Multi-center child pulmonary embolism online data platform system

Abstract

The invention discloses a multi-center child pulmonary embolism online data platform system, and belongs to the technical field of data processing. A data collection step: collecting electronic medical records, image data, laboratory results and treatment scheme multi-source data uploaded by each center, and unifying data formats and standardizing processing flows; a machine learning and artificial intelligence algorithm step: carrying out deep analysis on mass data, identifying risk factors, pathogenesis and treatment response modes of children pulmonary embolism, and supporting personalized medical decisions; a dynamic risk assessment tool step: developing a dynamic risk assessment tool according to the latest research results and clinical guidelines, and providing real-time treatment suggestions and prognosis assessment for doctors; a data security and privacy protection step: adopting a data encryption step and an anonymization step to ensure the security and privacy protection of patient information; and a multi-center cooperative network step: establishing a children pulmonary embolism research and treatment cooperative network nationwide through a data sharing and remote cooperation mechanism.

Description

Technical Field

[0001] The present invention belongs to the technical field of data processing and relates to a multi-center pediatric pulmonary embolism online data platform system. Background Art

[0002] Due to the unique pathophysiological characteristics of children, pulmonary embolism has hidden symptoms and is easily ignored, which often leads to missed diagnosis and misdiagnosis, resulting in a prolonged course of the disease, poor prognosis, and high mortality and disability rates. At present, there are few international studies on pediatric pulmonary embolism, with limited data, mainly case reports and single-center retrospective studies, and a lack of multicenter large-sample cohort studies. Due to the inconsistency of databases in various units, it is difficult to provide high-quality research evidence for clinical work, making it impossible for pediatricians to identify pulmonary embolism events early. There is also a lack of sufficient summary of treatment experience, especially a lack of research data and reports from China. Summary of the invention

[0003] In view of the problems in the prior art, the present invention provides a multi-center pediatric pulmonary embolism online data platform system.

[0004] A multi-center pediatric pulmonary embolism online data platform system, comprising the following steps:

[0005] The data collection step is to collect multi-source data such as electronic medical records, imaging data, laboratory results, and treatment plans uploaded by each center, unify the data format and standardize the processing process.

[0006] Machine learning and artificial intelligence algorithm steps conduct in-depth analysis of massive data to identify risk factors, pathogenesis, and treatment response patterns of pulmonary embolism in children, and support personalized medical decision-making.

[0007] Dynamic risk assessment tool steps: Develop a dynamic risk assessment tool based on the latest research results and clinical guidelines to provide doctors with real-time treatment recommendations and prognosis assessments.

[0008] Data security and privacy protection steps use data encryption and anonymization technology to ensure the security and privacy of patient information.

[0009] The multi-center collaborative network step is to establish a nationwide collaborative network for research and treatment of pediatric pulmonary embolism through data sharing and remote collaboration mechanisms.

[0010] The advantages of the present invention are: by means of the remote network resource advantage of the telemedicine center of Peking Union Medical College Hospital, the basic information, clinical information, imaging information, laboratory test information, therapeutic drug information, biological specimen information and case follow-up information of children's pulmonary embolism cases enrolled in each hospital are collected through the hospital electronic medical record (EMR), physical examination information system (PEIS), medical image archiving and communication system (PACS), laboratory information system (LIS), outpatient system, inpatient system, etc. The information is imported into the two-dimensional structured database of the system in real time, and is automatically connected with the biological specimen omics (genomics, metabolomics, etc.) database, so as to realize the automation, networking and real-time of the information collection and tracking and follow-up of the cohort population, and meet the remote data access under various operating systems.

[0011] The present invention has significant technical effects in medical practice and scientific research. First, the system realizes the automatic collection and integration of case data from various hospitals by integrating multiple platforms such as electronic medical records (EMR), physical examination information system (PEIS), medical image archiving and communication system (PACS), laboratory information system (LIS), etc., thereby improving the efficiency and accuracy of data collection and ensuring the standardization and consistency of multi-source data. The present invention can transmit and update case data in real time, store the information in a two-dimensional structured database, and automatically connect with the biological specimen omics database (such as genomics, metabolomics), supporting large-scale data storage, intelligent analysis and remote data access. This process breaks down the data barriers between hospitals with the help of the network resources of the telemedicine center and supports multi-center case sharing and collaborative research.

[0012] Through automated data processing and standardized case information entry, the present invention reduces the risk of manual errors and significantly improves the efficiency of data transmission and management. At the same time, the present invention can support real-time clinical data collection and tracking and follow-up of cohort populations, achieving long-term health monitoring and treatment effect evaluation of patients. With the help of advanced artificial intelligence and machine learning algorithms, the present invention can conduct in-depth analysis of massive case data, identify risk factors, pathogenesis and treatment response patterns of pediatric pulmonary embolism, provide doctors with personalized diagnosis and treatment recommendations, and improve the accuracy and efficiency of clinical decision-making.

[0013] The present invention is compatible with a variety of operating systems, ensuring that hospitals and research institutions can seamlessly access and operate on different devices and platforms, greatly enhancing the flexibility of the present invention. The present invention complies with international standards in terms of data encryption and privacy protection, and adopts strict security measures and authority control to ensure the security and legality of patient information. At the same time, the present invention supports multi-center collaborative research across hospitals and regions, promotes the progress of large-scale clinical trials and epidemiological studies, and especially provides an important data basis for clinical research on pediatric pulmonary embolism. Its automated patient screening and enrollment functions provide strong support for the conduct of clinical trials, and ensure the systematic and comprehensive management of patients through a continuous follow-up mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. As shown in the figure:

[0015] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] Example 1: Figure 1 As shown, a multi-center pediatric pulmonary embolism online data platform system optimizes the diagnosis, treatment and research process of pediatric pulmonary embolism, integrates the clinical data of major medical centers across the country, and forms a comprehensive, efficient and safe pediatric pulmonary embolism information sharing and analysis system. The user-friendly interface supports multi-role access by doctors, researchers and medical managers, and realizes full-process automated management from data collection, processing to analysis.

[0018] A multi-center pediatric pulmonary embolism online data platform system, comprising the following steps:

[0019] The data collection step is to collect multi-source data such as electronic medical records, imaging data, laboratory results, and treatment plans uploaded by each center, unify the data format and standardize the processing process.

[0020] Machine learning and artificial intelligence algorithm steps conduct in-depth analysis of massive data to identify risk factors, pathogenesis, and treatment response patterns of pulmonary embolism in children, and support personalized medical decision-making.

[0021] Specific steps for in-depth analysis of massive data:

[0022] In the multi-center pediatric pulmonary embolism online data platform system (or system and platform), in-depth analysis of massive data is achieved through a series of precise and systematic steps to identify risk factors, pathogenesis and treatment response patterns to support personalized medical decision-making. First, the platform collects multi-source data such as electronic medical records (EMR), medical images, laboratory results, genomic data and treatment plans from each center, and performs unified data formatting and standardization to ensure the compatibility and consistency of data from different hospitals. The data preprocessing stage includes data cleaning, outlier processing, missing data completion and feature extraction. Through this series of steps, characteristic variables that are crucial to the prediction of pulmonary embolism risk are extracted. After feature extraction, the system performs exploratory data analysis (EDA) to preliminarily identify potential risk factors and screen out variables closely related to pediatric pulmonary embolism through statistical analysis, correlation calculation and visualization.

[0023] Subsequently, the system performs feature selection and dimensionality reduction, screens key features through algorithms such as Lasso regression and random forest, and uses principal component analysis (PCA) and other techniques to reduce dimensionality and reduce model complexity. Based on the preprocessed data, the platform trains a variety of machine learning models. For labeled data, the platform uses classification algorithms (such as logistic regression, support vector machine, random forest, etc.) for risk prediction, or regression algorithms for prognostic analysis; for unlabeled data, the platform uses unsupervised learning algorithms (such as K-means clustering, DBSCAN, etc.) to identify patient subgroups and potential pathological patterns. At the same time, the system analyzes imaging data through convolutional neural networks (CNN), combines time series models (such as LSTM) to process changes in laboratory indicators, and provides disease progression predictions.

[0024] After model training, the system evaluates model performance through multiple indicators such as accuracy, recall, AUC, etc., and optimizes hyperparameters through grid search or random search to ensure the stability of the model on different data sets. To ensure the interpretability of the results, the system uses model interpretation tools such as SHAP value and LIME to analyze the contribution of each feature to model prediction, help doctors understand potential risk factors and their weights, and analyze the causal relationship between features through causal inference models. In addition, the system supports personalized treatment decisions through subgroup analysis and personalized prediction models, provides predictions of different patients' responses to different treatment options, and guides doctors to develop the best treatment plan.

[0025] The system also has real-time data update and model retraining functions, which can continuously optimize and retrain the model as new data is input to ensure the cutting-edge and accuracy of the analysis results. With the help of adaptive learning mechanisms, the platform automatically adjusts the algorithm parameters to adapt to the different data characteristics of each center. The platform also supports multi-center collaboration, provides data sharing functions, promotes cross-hospital collaborative research, and helps researchers conduct data extraction, hypothesis verification and large-scale clinical research through scientific research support tools. In addition, the platform supports long-term cohort follow-up analysis to help doctors track changes in patients' conditions in real time and evaluate treatment effects. Through these steps, the platform can comprehensively analyze massive data, identify risk factors, pathogenesis and treatment response patterns for pediatric pulmonary embolism, and provide strong support for personalized medicine and precision diagnosis and treatment.

[0026] Dynamic risk assessment tool steps: Develop a dynamic risk assessment tool based on the latest research results and clinical guidelines to provide doctors with real-time treatment recommendations and prognosis assessments.

[0027] Specific content of the dynamic risk assessment tool:

[0028] The dynamic risk assessment tool is an intelligent system based on the latest clinical guidelines and research results, designed to provide doctors with real-time treatment recommendations and prognosis assessments. Its core functions include automatically collecting patient information from multiple sources of data, such as electronic medical records, imaging data, laboratory test results, and genomic information, and updating and inputting data in real time. The system first standardizes and cleans the data to ensure that the data is accurate and consistent, and then combines traditional statistical models and machine learning algorithms (such as random forests, deep neural networks, etc.) to accurately assess the patient's personalized risk. Based on the pre-trained risk scoring algorithm, the tool can generate disease risk assessment results, such as cardiovascular disease, diabetes, pulmonary embolism, etc., to support doctors in judging the patient's disease risk. The system also embeds the latest clinical guidelines (such as the American Heart Association, the American Diabetes Association, etc.) and the latest medical research results to automatically generate standard treatment recommendations. The intelligent treatment recommendation engine combines the patient's risk score and clinical data to provide personalized treatment options, such as drug selection, dose adjustment, surgical recommendations, or lifestyle changes, to help doctors develop the best treatment plan.

[0029] The dynamic risk assessment tool also supports the comparison function of multiple treatment options, performs risk-benefit analysis on each option, and helps doctors weigh the effects of different options. The dynamic risk assessment tool can also predict the patient's long-term prognosis, including recurrence risk, survival rate, and probability of complications, and generate a personalized health management plan. The system predicts the treatment effect through time series analysis and adjusts the follow-up plan in real time according to changes in the patient's condition. The dynamic risk assessment tool has a highly intuitive user interface. Doctors can view the patient's risk assessment results, trend analysis, and treatment recommendations in real time through visual reports. The dynamic risk assessment tool supports multi-device operation and can be run on desktop computers and mobile devices, making it convenient for doctors to use in different scenarios.

[0030] In terms of data security and privacy protection, the dynamic risk assessment tool follows strict encryption standards to ensure the security of patient data during transmission and storage, while protecting patient privacy through data anonymization, in compliance with international privacy protection regulations such as HIPAA and GDPR. The system also has permission management functions to ensure that only authorized medical personnel can access patient data. In addition, the dynamic risk assessment tool also supports multi-center collaboration, allowing different medical institutions to share data and promote cross-hospital collaborative research and clinical trials. Researchers can conduct large-scale clinical research based on the risk assessment results of the dynamic risk assessment tool to further explore new disease risk factors and treatment options.

[0031] The dynamic risk assessment tool has an adaptive learning function. By continuously collecting clinical application feedback and new data, it automatically updates and optimizes the algorithm to ensure its accuracy and cutting-edge nature. Doctors' feedback will also be incorporated into the system optimization process to continuously improve the user experience. The system also provides a modular design to support risk assessment of multiple diseases, such as cardiovascular disease, cancer, diabetes, etc. Doctors can adjust assessment parameters according to needs and personalize operating procedures to suit specific patient groups. Through these steps, the dynamic risk assessment tool provides doctors with a powerful intelligent support system to help them quickly and accurately conduct risk assessments, treatment decisions, and prognosis predictions, ultimately improving patients' clinical management effects and health outcomes.

[0032] Data security and privacy protection steps use data encryption and anonymization technology to ensure the security and privacy of patient information.

[0033] Specific content of data encryption and anonymization technology:

[0034] Data encryption and anonymization technology are important means to protect the security and privacy of patient information and are widely used in the medical field. Data encryption technology converts data into unreadable ciphertext to ensure that even if the data is intercepted or leaked during transmission or storage, unauthorized persons cannot interpret it. Among them, symmetric encryption uses the same key to encrypt and decrypt data. Common algorithms include AES (Advanced Encryption Standard), which is suitable for fast encryption of large-scale data. Asymmetric encryption** uses public key encryption and private key decryption. Common algorithms are RSA and elliptic curve cryptography (ECC). It is usually used to protect sensitive information in network communications, such as medical data transmission. In order to ensure the security of data during network transmission, transport layer encryption (TLS / SSL) establishes an encrypted communication channel for data transmission. It is widely used in network communications of electronic health records (EHR), telemedicine platforms and other medical information systems. Database encryption is used to encrypt and store patient data in a database, such as automatically encrypting and decrypting stored data through transparent data encryption (TDE) to ensure that sensitive information remains safe in the event of a data leak.

[0035] Data anonymization technology protects patient privacy by removing or blurring patients' personal identity information to prevent data from being directly associated with individuals. De-identification technology removes or hides direct identifiers in data, such as names, ID numbers, etc. It is widely used in data sharing and research to ensure that unrelated people cannot identify patients. Pseudo-anonymization generates alternative identifiers (such as random numbers) instead of real identifiers, which are used when sharing or storing data to ensure that data is difficult to directly associate with patients, but the patient's identity can be restored through re-identification under certain conditions. Data generalization further reduces the risk of individual identification by converting precise data into a more ambiguous form, such as converting detailed dates of birth into age groups. It is particularly suitable for analysis or public reporting of large-scale data. Data perturbation protects personal privacy while maintaining the integrity of the statistical characteristics of the data by adding random noise or small changes to the data. It is suitable for research and data analysis. Another technology is synthetic data, which is to generate virtual data sets based on real data. It has similar statistical characteristics to real data but does not contain any actual patient information. It is often used in teaching or non-clinical research.

[0036] In practical applications, the best way to ensure data privacy and security is to use a combination of encryption and anonymization technologies. For example, patient data is encrypted using TLS during transmission, database encryption is applied during storage, and privacy risks are reduced by de-identification or pseudo-anonymization during sharing and analysis. For large-scale data for scientific research purposes, generalized or synthetic data can maintain the practicality of the data while protecting privacy. In addition, these technologies must comply with international privacy protection regulations such as HIPAA and GDPR to ensure the legal use of medical data. The medical system should monitor and record data access behavior through access control, audit mechanisms, and permission management to prevent unauthorized operations. With the cooperation of encryption and anonymization technologies, medical data can be effectively protected and provided to clinical researchers or data analysts when needed, ensuring both security and practicality. These measures have greatly improved the level of privacy protection of patient data in the medical system, while ensuring the operability and research value of the data.

[0037] The multi-center collaborative network step is to establish a nationwide collaborative network for research and treatment of pediatric pulmonary embolism through data sharing and remote collaboration mechanisms.

[0038] Example 2: Figure 1 As shown, a multi-center pediatric pulmonary embolism online data platform system has:

[0039] Data integration capability: The platform system can automatically collect multi-source data such as electronic medical records, imaging materials, laboratory results and treatment plans uploaded by various centers, and ensure the consistency and comparability of data through a unified data format and standardized processing flow.

[0040] Advanced data analysis and mining: Introduce machine learning and artificial intelligence algorithms to conduct in-depth analysis of massive data, identify risk factors, pathogenesis and treatment response patterns of pediatric pulmonary embolism, and support personalized medical decision-making.

[0041] Dynamic risk assessment tool: Combining the latest research results and clinical guidelines, a dynamic risk assessment tool is developed to provide doctors with real-time treatment recommendations and prognosis assessments, helping to reduce the incidence of pulmonary embolism in children and improve the success rate of treatment.

[0042] Data security and privacy protection: Use advanced data encryption and anonymization technologies to ensure the security and privacy of patient information, while complying with relevant laws, regulations and ethical principles.

[0043] Multi-center collaborative network: Through efficient data sharing and remote collaboration mechanisms, a nationwide collaborative network for pediatric pulmonary embolism research and treatment will be established to promote the exchange of medical knowledge and the promotion of best practices.

[0044] A multi-center pediatric pulmonary embolism online data platform system to enhance the representativeness and diversity of data: Through multi-center collaboration, patient data from different regions, races, and backgrounds can be collected, which helps to increase the universality and applicability of research. This diversity is crucial for understanding global differences in the disease, improving treatment methods, and discovering new disease patterns.

[0045] Improve research efficiency and quality: Multi-center data platforms allow researchers to share data and resources, which not only avoids duplication of work but also improves research efficiency. In addition, such collaboration helps to achieve larger-scale studies and provide more powerful statistical analyses, thereby increasing the reliability of research findings.

[0046] Promoting innovation and interdisciplinary research: The multi-center platform encourages collaboration among experts in different fields, thus promoting interdisciplinary research. This interdisciplinary collaboration can reveal the complex mechanisms of diseases and promote the development of new treatments and technologies.

[0047] Accelerate the progress of clinical trials and research: On a multi-center platform, a sufficient number of research subjects can be recruited for clinical trials more quickly, accelerating the progress of research. In addition, it also provides a more extensive infrastructure and resources for research, helping to meet the challenges of large-scale clinical trials.

[0048] Improve the personalization of patient treatment: The accumulation and analysis of multi-center data helps to identify individual differences, which is crucial for the development of personalized medicine. By understanding how different patients respond to treatment, doctors can provide patients with more accurate and effective treatment plans.

[0049] Example 3: Figure 1 As shown, a multi-center pediatric pulmonary embolism online data platform system includes the following steps: entering new patients (basic information) through research center 1, research center 2, research center 3, research center 4 and research center N, obtaining the new patient's unique identification number through the multi-center research collaboration module, judging the "inclusion criteria", if "no", excluding the patient from the database, if "yes", judging the "exclusion criteria", if "no", the patient is included, if "yes", the patient is excluded from the database, completing the information, combining treatment information, initial drug treatment, sequential drug treatment, follow-up information, follow-up endpoint events, pulmonary function and cardiac ultrasound to export data, and exporting baseline data, general information, operations and complications during hospitalization, personal history, past history, clinical manifestations, hemodynamic evaluation when pulmonary embolism occurs, imaging evaluation, thrombotic disease-related indicators, rheumatic disease-related indicators, infectious disease-related indicators and blood system disease-related indicators. It also includes the following steps:

[0050] Step 1) Register and log in to the multi-center pediatric pulmonary embolism online data platform system. The main features of the platform system are:

[0051] Data center server, establishes and maintains standard variables and standard dictionaries;

[0052] The data center server collects the original data from each data source and stores it in the corresponding database, which includes: hospital electronic medical record (EMR), physical examination information system (PEIS), medical image archiving and communication system (PACS), laboratory information system (LIS), outpatient system, inpatient system, and each data corresponds to a unique code. The data center server can process and utilize the input data.

[0053] Step 2) Screen patients who meet the following criteria and obtain data on their electronic medical records, physical examination information systems, medical images, communication details, laboratory information, and treatment information.

[0054] Diagnostic criteria: Children under 18 years old, clinically excluded from other cardiopulmonary diseases, still have symptoms such as dyspnea, chest pain, syncope or shock, hemoptysis, asymmetric swelling and pain in both lower limbs, especially those with risk factors for pediatric pulmonary embolism such as central venous catheterization, infection, surgery, malignant tumors, trauma, heart disease, nephrotic syndrome, obesity, and patients with one of the following.

[0055] Direct signs of CTPA: low-density filling defects are found in the pulmonary arteries, which are partially or completely surrounded by opaque blood flow (track sign) or are complete filling defects, and the distal blood vessels are not visualized; indirect signs of CTPA: wedge-shaped increased density shadows in the lung field, strip-like high-density areas or discoid atelectasis, dilation of the central pulmonary artery and reduction or disappearance of distal vascular branches.

[0056] Magnetic resonance imaging showed filling defect in the pulmonary artery cavity and disappearance of distal signals.

[0057] Radionuclide lung ventilation / perfusion scan shows local pulmonary perfusion defects in multiple or one lung lobe segment, good or abnormal ventilation, and mismatched ventilation-perfusion ratio.

[0058] Echocardiography revealed proximal pulmonary artery thrombus or right heart thrombus.

[0059] 3) Fill in the following information into the corresponding items of the platform system:

[0060] ①Basic information:

[0061] Organize and record the basic information of the selected children, including age, gender, ethnicity, height, weight, admission date, discharge date, etc.

[0062] ②Risk factors include:

[0063] Concomitant diseases: including active malignant tumors (with local or distant metastasis and / or malignant tumor treatment within 6 months), hyperhomocysteinemia, chronic obstructive pulmonary disease, hypertension, coronary atherosclerotic heart disease, respiratory failure, inflammatory bowel disease, rheumatic diseases, diabetes, myocardial infarction, heart failure, varicose veins, abnormal renal function, lung infection, sepsis, stroke (<1 month), etc.;

[0064] Surgery, trauma and operation conditions: mainly including type of surgery (arthroscopic surgery, laparoscopic surgery, joint replacement, minor surgery, other major surgery), duration of surgery, central venous catheterization, various types of fractures, pelvic / lower limb fractures, multiple trauma, and acute spinal cord injury;

[0065] Past medical history: history of venous thromboembolism and unexplained habitual abortion;

[0066] Other risk factors: bed rest / immobilization (72 hours and >72 hours), known thrombotic tendency (including antithrombin deficiency, positive lupus anticoagulant, protein C or S deficiency, Leiden V factor mutation, prothrombin G20210A mutation, antiphospholipid antibody syndrome, etc.), pregnancy or postpartum status, recent use of oral contraceptives or hormone replacement therapy, lower limb swelling, etc.

[0067] ③ Inspection and examination information includes:

[0068] Laboratory: blood routine, blood biochemistry, coagulation function, protein C activity, protein S activity, etc.;

[0069] Imaging: electrocardiogram, echocardiogram, CT pulmonary angiography, radionuclide pulmonary ventilation / perfusion (V / Q) imaging, magnetic resonance pulmonary angiography, pulmonary angiography, radionuclide venous imaging, CT venography, venography.

[0070] Step 4), the platform system will automatically prompt the names provided by the relevant standard dictionaries for the diseases involved in the risk factors. The standard dictionaries include: "International Statistical Classification of Diseases and Related Health Problems" ICD10, "Chinese Pharmacopoeia" or positive signs; the standard variable maintenance includes: standard item names, codes and classifications; the standard dictionary maintenance, according to the "International Statistical Classification of Diseases and Related Health Problems" ICD10 or "Chinese Pharmacopoeia", standardizes and structures the original data.

[0071] Step 5) The platform system performs deduplication processing on the basic information of the child, including: deduplication of ID number; converting the text data in the examination result table into structured data; and standardizing the examination result names and examination result name codes in the examination result table according to the standard dictionary of the data center server.

[0072] Step 6) Cohort formation: select the original data as the research variable, set the inclusion and exclusion criteria, and set the outcome variable to generate the scientific research cohort; the inclusion criteria refer to a series of indicators or conditions selected from patients who meet the diagnostic criteria; the exclusion criteria refer to the exclusion of several indicators that will interfere with the accuracy of the results; the outcome variable is also called the result variable, abbreviated as the outcome; it refers to the expected outcome event that will occur in the follow-up observation, that is, the event that the researcher hopes to track and observe; the scientific research cohort refers to the selection of the required research subjects in a specified population, based on whether they are currently or in a certain period in the past exposed to a risk factor to be studied, forming a data matrix; the data statistics platform system is used to count and display data.

[0073] Example 4: Figure 1 As shown, a multi-center pediatric pulmonary embolism online data platform system contains the following steps:

[0074] Step 1: Register and log in to the Multi-center Pediatric Pulmonary Embolism Online Data Platform System.

[0075] Step 2: Fill in the following information in the corresponding items of the platform system:

[0076] ①Basic information:

[0077] Organize and record the basic information of the selected children, including age, gender, ethnicity, height, weight, admission date, discharge date, etc.

[0078] ②Risk factors

[0079] ■Comorbidities: including active malignant tumors (with local or distant metastasis and / or malignant tumor treatment within 6 months), hyperhomocysteinemia, chronic obstructive pulmonary disease, hypertension, coronary atherosclerotic heart disease, respiratory failure, inflammatory bowel disease, rheumatic diseases, diabetes, myocardial infarction, heart failure, varicose veins, abnormal renal function, lung infection, sepsis, stroke (<1 month), etc.;

[0080] ■ Surgery, trauma and operation status: mainly including type of surgery (arthroscopic surgery, laparoscopic surgery, joint replacement, minor surgery, other major surgery), duration of surgery, central venous catheterization, various types of fractures, pelvic / lower limb fractures, multiple trauma, acute spinal cord injury;

[0081] ■Past medical history: history of venous thromboembolism, history of unexplained habitual miscarriage;

[0082] ■Other risk factors: bed rest / immobilization (72h and >72h), known thrombotic tendency (including antithrombin deficiency, positive lupus anticoagulant, protein C or S deficiency, Leiden V factor mutation, prothrombin G20210A mutation, antiphospholipid antibody syndrome, etc.), pregnancy or postpartum status, recent oral contraceptive use or hormone replacement therapy, lower limb swelling, etc.

[0083] ③Inspection information

[0084] ■Laboratory: blood routine, blood biochemistry, coagulation function, protein C activity, protein S activity, etc.;

[0085] ■Imaging: electrocardiogram, echocardiogram, CT pulmonary angiography, radionuclide pulmonary ventilation / perfusion (V / Q) imaging, magnetic resonance pulmonary angiography, pulmonary angiography, radionuclide venous imaging, CT venography, venography.

[0086] Step 3: Enter the disease involved and select the standard disease name from the names provided in the automatically prompted relevant standard dictionary.

[0087] Step 4: The platform system removes duplicates from the basic information of the child and standardizes the text data in the examination result table.

[0088] Step 5: Queue formation and data export.

[0089] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-center pediatric pulmonary embolism online data platform system, characterized in that: Contains the following steps: Data collection steps: collect multi-source data such as electronic medical records, imaging materials, laboratory results, and treatment plans uploaded by each center, unify the data format and standardize the processing process; Machine learning and artificial intelligence algorithm steps to conduct in-depth analysis of massive data to identify risk factors, pathogenesis and treatment response patterns of pediatric pulmonary embolism and support personalized medical decision-making; Dynamic risk assessment tool step: Develop a dynamic risk assessment tool based on the latest research results and clinical guidelines to provide doctors with real-time treatment recommendations and prognosis assessments; Data security and privacy protection steps, using data encryption and anonymization steps to ensure the security and privacy of patient information; The multi-center collaborative network step is to establish a nationwide collaborative network for research and treatment of pediatric pulmonary embolism through data sharing and remote collaboration mechanisms.

2. A multi-center pediatric pulmonary embolism online data platform system according to claim 1, characterized in that: Specific steps for in-depth analysis of massive data: Collect electronic medical records (EMR), medical images, laboratory results, genomic data and treatment plan data from various centers, and perform unified data formatting and standardization to ensure compatibility and consistency of data from different hospitals. It includes the data preprocessing stage: data cleaning, outlier processing, missing data completion and feature extraction, extracting characteristic variables that are crucial for pulmonary embolism risk prediction. After feature extraction, exploratory data analysis (EDA) is performed to preliminarily identify potential risk factors through statistical analysis, correlation calculation and visualization, and screen out variables closely related to pediatric pulmonary embolism. Perform feature selection and dimensionality reduction, select key features through Lasso regression and random forest algorithms, and use principal component analysis (PCA) technology to reduce dimensionality and reduce model complexity. Based on the preprocessed data, a variety of machine learning models are trained. For labeled data, classification algorithms (logistic regression, support vector machine, random forest) are used for risk prediction, or regression algorithms are used for prognosis analysis. For unlabeled data, the platform uses unsupervised learning algorithms (K-means clustering, DBSCAN) to identify patient subgroups and potential pathological patterns, analyzes imaging data through convolutional neural networks (CNN), and combines time series models (LSTM) to process changes in laboratory indicators to provide disease progression predictions. After the machine learning model is trained, the model performance is evaluated through multiple indicators such as accuracy, recall rate, and AUC. Hyperparameters are optimized through grid search or random search to ensure the stability of the model on different data sets. The model interpretation tools SHAP value and LIME are used to analyze the contribution of each feature to the model prediction, helping doctors understand potential risk factors and their weights. The causal relationship between features is analyzed through causal inference models. Personalized treatment decisions are supported through subgroup analysis and personalized prediction models. Predictions of different patients' responses to different treatment options are provided to guide doctors in formulating the best treatment plan.

3. A multi-center pediatric pulmonary embolism online data platform system according to claim 1, characterized in that: The steps of the dynamic risk assessment tool include automatically collecting patient information, electronic medical records, imaging data, laboratory test results and genomic information from multiple sources, updating and inputting data in real time, ensuring data accuracy and consistency through data standardization and cleaning, and then combining traditional statistical models and machine learning algorithms (random forests, deep neural networks) to accurately assess the patient's personalized risks. Based on the pre-trained risk scoring algorithm, the tool can generate disease risk assessment results for cardiovascular disease, diabetes, and pulmonary embolism, supporting doctors in judging patients' disease risks.

4. A multi-center pediatric pulmonary embolism online data platform system according to claim 1, characterized in that: Data security and privacy protection steps include data encryption and anonymization steps. By converting data into unreadable ciphertext, it ensures that even if the data is intercepted or leaked during transmission or storage, unauthorized persons cannot interpret it. The data encryption step adopts symmetric encryption, using the same key to encrypt and decrypt data, including AES (Advanced Encryption Standard). Asymmetric encryption** uses public key encryption and private key decryption. The algorithms are RSA and elliptic curve cryptography (ECC), which are used to protect sensitive information in network communications and ensure the security of data during network transmission. Transport layer encryption (TLS / SSL) establishes an encrypted communication channel for data transmission. Database encryption is used to encrypt and store patient data in the database. Transparent data encryption (TDE) automatically encrypts and decrypts stored data to ensure that sensitive information remains safe in the event of a data leak. The data anonymization step removes or blurs the patient's personal identity information to prevent the data from being directly associated with the individual. The de-identification step removes or hides the direct identifiers in the data, such as name and ID number. Pseudo-anonymization generates alternative identifiers (random numbers) instead of real identifiers. Data generalization further reduces the risk of individual identification by converting precise data into a more vague form, such as detailed dates of birth into age groups. Data perturbation protects personal privacy while maintaining the statistical integrity of the data by adding random noise or small changes to the data.

5. A multi-center pediatric pulmonary embolism online data platform system according to claim 2, characterized in that: In-depth analysis of massive data also includes real-time data updates and model retraining steps, which can continuously optimize and retrain models as new data is input to ensure the cutting-edge nature and accuracy of the analysis results. With the help of an adaptive learning mechanism, the platform automatically adjusts algorithm parameters to adapt to the different data characteristics of each center. It also includes multi-center collaboration steps, provides data sharing functions, promotes cross-hospital collaborative research, and helps researchers extract data, verify hypotheses, and conduct large-scale clinical research through scientific research support tools, supports long-term cohort follow-up analysis, and helps doctors track changes in patients' conditions in real time and evaluate treatment effects.

Images