result part is affixed with a "physical examination identifier", the cardiovascular disease target value is affixed with a "diagnosis identifier", and the second condition semantic description information is affixed with an "image semantic identifier". Then, based on these different contents with identifiers, the server begins to build contextual deduction information. It will analyze that in Ms. Wang's case, the palpitation with a "symptom identifier" is accompanied by mild dizziness symptoms, combined with the slight murmur of the heart auscultation with a "physical examination identifier", and then consider the semantic description of compensatory thickening of the left ventricular myocardium with an "image semantic identifier" and the cardiovascular disease target value of mild coronary heart disease with a "diagnosis identifier". Through logical analysis and medical knowledge, it deduces the development of Ms. Wang's condition at different stages, possible influencing factors, etc., thereby constructing a contextual deduction information about Ms. Wang. Similarly, for the desensitized archived medical record data of Mr. Zhang and Mr. Zhao, the server will also use the corresponding second condition semantic description information, second medical record document data and cardiovascular disease target value in the above manner, and use the first identifier to distinguish different contents, and then construct contextual deduction information about Mr. Zhang and Mr. Zhao respectively. In this way, the server obtains at least two contextual deduction information corresponding to at least two desensitized archived medical record data. After the server obtains the contextual deduction information about Ms. Wang, Mr. Zhang and Mr. Zhao, etc., it will then integrate this information. During the integration process, in order to clearly distinguish the contextual deduction information of different patients, the server will use a second identifier for distinguishing the contents of different medical record data.For example, Ms. Wang's contextual deduction information is labeled as a whole with "Ms. Wang identifier", Mr. Zhang's is labeled with "Mr. Zhang identifier", and Mr. Zhao's is labeled with "Mr. Zhao identifier". Then, the server merges these contextual deduction information with different identifiers. Just like putting different puzzle pieces together, it will arrange and combine these information in a certain order and logic so that they can together form a more comprehensive and macro information set, which is the contextual integration information. In this contextual integration information, through the second identifiers such as "Ms. Wang identifier", "Mr. Zhang identifier", and "Mr. Zhao identifier", it can be clearly distinguished which part of the information is about which patient, so as to facilitate subsequent analysis and processing. At the same time, the server also needs to construct related information based on Mr. Li's medical record data to be evaluated. The server first extracts Mr. Li's first condition semantic description information, assuming that it is obtained by mapping the features of his electrocardiogram image, "the electrical activity and structure of the heart as a whole present a certain specific state, and the myocardium may have a certain degree of functional abnormality". Then comes the first medical record document data, which contains Mr. Li's symptom description (palpitation, chest tightness, etc.), physical examination results (heart rate may be slightly irregular, etc.), and other contents. Similarly, the server will use the first identifier to distinguish these different contents. For example, a "symptom identifier" is attached to Mr. Li's symptom description part, a "regularity identifier" is attached to the physical examination result part (here it is assumed that this is used to distinguish the physical examination result part, to distinguish it from the previous "physical examination identifier" used for other patient medical records), and an "image semantic identifier" is attached to the first condition semantic description information. Then, based on these different contents with identifiers, the server analyzes the relationship between them, such as the association between symptom description and physical examination results, the association between symptom description and the first condition semantic description information, and the association between physical examination results and the first condition semantic description information, etc., to construct a piece of related information about Mr. Li. Finally, the server will generate condition related information based on the context integration information and related information that have been constructed. The server will organically combine the context integration information and related information. It will compare, analyze and integrate the information about the disease progression, influencing factors and other aspects of patients such as Ms. Wang, Mr. Zhang and Mr. Zhao in the context-integrated information with the information about Mr. Li's symptoms, physical examination results, semantic description of the disease and other aspects in the associated information.For example, by comparing Ms. Wang's palpitations with mild dizziness with Mr. Li's palpitations and chest tightness, we analyze the similarities and differences in symptom manifestations and the possible direction of cardiovascular disease development; by comparing Ms. Wang's physical examination results of slight murmurs on heart auscultation with Mr. Li's physical examination results of slightly irregular heart rate, we explore the similarities and differences between the two in the physical state of the cardiovascular system; by comparing the semantic description of Ms. Wang's compensatory thickening of the left ventricular myocardium with the semantic description of Mr. Li's heart electrical activity and structure presenting a certain state as a whole, we study the similarities and differences between the two in the internal state of the cardiovascular system. Then, the server will organize the results of these comparisons, analyses and fusions to form a comprehensive and integrated disease association information. This disease association information covers various associations between Mr. Li and other matching patients in terms of symptoms, physical examination results, image semantic descriptions, cardiovascular disease target values, etc., providing a solid foundation for the subsequent accurate identification of cardiovascular disease target values ​​in the medical record data to be evaluated.

[0026] In an embodiment of the present invention, the first multidimensional feature includes a target image feature and a first semantic description feature; matching at least two desensitized archived medical record data that match the medical record data to be evaluated according to the first multidimensional feature can be implemented through the following example.

[0027] Acquire a pre-established image feature matching pool and a semantic description feature matching pool, wherein the image feature matching pool includes a correlation relationship between each pending medical record data and the image feature corresponding to the pending medical record data, and the semantic description feature matching pool includes a correlation relationship between each pending medical record data and the semantic description feature corresponding to the pending medical record data; According to the target image features, the image feature matching pool and the semantic description feature matching pool are matched respectively, and according to the first semantic description features, the image feature matching pool and the semantic description feature matching pool are matched respectively to obtain at least two target pending medical record data that match the medical record data to be evaluated; According to the medical record matching coefficient between the at least two target pending medical record data and the medical record data to be evaluated, the desensitized archived medical record data is determined from the at least two target pending medical record data.

[0028] In an embodiment of the present invention, illustratively, in the hospital's information management system, the server undertakes the task of processing and analyzing massive medical record data in order to accurately evaluate cardiovascular diseases. In order to achieve efficient and accurate matching of medical record data, the server pre-establishes an image feature matching pool and a semantic description feature matching pool. The construction of the image feature matching pool is a long and complex process. The server will collect a large amount of pending medical record data of previous patients, and these pending medical record data cover various examination images related to cardiovascular diseases, such as electrocardiograms, cardiac color Doppler ultrasound, coronary angiography and other imaging materials. For each pending medical record data, the server will extract its corresponding image features through a professional image analysis algorithm. For example, from a pending electrocardiogram medical record data, the server will extract features such as the morphology, time limit, and voltage of the P wave, QRS complex, and T wave; from the pending cardiac color Doppler ultrasound medical record data, it will extract image features such as the size of the heart cavity, the thickness of the ventricular wall, and the blood flow velocity. Then, each pending medical record data is established with its corresponding image feature and stored in the image feature matching pool. The same is true for the establishment of a semantic description feature matching pool. The server will analyze and process the text description part of a large amount of pending medical record data and extract the semantic description features. These text descriptions include the patient's self-report of symptoms (such as the specific feelings of palpitations and chest tightness, the frequency of attacks, the duration, etc.), the doctor's physical examination records (such as auscultation results, heart rate, blood pressure, etc.) and the initial diagnosis impression. Through natural language processing technology, the server can extract key semantic description features from it, such as "frequent attacks of palpitations accompanied by mild chest pain" and "slightly fast and irregular heart rate". Similarly, after establishing an association relationship between each pending medical record data and its corresponding semantic description feature, it is stored in the semantic description feature matching pool. Taking the previously mentioned Mr. Li's pending medical record data as an example, the server has extracted the target image feature and the first semantic description feature from this medical record data. For the target image feature, it is assumed that the target image feature in Mr. Li's pending medical record data is from a recent electrocardiogram image, and its features include slightly wider P wave morphology, normal QRS complex duration but slightly lower voltage, etc. The server will match these target image features with the data in the image feature matching pool. It will traverse each pending medical record data and its corresponding image features in the image feature matching pool, and determine the degree of matching by calculating the similarity between the features. For example, for the electrocardiogram image features in a pending medical record data, if its P wave morphology also has a similar slightly wider situation, and the QRS wave group duration and voltage are similar to Mr. Li's target image features, then this pending medical record data has a higher degree of matching in terms of image feature matching. At the same time, the server will also match Mr. Li's target image features with the semantic description feature matching pool.Although the semantic description feature matching pool mainly stores semantic description features, since there may be a certain correlation between the image features and semantic description features of different medical record data, matching can also be performed in this pool through a certain mapping relationship. For example, if the semantic description features of a pending medical record data contain a description of "abnormal cardiac electrical activity may exist", and its corresponding electrocardiogram image features are similar to Mr. Li's target image features, then this pending medical record data will also be considered to have a certain degree of matching in this regard. For the first semantic description feature, assume that the first semantic description feature in Mr. Li's pending medical record data includes "palpitation symptoms occur 2-3 times a week, each lasting about 10-15 minutes, occasionally accompanied by mild chest pain" and other content. The server will also match these first semantic description features with the image feature matching pool and the semantic description feature matching pool respectively. When matching with the image feature matching pool, it will look for pending medical record data whose image features correspond to the patient's symptom description similar to Mr. Li's first semantic description feature. For example, if a patient in a pending medical record also has symptoms of palpitations, and the frequency, duration and accompanying symptoms are similar to those of Mr. Li, then this pending medical record will be considered to have a matching degree in this respect. When matching with the semantic description feature matching pool, the similarity of the semantic description features will be directly compared through natural language processing technology to find the pending medical record data that is closest to Mr. Li's first semantic description feature. Through the above two-way matching process in the two matching pools, the server finally obtains at least two target pending medical record data that match Mr. Li's pending medical record data. For example, the pending medical record data of patients such as Ms. Wang, Mr. Zhang and Mr. Zhao are matched, and their medical record data have certain similarities with Mr. Li's pending medical record data in terms of image features and semantic description features. After the server obtains the target pending medical record data of Ms. Wang, Mr. Zhang and Mr. Zhao that match Mr. Li's pending medical record data, it will further determine the desensitized archived medical record data by calculating the medical record matching coefficient. For each target pending medical record data, the server will calculate the medical record matching coefficient between it and Mr. Li's pending medical record data. The calculation of this matching coefficient is a comprehensive consideration process. Taking Ms. Wang's target pending medical record data as an example, the server will calculate the matching coefficient from multiple aspects. In terms of image features, the electrocardiogram, cardiac color Doppler ultrasound and other image features in Ms. Wang's pending medical record data will be compared with the corresponding image features in Mr. Li's pending medical record data. For example, by comparing the morphology, time limit, voltage and other features of the P wave, QRS complex, and T wave of the electrocardiogram, the similarity of the image features is calculated through a certain mathematical formula (such as the cosine similarity formula, etc.), and this similarity is used as part of the matching coefficient of the image features. In terms of semantic description features, the semantic description features such as symptom descriptions and physical examination records in Ms. Wang's pending medical record data will be compared with the first semantic description feature in Mr. Li's pending medical record data.For example, the frequency, duration, and accompanying symptoms of symptoms such as palpitations and chest tightness are compared, and the similarity of semantic description features is calculated through natural language processing technology, and this similarity is used as part of the matching coefficient of semantic description features. Then, the matching coefficient of image features and the matching coefficient of semantic description features are combined according to certain weights (assuming that the matching coefficient weight of image features is 0.6 and the matching coefficient weight of semantic description features is 0.4), and the medical record matching coefficient between Ms. Wang's target pending medical record data and Mr. Li's pending medical record data is obtained. Similarly, the server will perform similar calculations on other target pending medical record data such as Mr. Zhang and Mr. Zhao to obtain the medical record matching coefficients between them and Mr. Li's pending medical record data. Finally, based on these medical record matching coefficients, the desensitized archived medical record data is determined from at least two target pending medical record data such as Ms. Wang, Mr. Zhang, and Mr. Zhao. For example, if Ms. Wang's medical record matching coefficient is the highest, then Ms. Wang's pending medical record data will be determined as desensitized archived medical record data, so as to facilitate the subsequent construction of disease association information and the determination of cardiovascular disease evaluation results.

[0029] In an embodiment of the present invention, the at least two target pending medical record data include at least two target pending medical record data obtained by matching each matching method; the desensitized archived medical record data is determined from the at least two target pending medical record data based on the medical record matching coefficient between the at least two target pending medical record data and the medical record data to be evaluated, which can be implemented through the following examples.

[0030] For the at least two target pending medical record data obtained by matching in each matching method, a matching coefficient between the medical record data to be evaluated and each target pending medical record data is calculated; For each target pending medical record data, calculating the arithmetic mean of the matching coefficients of the target pending medical record data in each matching mode with the medical record data to be evaluated; The desensitized archived medical record data is determined from the at least two target pending medical record data according to the arithmetic mean of the matching coefficients.

[0031] In an embodiment of the present invention, illustratively, in the hospital's information management system, the server continuously processes medical records data related to cardiovascular diseases. Taking the previously mentioned Mr. Li's pending medical records data as an example, the server has obtained a variety of target pending medical records data through different matching methods, such as the target pending medical records data of patients such as Ms. Wang and Mr. Zhang through the image feature matching pool, and the target pending medical records data of patients such as Ms. Zhao and Mr. Liu through the semantic description feature matching pool. For the target pending medical records data of Ms. Wang obtained by matching the image feature matching pool, the server calculates the matching coefficient between Mr. Li's pending medical records data and it. The server will compare the image features of the two in detail, such as the P wave morphology, QRS wave group duration and voltage of Mr. Li's electrocardiogram image, and the corresponding electrocardiogram image features in Ms. Wang's pending medical records data. Through professional algorithms (such as calculation methods based on feature vector distance), the similarity between the two in terms of image features is accurately calculated, which is used as a partial matching coefficient under the image feature matching method. At the same time, for the semantic description features of Ms. Wang's target pending medical record data, the server compares the first semantic description features of Mr. Li's pending medical record data, such as the frequency and duration of palpitations and chest tightness, with the semantic description features such as similar symptom descriptions and physical examination records in Ms. Wang's pending medical record data. The similarity of the semantic description features is calculated using natural language processing technology as the partial matching coefficient under the semantic description feature matching method. Then, the partial matching coefficients of these two aspects (image feature and semantic description feature matching method) are combined, and according to a certain weight (assuming that the image feature weight is 0.6 and the semantic description feature weight is 0.4), the complete matching coefficient of Mr. Li's pending medical record data and Ms. Wang's target pending medical record data under this matching method is obtained. Similarly, the server will also perform the above-mentioned similar matching coefficient calculation process for other target pending medical record data obtained through different matching methods, such as the target pending medical record data of Mr. Zhang, Ms. Zhao, and Mr. Liu. Still taking Ms. Wang's target pending medical record data as an example, suppose that the matching coefficient calculated above with Mr. Li's pending medical record data under the image feature matching method is 0.7, and the matching coefficient under the semantic description feature matching method is 0.6. Then the server will perform arithmetic average calculation on these two matching coefficients (0.7 and 0.6), that is, (0.7+0.6) ÷ 2 = 0.65, and obtain the arithmetic average of the matching coefficients of Ms. Wang's target pending medical record data with Mr. Li's pending medical record data under these two matching methods. For other target pending medical record data such as Mr. Zhang, Ms. Zhao, and Mr. Liu, the server will also perform the same arithmetic average calculation on their respective matching coefficients with Mr. Li's pending medical record data under different matching methods, and obtain the arithmetic average of their respective matching coefficients. After calculating the arithmetic average of the matching coefficients of each target pending medical record data, the server determines the desensitized archived medical record data based on this.For example, the arithmetic mean of the matching coefficient of Ms. Wang's target pending medical record data is 0.65, Mr. Zhang's is 0.58, Ms. Zhao's is 0.62, and Mr. Liu's is 0.55. The server will compare these arithmetic means and find that Ms. Wang's matching coefficient has the highest arithmetic mean. Therefore, the server will determine that Ms. Wang's target pending medical record data is desensitized archived medical record data, so that this desensitized archived medical record data can be used to construct disease-related information and determine the final cardiovascular disease assessment results.

[0032] In the embodiment of the present invention, the multidimensional feature extraction of the medical record data to be evaluated to obtain the first multidimensional feature can be implemented through the following example.

[0033] Extracting at least two cardiovascular disease-related images from the medical record data to be evaluated, and dividing the at least two cardiovascular disease-related images into at least two data sets; Acquire target cardiovascular disease related images from the at least two data sets, and extract image features from the target cardiovascular disease related images to obtain first medical record image features; Acquire first medical record document data corresponding to the medical record data to be evaluated, and extract semantic description features from the first medical record document data to obtain first semantic description features; The first multidimensional feature is obtained according to the first medical record image feature and the first semantic description feature.

[0034] In an embodiment of the present invention, illustratively, in the hospital's information management system, the server is responsible for processing a large amount of medical record data for cardiovascular disease assessment. Taking Mr. Li's medical record data to be evaluated as an example, this medical record data contains rich information, including a variety of cardiovascular disease-related images. The server will extract at least two such images from it, such as electrocardiogram images and cardiac ultrasound images. Electrocardiogram images can reflect the electrical activity of the heart, and many cardiovascular problems can be seen through the waveform changes above; cardiac ultrasound images can clearly show the structure, size and blood flow of the heart. After extraction, the server will reasonably divide these images to form at least two data sets. For example, the electrocardiogram image is divided into one data set, which is managed separately for subsequent targeted analysis; the cardiac ultrasound image is divided into another data set, so that different types of image data can be processed more systematically. After the server divides the data set, it will select the target cardiovascular disease-related image from it for further analysis. Assume that the server selects the electrocardiogram image of Mr. Li during a recent panic attack from the electrocardiogram image data set as the target cardiovascular disease-related image. Then, the server uses a professional image analysis algorithm to extract features from the target image. For electrocardiogram images, it will extract features such as the shape of the P wave (high or low), the duration of the QRS complex (whether it is within the normal range), and the inversion of the T wave. These features are combined to form the first medical record image features, which can reflect the specific conditions of Mr. Li's cardiac electrical activity from different angles and provide important basis for subsequent disease assessment. At the same time, the server will also obtain the first medical record document data corresponding to Mr. Li's medical record data to be evaluated. These document data are recorded in detail by the doctor during the consultation with Mr. Li, including the symptoms described by Mr. Li himself (such as the feeling of panic, the degree of chest tightness, the frequency of attacks, etc.), the doctor's physical examination results (such as auscultation of the heart for murmurs, whether the heart rate is regular, etc.) and the initial diagnosis impression. The server then uses natural language processing technology to extract semantic description features from these first medical record document data. For example, from Mr. Li's description of the symptoms of panic "two or three times a week, each lasting about ten minutes, and feeling a strong heartbeat during the attack", the server can extract semantic description features such as "moderate frequency of panic attacks, medium duration, and high heartbeat intensity during the attack". These first semantic description features can reflect the characteristics of Mr. Li's condition in a more intuitive way. Finally, the server will obtain the first multidimensional features based on the first medical image features and the first semantic description features extracted previously. For example, the first medical image features such as P wave morphology and QRS complex duration extracted from the electrocardiogram image are combined with the first semantic description features such as "moderate frequency of panic attacks, medium duration, and high heartbeat intensity during attacks" extracted from the medical record document data.This forms a first multidimensional feature that covers both image and semantic information, which can describe Mr. Li's cardiovascular disease-related conditions more comprehensively and accurately, and lay the foundation for subsequent comparison and analysis with desensitized archived medical record data and other operations.

[0035] In the embodiment of the present invention, the identification of the cardiovascular disease target value of the medical record data to be evaluated based on the disease condition association information can be implemented through the following examples.

[0036] Performing time series feature mapping on the disease-related information to obtain disease-related time series features, wherein the time series features are input parameters corresponding to a cardiovascular disease recognition model; Placing the disease-associated time series features into the cardiovascular disease recognition model, the cardiovascular disease recognition model is obtained by maintaining the initial network parameters of a preset classification model, and updating the iterative network parameters of the classification model according to the sample disease-associated time series features, the iterative network parameters being related to the PEFT fine-tuning component added to the classification model; The target cardiovascular disease target value of the medical record data to be evaluated output by the cardiovascular disease recognition model is obtained.

[0037] In an embodiment of the present invention, exemplarily, in the hospital's information management system, the server has completed the construction of the disease-related information. For example, for Mr. Li's medical record data to be evaluated, a detailed disease-related information is obtained by performing multi-faceted correlation analysis with multiple desensitized archived medical record data. This disease-related information contains the comparison of Mr. Li with other similar cases in terms of symptoms, examination results, disease progression, etc. Now, the server needs to perform time series feature mapping on this disease-related information. The server will sort out and analyze the contents of the disease-related information in chronological order according to the preset algorithms and rules. For example, for the frequency of Mr. Li's palpitation symptoms, the time changes from the initial onset to the most recent onset; and the change trends of his various examination results (such as electrocardiogram, cardiac color ultrasound, etc.) at different time points. Through such processing, the originally relatively static disease-related information is converted into disease-related time series features with time series features. These time series features can more clearly reflect the development dynamics of the disease over time, which is exactly what the cardiovascular disease recognition model needs to enter, and prepare for the subsequent accurate recognition of cardiovascular disease target values. After obtaining the disease-associated time series features, the server will put them into the cardiovascular disease recognition model for analysis. This cardiovascular disease recognition model is specially trained. Initially, it is based on a preset classification model, such as a commonly used neural network classification model. During the training process, the server will keep the initial network parameters of the preset classification model unchanged, in order to retain some basic characteristics and structures of the model. Then, the server will use a large number of sample disease-associated time series features to update the iterative network parameters of the classification model. These sample disease-associated time series features are obtained by similarly processing the medical records of many patients with confirmed cardiovascular diseases in the past (such as constructing disease-associated information and performing time series feature mapping). When updating the iterative network parameters, the key is to add the PEFT fine-tuning component. This component uses specific technologies (such as LoRA technology), which can fine-tune the iterative network parameters of the classification model according to the error between the sample disease-associated time series features and the actual cardiovascular disease target value, so that the model can more accurately adapt to the task of cardiovascular disease recognition. When the server puts the time series features associated with Mr. Li's condition into this specially trained cardiovascular disease recognition model, the model can analyze Mr. Li's condition based on its internal optimized network parameters and algorithms. Finally, the server waits for the output results of the cardiovascular disease recognition model. After complex calculations and analysis within the model, based on the time series features associated with Mr. Li's condition and the network parameters optimized by previous training, the cardiovascular disease recognition model will output a target cardiovascular disease target value for Mr. Li's medical record data to be evaluated.For example, if the model outputs "moderate coronary heart disease", then this is the cardiovascular disease target value determined by the server for Mr. Li's medical record data to be evaluated after a series of complex processing. This value will provide an important basis for the subsequent formulation of targeted treatment plans and intervention measures.

[0038] In an embodiment of the present invention, the cardiovascular disease recognition model is obtained in the following manner and can be implemented through the following examples.

[0039] Obtaining first sample medical record data image features, first sample medical record document data, and first sample cardiovascular disease target values ​​corresponding to the first sample medical record data, and second sample medical record data image features, second sample medical record document data, and second sample cardiovascular disease target values ​​corresponding to the second sample medical record data; According to a preset initial feature conversion component, feature conversion operations are performed on the first sample medical record data image feature and the second sample medical record data image feature respectively to obtain first sample condition semantic description information and second sample condition semantic description information, and sample condition association information is constructed according to the first sample condition semantic description information, the first sample medical record document data and the first sample cardiovascular disease target value, and the second sample condition semantic description information and the second sample medical record document data; Performing time series feature mapping on the sample condition-related information to obtain the sample condition-related time series features; Adding the PEFT fine-tuning component to the classification model, so as to add the iterative network parameters to the classification model through the PEFT fine-tuning component; The initial network parameters of the classification model are maintained as they are, and the iterative network parameters of the classification model are updated according to the sample condition-related time series characteristics and the second sample cardiovascular disease target value to obtain the cardiovascular disease recognition model.

[0040] In an embodiment of the present invention, illustratively, in the huge medical record database of the hospital, the server undertakes the important task of collecting and organizing data to build a cardiovascular disease recognition model. Taking the medical record data of a large number of patients with cardiovascular diseases in the past as an example, the server will first select a part of it as the first sample medical record data, and then select another part as the second sample medical record data. For the first sample medical record data, it is assumed that it is a medical record of a patient named Mr. Wang. The server extracts relevant information from Mr. Wang’s medical record: Image features of the first sample medical record data: From the electrocardiogram image taken by Mr. Wang, the server extracts image features such as normal but slightly wide P wave morphology, normal QRS complex duration but slightly lower voltage; from the cardiac color ultrasound image, extract features such as the thickness of the left ventricular wall is slightly thicker than the normal range. These image features can reflect the electrical activity and structural condition of Mr. Wang’s heart. The first sample medical record document data: This part includes the symptoms described by Mr. Wang himself, such as palpitations that occur two or three times a week, each lasting about ten minutes, and a feeling of chest tightness during the attack; the doctor's physical examination results, such as auscultation of the heart with slight murmurs, a slightly faster heart rate but basically regular; and the doctor's initial diagnosis impression, etc., which record in detail the situation when Mr. Wang visited the doctor. The first sample cardiovascular disease target value: After a series of detailed examinations and diagnostic processes, Mr. Wang was finally diagnosed with mild coronary heart disease, which is his first sample cardiovascular disease target value. Similarly, for the second sample medical record data, such as Ms. Li's medical record. The server also performs similar extractions: The second sample medical record data image features: From Ms. Li's electrocardiogram image, it is extracted that the T wave is slightly inverted and the QRS complex duration is slightly prolonged; from the cardiac color ultrasound image, it is found that the right atrium is slightly enlarged. The second sample medical record document data: Ms. Li described that her symptoms of palpitations occurred several times a day, each lasting for a few minutes, accompanied by mild dizziness; the doctor found that her blood pressure was slightly high and there was no obvious murmur during heart auscultation; and the doctor gave the initial diagnosis, etc. The second sample cardiovascular disease target value: Ms. Li was eventually diagnosed with early myocardial ischemia, which is her second sample cardiovascular disease target value. After the server obtains the various features of the above two sets of sample medical record data, it begins to process them using the preset initial feature conversion component. For the image features of Mr. Wang’s first sample medical record data, the server inputs them into the initial feature conversion component. Taking the electrocardiogram image features as an example, the initial feature conversion component will convert the feature that the P wave morphology is normal but slightly wide into the first sample condition semantic description information such as "there may be slight changes in the atrial depolarization process. The P wave morphology is normal but has a slightly wider trend, which may indicate subtle abnormalities in the early stage of cardiac electrical activity" through built-in algorithms and rules; for the cardiac color ultrasound image feature that the left ventricular wall thickness is slightly thicker than the normal range, it is converted into semantic descriptions such as "there may be a certain degree of compensatory thickening of the left ventricular myocardium, indicating that the cardiovascular system may be under certain pressure".Similarly, for the image features of Ms. Li's second sample medical record data, after being processed by the initial feature conversion component, the feature of slightly inverted T waves may be converted into the second sample condition semantic description information such as "there may be local mild ischemia in the myocardium, which is reflected in the electrocardiogram as inverted T waves"; the feature of slightly enlarged right atrium is converted into semantic descriptions such as "the right atrium structure is slightly enlarged, which may affect cardiac hemodynamics, and attention should be paid to whether it is related to cardiovascular disease". Then, the server constructs sample condition association information based on these converted first sample condition semantic description information, first sample medical record document data and first sample cardiovascular disease target value, as well as second sample condition semantic description information, second sample medical record document data and second sample cardiovascular disease target value. Taking Mr. Wang and Ms. Li as examples, the server will analyze the relationship between Mr. Wang's first sample medical record document data, such as the onset of panic symptoms and slight murmurs on heart auscultation, and his first sample condition semantic description information (such as subtle abnormalities in the early stage of cardiac electrical activity, and the cardiovascular system may be under certain pressure), and his first sample cardiovascular disease target value of mild coronary heart disease; at the same time, it will compare Ms. Li's second sample medical record document data, such as the number of times a day the panic symptoms occur, accompanied by slight dizziness, and her second sample condition semantic description information (such as the possible local mild ischemia of the myocardium, the slightly larger right atrial structure may affect the hemodynamics of the heart), and her early myocardial ischemia, the relationship between the second sample cardiovascular disease target value. By comprehensively considering these factors, the server constructs a sample condition association information that includes the association between different sample medical record data such as Mr. Wang and Ms. Li in terms of symptoms, image semantic description, cardiovascular disease diagnosis, etc. After the server constructs the sample condition association information, it will then perform time series feature mapping on it. For Mr. Wang's sample disease-related information, the server will sort out the changes in the frequency of his palpitation symptoms in chronological order, such as the change from once a week at the beginning of the disease to two or three times a week later; as well as the change trends of his various examination results (such as electrocardiogram, cardiac ultrasound, etc.) at different time points, such as the changes of P wave morphology, QRS complex voltage and other features on the electrocardiogram over time. Similarly, for Ms. Li's sample disease-related information, the server will also sort out the change process of her palpitation symptom frequency from several attacks a day to a decrease after treatment; as well as the change trends of her various examination results at different time points, such as whether the T wave inversion on the electrocardiogram has improved. By performing a comprehensive time series feature mapping of the sample disease-related information of different sample medical records in this way, the server obtains the sample disease-related time series features. These time series features can more clearly reflect the development dynamics of the disease in the sample medical record data over time, providing an important basis for subsequent model training. After preparing the sample disease-related time series features, the server needs to adjust the preset classification model.Assume that the preset classification model is a commonly used neural network model, such as a multi-layer perceptron (MLP) model. The server adds the PEFT fine-tuning component to the classification model. The PEFT fine-tuning component is implemented using a specific technology, such as the LoRA technology. After the PEFT fine-tuning component is added to the classification model, iterative network parameters can be added to the classification model through this component. These iterative network parameters are the parameters required for updating the classification model in the future according to the sample disease-related time series characteristics and the corresponding cardiovascular disease target values. They can continuously optimize the classification model to adapt to the task of cardiovascular disease identification. After the server adds the PEFT fine-tuning component to the classification model, it starts to update the classification model to obtain a cardiovascular disease identification model. In this process, the server keeps the initial network parameters of the classification model unchanged, in order to retain some of the original basic characteristics and structures of the classification model so that it will not lose its original advantages due to over-adjustment. Then, the server updates the iterative network parameters of the classification model according to the obtained sample disease-related time series characteristics and the second sample cardiovascular disease target value. Taking the sample medical records of Mr. Wang and Ms. Li as an example, the server will input the sample disease-related time series features obtained from the previous processing of many sample medical records of Mr. Wang, Ms. Li and others into the classification model with the PEFT fine-tuning component added, and use the second sample cardiovascular disease target value of Ms. Li and others as the target output. By comparing the error between the predicted value output by the model and the actual second sample cardiovascular disease target value, the server will adjust and update the iterative network parameters of the classification model based on this error using a specific algorithm (such as the stochastic gradient descent algorithm). After many such iterative update processes, as the sample medical record data continues to increase and process, the iterative network parameters of the classification model are gradually optimized, and finally a cardiovascular disease recognition model that can accurately identify cardiovascular diseases is obtained. This model can be used for subsequent operations such as identification of cardiovascular disease target values ​​for new medical record data to be evaluated.

[0041] In the embodiment of the present invention, before performing the feature conversion operation on the image feature according to the preset initial feature conversion component, the following implementation manner is also provided.

[0042] Acquire sample cardiovascular disease-related images and sample medical record document data corresponding to the sample cardiovascular disease-related images; Extracting features from the sample cardiovascular disease-related images to obtain sample image features, and placing the sample image features into a preparatory feature conversion component, so that the preparatory feature conversion component converts the sample image features into an input semantic feature domain of the classification model; Obtaining a sample semantic description feature corresponding to a feature conversion request, and placing the sample semantic description feature and the target sample condition semantic description feature output by the feature conversion component into the classification model, wherein the network parameters of the classification model remain unchanged; The sample inferred medical record document data output by the classification model is obtained, and the preliminary feature conversion component is trained according to the sample medical record document data and the sample inferred medical record document data to obtain the initial feature conversion component.

[0043] In an embodiment of the present invention, illustratively, in the hospital's information management system, the server is responsible for preparing data for building an accurate cardiovascular disease recognition model. First, the server selects a large number of sample cardiovascular disease-related images and corresponding sample medical record document data from a massive medical record database. For example, a case of Mr. Chen is selected, and his sample cardiovascular disease-related images include multiple electrocardiogram images and cardiac ultrasound images. The corresponding sample medical record document data records Mr. Chen's symptoms in detail, such as irregular attacks of palpitations, each attack lasts for different durations, and sometimes accompanied by mild chest pain; the doctor's physical examination results, such as occasional slight murmurs when auscultating the heart, and a roughly normal heart rate; and the doctor's preliminary diagnostic impression based on these situations. After the server obtains Mr. Chen's electrocardiogram and cardiac ultrasound and other sample cardiovascular disease-related images, it will use a professional image analysis algorithm to extract features. For electrocardiogram images, the server extracts sample image features such as the morphology of the P wave, the duration and voltage of the QRS complex, and the inversion of the T wave. From the cardiac ultrasound images, sample image features such as the size of the heart cavity, the thickness of the ventricular wall, and the blood flow velocity are extracted. Then, these extracted sample image features are placed in the preparatory feature conversion component. The function of this preparatory feature conversion component is to convert the sample image features from the original image data format to the input semantic feature domain that the classification model can understand, just like translating one "language" into another "language" so that accurate analysis and processing can be performed in the classification model later. In this process, the server will also obtain sample semantic description features corresponding to the sample cardiovascular disease-related images. For example, from Mr. Chen's medical record document data, natural language processing technology is used to extract sample semantic description features such as "the panic attack is unstable and there may be abnormal myocardial electrical activity". This is a feature extraction based on the text description of Mr. Chen's condition. At the same time, the preparatory feature conversion component will convert the input sample image features and output the target sample condition semantic description features, such as converting the electrocardiogram image features into a more semantic description of the heart's electrical activity state. Then, the obtained sample semantic description features and the target sample condition semantic description features output by the preparatory feature conversion component are placed in the classification model together, and in this process, the network parameters of the classification model are kept intact, that is, the original structure and parameter settings of the classification model are not changed, so as to simply examine the effect of the preparatory feature conversion component. After receiving the sample semantic description features and the target sample condition semantic description features, the classification model will perform operations based on its internal algorithms and logic, and output the sample inferred medical record document data. For example, for Mr. Chen's case, the classification model may output an inferred description of Mr. Chen's condition, such as the prediction of the frequency and duration of palpitation symptoms, and the inference of the results of cardiac auscultation.Then, the server will train the preliminary feature conversion component based on the difference between Mr. Chen's original sample medical record document data and the sample inferred medical record document data output by the classification model. If the prediction of the frequency of panic attacks in the sample inferred medical record document data does not match the actual situation in the original sample medical record document data, the server will adjust the preliminary feature conversion component based on this difference through a specific training algorithm (such as the back propagation algorithm, etc.) to continuously optimize it. After a large number of such training processes, using data from many different sample cases, the initial feature conversion component that can accurately perform feature conversion is finally trained, laying the foundation for the subsequent construction of a cardiovascular disease recognition model.

[0044] In the embodiment of the present invention, the updating of the iterative network parameters of the classification model according to the sample condition-related time series characteristics to obtain the cardiovascular disease recognition model can be implemented through the following examples.

[0045] Placing the sample disease-associated time series features into the classification model to which the PEFT fine-tuning component is added, and obtaining the sample inferred cardiovascular disease target value output by the classification model; According to the error between the sample inferred cardiovascular disease target value and the second sample cardiovascular disease target value, updating the iterative network parameters of the classification model to obtain the cardiovascular disease recognition model; The embodiments of the present invention also provide the following implementation modes.

[0046] The component parameters of the initial feature conversion component are updated according to the error between the sample inferred cardiovascular disease target value and the second sample cardiovascular disease target value.

[0047] In the embodiment of the present invention, illustratively, the server has prepared the sample disease-associated time series features through a series of previous steps, and these features clearly reflect the development dynamics of the disease in the sample medical record data over time. The server places these sample disease-associated time series features into the classification model to which the PEFT fine-tuning component has been added. Assuming that this classification model is a carefully designed neural network model, the PEFT fine-tuning component uses advanced LoRA technology, which can effectively fine-tune and optimize the model. When the sample disease-associated time series features are input into this model, the model begins to perform calculations based on its internal complex algorithms and structures, as well as the fine-tuning capabilities brought by the PEFT fine-tuning component. After a series of calculations, the model finally outputs a sample inferred cardiovascular disease target value. For example, for the sample disease-associated time series features corresponding to a sample medical record data, the sample inferred cardiovascular disease target value output by the model may be "moderate coronary heart disease", which is the cardiovascular disease condition corresponding to the sample medical record data inferred by the model based on the input time series features and its own parameters and algorithms. After obtaining the sample inferred cardiovascular disease target value output by the model, the server will compare it with the known second sample cardiovascular disease target value. The second sample cardiovascular disease target value was clearly determined when the sample medical record data was initially collected. It is the real cardiovascular disease condition obtained after detailed diagnosis of the sample medical record data. For example, the second sample cardiovascular disease target value corresponding to the sample medical record data just now is "mild coronary heart disease", while the sample inferred cardiovascular disease target value output by the model is "moderate coronary heart disease", which has a certain error. The server will update the iterative network parameters of the classification model based on this error through a specific algorithm (such as the stochastic gradient descent algorithm, etc.). These iterative network parameters are related to the PEFT fine-tuning component. By continuously adjusting these parameters, the model can more accurately identify cardiovascular diseases. After each update of the iterative network parameters, the server will again input the sample condition-related time series features into the updated model, obtain a new sample inferred cardiovascular disease target value, and then continue to compare it with the second sample cardiovascular disease target value, and repeat multiple iterations. After repeated training and iterative updates of a large number of sample medical record data, the iterative network parameters of the classification model are gradually optimized, and finally a cardiovascular disease recognition model that can accurately identify cardiovascular diseases is obtained. At the same time, the server will also update the component parameters of the initial feature conversion component based on the error between the sample inferred cardiovascular disease target value and the second sample cardiovascular disease target value. The initial feature conversion component plays an important role in converting the sample image features into semantic features that are easier for the classification model to understand in the previous steps. When it is found that there is an error between the sample inferred cardiovascular disease target value and the second sample cardiovascular disease target value, it means that there may be certain problems in the feature conversion link.For example, if the error is large, it may be because the initial feature conversion component is not accurate enough in converting sample image features into semantic features, causing the subsequent classification model to make incorrect inferences based on these inaccurate semantic features. Therefore, the server will adjust and update the component parameters of the initial feature conversion component based on this error through specific algorithms (such as back propagation algorithms, etc.). By continuously optimizing the parameters of the initial feature conversion component, it can perform feature conversion more accurately, thereby providing more accurate input for the cardiovascular disease recognition model and further improving the accuracy of the entire cardiovascular disease recognition system.

[0048] In addition, in the cardiovascular disease assessment and management method based on big data, there are many detailed and interrelated steps involved, which will be integrated and explained below. First, the server plays an important role in obtaining the original medical record data of the target patient. It establishes a data interface connection with the electronic medical record system of the medical institution. This data interface uses a secure encryption protocol to ensure the security and confidentiality of data transmission, so as to obtain the original medical record data generated by the target patient in the medical institution in real time. At the same time, in order to more comprehensively evaluate the patient's cardiovascular disease status, the server will also obtain the target patient's real-time physiological data from wearable devices, such as heart rate, blood pressure and other information, and integrate it into the original medical record data as a supplementary data source. Then, the acquired original medical record data is subjected to data standardization to obtain the medical record data to be evaluated. According to the preset cardiovascular disease-related data standard specifications, the format of each indicator data in the original medical record data is unified. For example, the date, time and other information recorded in different formats are converted into a unified standard format. Different types of data have different processing methods. For image data, resolution adjustment, contrast enhancement and other operations will be performed; for text data, vocabulary standardization and grammar standardization will be implemented. Through these operations, it is ensured that the obtained medical record data to be evaluated are consistent and comparable, so that the comparison and analysis operation can be carried out smoothly in the future. In the process of comparing and analyzing the medical record data to be evaluated with at least two desensitized archived medical record data to obtain the cardiovascular disease evaluation results, there are a series of key steps. Before performing the matching operation, the first multidimensional feature of the medical record data to be evaluated needs to be processed. Specifically, it is necessary to perform a data cleaning operation on it, remove the abnormal values, duplicate data and data features unrelated to the cardiovascular disease evaluation in the first multidimensional feature, and then use the data normalization method to process the first multidimensional feature after cleaning, and map the values ​​of each data dimension feature to the preset standard interval, so as to improve the accuracy and efficiency of the subsequent matching with the desensitized archived medical record data. When matching, multiple complex sub-steps are involved. First, at least two cardiovascular disease-related images are extracted from the medical record data to be evaluated, and divided into at least two data sets, and then the target cardiovascular disease-related images are obtained from these data sets, and the image features are extracted to obtain the first medical record image features, and the first medical record document data corresponding to the medical record data to be evaluated is obtained at the same time, and the semantic description feature is extracted to obtain the first semantic description feature, and then the first multidimensional feature is obtained based on the two. Afterwards, the pre-established image feature matching pool and semantic description feature matching pool are obtained, and matching operations are performed according to specific rules, that is, the two matching pools are matched respectively according to the target image features and the first semantic description features in the first multidimensional features to obtain at least two target pending medical record data that match the medical record data to be evaluated.For these target pending medical record data, the matching coefficients of the medical record data to be evaluated and each target pending medical record data are calculated, and the arithmetic mean of the matching coefficients of each target pending medical record data with the medical record data to be evaluated in each matching method is further calculated, and finally the desensitized archived medical record data are determined from at least two target pending medical record data according to the mean value. After the desensitized archived medical record data are determined, the disease association information is constructed according to the second multidimensional features corresponding to each desensitized archived medical record data, the cardiovascular disease target value configured for each desensitized archived medical record data, and the first multidimensional features of the medical record data to be evaluated. In this process, if the number of the first medical record image features is at least two, it is necessary to perform a feature integration operation on them to obtain the target image features, and then convert the target image features to the semantic feature domain through a preset feature conversion component to obtain the first disease semantic description information, and at the same time, the second medical record image features corresponding to each desensitized archived medical record data are respectively subjected to feature mapping processing to obtain the second disease semantic description information corresponding to each desensitized archived medical record data. Next, obtain the first medical record document data corresponding to the medical record data to be evaluated and the second medical record document data corresponding to each desensitized archived medical record data, and construct contextual deduction information based on the first identifier used to distinguish the different contents included in the same medical record data, and the second condition semantic description information, the second medical record document data and the cardiovascular disease target value corresponding to the same desensitized archived medical record data, so as to obtain at least two contextual deduction information corresponding to at least two desensitized archived medical record data, integrate these contextual deduction information to obtain contextual integration information, and then construct association information based on the first identifier, the first condition semantic description information and the first medical record document data, and finally generate condition association information based on the contextual integration information and the association information. After obtaining the condition association information, it is necessary to map the time series features to obtain the condition association time series features to provide input parameters for the cardiovascular disease recognition model. The specific operation is to use the time window sliding technology to segment the disease-related information, divide it into multiple time window segments, and use the preset time series feature extraction algorithm (which can be implemented based on recursive neural networks or long short-term memory networks) to extract the corresponding time series features for each time window segment. Then, the time series features extracted from each time window segment are integrated and spliced ​​to form disease-related time series features, which fully reflect the characteristic changes of disease-related information in the time dimension. There is also a rigorous process for obtaining cardiovascular disease recognition models. First, obtain the first sample medical record data image features, the first sample medical record document data, and the first sample cardiovascular disease target value corresponding to the first sample medical record data, as well as the second sample medical record data image features, the second sample medical record document data, and the second sample cardiovascular disease target value corresponding to the second sample medical record data.Then, according to the preset initial feature conversion component, the image features of the first sample medical record data and the image features of the second sample medical record data are respectively converted to obtain the semantic description information of the first sample condition and the semantic description information of the second sample condition, and the sample condition association information is constructed based on this information, and the time series feature mapping is performed to obtain the sample condition association time series feature. Then, the PEFT fine-tuning component (implemented using LoRA technology) is added to the classification model to add iterative network parameters to the classification model through the component. In the process of updating the iterative network parameters using the PEFT fine-tuning component of LoRA technology, different rank values ​​are set for experiments, and the error between the sample inferred cardiovascular disease target value and the second sample cardiovascular disease target value under different rank values ​​is compared to determine the optimal rank value, and the PEFT fine-tuning component is configured according to the optimal rank value to complete the update and optimization of the cardiovascular disease recognition model, thereby improving its accuracy and generalization ability for the input of different condition association time series features. At the same time, in the process of updating the iterative network parameters of the classification model according to the sample disease-associated temporal characteristics to obtain the cardiovascular disease recognition model, it is also necessary to update the component parameters of the initial feature conversion component according to the error between the sample inferred cardiovascular disease target value and the second sample cardiovascular disease target value. Specifically, the stochastic gradient descent algorithm is used to calculate the gradient values ​​of the component parameters of the initial feature conversion component according to the error, and then the components of the initial feature conversion component are adjusted and updated according to the gradient value according to the preset learning rate, so that the initial feature conversion component can more accurately convert the image features into features of the semantic feature domain suitable for the input of the classification model in subsequent processing. Finally, according to the cardiovascular disease assessment results, the corresponding cardiovascular disease intervention strategy is selected in the preset cardiovascular disease intervention database. First, the key information such as the severity of the disease and the development trend in the cardiovascular disease assessment results are analyzed, and the preset cardiovascular disease intervention database is screened according to these key information. All intervention strategies suitable for the disease type are first screened out, and then the scope is further narrowed down according to the severity of the disease, and the cardiovascular disease intervention strategy that best suits the current assessment results is selected. In addition, in the selection process, the individual differences of patients, such as age, gender, and underlying diseases, are fully considered, and the selected intervention strategies are personalized and optimized. For example, if the patient is an elderly person and the assessment results show that cardiovascular disease is in the early stages, the frequency of regular physical examinations should be appropriately increased, and dietary recommendations should be adjusted to increase the proportion of food rich in dietary fiber and unsaturated fatty acids based on the selected intervention strategy; for female patients, according to the characteristics of their physiological cycle, exercise time and intensity should be reasonably arranged in the intervention strategy to avoid high-intensity exercise during the menstrual period; for patients with other underlying diseases such as diabetes, the use of drugs, dietary control and other aspects of the cardiovascular disease intervention strategy should be coordinated and adjusted in combination with the requirements for diabetes disease control to achieve the effect of comprehensive treatment and prevention.Through this series of complete and detailed steps, the cardiovascular disease assessment and management method based on big data can more accurately and comprehensively assess the patient's cardiovascular disease status and develop more targeted and effective intervention strategies.

[0049] The embodiment of the present invention provides a computer device 100, which includes a processor and a non-volatile memory storing computer instructions. When the computer instructions are executed by the processor, the computer device 100 executes the aforementioned cardiovascular disease assessment and management method based on big data. Figure 2 As shown, Figure 2 The block diagram of the computer device 100 provided in the embodiment of the present invention. The computer device 100 includes a memory 111, a processor 112 and a communication unit 113. To achieve data transmission or interaction, the memory 111, the processor 112 and the communication unit 113 are directly or indirectly electrically connected to each other. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.

[0050] For illustrative purposes, the foregoing description is made with reference to specific embodiments. However, the above illustrative discussion is not intended to be exhaustive or to limit the present disclosure to the precise form disclosed. Numerous modifications and variations are possible in accordance with the above teachings. These embodiments are selected and described in order to best illustrate the principles of the present disclosure and its practical application, so that those skilled in the art can best utilize the present disclosure and utilize various embodiments with different modifications to suit the intended specific application.

Claims

1. A cardiovascular disease assessment and management method based on big data, characterized in that: include: Obtain original medical records of target patients; Performing data standardization processing on the original medical record data to obtain medical record data to be evaluated; Compare and analyze the medical record data to be evaluated with at least two desensitized archived medical record data to obtain the cardiovascular disease evaluation results corresponding to the medical record data to be evaluated; According to the cardiovascular disease assessment results, a corresponding cardiovascular disease intervention strategy is selected from a preset cardiovascular disease intervention database.

2. The method according to claim 1, characterized in that: The step of comparing and analyzing the medical record data to be evaluated with at least two desensitized archived medical record data to obtain a cardiovascular disease evaluation result corresponding to the medical record data to be evaluated includes: Acquire medical record data to be evaluated, extract at least two cardiovascular disease-related images from the medical record data to be evaluated, and divide the at least two cardiovascular disease-related images into at least two data sets; Acquire target cardiovascular disease related images from the at least two data sets, and extract image features from the target cardiovascular disease related images to obtain first medical record image features; Acquire first medical record document data corresponding to the medical record data to be evaluated, and extract semantic description features from the first medical record document data to obtain first semantic description features; Obtaining a first multidimensional feature according to the first medical record image feature and the first semantic description feature, wherein the first multidimensional feature includes a target image feature and a first semantic description feature; Acquire a pre-established image feature matching pool and a semantic description feature matching pool, wherein the image feature matching pool includes a correlation relationship between each pending medical record data and the image feature corresponding to the pending medical record data, and the semantic description feature matching pool includes a correlation relationship between each pending medical record data and the semantic description feature corresponding to the pending medical record data; According to the target image features, the image feature matching pool and the semantic description feature matching pool are matched respectively, and according to the first semantic description features, the image feature matching pool and the semantic description feature matching pool are matched respectively to obtain at least two target pending medical record data that match the medical record data to be evaluated; the at least two target pending medical record data include at least two target pending medical record data obtained by matching each matching method; For the at least two target pending medical record data obtained by matching in each matching method, a matching coefficient between the medical record data to be evaluated and each target pending medical record data is calculated; For each target pending medical record data, calculating the arithmetic mean of the matching coefficients of the target pending medical record data in each matching mode with the medical record data to be evaluated; Determine the desensitized archived medical record data from the at least two target pending medical record data according to the arithmetic mean of the matching coefficients, each of the desensitized archived medical record data being configured with a cardiovascular disease target value; Constructing disease association information according to the second multidimensional features corresponding to each of the desensitized archived medical record data, the cardiovascular disease target value configured for each of the desensitized archived medical record data, and the first multidimensional features of the medical record data to be evaluated; According to the disease association information, the cardiovascular disease target value of the medical record data to be evaluated is identified to obtain a cardiovascular disease evaluation result.

3. The method according to claim 2, characterized in that The first multidimensional feature includes a first medical record image feature, the second multidimensional feature includes a second medical record image feature, and the number of the first medical record image features is at least two; the condition association information is constructed according to the second multidimensional feature corresponding to each of the desensitized archived medical record data, the cardiovascular disease target value configured for each of the desensitized archived medical record data, and the first multidimensional feature of the medical record data to be evaluated, including: Performing a feature integration operation on at least two of the first medical record image features to obtain a target image feature; The target image feature is converted into a semantic feature domain through a preset feature conversion component to obtain first semantic description information of the condition, and the second medical record image feature corresponding to each of the desensitized archived medical record data is subjected to feature mapping processing to obtain second semantic description information of the condition corresponding to each of the desensitized archived medical record data; Obtaining first medical record document data corresponding to the medical record data to be evaluated and second medical record document data corresponding to each of the desensitized archived medical record data; Constructing context deduction information according to a first identifier for distinguishing different contents included in the same medical record data, and the second condition semantic description information, the second medical record document data, and the cardiovascular disease target value corresponding to the same desensitized archived medical record data, so as to obtain at least two context deduction information corresponding to at least two desensitized archived medical record data; Integrating the at least two pieces of context deduction information to obtain context integration information, wherein at least two pieces of context deduction information in the context integration information are divided by a second identifier for distinguishing contents of different medical record data; Constructing association information according to the first identifier, the first condition semantic description information and the first medical record document data; The disease condition related information is generated according to the context integration information and the related information.

4. The method according to claim 2, characterized in that: The step of identifying the cardiovascular disease target value of the medical record data to be evaluated according to the disease association information includes: Performing time series feature mapping on the disease-related information to obtain disease-related time series features, wherein the time series features are input parameters corresponding to a cardiovascular disease recognition model; Placing the disease-associated time series features into the cardiovascular disease recognition model, the cardiovascular disease recognition model is obtained by maintaining the initial network parameters of a preset classification model, and updating the iterative network parameters of the classification model according to the sample disease-associated time series features, the iterative network parameters being related to the PEFT fine-tuning component added to the classification model; Obtaining a target cardiovascular disease target value of the medical record data to be evaluated output by the cardiovascular disease recognition model; The cardiovascular disease recognition model is obtained by the following methods, including: Obtaining first sample medical record data image features, first sample medical record document data, and first sample cardiovascular disease target values ​​corresponding to the first sample medical record data, and second sample medical record data image features, second sample medical record document data, and second sample cardiovascular disease target values ​​corresponding to the second sample medical record data; According to a preset initial feature conversion component, feature conversion operations are performed on the first sample medical record data image feature and the second sample medical record data image feature respectively to obtain first sample condition semantic description information and second sample condition semantic description information, and sample condition association information is constructed according to the first sample condition semantic description information, the first sample medical record document data and the first sample cardiovascular disease target value, and the second sample condition semantic description information and the second sample medical record document data; Performing time series feature mapping on the sample condition-related information to obtain the sample condition-related time series features; Adding the PEFT fine-tuning component to the classification model, so as to add the iterative network parameters to the classification model through the PEFT fine-tuning component; The initial network parameters of the classification model are maintained as they are, and the iterative network parameters of the classification model are updated according to the sample condition-related time series characteristics and the second sample cardiovascular disease target value to obtain the cardiovascular disease recognition model.

5. The method according to claim 4, characterized in that Before performing a feature conversion operation on the image feature according to the preset initial feature conversion component, the method further includes: Acquire sample cardiovascular disease-related images and sample medical record document data corresponding to the sample cardiovascular disease-related images; Extracting features from the sample cardiovascular disease-related images to obtain sample image features, and placing the sample image features into a preparatory feature conversion component, so that the preparatory feature conversion component converts the sample image features into an input semantic feature domain of the classification model; Obtaining a sample semantic description feature corresponding to a feature conversion request, and placing the sample semantic description feature and the target sample condition semantic description feature output by the feature conversion component into the classification model, wherein the network parameters of the classification model remain unchanged; The sample inferred medical record document data output by the classification model is obtained, and the preliminary feature conversion component is trained according to the sample medical record document data and the sample inferred medical record document data to obtain the initial feature conversion component.

6. The method according to claim 4, characterized in that The updating of the iterative network parameters of the classification model according to the sample disease-associated time series characteristics to obtain the cardiovascular disease recognition model includes: Placing the sample disease-associated time series features into the classification model to which the PEFT fine-tuning component is added, and obtaining the sample inferred cardiovascular disease target value output by the classification model; According to the error between the sample inferred cardiovascular disease target value and the second sample cardiovascular disease target value, updating the iterative network parameters of the classification model to obtain the cardiovascular disease recognition model; The method further comprises: The component parameters of the initial feature conversion component are updated according to the error between the sample inferred cardiovascular disease target value and the second sample cardiovascular disease target value.

7. The method according to claim 6, characterized in that The updating of the component parameters of the initial feature conversion component includes: Using a stochastic gradient descent algorithm, according to the error between the sample inferred cardiovascular disease target value and the second sample cardiovascular disease target value, the gradient value of each component parameter of the initial feature conversion component is calculated; According to the preset learning rate, the component parameters of the initial feature conversion component are adjusted and updated according to the gradient value.

8. The method according to claim 6, characterized in that The classification model is a neural network model, the PEFT fine-tuning component is implemented using LoRA technology, the iterative network parameters of the classification model are updated according to the sample disease-related time series characteristics and the second sample cardiovascular disease target value to obtain the cardiovascular disease recognition model, and further includes: In the process of updating the iterative network parameters by using the PEFT fine-tuning component of the LoRA technology, different rank values ​​are set for testing, and the optimal rank value is determined by comparing the error between the sample inferred cardiovascular disease target value output by the classification model under different rank values ​​and the second sample cardiovascular disease target value; The PEFT fine-tuning component is configured according to the optimal rank value to complete the update optimization of the cardiovascular disease recognition model.

9. The method according to claim 2, characterized in that: Before matching at least two desensitized archived medical record data that match the medical record data to be evaluated according to the first multidimensional feature, the method further includes: Performing a data cleaning operation on the first multidimensional feature of the medical record data to be evaluated, removing abnormal values, duplicate data, and data features irrelevant to cardiovascular disease evaluation in the first multidimensional feature; The data normalization method is used to process the cleaned first multidimensional features, and the value of each data dimension feature is mapped to a preset standard range.

10. A server system, characterized in that: The method comprises a server, wherein the server is used to execute the method described in any one of claims 1 to 9.

Images