Sepsis detection system
The sepsis detection system, which integrates data acquisition and machine learning algorithms, solves the delay and subjectivity problems of existing diagnostic methods, realizes real-time, personalized risk assessment and early warning of sepsis, and improves diagnostic accuracy and treatment effects.
Patent Information
- Application Number
- CN202510713696.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-30
- Publication Date
- 2025-09-05
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing sepsis diagnostic methods are characterized by diagnostic delays, reliance on a single marker, highly subjective clinical judgment, and a lack of personalized assessment and real-time monitoring, which makes early diagnosis difficult and fails to meet rapid and accurate clinical needs.
A sepsis detection system was designed, which integrates data acquisition module, data transmission module, data processing module, user display module and alarm module. It conducts comprehensive analysis of multiple physiological parameters and biomarker data, uses machine learning algorithms to perform real-time risk assessment, and triggers alarms when the risk is high.
It realizes real-time and personalized risk assessment of sepsis, improves diagnostic accuracy, reduces missed diagnosis rate, provides timely treatment recommendations, and improves patient prognosis.
Smart Images
Figure CN120585276A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of sepsis detection, and in particular to a sepsis detection system. Background Art
[0002] Sepsis is a systemic inflammatory response syndrome triggered by infection, often accompanied by multi-organ dysfunction. It is one of the leading causes of death worldwide, particularly in intensive care units (ICUs) and emergency departments, where the mortality rate is very high. Despite significant advances in medical technology in recent years, the early diagnosis and effective treatment of sepsis remain significant challenges.
[0003] Symptoms of sepsis typically include abnormal body temperature, increased heart rate, rapid breathing, and decreased blood pressure. These symptoms are often similar to those of other common infectious diseases (such as influenza and pneumonia). Therefore, clinicians find it difficult to make accurate early judgments when faced with patients with similar symptoms. Existing sepsis diagnostic methods mainly rely on clinical experience and traditional laboratory testing methods, which often have the following shortcomings:
[0004] Diagnostic Delays: The early symptoms of sepsis are vague and similar to those of many other infectious diseases. Existing diagnostic methods, such as blood cultures and imaging studies, often require time to produce results. During this time, a sepsis patient's condition can rapidly deteriorate, increasing the difficulty of treatment and the risk of death.
[0005] Reliance on a single marker: Current sepsis diagnosis mostly relies on a single biomarker, such as white blood cell count (WBC), C-reactive protein (CRP), and procalcitonin (PCT). Although these indicators can reflect the inflammatory response to a certain extent, due to the complexity and variability of sepsis, a single marker often cannot fully reflect the patient's condition changes.
[0006] Clinical judgment is highly subjective: Currently, the diagnosis of sepsis often relies on the clinical judgment of physicians, whose experience and judgment may be limited by individual differences and the complexity of specific cases. This subjective judgment may lead to missed or misdiagnosis of early sepsis, affecting treatment effectiveness.
[0007] Lack of personalized assessment: The course and risk of sepsis vary among individual patients. Current diagnostic methods lack the ability to conduct personalized assessments based on the patient's specific circumstances (such as age, underlying diseases, and immune status), resulting in insufficient early warning for high-risk patients.
[0008] Lack of real-time monitoring: Traditional sepsis diagnostic methods rely on static laboratory data, failing to provide real-time dynamic monitoring. Sepsis often develops rapidly and fluctuates significantly, making it difficult for traditional methods to capture changing trends and provide early warning.
[0009] Therefore, although existing technologies can provide diagnostic information of sepsis to a certain extent, these methods rely on single markers or delayed diagnosis, cannot fully meet the clinical needs for rapid and accurate diagnosis of sepsis, and lack consideration of individual differences among patients.
[0010] Therefore, we urgently need to design a sepsis detection system to solve the above problems. Summary of the Invention
[0011] In one aspect, the present invention provides a sepsis detection system, comprising the following modules:
[0012] Data acquisition module: used to collect patients' physiological parameters, laboratory test data and biomarker data, including but not limited to white blood cell count (WBC), C-reactive protein (CRP), procalcitonin (PCT), interleukin-6, body temperature (T), heart rate (HR), respiratory rate (RR), blood pressure (BP), and pulse oxygen saturation (SPO2);
[0013] The data acquisition module acquires data through the following devices and means: body temperature (T) is monitored in real time by a body temperature sensor (such as a digital thermometer, a wearable temperature sensor, etc.);
[0014] Heart rate HR is obtained through electrocardiogram (ECG) equipment or heart rate monitoring equipment (such as smart bracelets, smart watches, heart rate monitoring belts, etc.);
[0015] Respiratory rate (RR) is obtained through a wearable respiratory monitor or chest strap sensor, and blood pressure (BP) is monitored by a sphygmomanometer (such as a manual sphygmomanometer or an electronic sphygmomanometer).
[0016] Biomarker data including white blood cell count (WBC), C-reactive protein (CRP), procalcitonin (PCT), etc. are obtained through laboratory blood testing equipment (e.g., hematology analyzers, immunoassay analyzers);
[0017] Data transmission module: transmits the collected data to the data processing module, supporting wireless or wired transmission; data processing module: performs comprehensive analysis on the transmitted data, calculates the sepsis risk score, and outputs the risk level;
[0018] User display module: displays the sepsis risk score, risk level and corresponding treatment recommendations based on the results output by the data processing module;
[0019] Alarm module: When the sepsis risk score exceeds the preset threshold, an alarm is automatically triggered and medical staff are notified to ensure timely treatment measures.
[0020] As a preferred technical solution of the present invention, the data acquisition module is used to collect blood test data, including white blood cell count (WBC), C-reactive protein (CRP) and procalcitonin (PCT), and the data processing module calculates the risk score R of sepsis according to the following formula sepsis .
[0021] R sepsis =w1·WBC+w2·CRP+w3·PCT
[0022] R sepsis : Sepsis risk score, indicating the likelihood of a patient developing sepsis; WBC: white blood cell count (unit: cells / μL), used to reflect the severity of infection; CRP: C-reactive protein concentration (unit: mg / L), a marker of inflammatory response in the body; PCT: procalcitonin concentration (unit: ng / mL), used as a biomarker for infectious diseases; w1, w2, w3: weight coefficients obtained through machine learning algorithms or historical data analysis, indicating the degree of influence of different biomarkers on the risk of sepsis.
[0023] As a preferred technical solution of the present invention, the data acquisition module also includes sensors for real-time monitoring of the patient's body temperature (T), heart rate (HR), respiratory rate (RR), and blood pressure (BP), and transmits the collected physiological parameters to the data processing module. The data processing module performs sepsis risk assessment according to the following formula:
[0024] As a preferred technical solution of the present invention, the data acquisition module also includes sensors for real-time monitoring of the patient's body temperature T, heart rate HR, respiratory rate RR, and blood pressure BP, and transmits the collected physiological parameters to the data processing module. The data processing module performs sepsis risk assessment according to the following formula:
[0025]
[0026] The sepsis risk score based on vital sign data represents the sepsis risk inferred from physiological parameters; T is body temperature, obtained in real time through a body temperature sensor; HR is heart rate, obtained through a heart rate monitoring device; RR is respiratory rate, obtained through a respiratory monitoring instrument; BP is blood pressure, obtained through a blood pressure monitoring device; a1, a2, a3, a4 are coefficients determined by clinical data, indicating the contribution of different physiological parameters to the risk of sepsis. GSC is the Glasgow Coma Scale.
[0027] As a preferred technical solution of the present invention, the data processing module calculates the total risk score of sepsis by integrating the biomarker data and the physical sign data. The formula is as follows:
[0028]
[0029] The total sepsis risk score represents the risk of sepsis after comprehensive consideration of biomarkers and physiological parameters; R sepsis : Sepsis risk score calculated from biomarkers; Sepsis risk score calculated from physical sign data; α: weight coefficient (0<α<1), indicating the relative importance of biomarkers in sepsis assessment, usually derived from clinical data or experiments, R mottling The mottling score is based on the extension of the mottling area on the legs, with a score of 0 indicating no mottling; a score of 1 indicating the mottling area is located in the center of the knee joint; a score of 2 indicating the mottling area does not extend beyond the upper edge of the kneecap; a score of 3 indicating the mottling area does not extend beyond the mid-thigh; a score of 4 indicating the mottling area does not extend beyond the inguinal fold; and a score of 5 indicating the mottling area extends beyond the inguinal fold area.
[0030] As a preferred technical solution of the present invention, the data processing module will also modify the risk score based on the patient's clinical history data (such as chronic diseases, drug allergy history, etc.), and the modification formula is:
[0031]
[0032] The modified sepsis risk score represents the final risk after combining all data and clinical history. Total sepsis risk score; γ: clinical data correction coefficient, which indicates the impact of clinical data on sepsis risk assessment; D: correction factor generated based on the patient's clinical history data, which indicates the impact of individual patient differences on sepsis risk.
[0033] As a preferred technical solution of the present invention, the data processing module trains historical clinical data through a machine learning algorithm to generate a sepsis risk assessment model, and generates a sepsis risk score based on the input real-time data. The formula is:
[0034]
[0035] F sepsis : Sepsis risk score, which represents the prediction result of real-time data based on the model trained based on historical data; β i : The coefficient obtained through machine learning algorithm training represents the impact of different features on the risk of sepsis; x i: Input features, usually the patient's physiological data, laboratory test data, etc.; n: The number of input features, which indicates the number of features used in the model.
[0036] As a preferred technical solution of the present invention, the data processing module converts the sepsis risk score Converted into risk level, the judgment formula is:
[0037]
[0038] The final sepsis risk score is used to determine the risk level of sepsis. This formula is used to convert the sepsis risk score into a "low risk," "intermediate risk," or "high risk" level so that doctors can decide on treatment options based on the risk level.
[0039] As a preferred technical solution of the present invention, the alarm module triggers an alarm based on the sepsis risk score and its changing trend. The warning triggering conditions are:
[0040]
[0041] The change value of the sepsis risk score indicates the change in the risk score compared to the previous moment. δ: The set alarm threshold. When the change in the sepsis risk score reaches this threshold, an alarm is triggered. AlertTriggered: Whether to trigger an alarm. If An alarm is triggered.
[0042] As a preferred technical solution of the present invention, the alarm module can provide multiple alarm modes, including sound, visual, and mobile phone push notifications. The alarm mode can be configured and adjusted according to user needs to ensure that medical staff are notified in a timely manner when the risk of sepsis is high.
[0043] As a preferred technical solution of the present invention, the data processing module has self-learning capabilities and can continuously optimize the sepsis assessment model based on new patient data to improve the accuracy of sepsis risk assessment. The self-learning formula is as follows:
[0044]
[0045] θ t : Model parameters (such as weight coefficients), representing the parameters of the model at the current moment; η: learning rate, controlling the step size of the model update; The gradient of the cost function indicates the direction and magnitude of the model parameter update, and the gradient descent method is used to minimize the cost function.
[0046] Beneficial effects
[0047] The sepsis detection system of the present invention provides a comprehensive and accurate sepsis risk assessment method by combining multiple patient physiological parameters, laboratory test data, and biomarkers. The system monitors patients' vital signs in real time and, combined with blood test results, uses intelligent algorithms for comprehensive analysis, providing clinicians with scientific diagnostic evidence. This multi-dimensional data integration and analysis method significantly improves the diagnostic accuracy of sepsis and enables earlier identification of potential sepsis risks, thereby providing patients with more timely intervention measures.
[0048] In addition, the present invention continuously optimizes clinical data through a self-learning algorithm, and can dynamically adjust the model according to newly collected patient data to improve the accuracy of sepsis risk assessment. The system can not only provide personalized risk assessments, but also trigger real-time alarms when the risk of sepsis is high, helping doctors make timely treatment decisions. Compared with existing technologies, the present invention has significant advantages in real-time monitoring, dynamic evaluation and personalized diagnosis, which can effectively reduce the missed diagnosis rate of sepsis, improve patient prognosis, and enhance treatment effects. Integrate clinical symptoms and signs, combined with the performance of the compensatory period of shock, skin, consciousness, urine volume, etc., in the integration of vital signs: T, RR, SPO2, HR, ABP, urine volume, whether it is spotted, limb skin temperature, etc., laboratory tests need to be combined with lactate, central venous SCVO2,
[0049] CO2-GAP, blood sugar, and infection indicators such as neutrophil count, neutrophil rate, IL-6, etc., as well as the obtained etiological results, location, drug resistance, etc. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0051] Figure 1 Figure 2 is a system block diagram of the sepsis detection system. DETAILED DESCRIPTION
[0052] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.
[0053] The following combines multiple embodiments, Figure 1 , the specific implementation methods of the present invention are described in detail.
[0054] The present invention provides a sepsis detection system that integrates physiological parameters, biomarker data, and machine learning models to perform real-time sepsis detection and risk assessment, and generate relevant treatment recommendations. Based on a patient's vital signs, biomarker data, and clinical history, the system provides accurate sepsis risk assessments, assisting physicians in early diagnosis and treatment. The following describes specific implementations of the invention, detailing data acquisition, data processing, and the application of algorithmic formulas, along with several specific examples.
[0055] 1. Composition of sepsis detection system
[0056] The sepsis detection system of this embodiment includes the following five modules:
[0057] Data acquisition module: This module collects the patient's physiological parameters (such as body temperature, heart rate, respiratory rate, blood pressure, etc.) and biomarkers (such as white blood cell count, C-reactive protein, procalcitonin, etc.) in real time through multiple sensors and laboratory equipment.
[0058] For example:
[0059] Body temperature is obtained through wearable temperature sensors;
[0060] Heart rate is obtained through a heart rate monitor belt or smart watch;
[0061] Respiratory rate is obtained through a chest strap sensor or a wearable respiratory monitoring device;
[0062] Blood pressure was obtained using an electronic sphygmomanometer;
[0063] Biomarkers (WBC, CRP, PCT) were detected by hematology analyzer.
[0064] Data transmission module: This module is used to transmit the collected data to the data processing module via wireless communication (such as Wi-Fi or Bluetooth) or wired method (such as USB), ensuring that the data can be transmitted in real time and provided for subsequent analysis.
[0065] Data processing module: After receiving the data, this module processes the collected data based on the algorithm model, calculates the sepsis risk score, and determines the risk level of sepsis based on the calculation results.
[0066] The formula for calculating the sepsis risk score is as follows:
[0067] R sepsis =w1·WBC+w2·CRP+w3·PCT
[0068] Among them, w1, w2, and w3 are weight coefficients obtained through machine learning algorithms, indicating the contribution of each biomarker to sepsis.
[0069] Risk assessment is performed using a formula based on physiological parameters (body temperature, heart rate, respiratory rate, blood pressure):
[0070]
[0071] Among them, a1, a2, a3, a4, and a5 are coefficients determined by clinical data, representing the impact of physiological parameters on the risk of sepsis.
[0072] The total risk score of sepsis was calculated comprehensively:
[0073]
[0074] Among them, α is the weight coefficient, which determines the relative importance of biomarkers and physiological parameters in risk assessment.
[0075] User display module: This module is used to visualize the results of the processing module and show the risk score and risk level of sepsis to users (such as doctors).
[0076] The information displayed includes: Risk score (such as ); risk level (low, medium, high); treatment recommendations (such as whether further testing is needed, whether immediate intervention is needed, etc.).
[0077] Alarm module: When the sepsis risk score exceeds the set threshold, the alarm module triggers an early warning according to the following formula:
[0078]
[0079] in, is the change value of sepsis risk score, and δ is the set alarm threshold.
[0080] 2. Application of algorithm model;
[0081] In the data processing module, the sepsis risk score is calculated using an algorithm model based on machine learning.
[0082] The following is the detailed calculation process: Weighted summation of biomarker data: Using blood test data (such as white blood cell count, C-reactive protein and procalcitonin), the sepsis risk score is calculated according to the following formula:
[0083] R sepsis =w1·WBC+w2·CRP+w3·PCT
[0084] Among them, w1, w2, and w3 are coefficients obtained through training of historical clinical data, indicating the contribution of different biomarkers to the risk of sepsis.
[0085] Weighted summation of vital sign data: Using real-time monitoring of physiological parameters such as body temperature, heart rate, respiratory rate, and blood pressure, the risk of sepsis is assessed according to the following formula:
[0086]
[0087] Among them, a1, a2, a3, and a4 are coefficients obtained from clinical data analysis.
[0088] Comprehensive risk assessment: Finally, the biomarker data and physiological parameter data are combined to obtain the overall risk score of sepsis:
[0089]
[0090] The total sepsis risk score represents the risk of sepsis after comprehensive consideration of biomarkers and physiological parameters; R sepsis : Sepsis risk score calculated from biomarkers; Sepsis risk score calculated from physical sign data; α: weight coefficient (0<α<1), indicating the relative importance of biomarkers in sepsis assessment, usually derived from clinical data or experiments, R mottling The mottling score is based on the extension of the mottling area on the legs, with a score of 0 indicating no mottling; a score of 1 indicating the mottling area is located in the center of the knee joint; a score of 2 indicating the mottling area does not extend beyond the upper edge of the kneecap; a score of 3 indicating the mottling area does not extend beyond the mid-thigh; a score of 4 indicating the mottling area does not extend beyond the inguinal fold; and a score of 5 indicating the mottling area extends beyond the inguinal fold area.
[0091] 3. Examples
[0092] Example 1: Sepsis risk assessment based on blood testing;
[0093] Data collection: The collected blood data included: white blood cell count: 10,000 cells / μL; C-reactive protein: 30 mg / L; procalcitonin: 1.5 ng / mL.
[0094] Data processing: Set weight coefficients w1 = 0.4, w2 = 0.3, w3 = 0.3. Calculate risk score according to the formula: R sepsis =0.4·10000+0.3·30+0.3·1.5=4000+9+0.45=4009.45.
[0095] Output: Based on the calculated risk score, the system assesses the patient's sepsis risk as "high risk".
[0096] The alarm module is triggered to notify medical staff.
[0097] Example 2: Sepsis risk assessment based on physiological parameters;
[0098] Data collection: The collected physiological parameter data include: body temperature: 38.5℃; heart rate: 110bpm; respiratory rate: 24 times per minute; blood pressure: 90 / 60mmHg.
[0099] Data processing: Set coefficients a1 = 2.5, a2 = 1.5, a3 = 2.0, a4 = -1.0.
[0100] The risk score is calculated according to the formula:
[0101]
[0102] Output result: Based on the calculated physical sign risk score, the system assesses the patient's sepsis risk as "medium risk".
[0103] Example 3: Comprehensive assessment of sepsis risk;
[0104] Data collection: Biomarker data: White blood cell count: 12000 cells / μL;
[0105] C-reactive protein: 40 mg / L; procalcitonin: 2.5 ng / mL.
[0106] Physiological parameter data: body temperature: 38.7℃; heart rate: 120bpm; respiratory rate: 22 times / minute; blood pressure: 85 / 55mmHg.
[0107] Data processing: Calculation of risk scores from biomarkers and physiological parameters:
[0108] R sepsis =0.4·12000+0.3·40+0.3·2.5=4800+12+0.75=4812.75;
[0109]
[0110] Output: The system assesses the risk of sepsis as "high risk" based on the calculated total risk score.
[0111] The alarm module is triggered to notify medical staff to take appropriate intervention measures.
[0112] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.
Claims
1. A sepsis detection system, characterized in that: The system includes the following modules: Data collection module: used to collect patient symptoms and signs, physiological parameters, laboratory test data and biomarker data, including but not limited to white blood cell count (WBC), C-reactive protein (CRP), procalcitonin (PCT), interleukin-6, body temperature (T), heart rate (HR), respiratory rate (RR), blood pressure (BP), and pulse oxygen saturation (SPO2); The data acquisition module acquires data through the following devices and means: body temperature T is monitored in real time by a body temperature sensor to monitor the changes in the patient's body temperature; Heart rate HR is obtained through electrocardiogram (ECG) equipment or heart rate monitoring equipment; Respiratory rate (RR) is obtained through a wearable respiratory monitor or chest strap sensor, and blood pressure (BP) is monitored by a sphygmomanometer. Biomarker data including white blood cell count (WBC), C-reactive protein (CRP), and procalcitonin (PCT) were obtained through laboratory blood testing equipment; Data transmission module: transmits the collected data to the data processing module, supporting wireless or wired transmission; Data processing module: performs comprehensive analysis on the transmitted data, calculates the risk score of sepsis, and outputs the risk level; User display module: displays the risk score, risk level and corresponding treatment recommendations of sepsis based on the results output by the data processing module; Alarm module: When the sepsis risk score exceeds the preset threshold, an alarm is automatically triggered and medical staff are notified so that timely treatment measures can be taken.
2. The sepsis detection system according to claim 1, characterized in that: The data acquisition module is used to collect blood test data, including white blood cell count (WBC), C-reactive protein (CRP), and procalcitonin (PCT). The data processing module calculates the risk score R of sepsis according to the following formula: sepsis , R sepsis =w1·WBC+w2·CRP+w3·PCT where R sepsis : Sepsis risk score, indicating the likelihood of a patient developing sepsis; WBC: white blood cell count, used to reflect the severity of infection; CRP: C-reactive protein concentration, a marker of inflammatory response in the body; PCT: procalcitonin concentration, used as a biomarker for infectious diseases; w1, w2, w3: weight coefficients obtained through machine learning algorithms or historical data analysis, indicating the degree of influence of different biomarkers on sepsis risk.
3. The sepsis detection system according to claim 1, wherein: The data acquisition module also includes sensors for real-time monitoring of the patient's body temperature T, heart rate HR, respiratory rate RR, and blood pressure BP, and transmits the collected physiological parameters to the data processing module, which performs sepsis risk assessment according to the following formula: The sepsis risk score based on vital sign data represents the sepsis risk inferred from physiological parameters; T is body temperature, obtained in real time through a body temperature sensor; HR is heart rate, obtained through a heart rate monitoring device; RR is respiratory rate, obtained through a respiratory monitoring instrument; BP is blood pressure, obtained through a blood pressure monitoring device; a1, a2, a3, a4 are coefficients determined by clinical data, indicating the contribution of different physiological parameters to the risk of sepsis. GSC is the Glasgow Coma Scale.
4. The sepsis detection system according to claim 3, characterized in that: The data processing module calculates the total risk score of sepsis by integrating the biomarker data and the physical sign data. The formula is as follows: The total sepsis risk score represents the risk of sepsis after comprehensive consideration of biomarkers and physiological parameters; R sepsis : Sepsis risk score calculated from biomarkers; Sepsis risk score calculated from physical sign data; α: weight coefficient (0<α<1), indicating the relative importance of biomarkers in sepsis assessment, usually derived from clinical data or experiments, R mottling The mottling score is based on the extension of the mottling area on the legs, with a score of 0 indicating no mottling; a score of 1 indicating the mottling area is located in the center of the knee joint; a score of 2 indicating the mottling area does not extend beyond the upper edge of the kneecap; a score of 3 indicating the mottling area does not extend beyond the mid-thigh; a score of 4 indicating the mottling area does not extend beyond the inguinal fold; and a score of 5 indicating the mottling area extends beyond the inguinal fold area.
5. The sepsis detection system according to claim 1, wherein: The data processing module will also modify the risk score based on the patient's clinical history data. The modification formula is: The modified sepsis risk score represents the final risk after combining all data and clinical history. Total sepsis risk score; γ: clinical data correction coefficient, which indicates the impact of clinical data on sepsis risk assessment; D: correction factor generated based on the patient's clinical history data, which indicates the impact of individual patient differences on sepsis risk.
6. The sepsis detection system according to claim 1, wherein: The data processing module trains historical clinical data through a machine learning algorithm to generate a sepsis risk assessment model, and generates a sepsis risk score based on the input real-time data. The formula is: F sepsis : Sepsis risk score, which represents the prediction result of the model trained based on historical data for real-time data; β i : The coefficient obtained by machine learning algorithm training represents the impact of different features on the risk of sepsis; x i : Input features, usually the patient's physiological data, laboratory test data, etc.; n: The number of input features, which indicates the number of features used in the model.
7. The sepsis detection system according to claim 1, wherein: The data processing module converts the sepsis risk score Converted into risk level, the judgment formula is: The final sepsis risk score is used to determine the risk level of sepsis. This formula is used to convert the sepsis risk score into a "low risk," "intermediate risk," or "high risk" level so that doctors can decide on treatment options based on the risk level.
8. The sepsis detection system according to claim 1, wherein: The alarm module triggers an alarm based on the sepsis risk score and its changing trend. The early warning trigger conditions are: The change value of the sepsis risk score indicates the change in the risk score compared to the previous moment. δ: The set alarm threshold. When the change in the sepsis risk score reaches this threshold, an alarm is triggered. AlertTriggered: Whether to trigger an alarm. If An alarm is triggered.
9. The sepsis detection system according to claim 1, wherein: The alarm module can provide various alarm modes, including sound, visual, and mobile phone push notifications, and the alarm mode can be configured and adjusted according to user needs to promptly notify medical staff when the risk of sepsis is high.
10. The sepsis detection system according to claim 1, wherein: The data processing module has self-learning capabilities and continuously optimizes the sepsis assessment model based on new patient data to improve the accuracy of sepsis risk assessment. The self-learning formula is as follows: θ t : Model parameters, representing the parameters of the model at the current moment; η: learning rate, controlling the step size of the model update; The gradient of the cost function indicates the direction and magnitude of the model parameter update, and the gradient descent method is used to minimize the cost function.