Vital sign monitoring data tracking acquisition and analysis system

Abstract

The invention discloses a tracking acquisition and analysis system for vital sign monitoring data, which relates to the technical field of vital sign monitoring and comprises a data processing module for receiving and processing real-time vital sign data flow from one or more vital sign monitoring devices; the monitoring scheme management module is used for automatically matching and activating a corresponding personalized monitoring scheme from a predefined scheme library according to the input operation type of the patient; the circadian rhythm analysis module is used for dividing the real-time vital sign data flow into a daytime active period data set and a night rest period data set; and the risk assessment and early warning module is used for generating instant risk assessment based on the real-time vital sign data flow, realizing all-around and intelligent vital sign monitoring and risk early warning of the postoperative patient, and realizing risk early warning by integrating the real-time vital sign data flow, a personalized monitoring scheme, circadian rhythm analysis and multi-dimensional risk assessment. Abnormal physiological status of a patient can be found in time, and postoperative rehabilitation safety is improved.

Description

Technical Field

[0001] This invention relates to the field of vital sign monitoring technology, specifically a system for tracking, collecting, and analyzing vital sign monitoring data. Background Technology

[0002] Current methods for monitoring patient vital signs primarily rely on contact devices or cameras. However, these devices have several limitations. For example, contact device monitoring can pose inconvenience and potential risks to medical staff and patients in radiation environments; camera monitoring is prone to obstruction of view, making it impossible to monitor the patient's condition in real time. Furthermore, for postoperative patient vital sign monitoring, clinical practice mainly relies on a combination of fixed bedside monitors and periodic nurse rounds. This model generally uses a uniform alarm threshold based on population averages, failing to fully consider the differences in the impact of different surgical types on physiological parameters, as well as the personalized risk boundaries brought about by individual patient characteristics (such as age and underlying diseases).

[0003] Meanwhile, traditional monitoring focuses on whether real-time data exceeds static thresholds, which is a kind of "point-based" and reactive alarm mechanism. It completely ignores the inherent circadian rhythm pattern of human vital signs, which contains rich health information. This pattern cannot effectively identify early compensatory states (such as non-dipper blood pressure) where the values ​​are still within the traditional safe range, but the circadian rhythm has undergone characteristic disorder. These are often early signs of infection, poor pain control, or slow deterioration of organ function.

[0004] In addition, alarm information is often simplistic and lacks integration with clinical context, leading to alarm fatigue and false alarm interference for medical staff. It is difficult to accurately identify truly high-risk patients from massive amounts of data, affecting the timeliness of early warning and the targeted nature of intervention.

[0005] Therefore, there is an urgent need for a monitoring system that can enable personalized treatment plans, integrate real-time data with long-term rhythm analysis, and provide intelligent comprehensive risk assessment to improve the predictability, accuracy, and clinical effectiveness of postoperative monitoring. Summary of the Invention

[0006] The purpose of this invention is to provide a system for tracking, collecting, and analyzing vital sign monitoring data, aiming to achieve comprehensive and intelligent monitoring of vital signs and risk warning for postoperative patients. By integrating real-time vital sign data streams, personalized monitoring plans, circadian rhythm analysis, and multi-dimensional risk assessment, the system can promptly detect abnormalities in the patient's physiological state, provide decision support for clinical intervention, improve the safety of postoperative rehabilitation, and solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a system for tracking, collecting, and analyzing vital sign monitoring data, comprising:

[0008] The monitoring plan management module is configured to execute a personalized configuration strategy based on multi-dimensional features, including retrieving a basic monitoring template from a pre-set plan library in response to the input patient surgery type; then parsing the patient's individual feature data, dynamically correcting the monitoring parameter set and alarm thresholds in the basic monitoring template, and generating and activating a personalized monitoring plan specific to the patient; the personalized monitoring plan defines key monitoring parameters, a first alarm threshold, and a day-night difference safety threshold for the patient.

[0009] The data processing module collects real-time data streams from the vital signs monitoring device and extracts targeted parameter data according to the definition of the personalized monitoring scheme.

[0010] The circadian rhythm analysis module is configured to perform long-term physiological pattern analysis, including dynamically labeling and dividing the real-time data stream into a daytime activity period dataset and a nighttime rest period dataset based on the patient's real-time activity status; calculating the statistical characteristic values ​​of key monitoring parameters defined in the personalized monitoring plan in the daytime activity period dataset and the nighttime rest period dataset respectively; calculating the rhythm difference value of the key monitoring parameters between day and night, comparing the rhythm difference value with the daytime-night difference safety threshold set in the personalized monitoring plan, and outputting the rhythm deviation result;

[0011] The risk assessment and early warning module is configured to perform a dual-track risk fusion assessment, including:

[0012] The first track of instant risk assessment is generated by comparing the real-time data stream with the first alarm threshold.

[0013] Based on the rhythm deviation results, a second track is generated to assess the rhythm risk of abnormal physiological patterns;

[0014] A comprehensive risk level is generated by combining the real-time risk assessment and the rhythmic risk assessment, and an early warning is triggered when the level exceeds a preset limit.

[0015] Preferably, the data processing module includes:

[0016] The data buffer unit is configured to receive and buffer real-time vital sign data streams from the one or more vital sign monitoring devices via a distributed message queue.

[0017] The data cleaning unit is configured to clean the buffered data stream, including handling missing values, identifying outliers based on clinical rules and statistical models, and aligning and synchronizing data points from different devices with timestamps.

[0018] The standardized unit is configured to transform cleaned data into a structured data stream that follows a unified data model, which at least includes patient identifiers, monitoring parameter types, parameter values, timestamps, and data quality labels.

[0019] Preferably, the method for generating the personalized monitoring scheme includes:

[0020] A basic treatment plan template is matched based on the patient's surgical type;

[0021] Based on the patient's individual characteristics, the key monitoring parameter set and the first alarm threshold in the basic protocol template are automatically adjusted. The individual characteristics include at least one of age, gender, body mass index, chronic disease history, and intraoperative special events. When the monitoring protocol management module makes dynamic corrections, if it identifies that the patient has a specific chronic disease history, it automatically relaxes or tightens the first alarm threshold of the corresponding parameter and adjusts the weight of the parameter in the risk assessment.

[0022] Preferably, the day-night difference safety threshold is set based on at least one of the following methods:

[0023] Universal thresholds based on clinical guidelines;

[0024] Adjusted threshold based on historical data statistics of the patient's group;

[0025] Personal baseline rhythm thresholds are generated based on patients' postoperative stable vital sign data.

[0026] Preferably, the risk assessment and early warning module further includes performing multi-parameter correlation analysis during the real-time risk assessment, that is, determining the risk level based on the numerical combination pattern of at least two vital sign parameters.

[0027] Preferably, the method combining the real-time risk assessment and the rhythmic risk assessment includes a hybrid model that combines a rule engine with a weighted decision system:

[0028] The rule engine is configured to: when any vital sign parameter reaches a preset critical alarm threshold, directly set the comprehensive risk level to the highest level;

[0029] The weighted decision system is configured to assign dynamic weights to the immediate risk assessment sub-score and the rhythmic risk assessment sub-score, the dynamic weights being adjusted based on at least one of the following factors: postoperative stage, surgical type, or patient baseline condition.

[0030] Preferably, when the risk assessment and early warning module triggers an early warning, it generates structured early warning information, which includes at least the patient identifier, comprehensive risk level, key abnormal parameters that caused the early warning and their values, and a list of treatment suggestions retrieved from the clinical knowledge base.

[0031] As a preferred option, distributing the corresponding level of early warning information specifically includes: selectively distributing it in a tiered and coordinated manner through one or more channels, such as bedside audio-visual alarms, nurse station terminal pop-ups, mobile device message pushes, hospital information system event records, and regional voice broadcasts, based on the comprehensive risk level.

[0032] As a preferred approach, generating rhythm risk assessments involves mapping the degree of deviation from rhythm differences to specific clinical risk indicators.

[0033] Preferably, a visualization and reporting module is also included, configured to generate a central monitoring dashboard containing circadian rhythm characteristics. The dashboard not only displays real-time vital sign values, but also displays a comparative view of the patient's current circadian rhythm trend and the safe threshold range of the circadian difference in a visual overlay, and highlights periods of abnormal rhythm.

[0034] In summary, the beneficial effects of this invention are:

[0035] This invention integrates personalized monitoring schemes with intelligent circadian rhythm analysis to achieve comprehensive and proactive risk monitoring of postoperative patients' vital signs, from "instantaneous abnormal values" to "long-term abnormal physiological patterns." It can dynamically customize monitoring priorities and alarm thresholds based on specific surgical types and individual patient characteristics. Furthermore, it innovatively introduces and quantifies the circadian rhythm changes of vital signs. By fusing instantaneous threshold alarms and rhythm disorder warnings, a comprehensive risk level is generated. This allows for earlier and more accurate identification of potential, slowly developing physiological function deterioration trends, going beyond the limitations of traditional methods that only detect acute and overt risks. This significantly improves the predictability and accuracy of postoperative risk warnings, buying valuable time for clinical intervention and ultimately effectively reducing the incidence of postoperative complications, thus improving patient safety and recovery quality. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the 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 invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram of the overall process framework of a vital sign monitoring data tracking, acquisition, and analysis system according to the present invention;

[0038] Figure 2 This is a schematic diagram of the workflow framework for circadian rhythm analysis in a vital sign monitoring data tracking, acquisition, and analysis system of the present invention;

[0039] Figure 3 This is a schematic diagram of the risk assessment and early warning process framework in a vital sign monitoring data tracking, acquisition and analysis system of the present invention;

[0040] Figure 4 This is a schematic diagram of the configuration interface for a personalized monitoring scheme in a vital sign monitoring data tracking, acquisition, and analysis system of the present invention.

[0041] Figure 5 This is a schematic diagram of the circadian rhythm analysis interface in a vital sign monitoring data tracking, acquisition, and analysis system of the present invention. Detailed Implementation

[0042] The present invention will now be described in further detail with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. These drawings are simplified schematic diagrams, which are only used to illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0043] To facilitate understanding of the present invention, a more complete description of the invention will be given below with reference to the accompanying drawings, which illustrate several embodiments of the invention. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of the invention will be more thorough and complete.

[0044] All features disclosed in this specification, or all steps in all disclosed methods or processes, may be combined in any way, except for mutually exclusive features and / or steps.

[0045] Any feature disclosed in this specification (including any appended claims, abstract, and drawings) may be replaced by other equivalent or similar features for a similar purpose, unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is merely one example of a series of equivalent or similar features.

[0046] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, a direct connection, or an indirect connection through an intermediate medium; they can refer to the internal communication of at least two elements or the interaction relationship of at least two elements, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0047] The following is combined with Figures 1-5 The present invention will be described in detail below. One embodiment of the present invention provides a system for tracking, collecting, and analyzing vital sign monitoring data. The system adopts a modular and layered architecture, mainly including the following layers:

[0048] Data acquisition layer: Through the medical IoT interface, it connects to multi-source vital sign monitoring devices, such as electrocardiogram monitors, blood pressure monitors, pulse oximeters, and respiratory sensors.

[0049] Data processing and analysis layer: contains core functional modules to realize data stream processing, scheme matching, rhythm analysis and risk assessment.

[0050] Early warning and output layer: Generates visual reports and real-time early warning information, and supports integration with hospital information systems, nurse station dashboards, mobile terminals, etc.

[0051] The data acquisition layer includes a data processing module, whose core task is to transform raw vital sign data streams from different sources and in various formats into high-quality, standardized, structured data that can be directly used by downstream analysis modules. The processing flow is as follows:

[0052] 1. Data access and buffering

[0053] The system accesses multi-source data through a set of device adapters. These adapters support common protocols for medical IoT, such as HL7, FHIR, MQTT, DICOM waveforms, as well as proprietary protocols from various vendors. Data continuously flows in as a high-concurrency stream, and the module first sends it to a distributed message queue (such as Apache Kafka or Pulsar) for buffering. This is intended to decouple data production and consumption, cope with instantaneous traffic spikes, and ensure that no data is lost.

[0054] 2. Data Cleaning and Alignment

[0055] This is a key step in improving data quality, mainly addressing three types of issues:

[0056] Handling missing values: For short-term, non-continuous data loss, linear interpolation or trend prediction interpolation based on recent historical patient data is used to fill in the missing data. For long-term, critical data loss, it is marked as a "data interruption event" and the monitoring plan management module is notified, which may trigger a device status warning.

[0057] Identifying and handling outliers:

[0058] Rule-based filtering: First, apply hard rules based on clinical knowledge for filtering, such as considering a heart rate <20 or >200 beats / min as a physically unreliable value.

[0059] Statistical and model filtering: Combining individual patient baselines and population statistical models, data points that deviate from the normal range but are not obviously physical errors are identified. Such data are not simply discarded, but are marked as "suspicious values" for the risk assessment module to make a comprehensive judgment in conjunction with other parameters.

[0060] Timestamp alignment and synchronization: Clock drift or transmission delays exist between different devices. The module uses the system master clock as a reference to calibrate and synchronize the timestamps of each data point, ensuring that all vital sign parameters are strictly aligned on the timeline, laying the foundation for subsequent correlation analysis and circadian rhythm calculation.

[0061] 3. Format standardization and unified data modeling

[0062] The cleaned data is transformed into a standardized data model that is unified within the system. This model defines the core entities and relationships:

[0063] Patient vital signs record: As a basic data unit, it includes patient ID, timestamp, parameter type (such as "systolic blood pressure"), parameter value, numerical unit, and data quality label, such as "normal", "interpolated", "suspicious".

[0064] Device metadata: Records the device ID, type, and status from which the data originates.

[0065] Contextual information: Associates the patient's currently active monitoring protocol ID, providing context for subsequent personalized analysis.

[0066] This process is accomplished through a format conversion engine that has built-in parsers for various devices, mapping raw messages or JSON data to a standard model.

[0067] 4. Real-time stream processing and distribution

[0068] The standardized data stream enters a stream processing engine, such as Apache Flink or Spark Streaming. Here, the data is partitioned by patient ID, enabling parallel processing at the patient level. The engine performs two core operations:

[0069] Windowed aggregation calculation: In order to smooth instantaneous fluctuations and extract trends, the system uses a sliding window, such as a 60-second window, for high-frequency data, such as heart rate per heartbeat. The calculation is performed every 10 seconds, and statistical features such as mean, maximum, and minimum values ​​within the window are generated in real time.

[0070] Real-time distribution: The processed, standardized data stream is broadcast in real time to two downstream channels:

[0071] Real-time analysis channel: pushes data to the risk assessment and early warning module with extremely low latency for instant threshold comparison.

[0072] Persistence and Batch Processing Channel: Store data in time series databases, such as InfluxDB and TimescaleDB, for the circadian rhythm analysis module to perform daily batch statistical calculations, and support historical data backtracking and visualization.

[0073] 5. Outputs and Interfaces

[0074] The module's final output is a continuous, stable, and high-quality stream of standard vital signs data. This data stream has the following characteristics:

[0075] Structured: Strictly adhere to a unified data model.

[0076] Traceability: Each data point comes with a clear data quality label and processing history.

[0077] Rich in context: Deeply integrated with patient information and monitoring protocols.

[0078] Low latency: End-to-end latency from data access to availability for risk assessment is controlled within seconds.

[0079] Through the above process, the data processing module transforms the chaotic raw data into a reliable information source, enabling subsequent personalized monitoring, rhythm analysis, and risk assessment to be carried out reliably and efficiently, providing a solid data foundation for building an intelligent and accurate postoperative patient risk early warning system.

[0080] The data processing and analysis layer includes a monitoring scheme management module and a circadian rhythm analysis module;

[0081] The monitoring protocol management module dynamically configures monitoring strategies based on the patient's unique surgical background and physiological state, ensuring that system resources are focused on the most relevant and highest-risk vital signs parameters. The specific method includes the following steps:

[0082] 1. The architecture and knowledge sources of the monitoring solution library

[0083] The protocol library is a structured clinical knowledge base designed following the "template-instance" principle.

[0084] Protocol Template: Based on clinical guidelines, expert consensus, and historical data analysis, a standard monitoring protocol template is predefined for each type of surgery, such as coronary artery bypass grafting, total hip replacement, and laparoscopic cholecystectomy.

[0085] Solution Content: Each template contains three core dimensions:

[0086] Key monitoring parameter set: This set clearly indicates parameters that require "intensive monitoring." This includes not only parameter types (such as heart rate) but may also be refined to specific indicators, such as the ST segment on an electrocardiogram (ECG) or diastolic blood pressure. For example, postoperative protocols in neurosurgery might prioritize monitoring intracranial pressure and level of consciousness, while postoperative protocols in orthopedic surgery might focus on lower limb oxygen saturation.

[0087] The first alarm threshold system sets dynamic, multi-level safety boundaries for each key parameter. Thresholds are typically divided into three levels:

[0088] Warning threshold (yellow): Indicates an early deviation from the normal trend, requiring attention from medical staff.

[0089] Severe threshold (orange): Indicates a clear deterioration in physiological condition, requiring preparation for intervention.

[0090] Critical threshold (red): Indicates a life-threatening situation requiring immediate intervention.

[0091] The threshold value may be a range and can be stratified according to the patient's age, such as the lower heart rate warning threshold for elderly patients, which can be fine-tuned.

[0092] Monitoring intensity and frequency guidance: implicitly guides the resource allocation of data processing and early warning modules, such as using higher data sampling rates or more sensitive analysis algorithms for key parameters.

[0093] 2. The logic for matching and generating personalized solutions, refer to... Figure 4

[0094] Once patient information is entered into the system, the module initiates a multi-stage intelligent matching process:

[0095] Phase 1: Core Matching. Using the surgical type as the primary keyword, the most matching basic treatment plan template is retrieved from the treatment plan database.

[0096] Phase Two: Parametric Adjustment. This is a crucial personalization step where the system automatically adjusts the thresholds and parameter sets in the selected template based on the patient's individual characteristics.

[0097] Demographic adjustment: Adjust relevant thresholds, such as respiratory rate and blood pressure, based on the patient's age, gender, and baseline weight (e.g., BMI).

[0098] Adjustments based on complications and medical history: If a patient has a history of chronic diseases, such as hypertension, diabetes, or chronic obstructive pulmonary disease, the system will lower the warning thresholds for related parameters accordingly. For example, the blood pressure warning value for hypertensive patients will be raised, and the monitoring parameters of associated complications will be prioritized.

[0099] For example, the weighting and alarm sensitivity of blood oxygen saturation monitoring will be further increased for thoracic surgery patients with a history of COPD.

[0100] Intraoperative event-based adjustments: If the system is connected to intraoperative records, special intraoperative events, such as massive bleeding or prolonged hypotension, will trigger the temporary addition or enhancement of monitoring of relevant indicators, such as hemoglobin trends and blood pressure stability, in the postoperative plan.

[0101] Phase Three: Protocol Activation and Binding. This phase generates the final, personalized monitoring protocol "tailor-made" for the patient and dynamically binds its unique identifier to the patient's real-time data stream. This protocol serves as "metadata," continuously guiding the focus of all downstream analysis modules.

[0102] 3. Dynamic Management and Clinical Interaction

[0103] This module is not a "one-time setup" but supports dynamic interaction and adjustment:

[0104] Clinical Flexibility: Nurses or doctors can manually fine-tune the threshold of a parameter or temporarily add / remove a monitoring focus based on the patient's real-time condition on the system interface. All manual adjustments are recorded and the reasons are noted.

[0105] Protocol Evolution and Learning: The system records the actual operational effectiveness of each personalized protocol, such as the accuracy and false alarm rate of alerts. This anonymized aggregated data can be used for subsequent protocol library optimization, forming a closed loop of "clinical practice - data feedback - knowledge update," enabling the protocol library to continuously evolve with use and better reflect real clinical needs.

[0106] 4. Output and downstream collaboration

[0107] The module's final output is a dynamic, structured, and executable personalized monitoring scheme object. This object is distributed in real time to:

[0108] Data processing module: guides the focus of its data quality verification, such as zero tolerance for missing data for key parameters.

[0109] The circadian rhythm analysis module informs users which key parameters need to be calculated to differentiate between day and night.

[0110] Risk assessment and early warning module: Provides a personalized "first alarm threshold" standard for making real-time risk judgments.

[0111] Through this module, the system ensures that the monitored behavior is highly consistent with the patient's individual risk profile, accurately guiding limited clinical attention resources to the places where problems are most likely to occur, and achieving a qualitative leap from "homogenized monitoring" to "personalized care".

[0112] refer to Figure 5 The circadian rhythm analysis module identifies normal patterns and abnormal deviations in vital signs throughout the day-night cycle, enabling earlier and more intelligent warnings of potential patient risks. This module is built upon the core medical principle that "healthy physiological states exhibit predictable diurnal fluctuations," as detailed below:

[0113] 1. Data partitioning strategy: Dual-mode driven time period identification

[0114] The module employs a dual-track strategy to precisely distinguish between the patient's active and resting periods, ensuring the accuracy of rhythm analysis.

[0115] Based on a fixed time window baseline: As the default and fundamental strategy, the system establishes initial day-night time labels according to the organization's regular work and rest schedule, such as 06:00-22:00 as daytime and 22:00-06:00 the next day as nighttime, or the sunrise and sunset times of the geographical location. This ensures the consistency and repeatability of the analysis.

[0116] Dynamic verification and correction based on biological behavioral signals: To improve individual adaptability, the module integrates patient behavioral and physiological data for verification and adjustment.

[0117] Activity status analysis: The system quantifies the patient's physical activity level using data from mattress pressure sensors or triaxial accelerometers on wearable devices. A sustained low activity level combined with a nighttime time window indicates a rest period; conversely, if the patient exhibits abnormally high activity levels during the night, the system may mark it as a "nocturnal agitation period" and trigger special attention.

[0118] Autonomic nervous system state inference: By analyzing the time-frequency domain indicators of heart rate variability, the patient's sympathetic-parasympathetic nervous system balance can be objectively inferred, thereby more accurately identifying the onset of sleep, sleep depth, and wakefulness stages.

[0119] Decision-making logic: Using a fixed time window as the framework and biological behavioral signals as the flesh and blood for fine-tuning. For example, even if it is daytime, if the patient's HRV analysis shows that they are in a deep sleep state, the data of the relevant time period can be temporarily classified as "abnormal daytime sleep period", and its rhythm analysis will be labeled and evaluated separately.

[0120] 2. Statistical Value Calculation: From Raw Data to Rhythm Feature Extraction

[0121] On the predefined daytime and nighttime datasets, the system performs multi-level statistical calculations to form rhythmic "fingerprints":

[0122] Basic statistics: Calculate the mean, median, standard deviation, minimum, and maximum values ​​for each vital sign parameter, especially key monitoring parameters, within their respective time periods. The mean represents the overall level, and the standard deviation reflects volatility.

[0123] Trend and distribution statistics: Calculate specific percentiles to understand the range of parameter distribution; calculate the slope of short-term trend lines using a sliding window to capture the direction of parameter change over time.

[0124] Rhythm feature derivation: generating clinically crucial derivative indicators, such as:

[0125] Nighttime blood pressure drop rate: (Daytime average systolic blood pressure - Nighttime average systolic blood pressure) / Daytime average systolic blood pressure * 100%, which is key to assessing the diurnal blood pressure pattern (dipper / non-dipper / reverse dipper).

[0126] Heart rate diurnal variability: the ratio or difference between the standard deviations of daytime and nighttime heart rates.

[0127] Apnea-Hypopnea Index (AHI): In nighttime data, a preliminary AHI index is automatically calculated based on the periodic interruptions in the breathing waveform to screen for sleep apnea.

[0128] 3. Day-night difference analysis: Establishment and comparison of personalized safety boundaries

[0129] This is the intelligent core of the module, which compares the calculated rhythm differences with personalized safety boundaries:

[0130] Difference calculation:

[0131] For continuous parameters (such as blood pressure and heart rate), calculate the percentage change or absolute difference mentioned above.

[0132] For pattern parameters (such as arrhythmia events), calculate the ratio of their nighttime to daytime occurrence frequency.

[0133] Personalized setting of differential safety thresholds: Threshold settings reflect the concept of individualized medicine and are divided into three levels:

[0134] Based on evidence-based medicine, the general threshold is: first, apply recognized clinical guidelines, for example, that healthy adults should have a 10%-20% lower nighttime systolic blood pressure than daytime blood pressure.

[0135] Population statistics-based correction: Adjust the general threshold according to the historical data distribution of patients' subgroups, such as those of the same age, type of surgery, and complication groups.

[0136] For example, the normal range for the percentage drop in nighttime blood pressure may be narrower in elderly patients or patients with heart failure.

[0137] Dynamic calibration based on patient baseline: the most ideal personalized approach. In the early stages when the patient's condition is relatively stable after surgery, such as 24-48 hours post-surgery, the system automatically learns the "personal baseline rhythm pattern" of their vital signs and uses this pattern as the "gold standard" for subsequent comparisons.

[0138] For example, if a patient's baseline shows that their nighttime heart rate typically drops by 15%, then in subsequent monitoring, a drop of less than 5% or more than 25% is considered a "significant rhythm deviation" for that individual.

[0139] Threshold comparison and anomaly detection: The calculated actual difference value is compared with the corresponding personalized security threshold range.

[0140] It can not only determine whether something is "exceeding the limit", but also the direction and degree of deviation (such as "insufficient nighttime blood pressure drop (non-dipper type)" vs. "excessive nighttime blood pressure drop (super-dipper type)"). This information is crucial for clinical diagnosis of etiology.

[0141] 4. Output: Structured rhythmic health report

[0142] The module's output is not a simple "normal / abnormal" flag, but a comprehensive rhythm analysis report, including:

[0143] Rhythm parameter table: daytime and nighttime statistical values ​​of key parameters, calculated differences and their comparison with personalized safety thresholds (normal, warning, abnormal).

[0144] Rhythm pattern labeling: Labeling patients with their overall rhythmic health status, such as "Blood pressure: non-dipper rhythm", "Heart rate: good diurnal rhythm", "Respiration: nocturnal periodic abnormalities indicate OSA risk".

[0145] Trend visualization data: Provides the front-end dashboard with the data needed to draw "24-hour vital signs trend charts" and overlay "ideal rhythm intervals".

[0146] Trigger event log: Records the specific time, parameters, and deviation details of any security threshold being triggered.

[0147] Through its circadian rhythm analysis module, the system represents a paradigm shift from monitoring whether values ​​at a specific moment are within a safe range to assessing the health of physiological patterns throughout the day. It can uncover slowly developing rhythmic disturbances masked by traditional threshold monitoring, providing clinicians with deeper insights into patient recovery quality, pain management effectiveness, and autonomic nervous system function recovery, truly achieving a 24 / 7, intelligent interpretation of the patient's physiological state.

[0148] The early warning and output layer includes a risk assessment and early warning module. By integrating multi-dimensional analysis results, it forms a comprehensive and hierarchical risk assessment of the patient's health status and triggers a precise and efficient early warning response mechanism at critical moments, transforming data insights into actionable clinical actions, as detailed below:

[0149] 1. Multi-dimensional risk assessment engine

[0150] Real-time risk assessment: Capturing "point-like" crises in real time. The engine continuously receives standardized data streams from the data processing module and performs millisecond-level comparisons based on the first alarm threshold provided by the monitoring scheme management module.

[0151] Dynamic risk assessment:

[0152] Single-parameter classification: For each vital sign parameter, the system assigns a risk level based on its value relative to the three thresholds of personalized early warning, severe, and critical.

[0153] Multi-parameter correlation analysis: This is key to intelligent operation. The engine not only looks at individual parameters, but also analyzes the patterns of parameter combinations. For example:

[0154] "Increased heart rate" + "decreased blood pressure" may indicate insufficient blood volume or early shock, and the risk is far higher than "increased heart rate" + "normal blood pressure" alone.

[0155] "Increased respiratory rate" + "decreased blood oxygen saturation" strongly indicate deterioration of respiratory function.

[0156] Trend risk assessment: By combining short-term (e.g., the past 5-10 minutes) data trends, identify the risk difference between "stabilizing at a critical value" and "rapidly deteriorating to a critical value", the latter of which is usually assigned a higher degree of urgency.

[0157] Output: Generates a dynamically updated real-time risk vector, including the current highest risk level, a list of major abnormal parameters, risk combination pattern labels (such as "hemodynamic instability pattern"), and trend direction.

[0158] Rhythm Risk Assessment: Identifies potential risks that are “pattern-based” or “trend-based.” This engine analyzes structured reports from the circadian rhythm module.

[0159] Evaluation dimensions:

[0160] Deviation assessment: A quantitative score is given based on the degree (mild, moderate, severe) of the diurnal difference value of each key parameter deviating from its personalized safety threshold and the number of parameters.

[0161] Clinical significance mapping: mapping abstract rhythm deviations to specific clinical risk assumptions.

[0162] For example:

[0163] "Non-dipper blood pressure rhythm" is mapped to "increased long-term risk of cardiovascular events" and "risk of insufficient organ perfusion".

[0164] “Significantly reduced nocturnal heart rate variability” was mapped to “poor recovery of autonomic nervous function” or “potentially poor pain control”.

[0165] "Dysronic breathing rhythm disorder during sleep" was mapped to "risk of hypoxemia" and "poor sleep quality affecting rehabilitation".

[0166] Generate a rhythm risk analysis report, including an overall rhythm health score, a list of key rhythm abnormalities, and their corresponding clinical risk warnings.

[0167] Its risk level usually focuses more on "medium- to long-term early warning" rather than "immediate emergency rescue".

[0168] 2. Intelligent Risk Integration and Comprehensive Risk Level Generation

[0169] Aimed at solving the problem of "how to make the best judgment when the immediate risk and rhythmic risk signals are inconsistent", a hybrid model combining a multi-expert weighted decision system and a clinical rule engine is adopted.

[0170] Rule engine (handling specific critical situations): sets a series of priority rules.

[0171] For example, "if any single parameter reaches the 'critical' threshold, the overall risk level will automatically rise to the highest level," ensuring a zero-delay response to clear threats to life.

[0172] Weighted decision system (handling complex and non-critical situations): Dynamic weights are assigned to immediate risk sub-scores and rhythmic risk sub-scores. These weights can be adjusted based on the following factors:

[0173] Postoperative stage: In the early postoperative period (within 24 hours), the weight of immediate risk is extremely high; after entering the recovery period, the weight of rhythmic risk gradually increases in order to monitor the quality of rehabilitation.

[0174] Type of surgery: After cardiac surgery, the weight of the risk of electrocardiographic rhythm may be increased.

[0175] Patient baseline condition: For patients with autonomic neuropathy, the weight and sensitivity of rhythm risk assessment are increased.

[0176] Comprehensive risk level generation: Through a fusion algorithm, a unified comprehensive risk level is output, typically divided into four or five levels.

[0177] Examples include: "Normal", "Low Risk - Observation", "Medium Risk - Attention", "High Risk - Intervention", and "Critical - Rescue".

[0178] This level is accompanied by a detailed "risk summary" explaining the main contributing factors, such as: "Due to the sudden onset of premature ventricular contractions (immediate risk) and the loss of blood pressure diurnal rhythm (rhythm risk), the overall assessment is 'high risk - intervention'."

[0179] 3. Tiered early warning triggering and precise distribution system;

[0180] When the overall risk level rises to the preset trigger threshold (such as "medium risk" or above), the system automatically triggers the early warning workflow.

[0181] Structured generation of early warning information: The system automatically generates complete early warning events containing the following elements:

[0182] Key identifiers: Patient ID, bed number, and unique alert number.

[0183] Risk assessment: Based on the overall risk level and risk type, such as "arrhythmia", "precursor to respiratory failure", and "rhythm disorder indicating infection risk".

[0184] Data evidence: List the key abnormal parameters that led to the warning and their values, trend chart snapshots, and key screenshots of rhythm analysis.

[0185] Clinical context: Patient's type of surgery, current postoperative days.

[0186] Intelligent suggestions: By accessing the clinical knowledge base, it provides a preliminary list of treatment suggestions, such as "Check if the drainage tube is patent", "Re-examine arterial blood gas", and "Consider evaluating the analgesia plan", providing decision support for medical staff.

[0187] A multi-channel, tiered distribution mechanism:

[0188] Channel compatibility:

[0189] Bedside / Nurse Station: Audible and visual alarms, large screen flashing red, computer pop-up (forced confirmation).

[0190] Mobile terminals: Through the hospital's safety messaging platform, structured early warning information is pushed to the dedicated PADs or mobile phones of the responsible nurses and attending physicians.

[0191] Hospital Information System Integration: Early warning events are written to the patient logs of the HIS / EHR system and can be proactively notified at the doctor's workstation.

[0192] Broadcast system (for the highest level of alarm): broadcasts voice messages in a specific area.

[0193] Tiered linkage: The warning level is linked to the scope and method of notification. A "critical" alarm may simultaneously call the bedside nurse, the resident physician, and the anesthesia recovery room; a "moderate" alarm may only notify the responsible nurse and be displayed on the nurse station screen.

[0194] 4. Output and Closed-Loop Management

[0195] The module's final output is a real-time, dynamic, and traceable risk monitoring closed loop:

[0196] Real-time dashboard data: Provides a dynamic view of the overall risk level and risk labels for all patients on the central monitoring screen.

[0197] Warning event stream: The continuously generated records of warning events constitute the timeline of proactive monitoring.

[0198] Historical risk trajectory: A complete "risk timeline" is generated for each patient, recording changes in their risk level, all triggered warnings and clinical responses. This is a valuable resource for medical record review, quality improvement and scientific research analysis.

[0199] Feedback learning signals: The system records the clinical response results of the warning, such as: after the warning, the nurse took action and the patient's parameters returned to normal.

[0200] This feedback can be used to optimize the parameters of the risk assessment model in the future, reduce false alarms, and enable the system to continuously evolve.

[0201] This module completes the entire value chain from "data perception" to "intelligent analysis" and then to "clinical action." It is not only an alarm device, but also an intelligent partner that assists in clinical decision-making, aiming to improve the predictability, accuracy and overall safety of postoperative monitoring.

[0202] The above description is merely a specific embodiment of the invention, but the scope of protection of the invention is not limited thereto. Any variations or substitutions conceived without inventive effort should be included within the scope of protection of the invention. Therefore, the scope of protection of the invention should be determined by the scope defined in the claims.

Claims

1. A system for tracking, collecting, and analyzing vital sign monitoring data, characterized in that: include: The monitoring plan management module is configured to execute a personalized configuration strategy based on multi-dimensional features, including generating a unique personalized monitoring plan by parsing the patient's individual feature data based on the patient's surgical type; the personalized monitoring plan includes the patient's key monitoring parameters, the first alarm threshold, and the day-night difference safety threshold; The data processing module collects real-time data streams of vital signs and extracts targeted parameter data based on personalized monitoring plans. The circadian rhythm analysis module is configured to perform long-term physiological pattern analysis, including dynamically labeling and dividing the real-time data stream into daytime activity period datasets and nighttime rest period datasets based on the patient's real-time activity status; for key monitoring parameters, calculating their statistical characteristic values ​​in the daytime activity period dataset and the nighttime rest period dataset to obtain the circadian rhythm difference value between day and night; and outputting the circadian rhythm deviation result based on the circadian rhythm difference value and the daytime-night difference safety threshold. The risk assessment and early warning module is configured to perform a dual-track risk fusion assessment, including: The first track of instant risk assessment is generated by comparing real-time data stream with the first alarm threshold. A second track is generated based on the rhythm deviation results to reflect the rhythm risk assessment of abnormal physiological patterns; A comprehensive risk level is generated by combining real-time risk assessment and rhythmic risk assessment, and an early warning is triggered when the level exceeds a preset limit.

2. The system for tracking, collecting, and analyzing vital sign monitoring data according to claim 1, characterized in that: The data processing module includes: The data buffer unit is configured to receive and buffer real-time vital sign data streams from the one or more vital sign monitoring devices via a distributed message queue. The data cleaning unit is configured to clean the buffered data stream, including handling missing values, identifying outliers based on clinical rules and statistical models, and aligning and synchronizing data points from different devices with timestamps. The standardized unit is configured to transform cleaned data into a structured data stream that follows a unified data model, which at least includes patient identifiers, monitoring parameter types, parameter values, timestamps, and data quality labels.

3. The system for tracking, collecting, and analyzing vital sign monitoring data according to claim 1, characterized in that: The method for generating the personalized monitoring plan includes: A basic treatment plan template is matched based on the patient's surgical type; Based on the patient's individual characteristics, the key monitoring parameter set and the first alarm threshold in the basic protocol template are automatically adjusted. The individual characteristics include at least one of age, gender, body mass index, chronic disease history, and intraoperative special events. When the monitoring protocol management module makes dynamic corrections, if it identifies that the patient has a specific chronic disease history, it automatically relaxes or tightens the first alarm threshold of the corresponding parameter and adjusts the weight of the parameter in the risk assessment.

4. The system for tracking, collecting, and analyzing vital sign monitoring data according to claim 3, characterized in that: The day-night difference safety threshold is set based on at least one of the following methods: Based on general rhythm thresholds for this type of surgery according to clinical guidelines; Adjusted threshold based on historical data statistics of the patient's group; Personal baseline rhythm thresholds are generated based on patients' postoperative stable vital sign data.

5. The system for tracking, collecting, and analyzing vital sign monitoring data according to claim 1, characterized in that: The risk assessment and early warning module also includes performing multi-parameter correlation analysis during the real-time risk assessment, that is, determining the risk level based on the numerical combination pattern of at least two vital sign parameters.

6. The system for tracking, collecting, and analyzing vital sign monitoring data according to claim 5, characterized in that: The method combining the real-time risk assessment and the rhythmic risk assessment includes a hybrid model that combines a rule engine with a weighted decision system: The rule engine is configured to: when any vital sign parameter reaches a preset critical alarm threshold, directly set the comprehensive risk level to the highest level; The weighted decision system is configured to assign dynamic weights to the immediate risk assessment sub-score and the rhythmic risk assessment sub-score, the dynamic weights being adjusted based on at least one of the following factors: postoperative stage, surgical type, or patient baseline condition.

7. The system for tracking, collecting, and analyzing vital sign monitoring data according to claim 6, characterized in that: When the risk assessment and early warning module triggers an early warning, it generates structured early warning information, which includes at least the patient identifier, comprehensive risk level, key abnormal parameters that led to the early warning and their values, and a list of treatment recommendations retrieved from the clinical knowledge base.

8. The system for tracking, collecting, and analyzing vital sign monitoring data according to claim 7, characterized in that: The distribution of warning information at the corresponding level specifically includes: based on the comprehensive risk level, selectively distributing warning information through one or more channels, such as bedside audio-visual alarms, nurse station terminal pop-ups, mobile device message pushes, hospital information system event records, and regional voice broadcasts, in a tiered and coordinated manner.

9. The system for tracking, collecting, and analyzing vital sign monitoring data according to claim 1, characterized in that: Generating rhythm risk assessments involves mapping the degree of deviation from rhythm differences to specific clinical risk indicators.

10. The system for tracking, collecting, and analyzing vital sign monitoring data according to claim 1, characterized in that: It also includes a visualization and reporting module, configured to generate a central monitoring dashboard that includes circadian rhythm features. The dashboard not only displays real-time vital sign values, but also displays a visual overlay of the patient's current circadian rhythm trend and a comparison view of the circadian difference safety threshold range, and highlights periods of abnormal rhythm.

Images