A structural causal model construction method and device for acute myocardial infarction monitoring

Abstract

The present invention relates to a method and device for constructing a structural causal model for acute myocardial infarction monitoring, which belongs to the field of patient data processing technology and solves the problem in the prior art that patient data cannot be monitored in real time and the predicted data is inaccurate. The method for constructing a structural causal model of the present invention includes: step 1, collecting and processing historical medical big data to obtain a structured data set; step 2, discretizing the structured data set according to medical knowledge; step 3, constructing an AMI causal structure based on the discretized data set and medical constraints and causal discovery methods; step 4, based on the discretized data set of step 2, using the Bayesian parameter estimation method to perform parameter learning on the AMI causal structure to obtain the final AMI structural causal model. The present invention can more accurately quantify features and identify significant influencing factors of AMI through effective information value, and assist in the prevention, diagnosis and treatment of AMI.

Description

Technical Field

[0001] The present invention belongs to the technical field of patient data processing, and in particular relates to a method and device for constructing a structural causal model for acute myocardial infarction monitoring. Background Art

[0002] With the rapid development of medical means, disease monitoring and evaluation models based on medical informationization have been vigorously developed, and data processing and early prediction and warning of corresponding diseases are carried out based on the models, such as: Chinese patents CN109949936B and CN110957036B.

[0003] Acute myocardial infarction (AMI) is a disease with a high rate of critical illness and mortality, often presenting acutely and with severe symptoms. Early diagnosis, early warning, and treatment before AMI develops and becomes severe are crucial for the patient's health and well-being. However, some AMI patients may not experience obvious early clinical symptoms, and patient data monitoring is not timely. This leads many patients to neglect daily health management and potentially miss the optimal treatment opportunity. Summary of the Invention

[0004] In view of the above problems, the present invention provides a method and device for constructing a structural causal model for acute myocardial infarction monitoring, which solves the problems in the prior art of being unable to monitor patient data in real time and inaccurate predicted data.

[0005] The present invention provides a method for constructing a structural causal model for acute myocardial infarction monitoring, which comprises the following steps:

[0006] Step 1: Collect and process historical medical big data to obtain a structured dataset for training the AMI structural causal model;

[0007] Among them, the steps for processing historical medical big data are:

[0008] Determine the rules for extracting historical medical big data information;

[0009] Structural data organization of historical medical big data based on information extraction rules;

[0010] Identify feature information with regular expressions through regular expressions;

[0011] Use natural language processing methods to process feature information in historical medical big data that does not have regular expression forms;

[0012] Finally, a structured data set is obtained;

[0013] Step 2: Discretize the structured data set according to medical knowledge to obtain a discretized data set;

[0014] Step 3: Based on the discretized data set, an AMI causal structure is constructed based on medical constraints and causal discovery methods. The variables in the AMI causal structure are AMI, associated factors that cause AMI, and manifestation factors caused by AMI. Directed edges represent the generation process from cause to effect between variables.

[0015] Step 4: Based on the discretized data set in step 2, use the Bayesian parameter estimation method to learn the parameters of the AMI causal structure and estimate the conditional probability parameters; based on the conditional probability parameters, the final AMI structural causal model is obtained.

[0016] Optionally, historical medical big data includes medical record information in real hospital scenarios; medical record information includes the patient's medical history information, the doctor's description of the patient's condition, and the diagnosis results based on the patient's actual situation; medical record information includes the patient's biochemical test index information, daily living habits and past medical history; the doctor's description of the patient's condition includes the patient's oral description and the doctor's records.

[0017] Optionally, the variables include AMI incidence information variables and corresponding characteristic variables of the AMI incidence information variables.

[0018] Optionally, medical prior knowledge is introduced into the AMI causal structure.

[0019] Optionally, the AMI causal structure is updated using an additive noise model and / or medical knowledge.

[0020] Optionally, the causal discovery method is a fast causal inference algorithm based on conditional independence or a fast greedy equivalence class search algorithm based on scores.

[0021] Another aspect of the present invention discloses an intelligent wearable device for diagnosing acute myocardial infarction for monitoring acute myocardial infarction, comprising a signal acquisition device, a signal processing device and a display processing unit; the signal processing device is provided with an AMI structural causal model, and the AMI structural causal model is constructed using the aforementioned structural causal model construction method.

[0022] Compared with the prior art, the present invention has at least the following beneficial effects:

[0023] (1) The structural causal model for acute myocardial infarction monitoring of the present invention can realize intelligent monitoring of AMI, provide an interpretable and generalizable intelligent data processing and monitoring method for AMI, improve the accuracy of prediction results, and establish a systematic understanding of AMI as a complex disease, which can provide guidance for the prevention and treatment of AMI. Based on the structural causal model, the effective information value can be used to more accurately quantify features and identify significant influencing factors of AMI, assisting in the prevention, diagnosis and treatment of AMI. More importantly, the structural causal model can also be used to predict the risk of an individual's disease in the short term.

[0024] (2) The wearable device of the present invention, in view of the characteristics of AMI, which is rapid and highly dangerous, can effectively improve the user's ability to predict AMI data through real-time detection in daily life, thereby allowing the user to seek medical treatment earlier and reduce the risk of complications and severe illness after recovery.

[0025] (3) The wearable device of the present invention can collect user's individual data in real time, and realize real-time monitoring and early warning of the risk of acute myocardial infarction.

[0026] (4) The wearable device of the present invention can provide decision support for doctors, relieve the pressure of emergency room visits, and improve the accuracy of AMI clinical diagnosis. In addition, the device is also suitable for daily monitoring of the general population or hypertensive patients, etc., in order to identify the risk of myocardial infarction in users and issue early warnings several days in advance, thereby achieving real-time monitoring and assessment of personal health, early screening and pre-diagnosis of diseases, and proactive early intervention.

[0027] (5) The wearable device of the present invention is equipped with a physiological indicator regular monitoring program based on medical knowledge, which enables the present invention to more accurately and early detect potential health problems, prompting users to seek medical treatment as soon as possible, or suggesting users to adjust their lifestyle habits.

[0028] (6) The wearable device of the present invention is highly applicable and can be used for daily risk monitoring of other diseases, simply by replacing the corresponding medical data. It can open up a new field of health management, realize cloud diagnosis and treatment, save traditional medical resources, and open up new medical methods.

[0029] (7) The wearable device of the present invention can use the collected data of all users to conduct big data analysis and integration through cloud servers, incorporate it into the training of structural causal models, and apply it to the research and analysis of the epidemiological characteristics of myocardial infarction in the future, providing new treatment experience and scientific basis for disease prediction, prevention and control. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The drawings are only for purposes of illustrating particular embodiments and are not to be considered limiting of the invention.

[0031] Figure 1 The present invention is a flowchart of the method for constructing a structural causal model for acute myocardial infarction monitoring.

[0032] Figure 2 This is a schematic diagram of the wearable device for intelligent diagnosis of acute myocardial infarction for monitoring acute myocardial infarction according to the present invention.

[0033] Figure 3 This is a causal structure diagram of the AMI structure in the actual case of the present invention.

[0034] Figure 4 Schematic diagram of the significance of the causal impact of relevant variables on AMI in the actual case of the present invention.

[0035] Figure 5 Schematic diagram of the significance of the causal impact of AMI on related variables in the actual case of the present invention. DETAILED DESCRIPTION

[0036] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. In addition, the present invention can also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited by the specific embodiments disclosed below.

[0037] A specific embodiment of the present invention, as Figure 1-Figure 5 , discloses an intelligent wearable device for diagnosing acute myocardial infarction based on a structural causal model, including a signal acquisition device, a signal processing device and a display processing unit;

[0038] Signal acquisition instruments include blood pressure measuring instruments, electrocardiogram measuring instruments, blood oxygen measuring instruments and pulse measuring instruments.

[0039] Blood pressure measuring instrument, used to collect the user's blood pressure signal in real time;

[0040] ECG meter, used to collect the user's ECG signals in real time;

[0041] A blood oxygen meter, used to collect the user's blood oxygen saturation signal in real time;

[0042] A pulse meter, used to collect the user's pulse signal in real time;

[0043] Signal processing device, including a big data analyzer and an AMI structural causal model, is used to process blood pressure signals, electrocardiogram signals, blood oxygen saturation signals, pulse signals, and manually input data to obtain causal data of acute myocardial infarction and AMI prediction risk prediction values;

[0044] The display processing unit is used for manual data input, displaying historical monitoring results, displaying causal data for acute myocardial infarction, and displaying early warning risk prediction values. It enables visualization of physiological indicators, allowing dynamic observation of their changing patterns and tracing back the time periods when abnormal values occurred.

[0045] Among them, the manually input data is the user's examination data in the hospital, living habits, past medical history and physical condition and / or clinical manifestations.

[0046] Furthermore, an OCR (optical character recognition module) photo recognition module is provided for users to upload medical test report data.

[0047] Furthermore, a cloud server and a communication module are provided, and the signal processing instrument is provided on the cloud server; the communication module is used for transmitting various measurement signals and causal data.

[0048] Preferably, the communication module is a 4G communication module.

[0049] Preferably, an alarm is also provided for issuing an alarm according to the early warning risk prediction value; when the early warning risk prediction value is higher than the risk threshold, an alarm is issued;

[0050] Furthermore, the alarm is a loudspeaker; further, when the alarm is issued, the alarm information is simultaneously uploaded to the cloud server.

[0051] The wearable device of the present invention can provide real-time early warning of AMI risk and, leveraging the interpretability of the AMI structural causal model, provide targeted recommendations. After successfully acquiring user data on the device side (blood pressure monitor, electrocardiogram, oximeter, and pulse monitor) and the display processing unit, the data is forwarded via an API to a signal processor (cloud server). After storage and analysis, the current AMI risk prediction value is provided. The wearable device also includes a medically based recommendation library. Based on various collected user measurement signals, manually entered data, and the predicted AMI risk value, the wearable device returns targeted recommendations to the user, guiding them to seek medical attention promptly. This allows for screening and control of potential risks before AMI occurs, preventing the disease process at an early stage and avoiding the serious consequences of an AMI attack.

[0052] Furthermore, the big data analyzer is used to collect and process historical medical big data to obtain a structured data set for training the AMI structural causal model. The historical medical big data includes medical records in real hospital scenarios.

[0053] Among them, medical record information includes the patient's medical history information, the doctor's description of the patient's condition, and the diagnosis results based on the patient's actual situation; medical record information includes the patient's biochemical test index information, daily living habits and past medical history, etc.; the doctor's description of the patient's condition includes the patient's oral statement and the doctor's records.

[0054] Furthermore, the big data analysis module processes historical medical big data. First, it summarizes the information extraction rules of historical medical big data, organizes the structured data according to the information rules, uses regular expressions to identify key information, and finally processes the complex text information through natural language processing methods to obtain a structured data set. The specific processing steps are as follows:

[0055] First, summarize the rules for extracting historical medical big data information:

[0056] For the natural language information in historical medical big data, which is derived from patients' oral statements and doctors' records, we analyze its regular characteristics and formulate information extraction rules. The extraction rules include two aspects: 1) confirming the normal value and / or abnormal value range of biochemical test indicators in the natural language information in patients' oral statements and doctors' records, as well as the standard values of various biochemical test indicators; 2) based on the terminology used by doctors to describe diseases in diagnosis, a type of disease may have multiple expressions in medicine. For example, abnormal cardiac function can be expressed as "cardiac function grade x (NYHA classification)", "cardiac function grade x (Killip) classification", "heart failure", etc. Based on the terminology, we confirm the description rules for various diseases in the natural language information in patients' oral statements and doctors' records.

[0057] Then, structured data is organized according to information rules: for biochemical test indicator information with fixed rules and clear structure, data processing methods in Python Pandas are used to organize structured data;

[0058] Regular expressions are used to identify key information: For information with fixed expressions and strong keyword positioning, regular expressions are used to identify keywords and obtain disease results, such as "chronic renal insufficiency" and "type 2 diabetes".

[0059] Finally, a structured data set was obtained by processing complex text information using natural language processing methods. For historical medical big data whose semantics cannot be identified by keywords alone and which do not have a fixed form of expression, natural language processing methods were used for processing and analysis. For example, if the medical record description contains semantic information such as "the patient has a history of smoking" or "the patient does not smoke" that cannot be directly identified by keywords, the natural language processing model is used to realize intelligent recognition of fuzzy information.

[0060] Furthermore, the natural language inference task method in the natural language processing method is used for analysis and processing. The natural language inference task method can judge the relationship between the natural language statement premise and the given hypothesis by identifying the natural language statement premise. Taking the specific case as the premise, the hypothesis is set that the case has a certain feature, and whether the hypothesis is valid is identified, thereby realizing the intelligent extraction of case features and the intelligent processing of large amounts of text information.

[0061] Preferably, an advanced deep language processing discriminant model based on BERT is used to process natural language inference tasks, and is applied to the recognition and processing of medical texts with complex expressions (i.e., medical big data whose semantics cannot be recognized by keywords alone and which do not have a fixed form of expression). It has good sentence processing capabilities. In order to enhance its recognition capabilities for long texts, after adjusting the task application mode, the model can complete the recognition task based on the prompt words, and extract feature information of complex semantic expressions such as "smoking", "chest pain" and "drinking" from the medical record text with good accuracy.

[0062] Furthermore, the big data analysis module discretizes the structured data set based on medical knowledge.

[0063] For example, in clinical diagnosis, doctors typically make a diagnosis based on clinical symptoms, biochemical test indicators, and electrocardiogram characteristics, taking into account the patient's individual characteristics and ultimately determining whether the patient has AMI based on medical knowledge. Therefore, it is necessary to combine medical knowledge to determine which diagnostic items in the dataset are suitable for AMI diagnosis. Furthermore, because continuous data such as some biochemical test indicator information has large ranges, which affects calculation results, the present invention, starting from the actual pathophysiological mechanism, discretizes the data information based on medical principles and references the medical classification basis of various indicator values. This discretization process produces a structured dataset containing AMI pathogenesis information variables and related factor characteristic variables (including demographics, lifestyle habits, medical history and physical condition, and clinical manifestations).

[0064] Among them, the AMI structural causal model is used to describe the complex causal relationship between AMI and AMI-related factors.

[0065] Furthermore, the establishment of the AMI structural causal model introduces the AMI causal structure.

[0066] Compared with other data-driven model methods, the AMI structural causal model method of the present invention emphasizes the causal logical relationship between variables rather than just considering correlation. For example, the causal relationship emphasizes the sequential logic of the time dimension of variables. For example, if variable A is the cause of variable B, then variable A occurs before variable B.

[0067] See also Figure 2The AMI structural causal model is constructed based on a directed acyclic causal graph (i.e., AMI causal structure) to describe the causal relationship between variables in a variable set. The variables in the AMI causal structure are AMI, related factors that cause AMI, and manifestation factors caused by AMI. Directed edges represent the generation process from cause to effect between variables.

[0068] It is understandable that the variable entries may be AMI incidence information variables and related characteristic variables (including demographics, lifestyle habits, past medical history, physical condition and clinical manifestations).

[0069] Furthermore, an AMI causal structure (i.e., a directed acyclic causal graph) is constructed based on the causal discovery algorithm.

[0070] Furthermore, when using the causal discovery algorithm to construct the AMI causal structure, medical prior knowledge is introduced into the AMI causal structure, which avoids incorrect causal relationships in advance and generally improves the accuracy of the structural causal model construction;

[0071] Specifically, when using the causal discovery algorithm to construct the AMI causal structure, the constraints of the causal discovery process are designed based on medical prior knowledge:

[0072] The variable entries are divided into categories according to demographics, lifestyle, medical history, physical condition, and clinical manifestations. On this basis, constraints are designed based on medical prior knowledge. The specific constraints are:

[0073] (1) It is prohibited to use variables other than demographics in variable entries to point to demographics or the causal relationship between demographics;

[0074] (2) Based on medical knowledge, summarize the temporal order of variables such as demographics, lifestyle, past medical history, physical condition, and clinical manifestations associated with the incidence of acute myocardial infarction (AMI) in historical medical big data, and prohibit causal relationships that violate the temporal order;

[0075] (3) The internal causal relationship between clinical manifestations such as chest pain and lung rales is prohibited.

[0076] Furthermore, combined with medical knowledge constraints, the AMI causal structure is constructed based on constraint conditions and causal discovery methods.

[0077] Specifically, the causal discovery method adopted by the present invention is a fast causal inference algorithm based on conditional independence or a fast greedy equivalence class search algorithm based on scores, which respectively mine the causal relationship between relevant factors from historical medical big data based on different rules and construct the AMI causal structure;

[0078] Among them, the fast causal inference algorithm examines causal relationships from historical medical big data based on the conditional independence rule and identifies the causal direction by identifying the V structure between variables;

[0079] Specifically, for the identification of V-structure causal relationships, the expression is:

[0080]

[0081] Where V represents a V structure; A and B represent variables A and B respectively; C represents the common adjacent variable C of variables A and B; its causal meaning is that variables A and B are the causes of the common adjacent variable C respectively; its statistical characteristics are that, without the given common adjacent variable C, variables A and variable B are independent; with the given common adjacent variable C, variables A and variable B are conditionally not independent.

[0082] Specifically, the fast causal inference algorithm determines whether the V structure exists by identifying the separation set.

[0083] The set is a set of variables that can make two variables meet the conditional independence, expressed as:

[0084]

[0085] Among them, A and B represent variable A and variable B respectively; X represents a set of variables. Given the variable set X, variable A and variable B are conditionally independent. Their conditional independence can be judged by statistical test methods such as chi-square test, thereby determining the range of the separation set.

[0086] The fast causal inference algorithm starts from an undirected graph with all variables connected, searches for a separation set between any two variables, and if so, deletes the causal direction between the two variables. It then traverses all variables and constructs a directed acyclic causal graph.

[0087] It can be understood that if the common adjacent variable C of two variables A and B does not belong to their separation set X, a V structure is determined. .

[0088] Based on this, the fast causal inference algorithm obtains the complete causal structure by deleting causal edges and determining the V structure.

[0089] Among them, the fast greedy equivalence class search algorithm obtains the optimal AMI causal structure based on the Bayesian information criterion and other scoring searches;

[0090] Specifically, the fast greedy equivalence search algorithm determines causal direction by identifying the V-structure relationships between variables. Starting from a causal graph without edges (assuming all variables are independent), the fast greedy equivalence search algorithm identifies the optimal AMI causal structure by assigning scores to the causal model.

[0091] Specifically, a fast greedy equivalence class search algorithm can evaluate causal structure using scores such as the Bayesian Information Criterion;

[0092] The expression of Bayesian Information Criterion score is:

[0093]

[0094] in, Indicates the rating value; k the degrees of freedom that represent the causal structure (i.e., the number of parameters used to describe the statistical characteristics of the causal structure, which increases with the number of causal directions in the causal structure); N represents the sample size (i.e. the amount of data used to train the model); L represents the likelihood function; Indicates the variables, Indicates the The causal variable of each variable, represents the probability, which is calculated based on the statistical characteristics of the variable given the cause variable in the causal structure. n Indicates the total number of variables.

[0095] Specifically, the fast greedy equivalence class search algorithm is divided into two stages: the first stage selects the direction with the best score based on the direction score and adds causal directions to the direction with the best score; the second stage identifies causal directions that can improve the score of the causal graph model by removing them, thereby obtaining a complete AMI causal structure.

[0096] Furthermore, the additive noise model and medical knowledge were used to further refine the AMI causal structure and identify causal directions that were not recognized by the two previous causal discovery algorithms, resulting in the final AMI causal structure. This solved the problem of Markov equivalence classes in causal discovery algorithms, which prevented some causal directions from being identified in structural causal models with identical edge relationships and collision structures.

[0097] The expression of the AMI causal structure of the additive noise model is:

[0098]

[0099] in, is the cause variable entry, is the result variable entry, is the residual term.

[0100] It can be understood that the additive noise model confirms the causal direction by testing the conditional independence between the noise term and the independent variable. For example, if the causal variable term For AMI, set the result variable entry For "chest pain", we can identify the residual term and the cause variable term The independence between variables can be used to determine the true causal direction, and the causal relationship between other variables can be determined by analogy.

[0101] Because the data used in this paper is discretized feature data based on medical knowledge, an improved additive noise model calculation method suitable for discrete variables is introduced. Furthermore, for causal directions that cannot be determined by the additive noise model, medical knowledge is incorporated to determine the causal relationship between variables, ultimately resulting in a complete structural causal model structure.

[0102] Furthermore, the parameters of the AMI causal structure are learned based on the above. The conditional probability parameters of the AMI structural causal model are estimated using the Bayesian parameter estimation method. , that is, to estimate the variables The parent node variable collection In the case of variables The conditional distribution parameters of are used to obtain the final AMI structural causal model.

[0103] Furthermore, based on the AMI structural causal model, according to various real-time measurement signals and manually input data, the conditional probability of AMI (i.e. ,in, Represents a given set of evidence, thereby achieving the inference of the probability of AMI and obtaining causal data of acute myocardial infarction and AMI prediction risk prediction value.

[0104] The present invention evaluates the predictive performance of the structural causal model established based on historical medical big data of AMI. Specifically, given the patient condition indicators of historical medical big data, the probability of AMI is inferred based on the AMI structural causal model, and compared with the true value. The F1 score, accuracy, recall rate, precision and other indicators of the AMI structural causal model prediction results are calculated. Based on these indicators, the present invention is compared with the traditional data-driven method, and the predicted data obtained by the present invention is more accurate.

[0105] Furthermore, the wearable intelligent diagnosis system for acute myocardial infarction also includes a real-time measurement data input and output module for collecting data signals from blood pressure measuring devices, blood oxygen measuring devices and pulse detection devices in real time, as well as allowing users to pre-set personal demographics, medical history and clinical symptoms in the software, and output patient risk information.

[0106] Optionally, the real-time measurement data input and output module is a mobile phone APP.

[0107] Furthermore, based on the AMI structural causal model, the effective information value (EI) of real-time measurement signals and manually input data (i.e., indicators) was calculated. The real-time measurement signals and manually input data were ranked according to the significance of the causal relationship between the occurrence of AMI based on the effective information value. Combined with the feasibility of data monitoring and transmission, and the medical connection between real-time measurement signals and manually input data and AMI symptoms, a comprehensive analysis and evaluation was conducted from three aspects, thereby preliminarily confirming the range of indicators that play a key role in AMI prediction.

[0108] When in use, the signal data measured in real time on the device side is transmitted to the device, and the pre-input information and actual measurement values are combined to perform real-time calculation of the probability of AMI based on the built-in AMI intelligent diagnosis model based on the structural causal model, and timely prompt the risk of disease.

[0109] Specifically, since the user's blood pressure curve shows different characteristics within 24 hours, with dynamic variability and circadian rhythm, the blood pressure measuring instrument of the wearable device automatically measures blood pressure every 15 to 30 minutes.

[0110] Furthermore, during the day, the user keeps the wearable device at the same height as the heart; at night, the user lies flat to ensure normal data collection. The use of the wearable device of the present invention for daily blood pressure monitoring is reasonable and feasible. Too frequent sampling intervals may affect the user's normal life and cause frequent compression and contraction of blood vessels, thereby causing measurement errors. Therefore, the present invention combines the actual situation of the blood pressure measuring instrument and the developed prototype to determine that the blood pressure timing monitoring interval of the blood pressure measuring instrument is every half hour within 24 hours, which helps to monitor the user's blood pressure in a long-term and lasting manner, thereby scientifically achieving disease prevention.

[0111] Furthermore, a five-minute interval was selected as the pulse meter (i.e., heart rate) monitoring time interval. The arithmetic mean of the heart rate over a 24-hour period was selected as the model input, which improved the accuracy of the results.

[0112] Blood pressure measuring instrument, used to collect the user's blood pressure signal in real time;

[0113] ECG meter, used to collect the user's ECG signals in real time;

[0114] A blood oxygen meter, used to collect the user's blood oxygen saturation signal in real time;

[0115] A pulse meter, used to collect the user's pulse signal in real time;

[0116] Furthermore, the wearable device samples blood oxygen every five minutes during the night, which improves the accuracy of the results; further, the arithmetic mean of all nighttime data on blood oxygen saturation eliminates random errors.

[0117] Based on this, the present invention can provide a relatively scientific and complete physiological indicator monitoring solution to better adapt the user's prediction function of the risk of acute myocardial infarction in daily scenarios. The specific content is as follows:

[0118] Blood pressure is measured every half hour, heart rate every five minutes, and blood oxygen saturation every five minutes at night. All of this data is transmitted from the device to a cloud platform via 4G. The cloud server preprocesses the data, taking the daytime and nighttime peak values for blood pressure and the arithmetic mean of all data collected throughout the day for both blood oxygen and heart rate. This processed data is then fed into the structural causal model.

[0119] In another embodiment of the present invention, the wearable device for intelligent diagnosis of acute myocardial infarction based on the structural causal model also includes an AMI-related causal effect significant factor identification module, which is used to provide guidance for the prevention and treatment of AMI and establish a systematic understanding of AMI.

[0120] Furthermore, the significance of the causal relationship between relevant factors in the AMI structural causal model and AMI was evaluated to identify factors with significant causal effects on AMI, providing guidance for both front-end AMI risk warning and subsequent AMI clinical treatment. The significance of the causal relationship between factors in the structural causal model was quantified using the effective information index.

[0121] Specifically, the effective information index is used to quantify the set probability intervention and equal probability intervention of the AMI structural causal model, and the KL divergence between the two probability intervention results is obtained. The KL divergence is used to describe the significance of the causal effect, and the expression is:

[0122]

[0123]

[0124]

[0125]

[0126] in, Representing variablesA For variables the extent of the impact; Representing variables The total number of states in the state space of ; Representing variables The state space of status; Representing variables The total number of states in the state space of Representing variables The state space of status; Indicates intervention in the state of the variable; represents probability; represents an intervention state with equal probability.

[0127] For example, to calculate the effect of AMI on chest pain For example, AMI is represented as " A ”, chest pain is expressed as “ ”.

[0128] The following application of the proposed wearable device for intelligent diagnosis of acute myocardial infarction is demonstrated through a real-world case study. The data was collected from the emergency department of Beijing Jishuitan Hospital, which contains medical records of 3,521 patients. The information includes biochemical tests, medical history, lifestyle habits, clinical diagnosis results, and other dimensions.

[0129] Step 1: Processing and analysis of medical big data;

[0130] The medical record dataset contains information on 3521 patients, of which the raw data of biochemical tests (partial) are shown in Table 1.

[0131] Table 1 Raw data of biochemical tests (partial)

[0132]

[0133] Information extraction was performed on the medical record dataset. Key-value pair matching, regular expression recognition, and natural language processing were used to extract biochemical test data, text information with fixed expressions, and text information without fixed expressions. Invalid entries such as missing items, duplicates, and invalid entries in the dataset were deleted to ultimately obtain structured data.

[0134] Based on the summary of medical knowledge items, by investigating medical background knowledge and analyzing the available information in the data, four categories of AMI-related characteristics that should be taken into consideration were identified, including demographics, lifestyle habits, medical history and physical condition, and clinical manifestations, with a total of 56 key items. Table 2 shows the important AMI-related items (part).

[0135] Table 2 Important items related to AMI (partial)

[0136]

[0137] Furthermore, the data were discretized and some data discretization standards (partial) were developed based on medical knowledge. Some of the standards are shown in Table 3.

[0138] Table 3 Data discretization processing standards (partial)

[0139]

[0140] A total of 3521 medical records described in digital form were obtained and can be used for calculation. The data processing results (partial) are shown in Table 4.

[0141] Table 4 Data processing results (partial)

[0142]

[0143] Step 2: Establish the structural causal model of AMI;

[0144] Based on the data obtained in step 1, a structural causal model was constructed using a causal discovery algorithm. First, the medical causal relationships within each item were analyzed to determine the constraints of medical prior knowledge. These constraints include: 1) prohibiting causal relationships pointing to demographics and between them; 2) prohibiting causal relationships between clinical manifestations; and 3) constructing the model according to the causal order of demographics → lifestyle → medical history and symptoms → AMI onset → clinical manifestations.

[0145] Next, a structural causal model is constructed based on a fast greedy equivalence class search algorithm.

[0146] Furthermore, based on the discrete additive noise model and the medical knowledge refined structural causal model structure, Table 5 shows the causal relationship (partial) identified based on the ANM model.

[0147] Table 5 Causal relationships identified based on the ANM model (partial)

[0148]

[0149] Finally, the structure of the structural causal model is obtained as follows Figure 3 shown.

[0150] Finally, the structural causal model was parameterized based on the existing structural causal model structure, and the conditional probability was estimated based on the Bayesian method to obtain the final AMI structural causal model. The model performance was verified on this basis. By comparing the inferred AMI probability of the model with the true value, the performance evaluation results of the AMI structural causal model were calculated and shown in Table 6.

[0151] Table 6 Performance evaluation results of AMI structural causal model

[0152]

[0153] Step 3: Identify the significant factors related to the causal effect of AMI;

[0154] According to the above parameterized structural causal model, calculate , quantify the significance of causal effects between AMI-related factors. After calculation, 11 factors were found to be significantly affected by the causal effect of AMI, and 9 factors were found to have a significant causal effect on the incidence of AMI, such as Figure 4 、 Figure 5 shown.

[0155] Step 4: Implement intelligent diagnosis of AMI based on wearable devices;

[0156] The above model is connected to smart wearable devices and corresponding mobile phone apps. Based on the blood pressure, heart rate, pulse and other information collected by the smart wearable devices on the signal acquisition platform, as well as other indicators such as personal information, living habits, and medical history manually entered by the user into the app, real-time intelligent diagnosis of AMI is achieved.

[0157] At the same time, the above-mentioned AMI system analysis results are integrated, and the significance of the causal effects of AMI-related factors is converted into "health tips" to remind users to develop good daily living habits and guide daily health management.

[0158] The detailed variable entries and corresponding abbreviations in the case are shown in Table 7.

[0159] Table 7 Variable entries and corresponding abbreviations

[0160]

[0161] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with this technical field within the technical scope disclosed by the present invention should be covered by the scope of protection of the present invention.

Claims

1. A method for constructing a structural causal model for acute myocardial infarction monitoring, characterized in that: The specific steps are as follows: Step 1: Collect and process historical medical big data to obtain a structured dataset for training the AMI structural causal model; Among them, the steps for processing historical medical big data are: Determine the rules for extracting historical medical big data information; Structural data organization of historical medical big data based on information extraction rules; Identify feature information with regular expressions through regular expressions; Use natural language processing methods to process feature information in historical medical big data that does not have regular expression forms; Finally, a structured data set is obtained; Step 2: Discretize the structured data set according to medical knowledge to obtain a discretized data set; Step 3: Based on the discretized dataset, an AMI causal structure is constructed based on the medical prior knowledge constraints, the causal discovery algorithm, the additive noise model, and medical knowledge. The variables in the AMI causal structure are AMI, the associated factors that cause AMI, and the manifestation factors caused by AMI. The directed edges represent the generation process from cause to effect between the variables. Step 4: Based on the discretized data set from step 2, use the Bayesian parameter estimation method to learn the parameters of the AMI causal structure and estimate the conditional probability parameters; based on the conditional probability parameters, the final AMI structural causal model is obtained; Among them, the variable entries are based on demographics, living habits, past medical history, physical condition and clinical manifestations, and constraints are designed based on medical prior knowledge. The constraints are: (1) It is prohibited to use variables other than demographics in variable entries to point to demographics or the causal relationship between demographics; (2) Based on medical knowledge, summarize the temporal order of the variable entries in demographics, lifestyle, medical history, physical condition, and clinical manifestations associated with the incidence of acute myocardial infarction (AMI) in historical medical big data, and prohibit causal relationships that violate the temporal order; (3) The internal causal relationship between the clinical manifestations of chest pain and lung rales is prohibited; Among them, the causal discovery algorithm is a fast causal inference algorithm based on conditional independence or a fast greedy equivalence class search algorithm based on scores; The fast causal inference algorithm examines causal relationships from historical medical big data based on conditional independence rules and identifies causal directions by identifying the V structure between variables; For the identification of V-structure causal relationship, the expression is: Where V represents a V structure; A and B represent variables A and B respectively; C represents the common adjacent variable C of variables A and B; its causal meaning is that variables A and B are the causes of the common adjacent variable C; The fast causal inference algorithm determines whether the V structure exists by identifying the separation set. The set is a set of variables that can make two variables meet the conditional independence, expressed as: Where X represents a set of variables; The expression of the AMI causal structure of the additive noise model is: in, is the cause variable entry, is the result variable entry, is the residual term.

2. The method for constructing a structural causal model according to claim 1, characterized in that: Historical medical big data includes medical record information in real hospital scenarios; medical record information includes the patient's medical history information, the doctor's description of the patient's condition, and the diagnosis results based on the patient's actual situation; medical record information includes the patient's biochemical test index information, daily living habits and past medical history; the doctor's description of the patient's condition includes the patient's oral description and the doctor's records.

3. The method for constructing a structural causal model according to claim 2, characterized in that: The variables include AMI incidence information variables and corresponding characteristic variables of the AMI incidence information variables.

4. The method for constructing a structural causal model according to any one of claims 1 to 3, characterized in that: Introducing medical prior knowledge into the causal structure of AMI.

5. The method for constructing a structural causal model according to any one of claims 1 to 3, characterized in that: Adopting additive noise models and / or medical knowledge to update the AMI causal structure.

6. The method for constructing a structural causal model according to claim 5, characterized in that: The causal discovery method of AMI causal structure is a fast causal inference algorithm based on conditional independence or a fast greedy equivalence class search algorithm based on scores.

7. An acute myocardial infarction intelligent diagnostic wearable device for acute myocardial infarction monitoring, characterized in that: The invention comprises a signal acquisition instrument, a signal processing instrument and a display processing unit; the signal processing instrument is provided with an AMI structural causal model, and the AMI structural causal model is constructed using the structural causal model construction method according to any one of claims 1 to 6.

Images