Systems and methods for screening and predicting sepsis
Hemodynamic data from arterial blood pressure waveforms are used in computational models to objectively screen for and predict sepsis, overcoming the limitations of current diagnostic methods by providing timely and accurate risk or probability scores.
Patent Information
- Application Number
- JP2025501856
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2022-07-15
- Filing Date
- 2023-07-12
- Publication Date
- 2025-08-01
AI Technical Summary
Current methods for diagnosing sepsis lack clear correlation with pathogenic infections, relying on subjective criteria like SIRS, making early and accurate diagnosis challenging, especially in emergency settings.
A system and method using hemodynamic data from arterial blood pressure waveforms to extract features, which are input into computational models for sepsis screening or probability prediction, utilizing equations and machine-learned models to generate scores indicating risk or probability of sepsis.
Enables early and objective sepsis screening and prediction without relying on SIRS criteria, facilitating timely intervention and reducing organ damage by providing actionable risk or probability scores.
Smart Images
Figure 2025524811000001_ABST
Abstract
Description
Technical Field
[0001] Cross - reference to Related Applications This application claims the benefit of priority of U.S. Provisional Patent Application No. 63 / 389,577, filed on July 15, 2022, entitled "Systems and Methods for Screening and Predicting Sepsis", the disclosure of which is incorporated herein by reference.
[0002] The present disclosure generally relates to systems and methods for screening and predicting sepsis, and more particularly, to systems and methods for screening and predicting sepsis using hemodynamic data.
Background Art
[0003] Sepsis is a condition of the body when the body is over - reacting to a pathogenic infection. Sepsis can lead to tissue damage, organ failure, and death. Prompt diagnosis and treatment are required to prevent progression.
[0004] A patient is diagnosed with sepsis when they exhibit a set of signs and symptoms associated with a septic response. Most pathogenic infections do not result in sepsis, and there is no clear correlation between a pathogenic infection and the onset of sepsis, so an infection alone is not sufficient to diagnose sepsis. One way to diagnose sepsis is the assessment of the systemic inflammatory response syndrome (SIRS). When at least two of the following criteria are met, namely, (1) fever or hypothermia, (2) elevated heart rate (tachycardia), (3) elevated respiratory rate (tachypnea), and (4) low or high white blood cell count (leukocytosis or leukopenia) or a high ratio of band cells (bandemia), it is diagnosed as positive.
[0005] When there are signs of organ failure, sepsis progresses to severe sepsis. Organ failure can be evaluated by the Sequential Organ Failure Assessment (SOFA). SOFA assesses pulmonary respiration, blood coagulation, liver function, brain function, cardiovascular function, and renal function, and a higher SOFA score indicates more severe organ failure. A high SOFA score associated with a sepsis diagnosis indicates the need to promptly treat inflammation and infection to reduce organ damage and prevent death.
Summary of the Invention
Means for Solving the Problems
[0006] A system and method for assessing sepsis can include the use of a sensor to generate waveform data corresponding to a signal corresponding to, proportional to, or derived from arterial blood pressure. A set of features of hemodynamic data can be extracted from the waveform data. The extracted set of features of hemodynamic data and / or other clinical information including, but not limited to, patient demographics, vital signs, and clinical test results can be utilized in a computational model for screening for sepsis or predicting the probability that a patient has sepsis.
[0007] In some implementations, the computational method is for screening for sepsis. The method includes receiving, from a sensor attached to a patient, waveform data corresponding to a signal corresponding to, proportional to, or derived from arterial blood pressure. The method includes extracting a set of features of hemodynamic data from the waveform data. The method includes inputting the extracted set of features of hemodynamic data into a predictive computational model to obtain a sepsis screening score. The predictive computational model is trained to screen for sepsis using the extracted set of features of hemodynamic data.
[0008] In some implementations, the set of features of the extracted hemodynamic data includes heart rate, the kurtosis of the pressure distribution, and the sample entropy of the time to reach systolic MAP.
[0009] In some implementations, the set of features of the extracted hemodynamic data includes heart rate, arterial tone factor, sample entropy of decay area, dynamic arterial elastance, and approximate entropy of systolic time.
[0010] In some implementations, the prediction calculation model utilizes an equation to obtain a sepsis screening score.
[0011] In some implementations, the equation is
Number
[0012] In some implementations, the equation is
Number
[0013] In some implementations, the sepsis screening score indicates the risk of developing sepsis. The method further includes the step of further evaluating the patient regarding complications of sepsis.
[0014] In some embodiments, the sepsis screening score indicates the risk of developing sepsis. The method further includes monitoring the patient for sepsis complications over a particular time period.
[0015] In some embodiments, the computational model is for predicting the probability that a patient will develop sepsis. The method includes receiving, from a sensor attached to the patient, waveform data corresponding to, proportional to, or derived from arterial blood pressure. The method includes extracting a set of features of the hemodynamic data from the waveform data. The method includes inputting the set of features of the extracted hemodynamic data into a predictive computational model to obtain a sepsis probability score. The predictive computational model is trained to predict sepsis using the set of features of the extracted hemodynamic data.
[0016] In some embodiments, the set of features of the extracted hemodynamic data includes diastolic blood pressure, heart rate, entropy of the interbeat interval, time from reaching systolic MAP to the dicrotic notch, and entropy of the standard deviation of the decay phase.
[0017] In some embodiments, the set of features of the extracted hemodynamic data includes heart rate, approximate entropy of the blood pressure waveform, sample entropy of the systolic area, approximate entropy of the systolic time, and approximate entropy of the systolic decay time.
[0018] In some embodiments, the formula is
Number
[0019] In some implementations, the formula is [Number] where hr is the heart rate, ApEnV is the approximate entropy of the blood pressure waveform, areaSampEn is the sample entropy of the systolic area, tSysApEn is the approximate entropy of the systolic time, and tDecApEn is the approximate entropy of the systolic decay time.
[0020] In some implementations, the sepsis probability score indicates that the patient has sepsis. The method further includes the step of further evaluating the patient with respect to sepsis complications to confirm the probability score.
[0021] In some implementations, the sepsis probability score indicates that the patient has sepsis. The method further includes the step of treating the patient to treat sepsis.
[0022] In some implementations, the method further includes the step of sensing arterial blood pressure using a sensor.
[0023] In some implementations, the sensor is an intra-arterial catheter and a disposable blood pressure transducer, a pressurized finger cuff and a light sensor, or an applanation tonometer.
[0024] In some implementations, the set of features of the extracted hemodynamic data includes heart rate, respiratory rate, cardiac output, stroke volume, stroke volume variability, vascular tone, contractility, afterload, systemic vascular resistance, systolic blood pressure, diastolic blood pressure, mean arterial pressure (MAP), kurtosis of the pressure distribution, left ventricular ejection time, time from reaching systolic MAP to the dicrotic notch, sample entropy of the time to reach systolic MAP, entropy of the interbeat interval, entropy of the standard deviation of the decay phase, [Number] , [Number] , [Number] , [Number] or [Number] includes at least one of them.
[0025] In some implementations, the method further includes the step of inputting the patient's clinical information into the model.
[0026] In some implementations, the patient's clinical information includes at least one of the patient's demographics, the patient's vital signs, and the patient's laboratory results.
[0027] In some implementations, the prediction calculation model is a regression-based model, a classification-based model, or an ensemble model.
[0028] In some implementations, the patient monitoring system is for screening sepsis based on the captured waveform data. The patient monitoring system includes a sensor and a computing and processing system operably connected to the sensor. The computing and processing system includes a processing system and a memory system including one or more applications, where the one or more applications cause the processor system to receive waveform data corresponding to, proportional to, or derived from an arterial blood pressure from the sensor attached to the patient, extract a set of features of the hemodynamic data from the waveform data, and input the set of features of the extracted hemodynamic data into a predictive calculation model to obtain a sepsis screening score. The predictive calculation model is trained to screen for sepsis using the set of features of the extracted hemodynamic data.
[0029] In some implementations, the set of features of the extracted hemodynamic data includes heart rate, kurtosis of the pressure distribution, and sample entropy of the time to reach systolic MAP.
[0030] In some implementations, the set of features of the extracted hemodynamic data includes heart rate, arterial stiffness factor, sample entropy of the decay area, dynamic arterial elastance, and approximate entropy of the systolic time.
[0031] In some implementations, the predictive calculation model utilizes an equation to obtain a sepsis screening score.
[0032] In some implementations, the equation is
Equation
[0033] In some implementations, the formula is [Number] where hr is the heart rate, avgK is the arterial tone factor, decAreaSampEn is the sample entropy of the attenuation area, dynEa is the dynamic arterial elastance, and tSysApEn is the approximate entropy of the systolic time.
[0034] In some implementations, one or more applications are further configured to instruct a processor system to display a sepsis screening score on a monitor operably connected to a computing system.
[0035] In some implementations, the sepsis screening score indicates the risk of developing sepsis. One or more applications are further configured to instruct a processor system to provide an alert indicating the risk when it is determined that the sepsis screening score indicates the risk of developing sepsis.
[0036] In some implementations, a patient monitoring system is for predicting whether a patient has sepsis based on the captured arterial pressure. The patient monitoring system includes a sensor and a computing system operably connected to the sensor. The computing system includes a processor system and a memory system including one or more applications, wherein the one or more applications are configured to instruct the processor system to receive waveform data corresponding to, proportional to, or derived from a signal corresponding to the arterial blood pressure from a sensor attached to the patient, extract a set of features of the hemodynamic data from the waveform data, and input the set of features of the extracted hemodynamic data into a prediction calculation model to obtain a sepsis probability score. The prediction calculation model is trained to predict sepsis using the set of features of the extracted hemodynamic data.
[0037] In some implementations, the set of features of the extracted hemodynamic data includes diastolic blood pressure, heart rate, entropy of the interbeat interval, time from reaching systolic MAP to the dicrotic notch, and entropy of the standard deviation of the decay phase.
[0038] In some implementations, the set of features of the extracted hemodynamic data includes heart rate, approximate entropy of the blood pressure waveform, sample entropy of the systolic area, approximate entropy of the systolic time, and approximate entropy of the systolic decay time.
[0039] In some implementations, the prediction calculation model utilizes an equation to obtain a sepsis probability score.
[0040] In some implementations, the equation is
Equation
[0041] In some implementations, the equation is
Equation
[0042] In some implementations, one or more applications are further configured to instruct a processor system to display a sepsis probability score on a monitor operably connected to the computational processing system.
[0043] In some implementations, the sepsis probability score indicates the risk of developing sepsis. One or more applications are further configured to instruct a processor system to provide an alert indicating that the patient has sepsis if the sepsis probability score is determined to indicate that the patient has sepsis.
[0044] In some implementations, the sensor is an intra-arterial catheter and disposable blood pressure transducer, a pressurized finger cuff and optical sensor, or a applanation tonometer.
[0045] In some implementations, the set of features of the extracted hemodynamic data includes heart rate, respiratory rate, cardiac output, stroke volume, stroke volume variability, vascular tone, contractility, afterload, systemic vascular resistance, systolic blood pressure, diastolic blood pressure, mean arterial pressure (MAP), skewness of the pressure distribution, left ventricular ejection time, time from reaching systolic MAP to incisura, sample entropy of the time to reach systolic MAP, entropy of the interbeat interval, entropy of the standard deviation of the decay phase,
Number
Number
Number
Number
Number
[0046] In some implementations, one or more applications are further configured to instruct a processor system to input clinical information of a patient into the model.
[0047] In some implementations, the patient's clinical information includes at least one of the patient's demographics, the patient's vital signs, and the patient's test results.
[0048] In some implementations, the predictive calculation model is a regression-based model, a classification-based model, or an ensemble model.
[0049] The description and claims will be more fully understood by reference to the following figures and data graphs, which are presented as examples of the disclosure and should not be construed as a complete description or interpretation of the scope of the disclosure.
Brief Description of the Drawings
[0050]
Figure 1
Figure 2
Figure 3
Modes for Carrying Out the Invention
[0051] The present disclosure details a system and method for evaluating sepsis using hemodynamic data derived from a continuous blood pressure sensor. Characteristics of the hemodynamic data are derived from blood pressure waveforms and can be utilized to screen for sepsis and / or predict the probability of sepsis. Accordingly, the system and method can screen for early identification of sepsis in a patient or can predict the probability that a patient has sepsis. In some implementations, the characteristics of the hemodynamic data are utilized in a trained computational model for screening for sepsis or predicting the probability of sepsis. In some implementations, the characteristics of the hemodynamic data are utilized in an equation for calculating a score for screening for sepsis or predicting the probability of sepsis.
[0052] The novel system and method provide screening for sepsis and / or prediction of the probability of sepsis utilizing hemodynamic data. Accordingly, sepsis can be initially screened for and / or diagnosed without analysis of SIRS criteria, which can be useful in situations where analysis of SIRS criteria is not readily available, such as in an emergency department. Depending on the situation, a patient may be screened for potential risk of developing sepsis and, when a high risk is indicated, the patient is monitored and / or further evaluated for sepsis. Depending on the situation, a patient is predicted to have sepsis and subsequent confirmatory analysis and / or treatment of sepsis is performed.
[0053] A method for screening for sepsis or predicting the probability of sepsis, which can be implemented as a computational process, is shown in FIG. 1. Method 100 measures waveform data corresponding to, proportional to, or derived from arterial blood pressure (101). Any method for measuring continuous arterial blood pressure can be utilized, including non-invasive and invasive methods. Thus, blood pressure can be measured by an intra-arterial catheter with a disposable blood pressure transducer (e.g., an in-artery blood pressure catheter), by a pressurized finger cuff and an optical sensor (e.g., volume clamp method), by applanation tonometry, or by any other means for obtaining a signal proportional to or derived from an arterial pressure waveform or arterial blood pressure. [[ID=……]] [[ID=……]]
[0054] [[ID=……]] Also, method 100 extracts features of hemodynamic data from the waveform data (103). Various features of hemodynamic data are useful for screening for sepsis or predicting the probability of sepsis. Generally, any feature of hemodynamic data that can provide predictive ability can be utilized. Some features of hemodynamic data are known to provide predictive ability. Features of hemodynamic data that can be extracted and utilized for predicting or screening for sepsis include heart rate, respiratory rate, cardiac output, stroke volume, stroke volume variability, vascular tone, contractility, afterload, systemic vascular resistance, systolic blood pressure, diastolic blood pressure, mean arterial pressure (MAP), kurtosis of the pressure distribution, left ventricular ejection time, time from reaching systolic MAP to the dicrotic notch, sample entropy of the time to reach systolic MAP, entropy of the inter-beat interval, entropy of the standard deviation of the decay phase, [[ID=……]] [[ID=……]]
Number
Number
Number
[0055] Furthermore, method 100 screens for sepsis or predicts the probability of sepsis (105) using the characteristics of the hemodynamic data extracted from the measured waveform data. In some implementations, the extracted characteristics are input into a prediction calculation model, and the model provides a result indicating the risk or probability of developing or having sepsis. In some implementations, the extracted characteristics are input into an equation, and the equation provides a result indicating the risk or probability of developing or having sepsis. In some implementations, the patient's clinical information is input into the model. The clinical information may include, but is not limited to, the patient's demographics, the patient's vital signs, and the patient's clinical test results.
[0056] To screen for sepsis or predict the probability of sepsis, a computational model can be trained with hemodynamic data collected from a cohort of patients diagnosed with sepsis. The hemodynamic data for each patient can be associated with the patient's sepsis diagnosis for training the model. Various computational models can be utilized, including but not limited to regression-based models or classification-based models. Regression-based models can include, but are not limited to, LASSO regression, ridge regression, k-nearest neighbors, elastic net, least angle regression (LAR), and random forest regression. Classification-based models can include, but are not limited to, logistic regression, support vector machine (SVM), decision tree, random forest, and naive Bayes. In some implementations, the model is regularized. In some implementations, the model can be ensembled from multiple models of one or more of the model types listed above.
[0057] To screen for sepsis or predict the probability of sepsis, an equation can be developed using hemodynamic data collected from a cohort of patients diagnosed with sepsis. Weights can be applied to the various hemodynamic data features within the equation to obtain a score that provides a diagnostic indicator for sepsis.
[0058] In one example, a machine-learned model was developed to screen for the identification of early sepsis using features extracted from arterial blood pressure waveforms. The machine-learned model developed an equation that utilized the following features: heart rate (hr), kurtosis (kurt) of blood pressure distribution, and sample entropy (sampEn) of the time to reach systolic MAP. In a particular implementation, the equation is calculated as follows.
Equation
[0059] In another example, a machine - learned model was developed to screen for the identification of early sepsis using features extracted from arterial blood pressure waveforms. The machine - learned model developed an equation that utilizes the following features: heart rate (hr), arterial stiffness factor (avgK), sample entropy of the attenuation area (decAreaSampEn), dynamic arterial elastance (dynEa), and approximate entropy of the systolic time (tSysApEn). In a particular implementation, the equation is calculated as follows.
Number
[0060] The screening score is given by a score in the range from 0 to 100 and indicates the early risk of developing sepsis. The higher the score, the higher the risk of developing sepsis. The selected features, feature weights, and score scaling are provided as examples for obtaining a sepsis screening score. Thus, the selected features, feature weights, and score scaling can be modified as understood in the art.
[0061] In some implementations, when a patient's calculated screening score indicates a risk of developing sepsis, the patient is screened further for sepsis complications. Further screening can include (but is not limited to) assessment of systemic inflammatory response syndrome (SIRS) criteria, blood lactate concentration, blood culture assessment for bacterial infections, assessment of organ function, and calculation of the sequential organ failure assessment (SOFA) score. In some implementations, when a patient's calculated screening score indicates a risk of developing sepsis, the patient is monitored by a clinician for a specific time period.
[0062] In one example, a machine - learned model was developed to predict the probability of sepsis using features extracted from arterial blood pressure waveforms. The machine - learned model developed an equation that utilizes the following features, namely, heart rate (hr), diastolic blood pressure (Dia), time from reaching systolic MAP to the dicrotic notch (timeMAP), entropy of the inter - beat interval (enIBI), entropy of the standard deviation of the decay phase (enDecay). In a particular implementation, the equation is calculated as follows.
Number
[0063] In another example, a machine - learned model was developed to predict the probability of sepsis using features extracted from arterial blood pressure waveforms. The machine - learned model developed an equation that utilizes the following features, namely, heart rate (hr), approximate entropy of the blood pressure waveform (ApEnV), sample entropy of the systolic area (areaSampEn), approximate entropy of the systolic time (tSysApEn), and approximate entropy of the systolic decay time (tDecApEn). In a particular implementation, the equation is calculated as follows.
Number
[0064] The probability score indicates the probability that an individual has sepsis, given as a percentage in the range from 0% to 100%. The higher the percentage, the greater the likelihood of having sepsis. The selected features, feature weights, and score scaling are provided as examples of obtaining a sepsis probability score. Thus, the selected features, feature weights, and score scaling can be modified as understood in the art.
[0065] In some embodiments, when a patient's calculated probability score indicates a high probability of sepsis, the patient is diagnosed with sepsis. In some embodiments, when a patient's calculated probability score indicates a high probability of sepsis, the patient is further screened to confirm the score result. Further screening may include, but is not limited to, assessment of systemic inflammatory response syndrome (SIRS) criteria, assessment of organ function, and calculation of the sequential organ failure assessment (SOFA) score. When a patient's calculated probability score indicates a high probability of sepsis, the patient is treated for sepsis. Treatment for sepsis includes, but is not limited to, administration of antibiotics, administration of intravenous fluids, administration of vasopressors, and surgery to remove abscesses, infected tissue, or necrotic tissue.
[0066] Specific examples of methods for screening for sepsis or predicting the probability of sepsis have been described above, but one of ordinary skill in the art will understand that the various steps of the method can be performed in different orders and that certain steps may be optional depending on the particular embodiment. Thus, it should be apparent that the various steps of the method can be appropriately used according to the requirements of a particular application. Further, any of the various methods for screening for sepsis or predicting the probability of sepsis suitable for the requirements of a given application can be utilized in various embodiments.
[0067] Feature Selection As described in the previous section, hemodynamic data, and / or other clinical information including, but not limited to, patient demographics, vital signs, and clinical test results are used as features for constructing a computational model that is subsequently used to screen for sepsis or predict the probability of sepsis. The features used to train the model can be selected in several ways. In some situations, the features are determined by which data provides a strong correlation with a sepsis diagnosis. In some situations, the features are determined using a computational model that can determine which feature or combination of features provides good predictive ability.
[0068] Features can be identified and / or selected in several ways. In some cases, relevant features are selected based on the clinical importance of sepsis and related diseases. In some cases, features are selected based on a high level of correlation with a result or performance to predict a measure of the result. Thus, the strength of the relationship between hemodynamic data and a sepsis diagnosis (e.g., SIRS criteria) can be determined. Many statistical methods are known for determining the strength of a correlation (e.g., a correlation coefficient), including linear correlation (Pearson's correlation coefficient), Kendall's rank correlation coefficient, and Spearman's rank correlation coefficient. In some cases, a computational model can identify a feature or combination of features based on their cost functions. Computational models for feature selection can use weights or coefficients based on their performance to identify features, including but not limited to LASSO, elastic net, and ridge regression. A computational model for identifying useful features can be a different (or the same) model as a prediction model used to provide early screening for sepsis or to predict the probability that a patient has sepsis. In some cases, a computational approach can explore all possible features and identify which features are the most sensitive. Some computational approaches for exploring features and identifying sensitivity include, but are not limited to, restrictive to recursive feature elimination, and information gain criteria. Depending on the situation, an ensemble approach that combines multiple models for feature selection and model development may be utilized. In any approach, an appropriate computational model that results in a manageable number of features can be selected. For example, constructing a prediction model from a large number of features may suffer from the problem of overfitting. Similarly, too few features may result in lower predictive power.
[0069] Computational Processing and Monitoring System A computing system for screening for sepsis or predicting the probability of sepsis according to various methods and processes of the present disclosure typically utilizes a processing system including one or more of a CPU, a GPU, and / or a neural processing engine. Waveform data corresponding to arterial blood pressure, or corresponding to a signal proportional to or derived from arterial blood pressure, can be recorded by a sensor. The sensors include, but are not limited to, an intra-arterial catheter, a disposable blood pressure transducer, a pressurized finger cuff and optical sensor, and an applanation tonometer. Further, features of hemodynamic data can be extracted from the waveform data to screen for sepsis or predict the probability of sepsis.
[0070] The computing system can be housed within a patient monitor in a direct connection between the monitor including the sensor and components or between components. Alternatively, the computing system can be housed separately from the patient monitor and / or components and receive the acquired waveform data via a wired or wireless connection (e.g., WiFi®, cellular, Bluetooth®, etc.). The computing system can be implemented on any suitable computing device such as, but not limited to, a patient monitor, a tablet, and / or a portable computer.
[0071] Exemplary computing processing systems that can be utilized to execute various methods and processes of the present disclosure are shown in FIGS. 2 and 3. FIG. 2 shows a computing system for screening for sepsis (e.g., early detection of the likelihood of developing sepsis), and FIG. 3 shows a computing system for predicting the probability that a patient is suffering from sepsis. The computing processing system 110 includes a processor system 112, an I / O interface 114, a memory system 116, and a sensor 118. As can be readily understood, the processor system 112, the I / O interface 114, and the memory system 116 can be implemented using any of a variety of components suitable for specific application requirements, including (but not limited to) a CPU, a GPU, an ISP, a DSP, a wireless modem (e.g., WiFi, Bluetooth modem), a serial interface, a volatile memory (e.g., DRAM), and / or a non-volatile memory (e.g., SRAM, and / or NANO flash).
[0072] The sensor 118 can be attached to a patient to sense the patient's waveform data corresponding to, proportional to, or derived from an arterial blood pressure. The sensor 118 is operably connected to the monitoring system 110 and the I / O interface 114, thereby enabling a visual representation of the arterial pressure waveform captured from the sensor to be provided. The sensor 118 can be a non-invasive or invasive pressure sensor. Thus, the sensor 118 can be an intra-arterial catheter with a disposable blood pressure transducer (e.g., an in-artery blood pressure catheter), a pressurized finger cuff and an optical sensor (e.g., volume clamp method), an applanation tonometer, or any other pressure sensor for obtaining an arterial pressure waveform.
[0073] In the illustrated example, the memory system 116 can store various data and models. The listed data and models are representative samples of what can be stored in the memory, and it should be understood that various memory systems may store some or all of the various listed data and models. Further, any combination of data and models can be stored, and in some implementations, various data, applications, and / or models are stored temporarily.
[0074] In some implementations, the memory system 116 can store waveform data 200 that can be obtained from the sensor 118. An application can extract hemodynamic data 202 from the waveform data 200, and the hemodynamic data 202 can also be stored in the memory system 116. The extracted hemodynamic data 202 can be utilized by a sepsis screening model 204 stored in the memory system 116. The processor system 112 is configured to execute the sepsis screening model 204 to generate a calculated score 206 indicative of an early screening of a patient for sepsis. Further, the waveform data 200 and / or the calculated score 206 can be displayed on a monitor or other screen via the I / O interface 114.
[0075] In some implementations, the memory system 116 can store waveform data 300 that can be obtained from the sensor 118. An application can extract hemodynamic data 302 from the waveform data 300, and the hemodynamic data 302 can also be stored in the memory system 116. The extracted hemodynamic data 302 can be utilized by a sepsis probability model 304 stored in the memory system 116. The processor system 112 is configured to execute the sepsis probability model 304 to generate a calculated score 306 indicative of the probability that a patient has sepsis. Further, the waveform data 300 and / or the calculated score 306 can be displayed on a monitor or other screen via the I / O interface 114.
[0076] Based on the calculated score, the monitoring system 110 can provide alerts of screening results and / or sepsis probability results to the clinician. In particular, if an individual is predicted to have sepsis, the alert can enable timely and effective intervention to prevent organ failure or other serious complications associated with sepsis.
[0077] Although a specific computing system has been described above with reference to FIGS. 2 and 3, it should be readily understood that the computing processes and / or other processes utilized to provide sepsis screening or prediction can be implemented on any of a variety of processing devices, including combinations of processing devices. Thus, it should be understood that the computing device is not limited to a particular monitoring system, computing system, and / or particular application and model. The computing device can be implemented using any combination of the systems described herein and / or modified versions of the systems described herein to execute the processes, combinations of processes, and / or modified versions of the processes described herein.
Description of Reference Numerals
[0078] 100 Method 110 Computing and Monitoring System 112 Processor System 114 I / O Interface 116 Memory System 118 Sensor 200 Waveform Data 202 Hemodynamic Data 204 Sepsis Screening Model 206 Calculated Score 300 Waveform Data 302 Hemodynamic Data 304 Sepsis Probability Model 306 Calculated Score
Claims
1. A calculation method for screening sepsis, comprising: receiving, from a sensor attached to a patient, waveform data corresponding to a signal corresponding to, proportional to, or derived from arterial blood pressure; extracting a set of features of hemodynamic data from the waveform data; inputting the set of extracted features of hemodynamic data into a predictive calculation model to obtain a sepsis screening score, wherein the predictive calculation model is trained to screen for sepsis using the set of extracted features of hemodynamic data; A calculation method comprising the steps of:
2. The calculation method according to claim 1, further comprising the step of sensing the arterial blood pressure using the sensor.
3. The calculation method according to claim 1 or 2, wherein the sensor is an arterial catheter and a disposable blood pressure transducer, a pressurized finger cuff and a light sensor, or an applanation tonometer.
4. The set of extracted features of hemodynamic data includes heart rate, respiratory rate, cardiac output, stroke volume, stroke volume variability, vascular tone, contractility, afterload, systemic vascular resistance, systolic blood pressure, diastolic blood pressure, mean arterial pressure (MAP), kurtosis of pressure distribution, left ventricular ejection time, time from reaching systolic MAP to incisura, sample entropy of time to reach systolic MAP, entropy of interbeat interval, entropy of standard deviation of decay phase, 【Number 1】 、 【Number 2】 、 【Number 3】 、 【Number 4】 , or 【Number 5】 The calculation method according to claim 1, 2, or 3, comprising at least one of the above.
5. The calculation method according to any one of claims 1 to 4, further comprising the step of inputting clinical information of the patient into the model.
6. The calculation method according to claim 5, wherein the clinical information of the patient includes at least one of patient demographics, patient vital signs, and patient test results.
7. The calculation method according to any one of claims 1 to 6, wherein the set of extracted features includes heart rate, kurtosis of pressure distribution, and sample entropy of time to reach systolic MAP.
8. The calculation method according to any one of claims 1 to 6, wherein the set of extracted features includes heart rate, arterial stiffness factor, sample entropy of decay area, dynamic arterial elastance, and approximate entropy of systolic time.
9. The calculation method according to any one of claims 1 to 8, wherein the prediction calculation model uses an equation for obtaining the screening score for sepsis.
10. The equation is 【Number 6】 where hr is the heart rate, kurt is the kurtosis of the pressure distribution, and sampEn is the sample entropy of the time to reach the systolic MAP. The calculation method according to claim 9.
11. The equation is 【Number 7】 where hr is the heart rate, avgK is the arterial tone factor, decAreaSampEn is the sample entropy of the decay area, dynEa is the dynamic arterial elastance, and tSysApEn is the approximate entropy of the time during systole. The calculation method according to claim 9.
12. The calculation method according to any one of claims 1 to 11, wherein the prediction calculation model is a regression-based model, a classification-based model, or an ensemble model.
13. The calculation method according to any one of claims 1 to 12, wherein the screening score for sepsis indicates the risk of developing sepsis, and the method further includes the step of further evaluating the patient regarding the complications of sepsis.
14. The screening score for sepsis indicates the risk of developing sepsis, and the method further includes the step of monitoring the patient regarding the complications of sepsis for a specific period. The calculation method according to any one of claims 1 to 13.
15. A calculation method for predicting the probability that a patient will develop sepsis, comprising: receiving waveform data corresponding to, proportional to, or derived from arterial blood pressure from a sensor attached to the patient; extracting a set of features of hemodynamic data from the waveform data; inputting the set of features of the extracted hemodynamic data into a prediction calculation model to obtain a sepsis probability score, wherein the prediction calculation model is trained to predict sepsis using the set of features of the extracted hemodynamic data; and the calculation method includes the above steps.
16. The calculation method according to claim 15, further comprising the step of sensing the arterial blood pressure using the sensor.
17. The calculation method according to claim 15 or 16, wherein the sensor is an intra-arterial catheter and a disposable blood pressure transducer, a pressurized finger cuff and an optical sensor, or an applanation tonometer.
18. The set of characteristics of the extracted hemodynamic data includes heart rate, respiratory rate, cardiac output, stroke volume, stroke volume variability, vascular tone, contractility, afterload, systemic vascular resistance, systolic blood pressure, diastolic blood pressure, mean arterial pressure (MAP), skewness of the pressure distribution, left ventricular ejection time, time from reaching systolic MAP to incisura, sample entropy of the time of reaching systolic MAP, entropy of the interbeat interval, entropy of the standard deviation of the decay phase, 【Number 8】 、 【Number 9】 、 【Number 10】 、 【Number 11】 , or 【Number 12】 The calculation method according to claim 15, 16, or 17, including at least one of.
19. The calculation method according to any one of claims 15 to 18, further including the step of inputting the patient's clinical information into the model.
20. The calculation method according to claim 19, wherein the patient's clinical information includes at least one of the patient's demographics, the patient's vital signs, and the patient's test results.
21. The calculation method according to any one of claims 15 to 20, wherein the set of characteristics of the extracted hemodynamic data includes diastolic blood pressure, heart rate, entropy of the interbeat interval, time from reaching systolic MAP to incisura, and entropy of the standard deviation of the decay phase.
22. The calculation method according to any one of claims 15 to 20, wherein the set of characteristics of the extracted hemodynamic data includes heart rate, approximate entropy of the blood pressure waveform, sample entropy of the systolic area, approximate entropy of the systolic time, and approximate entropy of the systolic decay time.
23. The calculation method according to any one of claims 15 to 22, wherein the prediction calculation model uses an equation for obtaining the probability score of sepsis.
24. The equation is 【Number 13】 , where Dia is diastolic blood pressure, hr is heart rate, enIBI is entropy of the interbeat interval, timeMAP is the time from reaching systolic MAP to incisura, and enDecay is entropy of the standard deviation of the decay phase. The calculation method according to claim 23.
25. The equation is 【Number 14】 where hr is the heart rate, ApEnV is the approximate entropy of the blood pressure waveform, areaSampEn is the sample entropy of the systolic area, tSysApEn is the approximate entropy of the systolic time, and tDecApEn is the approximate entropy of the systolic decay time, the calculation method according to claim 23.
26. The calculation method according to any one of claims 15 to 25, wherein the prediction calculation model is a regression-based model, a classification-based model, or an ensemble model.
27. The probability score of sepsis indicates that the patient has sepsis, The method further includes the step of further evaluating the patient regarding the complications of sepsis to confirm the probability score. The calculation method according to any one of claims 15 to 26.
28. The probability score of sepsis indicates that the patient has sepsis, The method further includes the step of treating the patient to treat the sepsis. The calculation method according to any one of claims 15 to 27.
29. A patient monitoring system for screening sepsis by the captured waveform data, comprising: a sensor; a calculation processing system operably connected to the sensor; The calculation processing system includes: a processor system; a memory system including one or more applications, the one or more applications causing the processor system to: receive waveform data corresponding to, proportional to, or derived from the arterial blood pressure from the sensor attached to the patient; extract a set of features of the hemodynamic data from the waveform data; input the set of features of the extracted hemodynamic data into a prediction calculation model to obtain a screening score for sepsis, the prediction calculation model being trained to screen for sepsis using the set of features of the extracted hemodynamic data; a memory system configured to instruct to perform; A patient monitoring system including.
30. The patient monitoring system according to claim 29, wherein the sensor is an intra-arterial catheter and a disposable blood pressure transducer, a pressurized finger cuff and an optical sensor, or a applanation tonometer.
31. The set of characteristics of the extracted hemodynamic data includes heart rate, respiratory rate, cardiac output, stroke volume, stroke volume variability, vascular tone, contractility, afterload, systemic vascular resistance, systolic blood pressure, diastolic blood pressure, mean arterial pressure (MAP), kurtosis of pressure distribution, left ventricular ejection time, time from reaching systolic MAP to dicrotic notch, sample entropy of the time to reach systolic MAP, entropy of heart rate intervals, entropy of the standard deviation of the decay phase, 【Number 15】 、 【Number 16】 、 【Number 17】 、 【Number 18】 , or 【Number 19】 The patient monitoring system according to claim 29 or 30, including at least one of them.
32. The one or more applications are further configured to instruct the processor system to Input the patient's clinical information into the model. The patient monitoring system according to claim 25, 26, or 27.
33. The patient's clinical information includes at least one of patient demographics, patient vital signs, and patient test results. The patient monitoring system according to claim 32.
34. The set of extracted features includes heart rate, kurtosis of pressure distribution, and sample entropy of the time to reach systolic MAP. The patient monitoring system according to any one of claims 29 to 33.
35. The set of extracted features includes heart rate, arterial tone factor, sample entropy of decay area, dynamic arterial elastance, and approximate entropy of systolic time. The patient monitoring system according to any one of claims 29 to 33.
36. The prediction calculation model uses an equation to obtain the sepsis screening score. The patient monitoring system according to any one of claims 25 to 29.
37. The equation is 【Number 20】 , where hr is the heart rate, kurt is the kurtosis of the pressure distribution, and sampEn is the sample entropy of the time to reach systolic MAP. The patient monitoring system according to claim 36.
38. The equation is 【Number 21】 , where hr is the heart rate, avgK is the arterial tone factor, decAreaSampEn is the sample entropy of the decay area, dynEa is the dynamic arterial elastance, and tSysApEn is the approximate entropy of systolic time. The patient monitoring system according to claim 36.
39. The patient monitoring system according to any one of claims 29 to 38, wherein the prediction calculation model is a regression-based model, a classification-based model, or an ensemble model.
40. The one or more applications are further configured to instruct the processor system to display the screening score for sepsis on a monitor operably connected to the calculation processing system, the patient monitoring system according to any one of claims 29 to 39.
41. The screening score for sepsis indicates the risk of developing sepsis, and the one or more applications are further configured to instruct the processor system to provide an alert indicating the risk when it is determined that the screening score for sepsis indicates the risk of developing sepsis, the patient monitoring system according to any one of claims 29 to 40.
42. A patient monitoring system for predicting whether a patient has sepsis based on the captured arterial pressure, a sensor, a calculation processing system operably connected to the sensor, and comprising, the calculation processing system a processor system, a memory system including one or more applications, the one or more applications being configured to instruct the processor system to receive, from the sensor attached to the patient, waveform data corresponding to a signal corresponding to, proportional to, or derived from the arterial blood pressure, extract a set of features of the hemodynamic data from the waveform data, input the set of features of the extracted hemodynamic data into a prediction calculation model to obtain a sepsis probability score, the prediction calculation model being trained to predict sepsis using the set of features of the extracted hemodynamic data, and inputting, a memory system configured to instruct to perform A patient monitoring system including.
43. The patient monitoring system according to claim 42, wherein the sensor is an arterial catheter and a disposable blood pressure transducer, a pressurized finger cuff and an optical sensor, or a applanation tonometer.
44. The set of features of the extracted hemodynamic data includes heart rate, respiratory rate, cardiac output, stroke volume, stroke volume variability, vascular tone, contractility, afterload, systemic vascular resistance, systolic blood pressure, diastolic blood pressure, mean arterial pressure (MAP), kurtosis of pressure distribution, left ventricular ejection time, time from reaching systolic MAP to incisura, sample entropy of the time of reaching systolic MAP, entropy of heart rate intervals, entropy of the standard deviation of the decay phase, 【Number 22】 、 【Number 23】 、 【24 Points】 、 【Number 25】 , or 【Number 26】 The patient monitoring system according to claim 42 or 43, comprising at least one of
45. The one or more applications are further configured to instruct the processor system to input the clinical information of the patient into the model, the patient monitoring system according to claim 42, 43, or 44.
46. The clinical information of the patient includes at least one of patient demographics, patient vital signs, and patient test results, the patient monitoring system according to claim 45.
47. The set of features of the extracted hemodynamic data includes diastolic blood pressure, heart rate, entropy of heart rate intervals, time from reaching systolic MAP to incisura, and entropy of the standard deviation of the decay phase, the patient monitoring system according to any one of claims 42 to 46.
48. The set of features of the extracted hemodynamic data includes heart rate, approximate entropy of blood pressure waveform, sample entropy of systolic area, approximate entropy of systolic time, and approximate entropy of systolic decay time, the patient monitoring system according to any one of claims 42 to 46.
49. The prediction calculation model uses an equation to obtain the probability score of sepsis, the patient monitoring system according to any one of claims 42 to 48.
50. The equation is 【Number 27】 , where Dia is diastolic blood pressure, hr is heart rate, enIBI is entropy of heart rate intervals, timeMAP is time from reaching systolic MAP to incisura, and enDecay is entropy of the standard deviation of the decay phase, the patient monitoring system according to claim 49.
51. The equation is 【Number 28】 where hr is the heart rate, ApEnV is the approximate entropy of the blood pressure waveform, areaSampEn is the sample entropy of the systolic area, tSysApEn is the approximate entropy of the systolic time, and tDecApEn is the approximate entropy of the systolic decay time, the patient monitoring system according to claim 49.
52. The patient monitoring system according to any one of claims 42 to 51, wherein the prediction calculation model is a regression-based model, a classification-based model, or an ensemble model.
53. The one or more applications cause the processor system to The patient monitoring system according to any one of claims 42 to 52, further configured to instruct the monitor operably connected to the calculation processing system to display the probability score of sepsis thereon.
54. The probability score of sepsis indicates the risk of developing sepsis, and the one or more applications cause the processor system to The patient monitoring system according to any one of claims 42 to 53, further configured to instruct the processor system to provide an alert indicating that the patient is suffering from sepsis when it is determined that the probability score of sepsis indicates that the patient is suffering from sepsis.