Search

CN-121421477-B - Hierarchical early warning system based on multiparameter vital sign detection

CN121421477BCN 121421477 BCN121421477 BCN 121421477BCN-121421477-B

Abstract

The invention relates to the technical field of medical early warning and discloses a grading early warning system based on multi-parameter vital sign detection. The system comprises a vital sign monitoring device, a parameter input feature extraction network, a dynamic risk assessment matrix and a monitoring system, wherein the vital sign monitoring device is used for collecting real-time physiological parameters such as heart rate, blood pressure, blood oxygen saturation and body temperature of a patient, the parameters are input into the feature extraction network to generate multidimensional physiological feature vectors, and the dynamic risk assessment matrix containing different time window physiological state change trends is constructed according to the multidimensional physiological feature vectors. And when the score exceeds a preset early warning trigger line, activating a corresponding early warning response mechanism. The system can realize comprehensive and dynamic evaluation of the physiological state of the patient, is suitable for emergency treatment, intensive care and chronic disease care scenes, and meets the requirements of clinical health risk monitoring and early warning.

Inventors

  • LI ANG
  • MA LIJIE
  • XU ZHEN

Assignees

  • 中国人民解放军总医院第八医学中心

Dates

Publication Date
20260508
Application Date
20251208

Claims (10)

  1. 1. Hierarchical early warning system based on multiparameter vital sign detects, characterized by comprising: collecting real-time physiological parameters of a patient through vital sign monitoring equipment, wherein the physiological parameters comprise heart rate, blood pressure, blood oxygen saturation and body temperature; Inputting the real-time physiological parameters into a feature extraction network to generate a multidimensional physiological feature vector; constructing a dynamic risk assessment matrix based on the multidimensional physiological feature vector, wherein the dynamic risk assessment matrix comprises physiological state change trends under different time windows; Dividing the dynamic risk assessment matrix into areas according to a preset grading early warning threshold value to obtain a plurality of risk grade intervals; dynamically adjusting the risk level interval by adopting an adaptive weight distribution strategy to generate a comprehensive risk score of the current patient; and when the comprehensive risk score exceeds a preset early warning trigger line, activating a corresponding early warning response mechanism.
  2. 2. The hierarchical early warning system based on multi-parameter vital sign detection of claim 1, wherein the feature extraction network comprises: performing time sequence standardization processing on the real-time physiological parameters to eliminate noise interference in the acquisition process; extracting local time sequence characteristics of the real-time physiological parameters through a convolutional neural network; And fusing the local time sequence features with global statistical features to generate the multidimensional physiological feature vector.
  3. 3. The hierarchical early warning system based on multi-parameter vital sign detection of claim 1, wherein the process of constructing the dynamic risk assessment matrix comprises: calculating the deviation degree of the physiological state according to the change rate of the multidimensional physiological feature vector; generating the physiological state change trend under the different time windows by combining the historical physiological data; And performing association mapping on the physiological state deviation degree and the physiological state change trend, and filling the dynamic risk assessment matrix.
  4. 4. The hierarchical early warning system based on multi-parameter vital sign detection of claim 1, wherein the hierarchical early warning threshold setting process includes: Determining normal fluctuation ranges of different physiological parameters based on clinical data statistics; dividing three level intervals of low risk, medium risk and high risk according to the normal fluctuation range; A dynamic boundary adjustment coefficient is set for each level interval to accommodate individual differences.
  5. 5. The multi-parameter vital sign detection-based hierarchical early warning system of claim 1, wherein the adaptive weight distribution strategy comprises: calculating the contribution weight of each physiological parameter according to the variance distribution of the multidimensional physiological feature vector; correcting the contribution weights in combination with patient history data; And smoothing the corrected contribution weight by adopting a moving average algorithm to generate the comprehensive risk score.
  6. 6. The hierarchical early warning system based on multi-parameter vital sign detection of claim 1, wherein the activation process of the early warning response mechanism comprises: triggering a first-level early warning response when the comprehensive risk score enters a high risk level interval; When the comprehensive risk score continuously exceeds an early warning trigger line, upgrading to a secondary early warning response; and automatically matching the corresponding clinical intervention plan according to the early warning response level.
  7. 7. The hierarchical early warning system based on multi-parameter vital sign detection of claim 6, wherein the matching process of the clinical intervention protocol comprises: Retrieving a treatment plan corresponding to the current risk level from a medical knowledge base; Screening exclusion conditions related to contraindications according to the electronic medical record of the patient; and generating a personalized intervention instruction set and pushing to the terminal equipment.
  8. 8. The multi-parameter vital sign detection-based hierarchical early warning system of claim 1, further comprising: establishing a cross-modal association relationship between the real-time physiological parameters and the medical image data; and when the comprehensive risk score is abnormal, automatically calling the associated imaging examination result to carry out auxiliary verification.
  9. 9. The multi-parameter vital sign detection-based hierarchical early warning system of claim 1, further comprising: Storing a version iteration record of the dynamic risk assessment matrix through a blockchain network; And carrying out non-tamperable evidence storage on the early warning triggering event by adopting a digital signature technology.
  10. 10. The multi-parameter vital sign detection-based hierarchical early warning system of claim 1, further comprising: Deploying an edge computing node to perform localized pretreatment on the real-time physiological parameter; And when the network delay exceeds a threshold value, enabling the local risk assessment model to generate a temporary early warning decision.

Description

Hierarchical early warning system based on multiparameter vital sign detection Technical Field The invention relates to the technical field of medical early warning, in particular to a grading early warning system based on multi-parameter vital sign detection. Background In a modern medical service system, real-time monitoring and risk early warning of vital signs of patients are important components in the clinical diagnosis and treatment process, and particularly in the scenes of emergency rescue, intensive care, chronic disease home care and the like, timely and accurate vital sign evaluation directly relates to diagnosis and treatment opportunity grasp and health security guarantee of the patients. At present, most of vital sign monitoring means adopted in clinic rely on independent monitoring of single or few physiological parameters, for example, heart rate data is obtained through a common heart rate monitor or blood pressure values are regularly measured through an electronic sphygmomanometer, and the monitoring modes can only capture the state of a single physiological index of a patient at a certain moment and cannot comprehensively reflect the overall change trend of the physiological function of the patient. In the risk assessment process, the conventional early warning system mostly adopts a fixed threshold value for judgment, namely, when a certain physiological parameter exceeds a preset fixed value, an early warning mechanism is triggered. This way of assessing a fixed threshold has obvious limitations, as the age, underlying disease, and physical condition of different patients have individual differences, and the same physiological parameter value may represent a completely different degree of health risk for different patients. For example, for healthy adults, a heart rate around 90 beats/min may still be in the normal range, but for elderly patients suffering from severe heart disease, the same heart rate value may already mean a higher health risk. In addition, when the existing early warning system comprehensively analyzes multiple parameters, an effective weight distribution strategy is lacked, dynamic adjustment cannot be performed according to importance differences of different physiological parameters under specific conditions, and accuracy and pertinence of risk assessment results are insufficient. In clinical practice, the traditional monitoring and early warning mode often causes two situations, namely, for some potential and slowly developed health risks, as the variation amplitude of a single parameter does not reach a fixed early warning threshold, a system cannot capture risk signals in time, so that the optimal intervention time is delayed, for some physiological parameter fluctuation caused by individual difference, the system can misjudge as a risk event, trigger unnecessary early warning response, not only increase the workload of medical staff, but also cause unnecessary panic of patients and families. Along with the continuous development of medical technology and the continuous improvement of the requirements of people on medical service quality, the existing vital sign monitoring and early warning mode based on single parameters and fixed thresholds is difficult to meet the requirements of clinical accurate, dynamic and comprehensive evaluation on the health risk of patients, and a novel system capable of integrating multi-parameter information and realizing dynamic risk evaluation and self-adaptive early warning is needed. Disclosure of Invention The invention aims to provide a grading early warning system based on multi-parameter vital sign detection so as to solve the problems in the background technology. To achieve the above object, the present invention provides a hierarchical early warning system based on multi-parameter vital sign detection, the system comprising: collecting real-time physiological parameters of a patient through vital sign monitoring equipment, wherein the physiological parameters comprise heart rate, blood pressure, blood oxygen saturation and body temperature; Inputting the real-time physiological parameters into a feature extraction network to generate a multidimensional physiological feature vector; constructing a dynamic risk assessment matrix based on the multidimensional physiological feature vector, wherein the dynamic risk assessment matrix comprises physiological state change trends under different time windows; Dividing the dynamic risk assessment matrix into areas according to a preset grading early warning threshold value to obtain a plurality of risk grade intervals; dynamically adjusting the risk level interval by adopting an adaptive weight distribution strategy to generate a comprehensive risk score of the current patient; and when the comprehensive risk score exceeds a preset early warning trigger line, activating a corresponding early warning response mechanism. Preferably, the feature extraction network includes: perform