Intelligent home detection data linkage management system for pregnant and lying-in women in large perinatal period

Abstract

The invention discloses a health data linkage management system for pregnant and lying-in women in a large perinatal period, and belongs to the technical field of intelligent data management. The system comprises a multi-source heterogeneous data access and standardization gateway which is used for connecting household equipment and a hospital system and standardizing original data through a deep cleaning engine; a data fusion and association unit of the central data processing server fuses multi-source data according to a time sequence by relying on a gestational week management unit, and a multi-modal health analysis engine is combined with a clinical rule and a machine learning model to carry out early warning, risk prediction and personalized analysis; and the multi-terminal interaction platform provides a fusion data view and a closed-loop management tool based on role permission. According to the invention, equipment and hospital information islands are broken through, and through high-quality data management and'rule + AI 'fusion analysis, accurate and foresight management and efficient cooperation of the full-cycle health state of pregnant and lying-in women are realized.

Description

Technical Field

[0001] This invention relates to the field of intelligent data management technology, specifically to a home-based intelligent monitoring and management system for pregnant women during the perinatal period. Background Technology

[0002] With the deepening development of the "Internet + Healthcare" model, home-based health management for pregnant and postpartum women is becoming an important part of perinatal care. Various home-use smart monitoring devices have emerged on the market, such as fetal heart rate monitors, blood glucose meters, blood pressure monitors, and scales, enabling pregnant and postpartum women to conveniently monitor their daily physiological indicators. However, existing technologies have the following significant problems in home-based health management for pregnant and postpartum women throughout the perinatal period (pre-conception, pregnancy, and postpartum): The problem of data silos in hospital information systems: Aside from home devices, core health data for pregnant and postpartum women originates from hospital information systems (HIS), laboratory information systems (LIS), and picture archiving and communication systems (PACS) at various levels of medical institutions. These systems, as well as with external health management platforms, suffer from severe issues of inconsistent interfaces and heterogeneous data formats (such as unstructured text reports and non-standardized coding). This makes it difficult for external systems to automatically and accurately acquire and understand key data from prenatal checkups, laboratory tests, and imaging reports within the hospital, creating another layer of "data silos" and hindering the construction of complete health records.

[0003] Device and data silos: Different brands and models of home monitoring devices use different proprietary communication protocols (such as Bluetooth and Wi-Fi) and data formats, lacking a unified data interface standard. This results in fragmented home monitoring data, which cannot be automatically and in real-time aggregated into a unified health management platform. Users need to manually enter the data, which is not only cumbersome but also prone to errors, making it difficult to form a continuous and complete personal health dataset.

[0004] Raw data quality and usability issues: Whether from hospital systems or home devices, raw data often contains noise, non-standard terminology, inconsistent units, or missing values. Directly using this data for in-depth analysis, especially AI-based analysis, can lead to decreased model accuracy or even erroneous conclusions. Existing solutions lack a unified, intelligent, and standardized data cleaning and processing flow, failing to provide a high-quality data foundation for advanced analytics.

[0005] The static and fragmented nature of maternal health records: Existing maternal health records are mostly limited to electronic records of in-hospital prenatal check-up data, or simply serve as a data repository, failing to effectively integrate multi-source information such as home self-testing data and online consultation records. More importantly, they lack the ability to intelligently correlate data based on the core temporal dimension of pregnancy—gestational week. The inability to automatically link data to specific gestational week points significantly diminishes the value of valuable continuous monitoring data, making it difficult for clinicians to intuitively grasp the dynamic and comprehensive picture of pregnancy development.

[0006] Limitations of analytical and early warning capabilities: Existing early warning systems are mostly based on simple threshold judgments, lacking the ability to deeply analyze complex physiological patterns, multi-indicator correlations, and long-term trends. Although artificial intelligence has shown potential in medical data analysis, its application in the perinatal period faces unique challenges such as high data quality requirements, strict model interpretability, and clear clinical responsibility definition. Therefore, a prudent and reliable intelligent analysis framework is needed that organically combines rule-driven and data-driven methods and is built upon high-quality data.

[0007] The problem of low efficiency in multi-party collaboration: Pregnant women, their families, community doctors, and hospital specialists need to work closely together in perinatal management. However, existing solutions lack a safe, efficient, and clearly defined data sharing and business collaboration mechanism. Poor information flow and difficulty in tracking the execution of medical orders prevent the formation of a closed-loop management system of "monitoring-assessment-intervention-feedback," hindering improvements in management efficiency and service quality.

[0008] Therefore, there is an urgent need for a comprehensive perinatal health management system that can break down equipment barriers, intelligently integrate multi-source data, and achieve dynamic early warning and closed-loop collaboration, so as to truly realize continuous and refined health management from the hospital to home. Summary of the Invention

[0009] To address the problems existing in the above-mentioned background technology, the present invention adopts the following technical solution: Firstly, this application provides a home-based intelligent monitoring and data linkage management system for pregnant women during the perinatal period, comprising: a multi-source heterogeneous data access and standardization gateway, used to connect various home-based intelligent monitoring devices and interface with data systems of different medical institutions, and to convert raw data from different protocols and formats into standardized health data with a unified structure; a central data processing server, communicatively connected to the multi-source heterogeneous data access and standardization gateway, the server comprising: a health record database, used to store the health data records of each pregnant woman throughout her entire life cycle; a data fusion and association unit, connected to the health record database, used to receive the standardized health data and associate and fuse it with the corresponding pregnant woman's records in the database, as well as a timeline based on gestational age; a multimodal health analysis engine, connected to the data fusion and association unit and the health record database, used to call preset clinical rules and / or machine learning models, combined with the pregnant woman's historical data, to perform anomaly judgment and risk assessment on the associated new data; and a multi-terminal interaction platform, communicatively connected to the central data processing server, used for providing interaction between pregnant women, doctor assistants, and attending physicians. Provide users with health data views, analysis results, and interactive functions that match their role permissions.

[0010] As a preferred embodiment of this application, the multi-source heterogeneous data access and standardization gateway includes: a protocol adaptation layer configured to support automatic identification and connection of Bluetooth and Wi-Fi communication protocols, and to support data interface docking with hospital information systems or offline data file reception; and a multi-source data parsing and cleaning engine configured to have built-in data parsing and standardization templates corresponding to different brands and models of testing equipment and different hospital data sources. The engine calls the corresponding template according to the received data source identification information, extracts data containing measurement indicators, values, units and measurement timestamps from the original data, and performs terminology standardization, value range normalization and logical consistency verification on the extracted data, and finally encapsulates it into a predefined unified data format message.

[0011] As a preferred embodiment of this application, the data fusion and association unit is configured to perform the following operations: receiving standardized health data containing a pregnant woman's identifier and a measurement timestamp; retrieving her health record and current gestational week information based on the pregnant woman's identifier; and classifying and storing the standardized health data according to the correspondence between the measurement timestamp and the current gestational week information for that gestational week. Centralized home monitoring data under intermediate nodes.

[0012] As a preferred embodiment of this application, the system further includes a gestational age management unit, which is configured to calculate and continuously maintain the current gestational age timeline of the pregnant woman based on the last menstrual period date or the date of a key prenatal check-up event input by the pregnant woman, and provide the timeline to the data fusion and association unit as a time series reference for data association.

[0013] As a preferred embodiment of this application, the multimodal health analysis engine includes: a rule base storing clinical threshold rules for physiological indicators related to gestational age ranges; and a logic judgment module used to compare the new data after data fusion and association unit processing with the thresholds corresponding to gestational age in the rule base. The system compares against established rules and / or retrieves historical data sequences of the indicator from the health record database for trend consistency comparison; the intelligent analysis module is equipped with a trained machine learning model for analyzing standardized historical data from the health record database. Real-time data sequences can be used for risk prediction, pattern recognition, or to generate personalized monitoring recommendations. When the output of the logic judgment module or intelligent analysis module meets the preset abnormal or risk conditions, an early warning or risk warning message is generated.

[0014] As a preferred embodiment of this application, the early warning or risk alert information generated by the multimodal health analysis engine is sent to the multi-terminal interaction platform and pushed in a differentiated manner to the pregnant women's terminal and the corresponding medical staff terminal according to the pre-configured subscription rules.

[0015] As a preferred embodiment of this application, the multi-terminal interaction platform includes: The access control module is used to define and manage the viewing, commenting, and operation permissions of different roles for health record data; The view generation module is used to dynamically organize the integrated multi-source data from the health record database according to the logged-in user's role and permissions, and combine it with the analysis conclusions generated by the multimodal health analysis engine to generate a customized interactive interface. Among them, the pregnant and postpartum view focuses on the intuitive display of personal health trends, personalized task reminders, and easy-to-understand health interpretations; the medical staff view focuses on the integrated display of multi-source data comparison, clinical decision support information, and patient management tools.

[0016] As a preferred embodiment of this application, the healthcare worker view provides an interface for issuing medical orders and planning clinical pathways, and the issued medical orders or plans are bound to the health records of specific pregnant women; the pregnant woman view provides an interface for confirming medical orders, providing execution feedback, and reporting symptom self-checks, and the resulting feedback information is sent back and updated to the health record database, forming a digital management closed loop.

[0017] In a second aspect, this application also provides an electronic device, including a processor, a memory, and a computer program stored in the memory, wherein the processor executes the computer program to implement the module functions of the system as described in the first aspect.

[0018] Thirdly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the module functions of the system as described in the first aspect.

[0019] Compared with the prior art, the present invention has the following beneficial effects: It achieves unified access and deep cleaning of multi-source heterogeneous data, breaking down data silos across the entire domain: Through innovative multi-source heterogeneous data access and standardized gateways, it is not only compatible with various home devices but also connects to heterogeneous information systems from different hospitals. Its built-in multi-source data parsing and cleaning engine can perform deep cleaning operations such as terminology standardization and value range normalization on unstructured reports and non-standard interface data, transforming data from diverse sources and of varying quality into high-quality structured data that can be directly used by the system. This solves the dual challenges of "device silos" and "hospital system silos."

[0020] A dynamic, structured health record centered on gestational week has been constructed to enhance the clinical value of the data. The system provides precise time-series benchmarks through gestational week management units, driving data fusion and correlation units to automatically and accurately link each piece of home monitoring data to a specific gestational week node in the pregnant woman's personal record. This gestational week-based structured storage method activates discrete data points into a continuous life map with clear clinical significance, greatly facilitating data retrieval, trend analysis, and comprehensive evaluation by gestational week stage, providing strong data support for clinical decision-making.

[0021] A multimodal intelligent analysis framework integrating rules and AI has been constructed to enhance the depth and foresight of risk insights: The system pioneered a multimodal health analysis engine that integrates a "rule base + logical judgment module" and a "machine learning model + intelligent analysis module." While ensuring strict adherence to clinical rules, an AI model trained on high-quality data is introduced to perform complex analytical tasks such as risk prediction and abnormal pattern recognition. This "rule-based, AI-enhanced" architecture leverages AI's advantages in handling nonlinear and high-dimensional relationships while ensuring the reliability and interpretability of the analysis results through rules, enabling earlier, more accurate, and more personalized risk insights into pregnancy complications.

[0022] It provides a high-quality data foundation for advanced analytics, unlocking the potential of data: the system's rigorous data cleaning and standardization processes at the front end ensure the quality of data imported into the health record database. This provides an indispensable, clean, and consistent data foundation for backend applications of advanced data analytics techniques such as machine learning, enabling data-driven intelligent analysis and significantly improving its accuracy and reliability, truly releasing the potential value of multi-source health data.

[0023] A clear multi-role collaborative closed loop with well-defined responsibilities has been established, improving management efficiency and user experience. Through the permission management and view generation modules of a multi-terminal interactive platform, customized and differentiated data views and operation interfaces are provided for different roles such as pregnant women, physician assistants, and attending physicians. Online medical order issuance, execution confirmation, and feedback are supported, fully recording and tracking the entire management process. This constructs a safe, efficient, and transparent multi-party collaborative workflow, forming a digital management closed loop of "monitoring-early warning-intervention-feedback," significantly improving the efficiency of doctor-patient communication and enhancing pregnant women's sense of participation and compliance. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a schematic diagram of the intelligent home monitoring and data linkage management system for pregnant women during the perinatal period, according to Embodiment 1 of the present invention. Figure 2 This is a schematic diagram of the structure of the multi-source heterogeneous data access and standardized gateway in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the operation flow of the data fusion and association unit in Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of the operation flow of the multimodal health analysis engine in Embodiment 1 of the present invention; Figure 5 This is a schematic diagram of the structure of the intelligent home monitoring and data linkage management system for pregnant women during the perinatal period, which is a second embodiment of the present invention. Detailed Implementation

[0026] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention. It should be noted that relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0027] Example 1 like Figure 1-4 As shown, this application provides a home-based intelligent monitoring data linkage management system for pregnant women during the perinatal period, including: a multi-source heterogeneous data access and standardized gateway, used to connect various home-based intelligent monitoring devices and interface with the data systems of different medical institutions, and to convert raw data from different protocols and formats into standardized health data with a unified structure; multi-protocol access refers to the simultaneous access of devices that support multiple communication protocols (such as Bluetooth, Wi-Fi, Zigbee, infrared, etc.), solving the "island" problem of different brand devices. A central data processing server, communicatively connected to the multi-source heterogeneous data access and standardization gateway, includes: a health record database for storing the entire health data records of each pregnant woman throughout her pregnancy; a data fusion and association unit connected to the health record database for receiving standardized health data and associating and fusing it with the corresponding pregnant woman's records and a gestational age-based timeline in the database; a multimodal health analysis engine connected to the data fusion and association unit and the health record database for calling pre-set clinical rules and / or machine learning models, combining the pregnant woman's historical data, and performing anomaly detection and risk assessment on the newly associated data; and a multi-terminal interaction platform, communicatively connected to the central data processing server, for providing interaction between pregnant women, physician assistants, and attending physicians. Provide users with health data views, analysis results, and interactive functions that match their role permissions.

[0028] Among them, such as Figure 2As shown, the multi-source heterogeneous data access and standardization gateway includes: a protocol adaptation layer, configured to support automatic identification and connection of communication protocols such as Bluetooth and Wi-Fi, while providing standardized data interfaces (such as HL7FHIR, Web Service API) and secure file transfer channels for receiving batch data files or pushing data in real time from the HIS, LIS, and other systems of cooperating hospitals; the protocol adaptation layer encapsulates the underlying connection details (such as device discovery, pairing, connection establishment, and data transmission) of different communication protocols such as Bluetooth (e.g., BLE) and Wi-Fi. It provides a unified "connection management" interface, so that upper-layer applications do not need to care about the physical protocol currently being used, which is the key to achieving "plug and play". A multi-source data parsing and cleaning engine, configured with a built-in data parsing and standardization template library corresponding to different brands and models of testing equipment and different hospital data sources. The engine calls the corresponding template according to the received data source identification information (such as device MAC address, hospital code + report type). One template corresponds to one device protocol. A unified data format message refers to a standard data structure defined internally by the system, such as a data packet in JSON or XML format (containing fields such as device ID, user ID, indicator type, value, unit, and timestamp). For home device data, the template guides the completion of basic parsing; for raw hospital data (such as PDF format prenatal check-up reports and text format test results), the template guides the engine to perform optical character recognition (OCR), key information extraction, medical terminology mapping (such as unifying "BP" as "blood pressure" and mapping "fasting blood glucose" to standard LOINC encoding), and performs unit unification (such as converting mg / dL to mmol / L) and logical verification (such as systolic blood pressure should be greater than diastolic blood pressure).

[0029] Finally, the engine encapsulates the cleaned data into a predefined, uniform data format message.

[0030] In this embodiment, the specific operating principle and process of multi-source heterogeneous data access and standardized gateway includes the following parallel and serial steps: A. Data access process for home-use smart testing devices: S1-A, Device Discovery and Connection: After the gateway starts, the device communication module of the protocol adaptation layer continuously scans for Bluetooth and Wi-Fi devices in the surrounding environment. When a pregnant woman turns on devices such as a smart blood pressure monitor (Bluetooth) or a blood glucose meter (Wi-Fi) at home, this module automatically identifies the device type and establishes a connection. Taking a blood pressure monitor as an example, when the pregnant woman opens the mobile app (which serves as one of the gateway's carriers) and the smart blood pressure monitor at home, the protocol adaptation layer initiates a Bluetooth scan, discovers the signal broadcast by the blood pressure monitor, and automatically completes pairing and connection establishment.

[0031] S2-A, Data Reception and Routing: After the blood pressure monitor completes the measurement, it sends a line of raw data (e.g., "BP:120 / 80,65,202310270830") via Bluetooth. The protocol adaptation layer receives this data and, along with identification information such as the device's MAC address, passes it to the multi-source data parsing and cleaning engine.

[0032] S3-A, Device Data Parsing and Initial Cleaning: The multi-source data parsing and cleaning engine calls the corresponding device data parsing template based on the device identification information (e.g., matching the MAC address to "a certain brand of type B blood pressure monitor"). This template tells the engine: in this data string, the first two numbers are the systolic and diastolic blood pressure, the third is the heart rate, and the last is the timestamp. Based on this, the engine parses the following: Systolic blood pressure: 120 mmHg, Diastolic blood pressure: 80 mmHg, Heart rate: 65 beats / minute, Measurement timestamp: 2023-10-27T08:30:00.

[0033] B. Data access process for medical institution data systems: S1-B, Hospital Data Reception: The medical data interface module of the protocol adaptation layer receives batch data files (such as JSON / XML data packets) or original documents (such as PDF format prenatal examination reports and text format test result lists) from the HIS, LIS and other systems of cooperating hospitals through standardized data interfaces (such as HL7 FHIR API) or secure file transfer channels (SFTP) in accordance with predetermined cycles or in real time.

[0034] S2-B, Hospital-Level Data Parsing and Deep Cleaning: The multi-source data parsing and cleaning engine calls the corresponding hospital data parsing and standardization template based on the data source identification information (such as "Hospital A_LIS_Blood Routine Report"). This template guides the engine to perform the following operations: Structured extraction: For unstructured PDF reports, trigger optical character recognition (OCR) and extract key fields (such as "white blood cell count: 8.5 ×10^9 / L"); for semi-structured text or non-standard interface data, locate data items using regular expressions or parsing rules.

[0035] Terminology standardization: Extracted non-standard descriptions are mapped to standard terms within the system. For example, "WBC", "white blood cells", and "white blood cell count" are uniformly mapped to the standard indicator code "LAB_LEUCOCYTE"; "BP" is mapped to "blood pressure".

[0036] Value range normalization: unifying numerical values ​​to standard units. For example, converting blood glucose values ​​from "mg / dL" to "mmol / L"; converting gestational week descriptions "28W+3D" to structured data "{"week": 28, "day": 3}".

[0037] Logical verification: Perform common-sense or medical logical verification on the data. For example, verify that "systolic blood pressure > diastolic blood pressure" or "whether the test result is within a reasonable range", and mark outliers or trigger a manual review process.

[0038] C. Unified encapsulation and forwarding process: S4. Data Fusion and Encapsulation: After the data has been processed by steps S3-A or S3-B, the multi-source data parsing and cleaning engine encapsulates all fields (indicator type, standard value, unit, standard timestamp, data source, etc.) according to a predefined unified data format message (such as a specific JSON Schema) to generate a standardized health data object with consistent structure and clear semantics.

[0039] S5. Secure Forwarding: The gateway securely uploads this standardized health data to the data fusion and association unit of the central data processing server via the HTTPS protocol.

[0040] Among them, the central data processing server includes: The health record database, as a structured data storage system (such as a database or data lake warehouse supporting a hybrid model), creates a master file for each pregnant woman, centered on a unique identifier (such as a patient ID). This file not only contains static information (such as name, age, parity, and due date), but more importantly, it is designed with a dynamic data structure centered on gestational week. For example, the database organization can be abstracted as: Patient ID (P123456) -> Gestational Week 24 -> { Data Source: { “Home Monitoring”: [Blood Pressure Data Points..., Blood Glucose Data Points...], “Hospital Prenatal Checkups”: [Ultrasound Reports..., Blood Routine Reports...], “Health Records”: [Consultation Records..., Symptom Logs...]}}. This multi-level index structure of “Patient-Gestational Week-Data Source-Indicator Type” enables the system to efficiently aggregate heterogeneous data from different sources (home, hospital) according to clinical stage (gestational week), providing underlying support for comprehensive analysis.

[0041] like Figure 3 As shown, the data fusion and association unit is configured to perform the following operations: 1. Receiving and Context Identification: Receiving standardized health data from a multi-source heterogeneous data access and standardization gateway. This data packet includes not only the pregnant woman identifier (patient_id), measurement timestamp, and indicator values, but also key data source identifiers (such as source: "home_device_bp_monitor" or source: "hospital_A_LIS") and cleaned standard medical terminology codes. 2. File Location and Timeline Space Alignment, including: Record location: Based on patient_id (e.g., P123456), accurately locate the root node of the pregnant woman's record in the health record database.

[0042] Timeline Alignment: Call the gestational week management service (a standalone or embedded microservice), passing in the patient_id and the measurement timestamp (e.g., 2023-10-27T08:30:00). Based on the authoritative gestational week timeline maintained by the pregnant woman, the service returns the precise gestational week information corresponding to this time point (e.g., {"week": 24, "day": 3}).

[0043] This step is crucial for unifying and aligning data generated at different times and locations to the core clinical timeline of "gestational week." 3. Multi-source data fusion and structured storage: Based on the calculated gestational week (e.g., "24 weeks"), data source identifiers (e.g., "hospital prenatal checkup"), and indicator type (e.g., "complete blood count - white blood cell count"), the system locates or creates corresponding data storage containers under the corresponding gestational week node in the archive. Standardized data points (e.g., {indicator: "white blood cell count", value: 8.5, unit: "10^9 / L", source: "hospital A_LIS", time: ...}) are stored as a new record in this container. For example, complete blood count results from the hospital's LIS and blood pressure data from a home device within the same gestational week are stored side-by-side under the "gestational week 24" archive branch. This process achieves spatiotemporal fusion of multi-source data, integrating discrete data points from home devices and hospital systems into a continuous health record organized by gestational week with a unified view.

[0044] The data fusion and association unit is the system's data hub. It relies on high-quality, standardized data provided by the upstream gateway and utilizes the time-series benchmark provided by the gestational age management service to execute core spatiotemporal alignment and fusion logic. Its output is a structured, time-series-clear "digital twin" health record that integrates multi-source data by gestational age. This high-quality fused record is the sole reliable data source for subsequent multimodal health analysis engines to perform reliable rule judgments and intelligent model inferences. It also forms the data foundation for the multi-terminal interactive platform to generate visualized health trend views by gestational age and data source.

[0045] The multimodal health analysis engine includes: a rule base storing clinical threshold rules for physiological indicators related to gestational age range; such as those in the "Guidelines for the Diagnosis and Treatment of Hypertension in Pregnancy". A predetermined blood pressure threshold. A logic judgment module is used to perform rapid judgments based on explicit rules; The intelligent analysis module is loaded with machine learning models (such as time-series prediction models and classification models) trained on historical health data. These models can analyze the combined changes of multiple indicators across gestational weeks in pregnant women to perform one or more of the following auxiliary tasks: 1) predicting the probability of an indicator exceeding a threshold within the next 1-2 weeks of gestation; 2) identifying specific blood pressure and urinary protein combinations associated with the development of gestational hypertension; 3) dynamically suggesting more personalized weight gain warning lines based on the pregnant woman's individual characteristics and historical data. The output of the intelligent analysis module serves as a risk warning and is submitted together with the warning conclusions from the logical judgment module for comprehensive reference by medical staff.

[0046] In the early warning generation step, the deterministic early warning from the integrated logic judgment module and the risk probability prompt from the intelligent analysis module generate a health report containing different confidence levels.

[0047] like Figure 4 As shown, the operating principle and process of the multimodal health analysis engine include: 1. Data triggering and context loading: Trigger: When the data fusion and association unit successfully archives and integrates one or more new standardized health data into the maternal health record, it will send an analysis trigger event to the multimodal health analysis engine.

[0048] Context loading: Upon receiving the event, the engine immediately loads the complete analysis context from the health record database based on the pregnant woman's identifier and gestational age information. This includes: New data that triggered this analysis (e.g., fasting blood glucose level 6.2 mmol / L, source: home blood glucose meter).

[0049] The authoritative gestational age information for the pregnant woman (e.g., 28 weeks + 3 days).

[0050] The archive contains historical multi-source data sequences that have been aligned and cleaned by gestational week (e.g., historical values ​​of blood sugar, blood pressure, and weight over the past 8 weeks, which may have come from multiple home measurements and hospital test reports).

[0051] The pregnant woman's static characteristics (such as age, pre-pregnancy BMI, history of gestational diabetes, etc.).

[0052] 2. Parallel multimodal analysis: The engine starts with a logic judgment module and an intelligent analysis module, which perform parallel analysis based on the loaded context data: A. Rule-based logical judgment (fast path): Rule matching: The logic judgment module retrieves and applies the corresponding clinical threshold rules from the rule base based on the current gestational week (28 weeks). For example, the rule is applied: "For 24-28 weeks of gestation, fasting blood glucose ≥5.3 mmol / L is abnormal."

[0053] Threshold and trend comparison: The module performs a dual judgment. Threshold comparison: The new data (6.2 mmol / L) was compared with the rule threshold (5.3 mmol / L) to confirm that the "absolute value exceeded the standard".

[0054] Trend comparison: Retrieve the pregnant woman's merged historical blood glucose sequence (e.g., [4.9, 5.0, 5.2, 5.8, 6.2]) and perform simple statistical trend analysis (such as calculating moving average and slope of change) to confirm the "rapid upward trend".

[0055] Output: Generate a definitive judgment conclusion (e.g., "Fasting blood glucose is elevated and is showing an upward trend").

[0056] B. Model-based intelligent analysis (deep path): Model Inference: The intelligent analysis module loads multi-dimensional, cross-gestational week, and multi-source fusion data sequences and inputs them into a pre-trained machine learning model (e.g., a time-series classification model for predicting the risk of blood sugar runaway).

[0057] Pattern recognition and prediction: The model analyzes joint patterns in the data and performs one or more of the following tasks: Risk assessment: Predict the probability that the pregnant woman's blood sugar will remain high or develop into a condition requiring insulin intervention within the next 1-2 weeks (e.g., high risk probability 85%).

[0058] Pattern detection: Identified a potential correlation between this blood sugar elevation and recent rapid weight gain or blood pressure fluctuations.

[0059] Personalized insights: Based on the pregnant woman's individual characteristics and historical responses, personalized suggestions are dynamically generated, such as "It is recommended to control the average daily carbohydrate intake to below XXX grams in the coming week."

[0060] Output: Generate risk probability alerts and decision support insights (e.g., "Risk of poor blood sugar control in the next two weeks: high (85%)", "Association pattern: Recent weekly weight gain exceeds recommended value").

[0061] 3. Conclusion Integration and Report Generation: Comprehensive analysis: The engine's core scheduler receives deterministic conclusions from the logic judgment module and probabilistic hints and insights from the intelligent analysis module.

[0062] Generate a tiered report: Based on a predefined fusion strategy, generate a structured health analysis report. The report clearly distinguishes between: Alert: Based on a clear rule violation that requires immediate attention (e.g., "Fasting blood glucose is too high, triggering a gestational diabetes monitoring alert").

[0063] Risk Indication: Probabilistic risk predictions based on AI model output (e.g., "Future risk of blood sugar out of control: high", with confidence level).

[0064] Advisory recommendations: Personalized management suggestions that combine rules and model outputs (e.g., "1. It is recommended to retest fasting blood glucose and glycated hemoglobin within 3 days; 2. It is recommended to schedule an appointment with the nutrition department; 3. Monitor and control weekly weight gain.").

[0065] Supporting evidence: The data includes trend curves from recent weeks and key historical data points used for analysis.

[0066] 4. Results Distribution and Driving: The generated health analysis report (including warnings, risk alerts, and recommendations) is immediately pushed to the multi-terminal interactive platform.

[0067] The platform pushes information in a differentiated and tiered manner based on the configuration of the permission management module: For pregnant and postpartum women: They may receive simplified reminders and advice (such as: "Your blood sugar level is high. Please pay attention to your diet and schedule a follow-up appointment as soon as possible").

[0068] Physician / assistant: Receive a professional view containing complete reports, data evidence, and AI analysis insights to support rapid clinical decision-making.

[0069] The multimodal health analysis engine serves as the intelligent decision-making center of the entire system. It relies heavily on high-quality, integrated time-series archives provided by data fusion and correlation units as its "fuel." Through an analysis paradigm of "rule-based baseline and AI enhancement," it not only achieves rapid response to immediate anomalies but also possesses the ability for forward-looking risk prediction and personalized insights. Its output is no longer a simple warning but rather tiered, evidence-based, and actionable decision support information that drives precise intervention, achieving a leap from passive monitoring to proactive, forward-looking health management.

[0070] The multi-terminal interaction platform includes: The access control module is used to define and manage the viewing, commenting, and operation permissions of different roles for health record data; The view generation module dynamically organizes integrated multi-source data from the health record database based on the logged-in user's role and permissions. It then combines this data with analysis conclusions generated by a multimodal health analysis engine to create a customized interactive interface. The pregnant / postpartum view focuses on intuitively displaying individual health trends, personalized task reminders, and easy-to-understand health interpretations. The healthcare worker view emphasizes the integrated display of multi-source data comparisons (such as side-by-side comparison of home self-test data and hospital test reports), clinical decision support information (including rule-based warnings and AI risk alerts), and efficient patient management tools. The healthcare worker view provides interfaces for issuing medical orders and planning clinical pathways, with issued orders or plans linked to the health records of specific pregnant / postpartum women. The pregnant / postpartum view provides interfaces for order confirmation, execution feedback, and symptom self-reporting. The resulting feedback information is transmitted back to and updated in the health record database, forming a closed-loop digital management system.

[0071] The operating principle and process of the multi-terminal interaction platform include: 1. User Login and Dynamic Authentication: The system predefines roles (e.g., pregnant women, family members, physician assistants, attending physicians, nutritionists, etc.) and fine-grained permissions (e.g., viewing scope, operable data types, and executable clinical actions). When a user logs in, the permission management module verifies their identity and role, and dynamically loads the permission set associated with that role and their affiliated institution / team.

[0072] Example: When physician assistant Xiao Li logs in, the system determines her role as "Obstetric Physician Assistant of a certain hospital" and loads her permissions: she can view the integrated health records of all pregnant women registered in the department, she can enter temporary vital signs, and she can process warning messages, but she cannot issue final diagnostic medical orders.

[0073] 2. Intelligent view generation and data organization: The view generation module initiates combined data requests to the data fusion and association unit and the multimodal health analysis engine based on the current user's role and permissions.

[0074] Example (continuing from Xiao Li's example): Xiao Li views the pregnant woman's pages (P123456). View generation module: The data fusion and association unit requests a dataset of data for the pregnant woman at the current and adjacent gestational weeks, which integrates home blood pressure / blood glucose and hospital test reports.

[0075] Request the latest analysis reports related to the pregnant woman from the multimodal health analysis engine (e.g., an "alert" about blood sugar and a "medium risk warning" about the risk of preeclampsia).

[0076] Based on the above information, a "Clinical Workbench View" is generated for Xiao Li: the top displays warning and risk alert cards to be processed; the left side is a timeline of multi-source data organized by gestational week, which can be switched with one click to view home data or hospital report details; the right side is a trend panel of the pregnant woman's core indicators.

[0077] 3. Interaction, decision-making, and closed-loop driven: Users interact within the generated view.

[0078] Example (deep closed loop): In the "Clinical Workbench View", Xiao Li saw a high-confidence warning: "Pregnant woman P123456, 28 weeks pregnant, fasting blood glucose continues to exceed the standard, triggering GDM management warning", and an AI risk warning: "Risk of blood glucose out of control in the next two weeks: high (80%), related factor: recent accelerated weight gain".

[0079] After marking the case, Xiao Li handed it over to the attending physician, Director Zhang. In his "Expert Decision View," Director Zhang could not only see the above information, but also access multi-indicator joint analysis charts over a longer time span, as well as summaries of auxiliary diagnosis and treatment suggestions based on similar cases.

[0080] Director Zhang issued a structured medical order through the integrated medical order issuance interface: "1. Repeat fasting blood glucose and glycated hemoglobin test tomorrow; 2. Schedule a nutrition clinic appointment for dietary assessment; 3. Increase home blood glucose monitoring frequency to 4 times daily." This medical order was stored in a structured manner by the system and automatically associated with the file P123456 and the alert event that triggered this medical order.

[0081] Pregnant woman P123456 immediately received a clear push notification in the "My Health Center" view of her mobile app: "Your doctor has created a new management plan for you, please check it." After clicking, she saw a simplified version of the medical instructions, a link to schedule a follow-up appointment, and an interface for recording feedback. She confirmed that she understood and made an appointment for the outpatient visit.

[0082] After a follow-up visit, new blood glucose data (potentially from the hospital's LIS or a home blood glucose meter) re-enters the system through multi-source heterogeneous data access and a standardized gateway, triggering a new round of analysis and evaluation. The status of medical orders (appointed, tested) is also automatically updated in the record, forming a complete digital closed loop of "monitoring -> analysis -> early warning -> clinical decision -> patient execution -> remonitoring".

[0083] Furthermore, to better understand the operating principle of this system, we will use the scenario of a pregnant woman, Ms. Li (28 weeks pregnant), using a smart fetal heart monitor that integrates hospital data as an example to illustrate the overall operation process of this system, including the following steps: First, multi-source data collection and fusion are performed: Home Data: Ms. Li used a smart fetal heart rate monitor and measured a fetal heart rate of 155 beats per minute. The data was processed through multi-source heterogeneous data access and a standardized gateway to generate a standardized message {Indicator: Fetal Heart Rate, Value: 155, Unit: bpm, Time: T1, Source: Home Device}.

[0084] Hospital Data: On the same day, Ms. Li's prenatal checkup fetal heart rate monitoring (NST) results were pushed to the gateway by the hospital system. The gateway's multi-source data parsing and cleaning engine parsed the PDF report, extracted the key conclusion "NST reactive, baseline fetal heart rate 150 bpm", and converted it into a standard message {Indicator: Fetal Heart Rate Monitoring Conclusion, Value: Reactive, Detailed Data: {...}, Time: T1, Source: Hospital A_PACS}.

[0085] Next, data fusion and association storage are performed: two standardized data sets are uploaded to the central data processing server. The data fusion and association unit locates Ms. Li's file based on her "Ms. Li ID" and calls the gestational age management unit to confirm that time T1 corresponds to "28 weeks + 1 day". Subsequently, the raw home fetal heart rate data points and the structured hospital fetal heart rate monitoring report are stored side-by-side in the "Gynecology 28 -> Fetal Heart Rate Monitoring" dataset in Ms. Li's file, with their respective sources marked.

[0086] Secondly, perform multimodal health analysis: the multimodal health analysis engine is triggered.

[0087] Logical judgment module: Calls the rule base to determine that the home fetal heart rate of 155 bpm and the baseline of 150 bpm reported by the hospital are both within the normal range (110-160 bpm), and the hospital's conclusion is "reactive" with no immediate abnormalities.

[0088] Intelligent Analysis Module: Loaded Ms. Li's fetal heart rate sequences from the past few weeks (integrating multiple home measurement points and previous hospital NST results), and analyzed their long-term variation trends using a model, finding no abnormal patterns. Engine Overall Judgment: Current status is normal, no warnings.

[0089] Finally, it enables collaborative viewing and value presentation by multiple roles: Ms. Li saw the notification "Today's fetal heart rate monitoring completed" on the app, and at the same time (28 weeks + 1 day), she could simultaneously view two pieces of information: "Self-monitored fetal heart rate: 155 beats / min" and "Hospital fetal monitoring result: normal", which gave her peace of mind.

[0090] Her attending physician, at the workstation, was reviewing Ms. Li's health records for 28 weeks when the system presented a fusion view: a timeline simultaneously displaying daily home fetal heart rate measurements (scattered trend line) and hospital NST examination nodes (marked with conclusions such as "reactive"). The doctor could clearly see the continuity and consistency of the monitoring data from inside and outside the hospital, greatly improving assessment efficiency and data utilization.

[0091] This example demonstrates how the system can deeply integrate, uniformly analyze, and collaboratively present data from fragmented home monitoring and discrete hospital examinations, thereby providing strong support for continuous health management.

[0092] This embodiment provides a perinatal maternal health data linkage management system. By constructing an integrated technical architecture of "multi-source heterogeneous data access and standardized gateway - central data processing server (including data fusion and association unit, multimodal health analysis engine) - multi-terminal interaction platform", it realizes intelligent governance, fusion analysis and collaborative management of maternal health data from multiple sources throughout the entire life cycle.

[0093] The core innovations and technological achievements of the system are reflected in: It achieves unified access and high-quality governance of comprehensive health data: Through multi-source heterogeneous data access and standardized gateways, the system is not only compatible with various home-use smart testing devices, but also capable of connecting to and parsing data from heterogeneous information systems of different medical institutions (such as HIS, LIS, and PACS). Its built-in deep cleaning engine, through technologies such as optical character recognition (OCR), medical terminology mapping, and value range normalization, transforms raw data from diverse sources and in different formats into high-quality, standardized health data, fundamentally breaking down the dual barriers of "device silos" and "hospital system silos."

[0094] A multi-source data fusion archive based on gestational week was constructed: the data fusion and association unit in the central data processing server, with clinical gestational week as the core timeline, accurately aligns and merges standardized data from home monitoring and hospital treatment, creating a continuous, unified, and traceable "digital twin" health record. This provides a unique and reliable data foundation for continuous health assessment across stages and scenarios.

[0095] A multimodal intelligent analysis paradigm integrating rules and AI has been established: the system's pioneering multimodal health analysis engine integrates a logical judgment module based on explicit clinical rules with an intelligent analysis module based on machine learning. It can not only perform rapid anomaly threshold and trend judgment, but also analyze multi-indicator, cross-time-series fused data through models to achieve early risk prediction, complex pattern recognition, and personalized management suggestion generation. This "rule-based, AI-enhanced" analysis model significantly improves the foresight, accuracy, and personalization of health risk assessment.

[0096] A data-driven, collaborative, closed-loop management ecosystem has been established: the multi-terminal interactive platform, based on fine-grained access control, provides customized views that deeply integrate multi-source data and intelligent analysis conclusions for different roles such as pregnant women, medical staff, and others. The platform supports the entire online process from intelligent early warning / risk alerts to structured medical orders and patient execution feedback. Ultimately, new data feedback triggers re-analysis, forming a digital management closed loop of "multi-source monitoring → intelligent analysis → collaborative decision-making → patient execution → effect re-evaluation".

[0097] In summary, this system, through technological innovation, expands the scope of health management from simple home monitoring to the integrated utilization of health data across both in-hospital and out-of-hospital environments; upgrades analytical capabilities from rule-based passive early warning to proactive prediction and insight combining rules and AI; and evolves the management model from fragmented recording to data-driven continuous closed-loop collaboration. This system significantly enhances the scientific rigor, accuracy, collaboration, and forward-looking nature of perinatal health management, providing strong technical support for reducing risks to mothers and fetuses, optimizing the allocation of medical resources, and innovating health service models. It has broad clinical application prospects and social value.

[0098] Example 2 Building upon Example 1, this example further introduces a gestational age management unit, serving as the core time-series engine of the entire perinatal maternal health data linkage management system, emphasizing its independence and central role. This unit is not a simple date calculator, but a microservice providing the system with a unified, accurate, and calibrable clinical time baseline. Its synergy with the other core modules described in Example 1 is crucial for the system to achieve accurate data fusion and scientific analysis. Specifically, as... Figure 5 As shown, the system also includes: The gestational age management unit, as an independent service module, is configured to calculate and continuously maintain the pregnant woman's current gestational age timeline based on the last menstrual period date or key prenatal checkup dates entered by the woman. More importantly, it provides this dynamic and authoritative timeline to the data fusion and association unit through a standardized application programming interface (API) as the temporal benchmark for all data associations, while also providing the necessary gestational age context for rule matching and trend analysis by the multimodal health analysis engine.

[0099] The gestational age management unit interacts closely with the core module described in Example 1, and their technical synergy is specifically reflected in the following aspects: 1. Collaboration with the Data Fusion and Association Unit (Providing a Fusion Benchmark): The gestational age management unit is the core service upon which the data fusion and association unit relies for spatiotemporal alignment operations. When the data fusion and association unit receives standardized health data from the gateway, it must immediately invoke the gestational age management unit's API, passing in the pregnant woman's identifier and measurement timestamp, to obtain the precise gestational age information corresponding to that time point (e.g., "{"week": 28, "day": 3}"). Only with this authoritative gestational age can the data fusion and association unit accurately categorize data from different sources (home devices, Hospital A, Hospital B) into the corresponding dataset under the same gestational age, achieving true multi-source data spatiotemporal fusion. Without accurate gestational age, fusion loses its clinical significance.

[0100] 2. Collaboration with the Multimodal Health Analysis Engine (Providing Analysis Context): The gestational age management unit provides essential analysis dimension parameters for the logical judgment module and intelligent analysis module of the multimodal health analysis engine.

[0101] For the logic judgment module: This module must retrieve the corresponding clinical threshold rule from the rule base based on the current gestational week. For example, the abnormal blood pressure thresholds are different at 28 weeks and 32 weeks of gestation. The precise gestational week provided by the gestational week management unit ensures the dynamic accuracy and gestational week specificity of clinical rule application.

[0102] For the intelligent analytics module: gestational age is a core temporal feature for machine learning models when performing risk prediction or pattern recognition. The model needs to know the gestational stage of the data points to understand the normal trajectory of physiological changes. The continuous timeline provided by the gestational age management unit is the foundation for the model to conduct meaningful time series analysis and cross-gestational age pattern mining.

[0103] Example: When analyzing a pregnant woman's blood pressure trend, if the gestational age is calculated incorrectly (e.g., 28 weeks is mistakenly judged as 24 weeks), the system may incorrectly apply a more lenient threshold, leading to missed warnings of preeclampsia risk; at the same time, the AI ​​model may also draw biased predictions based on incorrect time-series features.

[0104] 3. Collaboration with the Health Record Database (Defining Storage Structure): The output of the gestational week management unit directly defines the core index structure and organizational logic of the health record database. The database adopts a multi-level index structure of "patient-gestational week-data source". This "gestational week" label is assigned by the data fusion and association unit based on the calculation results of the gestational week management unit when writing data. This makes efficient querying, aggregation, and backtracking by gestational week possible, laying the foundation for efficient data retrieval and utilization throughout the system.

[0105] 4. Collaboration with Multi-Terminal Interactive Platforms (Unified Display Axis): Ensuring all user interfaces are based on the same time reference system for synchronized understanding. Whether it's the trend chart on the doctor's workstation with "gestational week" as the X-axis, or the task reminders triggered by gestational week in the pregnancy and postpartum app (such as "24-28 weeks of pregnancy: glucose tolerance screening"), the time reference is uniformly derived from the gestational week management unit. This eliminates collaboration confusion caused by different time calculation methods and ensures consistency in doctor-patient communication and health guidance.

[0106] The gestational age management unit emphasized in this embodiment decouples, modularizes, and service-oriented the core clinical concept of gestational age from implicit business logic. As the system's "temporal hub," it transforms biological time into machine-manageable data services, permeating the entire chain of data fusion, intelligent analysis, storage, retrieval, and interactive display. This not only enhances the system's scientific rigor and accuracy but also serves as an indispensable technological infrastructure for achieving personalized and precise perinatal health management, transforming the system from a data processor into an intelligent partner that understands clinical temporal patterns.

[0107] Example 3 The present invention also provides an electronic device, including: a processor, a transmitting device, an input device, an output device, and a memory. The processor may be implemented using a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit, or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory may be implemented using a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), and is used to store computer program code. The computer program code includes computer instructions. When the processor executes the computer instructions, the electronic device performs the module functions as described above in any of the possible implementation methods.

[0108] Example 4 The present invention also provides a computer-readable storage medium storing a computer program, the computer program including program instructions, which, when executed by a processor of an electronic device, cause the processor to perform module functions as described in any of the above possible implementations.

[0109] In the description of this specification, the references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0110] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A home-based intelligent monitoring and data linkage management system for pregnant women during the perinatal period, characterized in that, include: Multi-source heterogeneous data access and standardized gateway is used to connect various home smart detection devices and interface with the data systems of different medical institutions, and to convert raw data from different protocols and formats into standardized health data with a unified structure; The central data processing server is communicatively connected to the multi-source heterogeneous data access and standardization gateway. The server includes a health record database, which stores the health data records of pregnant women throughout their entire life cycle, on a per-pregnancy basis. A data fusion and association unit, connected to the health record database, receives the standardized health data and associates and fuses it with the corresponding pregnant woman's records and a gestational age-based timeline in the database. A multimodal health analysis engine, connected to the data fusion and association unit and the health record database, invokes pre-set clinical rules and / or machine learning models, combining the pregnant woman's historical data to perform anomaly detection and risk assessment on the newly associated data. A multi-terminal interaction platform, communicating with the central data processing server, provides a platform for pregnant women, physician assistants, and attending physicians. Provide users with health data views, analysis results, and interactive functions that match their role permissions.

2. The system according to claim 1, characterized in that, The multi-source heterogeneous data access and standardization gateway includes: a protocol adaptation layer configured to support automatic identification and connection of Bluetooth and Wi-Fi communication protocols, and to support data interface docking with hospital information systems or offline data file reception; and a multi-source data parsing and cleaning engine configured to have built-in data parsing and standardization templates corresponding to different brands and models of testing equipment and different hospital data sources. The engine calls the corresponding template according to the received data source identification information, extracts data containing measurement indicators, values, units and measurement timestamps from the raw data, and performs terminology standardization, value range normalization and logical consistency verification on the extracted data, and finally encapsulates it into a predefined unified data format message.

3. The system according to claim 1, characterized in that, The data fusion and association unit is configured to perform the following operations: receive standardized health data containing a pregnant woman's identifier and a measurement timestamp; retrieve her health record and current gestational week information based on the pregnant woman's identifier; and classify and store the standardized health data into a home monitoring dataset under the corresponding gestational week time node according to the correspondence between the measurement timestamp and the current gestational week information.

4. The system according to claim 3, characterized in that, The system also includes a gestational age management unit, which is configured to calculate and continuously maintain the current gestational age timeline of the pregnant woman based on the last menstrual period date or the date of a key prenatal check-up event input by the pregnant woman, and provide the timeline to the data fusion and association unit as a time series reference for data association.

5. The system according to claim 1, characterized in that, The multimodal health analysis engine includes: a rule base storing clinical threshold rules for physiological indicators associated with gestational age ranges; and a logic judgment module used to compare the new data after data fusion and association with the thresholds corresponding to gestational age in the rule base. The system compares against established rules and / or retrieves historical data sequences of the indicator from the health record database for trend consistency comparison; the intelligent analysis module is equipped with a trained machine learning model for analyzing standardized historical data from the health record database. Real-time data sequences can be used for risk prediction, pattern recognition, or to generate personalized monitoring recommendations. When the output of the logic judgment module or intelligent analysis module meets the preset abnormal or risk conditions, an early warning or risk warning message is generated.

6. The system according to claim 5, characterized in that, The warning or risk alert information generated by the multimodal health analysis engine is sent to the multi-terminal interaction platform and pushed to the pregnant women's terminals and the corresponding medical staff terminals in a differentiated manner according to the pre-configured subscription rules.

7. The system according to claim 1, characterized in that, The multi-terminal interaction platform includes: The access control module is used to define and manage the viewing, commenting, and operation permissions of different roles for health record data; The view generation module is used to dynamically organize the integrated multi-source data from the health record database according to the logged-in user's role and permissions, and combine it with the analysis conclusions generated by the multimodal health analysis engine to generate a customized interactive interface. Among them, the pregnant and postpartum view focuses on the intuitive display of personal health trends, personalized task reminders, and easy-to-understand health interpretations; the medical staff view focuses on the integrated display of multi-source data comparison, clinical decision support information, and patient management tools.

8. The system according to claim 7, characterized in that, The healthcare worker view provides an interface for issuing medical orders and planning clinical pathways. The issued medical orders or plans are linked to the health records of specific pregnant women. The pregnant woman view provides an interface for confirming medical orders, providing execution feedback, and reporting symptom self-checks. The feedback information generated is sent back and updated to the health record database, forming a digital management closed loop.

9. An electronic device comprising a processor, a memory, and a computer program stored in the memory, characterized in that, When the processor executes the computer program, it implements the module functions of the system as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the module functions of the system as described in any one of claims 1 to 8.

Images