Pediatric sepsis risk early warning method and system based on leukocyte dynamic trajectory
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGDONG WOMEN & CHILDREN HOSPITAL
- Filing Date
- 2026-03-18
- Publication Date
- 2026-07-03
AI Technical Summary
Current technology cannot achieve continuous and dynamic monitoring of white blood cell counts in pediatric sepsis patients, resulting in delayed risk assessment and a lack of automated and precise early warning tools.
By acquiring continuous white blood cell count data of target pediatric patients, a dynamic trajectory classification model of white blood cells is constructed using the group basis trajectory modeling method. Typical trajectory categories are identified, and the risk level of sepsis is automatically determined based on the trajectory pattern to generate early warning information.
It enables early dynamic risk monitoring, makes assessments more objective and accurate, reduces the burden on medical staff, improves clinical work efficiency, has a high degree of integration, and is easy to promote.
Smart Images
Figure CN122337581A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a method and system for early warning of pediatric sepsis risk based on the dynamic trajectory of leukocytes, belonging to the interdisciplinary field of medical information technology and critical care medicine. Background Technology
[0002] Currently, the clinical risk assessment of pediatric sepsis mainly relies on the following methods:
[0003] 1) Static biomarker single-point detection: such as white blood cell (WBC) count at a single or discrete time point, procalcitonin (PCT), C-reactive protein (CRP), etc. These indicators only reflect transient conditions and cannot depict the continuous evolution of immune status, resulting in a significant lag in early warning.
[0004] 2) Composite scoring systems: such as the Pediatric Critical Care Identification System (PCIS) or the Pediatric Sequential Organ Failure Assessment (pSOFA). These scores are mostly based on multiple parameters at a certain time point, which is cumbersome to calculate and not sensitive to early and subtle dynamic changes.
[0005] 3) Clinical experience judgment: It relies heavily on the personal experience of medical staff and lacks objective and unified quantitative standards, which can easily lead to inconsistent assessments and missed judgments.
[0006] The core problem with the aforementioned methods lies in their "static" and "fragmented" nature, failing to achieve continuous, dynamic, and systematic monitoring of the child's immune status (with white blood cell count as one of the core indicators). The immune response of children with sepsis is complex and variable, and the trajectory of their white blood cell count over time (such as persistently low values, gradual decline, or persistently high values) may contain richer prognostic information than a single numerical value. However, current technology lacks a tool capable of automatically extracting and analyzing this dynamic trajectory and transforming it into an intuitive, real-time risk warning. Therefore, there is an urgent need for a new technological solution that can automatically, dynamically, and accurately assess the risk of pediatric sepsis. Summary of the Invention
[0007] In view of this, the present invention provides a method, system, computer device and storage medium for early warning of pediatric sepsis risk based on the dynamic trajectory of white blood cells. It can automatically analyze the continuous white blood cell count data after the patient is admitted to the hospital, identify its characteristic dynamic change trajectory, and automatically determine the risk level based on the trajectory pattern, and issue an early warning in a timely manner.
[0008] The first objective of this invention is to provide a method for early warning of pediatric sepsis risk based on the dynamic trajectory of leukocytes.
[0009] The second objective of this invention is to provide a pediatric sepsis risk early warning system based on the dynamic trajectory of leukocytes.
[0010] A third objective of this invention is to provide a computer device.
[0011] A fourth objective of this invention is to provide a computer-readable storage medium.
[0012] The first objective of this invention can be achieved by adopting the following technical solution:
[0013] A method for early warning of pediatric sepsis risk based on dynamic leukocyte trajectories, the method comprising:
[0014] Obtain time-series white blood cell count data of the target pediatric patient for N consecutive days after admission or diagnosis of sepsis to form the raw data sequence;
[0015] The original data sequences are preprocessed to eliminate baseline differences between individuals;
[0016] The preprocessed data sequence is input into a pre-trained leukocyte dynamic trajectory classification model to output the probability of the target pediatric patient belonging to each category, and the category with the highest probability is determined as the current trajectory category of the target pediatric patient;
[0017] According to the preset trajectory category-risk level mapping rules, the current trajectory category is mapped to the corresponding sepsis risk level;
[0018] Whether to generate an early warning message depends on the sepsis risk level.
[0019] Furthermore, the preprocessing of the original data sequence includes:
[0020] The original data sequence is standardized or normalized.
[0021] Furthermore, the leukocyte dynamic trajectory classification model is constructed using a base case trajectory modeling method. Based on time-series leukocyte count data from a historical cohort of pediatric sepsis patients, it identifies K significantly different typical trajectory categories. The training process of the leukocyte dynamic trajectory classification model includes:
[0022] Obtain time-series white blood cell count data from a historical cohort of pediatric sepsis patients;
[0023] Time-series white blood cell count data of historical pediatric sepsis patients were cleaned to remove outliers; for missing data, linear interpolation or carry-over from the last observation was used to fill in the gaps.
[0024] Using the cleaned time-series white blood cell count data as input features, the maximum likelihood estimation method was used to fit the group basis trajectory model.
[0025] Simultaneously, we examine the number of groups and the order of the trajectory function to identify potential heterogeneous trajectory subgroups in the data, and characterize the changing trends of different forms through the order of the polynomial.
[0026] Candidate models were screened by comprehensively using three indicators: Bayesian information criterion, average posterior probability, and clinical interpretability, in order to determine the optimal model. The parameters of the trained optimal model were then solidified to obtain a trained white blood cell dynamic trajectory classification model.
[0027] Furthermore, the typical trajectory categories are three: persistently high value, gradually decreasing value, and persistently low value.
[0028] The term "persistently high value" refers to a white blood cell count that remains at a high level throughout the entire observation period.
[0029] The term "gradual decline" refers to a white blood cell count that starts from a mid-to-high baseline level and shows a gradual downward trend over time.
[0030] The term "persistently low value" refers to a white blood cell count that remains at a low level throughout the observation period.
[0031] Furthermore, the step of determining whether to generate early warning information based on the sepsis risk level includes:
[0032] If the sepsis risk level is intermediate or low, continue monitoring;
[0033] If the sepsis risk level is high, an early warning message will be generated.
[0034] Furthermore, the method also includes:
[0035] The warning information and corresponding supporting data are pushed to at least one of the doctor's workstation, nurse's station, and mobile terminal and highlighted. The supporting data includes trajectory map and risk level.
[0036] The second objective of this invention can be achieved by adopting the following technical solution:
[0037] A pediatric sepsis risk early warning system based on leukocyte dynamic trajectory, the system comprising:
[0038] The data interface module is used to acquire time-series white blood cell count data of the target pediatric patient for N consecutive days after admission or diagnosis of sepsis, forming the raw data sequence;
[0039] The preprocessing module is used to preprocess the raw data sequence to eliminate baseline differences between individuals;
[0040] The core analysis module is used to input the preprocessed data sequence into the pre-trained white blood cell dynamic trajectory classification model to output the probability of the target pediatric patient belonging to each category, and to determine the category with the highest probability as the current trajectory category of the target pediatric patient.
[0041] The risk level determination module is used to map the current trajectory category to the corresponding sepsis risk level according to the preset trajectory category-risk level mapping rules;
[0042] The early warning generation module is used to determine whether to generate early warning information based on the sepsis risk level.
[0043] Furthermore, the system also includes:
[0044] The push module is used to push early warning information and corresponding supporting data to at least one of the doctor's workstation, nurse's station, and mobile terminal, and to highlight it. The supporting data includes trajectory map and risk level.
[0045] The third objective of this invention can be achieved by adopting the following technical solution:
[0046] A computer device includes a processor and a memory for storing a processor-executable program, wherein when the processor executes the program stored in the memory, it implements the above-described method for early warning of pediatric sepsis risk.
[0047] The fourth objective of this invention can be achieved by adopting the following technical solution:
[0048] A computer-readable storage medium storing a program that, when executed by a processor, implements the above-described method for early warning of pediatric sepsis risk.
[0049] The present invention has the following advantages over the prior art:
[0050] 1. This invention enables dynamic risk monitoring and earlier warning, abandons single-point value judgment, focuses on the continuous change trajectory of white blood cells, and can capture the early dynamic trend of immune status deterioration.
[0051] 2. The evaluation of this invention is more objective and accurate. It uses a machine learning model based on big data for automated classification, avoiding subjective experience differences.
[0052] 3. This invention can improve clinical work efficiency, realize fully automated data capture, analysis and early warning, reduce the burden of medical staff repeatedly reviewing and comparing data, and enable them to focus more on treatment decisions.
[0053] 4. This invention has a high degree of integration and strong practicality. It can be used as an independent system or module and seamlessly integrated into the existing hospital information platform without changing the existing clinical workflow, making it easy to promote and use. Attached Figure Description
[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.
[0055] Figure 1 This is an overall architecture diagram of the pediatric sepsis risk warning method based on the dynamic trajectory of leukocytes in Embodiment 1 of the present invention.
[0056] Figure 2 This is a flowchart of the pediatric sepsis risk warning method based on the dynamic trajectory of leukocytes in Embodiment 1 of the present invention.
[0057] Figure 3 This is a trend chart of three types of dynamic trajectories of white blood cells in Embodiment 2 of the present invention.
[0058] Figure 4 This is a Kaplan-Meier 28-day survival curve diagram corresponding to the three types of leukocyte dynamic trajectory in Example 2 of the present invention.
[0059] Figure 5 This is a structural block diagram of the pediatric sepsis risk early warning system based on the dynamic trajectory of leukocytes according to Embodiment 3 of the present invention.
[0060] Figure 6 This is a structural block diagram of the computer device according to Embodiment 4 of the present invention. Detailed Implementation
[0061] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.
[0062] Example 1:
[0063] like Figure 1 and Figure 2 As shown in the figure, this embodiment provides a method for early warning of pediatric sepsis risk based on the dynamic trajectory of white blood cells. The method includes the following steps:
[0064] S201. Obtain time-series white blood cell count data for the target pediatric patient for N consecutive days after admission or diagnosis of sepsis, and form the raw data sequence.
[0065] This embodiment connects to hospital data sources, such as Hospital Information System (HIS) and Laboratory Information System (LIS), through a data interface to obtain time-series white blood cell count data of the target pediatric patient for N consecutive days after admission or diagnosis of sepsis, forming the original data sequence.
[0066] S202. Preprocess the original data sequence to eliminate baseline differences between individuals.
[0067] To eliminate inter-individual baseline shifts caused by differences in age, physiological basis, or testing instruments among different patients, and to ensure that the model can accurately identify the dynamic trend of white blood cell changes rather than absolute values, the system first standardizes the acquired raw time series data.
[0068] It should be noted that the standardization process in this embodiment aims to map the original data to a unified scale space, making data of different magnitudes comparable. In one specific implementation, the Z-score standardization method is used, and its calculation formula is as follows:
[0069] x′ = (x − μ) / σ;
[0070] in, This represents the raw white blood cell count. This is the mean of the child's current time series (e.g., data from day 1 to day 5 after admission). Its standard deviation. The data sequence after this processing. It follows a distribution with a mean of 0 and a standard deviation of 1, preserving the fluctuation characteristics of the original data while eliminating the interference of individual absolute levels.
[0071] It is understood that in other implementations, linear transformation methods such as Min-Max normalization can also be used, as long as the processing method is consistent with the preprocessing method used in the model training stage.
[0072] S203. Input the preprocessed data sequence into the pre-trained leukocyte dynamic trajectory classification model to output the probability of the target pediatric patient belonging to each category, and determine the category with the highest probability as the current trajectory category of the target pediatric patient.
[0073] The leukocyte dynamic trajectory classification model in this embodiment is constructed using the Group Basis Trajectory Modeling (GBTM) method. Based on time-series leukocyte count data from a historical cohort of pediatric sepsis patients, it identifies K typical trajectory categories with significant differences. The training process includes the following steps:
[0074] 1) Data preparation and cleaning. A total of 887 pediatric sepsis patients were collected from the hospital information system (HIS / EMR). White blood cell count data for 7 consecutive days after admission were extracted. These data were cleaned to remove outliers caused by clinical events (such as surgery, blood transfusion, etc.). For missing data, linear interpolation or carry-over of last observation was used to fill in the missing data.
[0075] 2) Model Fitting and Parameter Optimization. Using cleaned white blood cell count data as input features, the maximum likelihood estimation method was employed to fit the base trajectory model. During model selection, the number of groups (K values ranging from 1 to 5) and the order of the trajectory function (linear, quadratic, and cubic polynomials were used respectively) were considered to identify potential heterogeneous trajectory subgroups in the data and to characterize the changing trends of different morphologies through the polynomial order.
[0076] 3) Model Selection and Validation. Candidate models were screened using three indicators: Bayesian Information Criterion (BIC), Average Posterior Probability (AvePP), and Clinical Interpretability. The BIC was used to balance goodness of fit and complexity, prioritizing models with smaller absolute values. The AvePP required that the average probability of patients within each trajectory group be greater than 0.7 to ensure good consistency within each group. Simultaneously, clinical experts were invited to evaluate whether the morphology of each trajectory group conformed to the pathophysiological patterns of pediatric sepsis.
[0077] 4) Determining the Optimal Model. Based on the above indicators, the optimal model was ultimately determined to be the "3-group cubic trajectory model." This model divides the historical patient population into three subgroups with significantly different dynamic evolution characteristics, each corresponding to a different clinical outcome risk. The trained optimal model parameters (including the number of groups, regression coefficients for each group, intercept term, and posterior probability calculation formula) were solidified and integrated into the system's data processing and analysis engine for trajectory classification of real-time acquired clinical data.
[0078] Among them, three typical trajectory categories are preferred: persistently high value category (category 1), gradually decreasing category (category 2), and persistently low value category (category 3). The specific descriptions of each category are as follows:
[0079] 1) Persistently high values refer to white blood cell counts that remain at a high level throughout the observation period, with a possible slight increase later. The observation period is within 7 days after admission.
[0080] 2) Gradually declining type refers to white blood cell counts that start from a mid-to-high baseline level and show a gradual downward trend over time.
[0081] 3) Persistently low values refer to white blood cell counts that remain at a low level throughout the observation period with minimal fluctuations.
[0082] The definition of "level" (high, medium-high, low) in the above trajectory categories is automatically formed by clustering based on the statistical distribution characteristics of the historical training dataset after standardization, rather than relying on a preset fixed absolute value threshold. In practical applications, the system achieves classification by comparing the similarity between the standardized trajectory pattern of the target pediatric patient and the standard trajectory of each category.
[0083] S204. Based on the preset trajectory category-risk level mapping rules, map the current trajectory category to the corresponding sepsis risk level.
[0084] According to the preset trajectory category-risk level mapping rule, if the current trajectory category is a persistently high value category, the corresponding sepsis risk level is high; if the current trajectory category is a gradually decreasing category, the corresponding sepsis risk level is low; and if the current trajectory category is a persistently low value category, the corresponding sepsis risk level is medium.
[0085] S205. Based on the sepsis risk level, decide whether to generate an early warning message.
[0086] In this embodiment, if the sepsis risk level is medium or low, monitoring continues; if the sepsis risk level is high, an early warning message is generated.
[0087] To enable doctors and nurses to be promptly informed of the early warning status of target pediatric patients, this embodiment may further include the following after generating the early warning information:
[0088] S206. Push the warning information and corresponding support data to at least one of the doctor's workstation, nurse's station, and mobile terminal, and highlight it.
[0089] In this embodiment, the supported data includes trajectory maps and risk levels, and the highlighting includes pop-ups, color warnings, and sound prompts.
[0090] Example 2:
[0091] This embodiment uses a 1-year-old child diagnosed with sepsis as an example to illustrate the steps of the pediatric sepsis risk warning method based on leukocyte dynamic trajectory:
[0092] S1. Automatically acquire time-series white blood cell count data from day 1 to day 7 via the data interface, forming the original data sequence: [18, 19, 20, 21, 20, 22, 23] (×10 9 / L).
[0093] S2. Standardize the original data sequence.
[0094] S3. Input the preprocessed data into the deployed and pre-trained leukocyte dynamic trajectory classification model. The model calculates that the probability of the sequence belonging to "Category 1 (persistently high value type)" is the highest (95%).
[0095] Figure 3 The chart shows trend graphs for three categories of white blood cell dynamic trajectories; Figure 4 The Kaplan-Meier 28-day survival curves corresponding to the three leukocyte dynamic trajectory categories are shown (with Log-rank test P value, Log-rank P=0.001), which can prove the strong correlation between the leukocyte dynamic trajectory classification model of this embodiment and the prognostic discrimination ability.
[0096] S4. Based on the preset trajectory category-risk level mapping rule ("Category 1" is mapped to "high risk"), the child is determined to be a high-risk patient for sepsis, and an early warning is immediately triggered.
[0097] S5. Generate a red alert with the following content: "Alert: Patient [ID], white blood cell dynamic trajectory type is 'persistently high value', sepsis risk level: high risk, it is recommended to closely monitor and evaluate organ function and treatment plan." This alert is pushed to the attending physician's mobile APP and the nurse station screen in real time.
[0098] It should be noted that although the above-described method operations are depicted in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all of the illustrated operations must be performed to achieve the desired result. On the contrary, the order of execution of the depicted steps can be changed. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.
[0099] Example 3:
[0100] like Figure 5 As shown, this embodiment provides a pediatric sepsis risk early warning system based on the dynamic trajectory of white blood cells. The system includes a data interface module 501, a preprocessing module 502, a core analysis module 503, a risk level determination module 504, and an early warning generation module 505. The specific functions of each module are as follows:
[0101] Data interface module 501 is used to acquire time-series white blood cell count data of the target pediatric patient for N consecutive days after admission or diagnosis of sepsis, forming a raw data sequence;
[0102] The preprocessing module 502 is used to preprocess the original data sequence to eliminate baseline differences between individuals;
[0103] The core analysis module 503 is used to input the preprocessed data sequence into the pre-trained leukocyte dynamic trajectory classification model to output the probability of the target pediatric patient belonging to each category, and to determine the category with the highest probability as the current trajectory category of the target pediatric patient.
[0104] The risk level determination module 504 is used to map the current trajectory category to the corresponding sepsis risk level according to the preset trajectory category-risk level mapping rules;
[0105] The early warning generation module 505 is used to determine whether to generate early warning information based on the sepsis risk level.
[0106] Furthermore, the pediatric sepsis risk early warning system of this embodiment may also include:
[0107] The push module 506 is used to push the warning information and corresponding supporting data to at least one of the doctor's workstation, nurse's station, and mobile terminal, and to highlight them. The supporting data includes trajectory map and risk level.
[0108] In this embodiment, the preprocessing module 502 and the core analysis module 503 constitute the system's data processing and analysis engine. The early warning generation module 505 can generate a visual early warning report. The push module 506 pushes the early warning information to clinical terminals (doctor's workstation, nurse's station) and attending physician's mobile terminal (mobile phone, etc.) through the hospital intranet or a dedicated protocol. The system also has a user interface for displaying patient lists, individual trajectory visualization charts, risk levels, and historical early warning records.
[0109] The specific implementation of each module in this embodiment can be found in Embodiment 1 above, and will not be repeated here. It should be noted that the system provided in this embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure can be divided into different functional modules to complete all or part of the functions described above.
[0110] Example 4:
[0111] This embodiment provides a computer device, such as... Figure 6As shown, it includes a processor 602, a memory, an input device 603, a display device 604, and a network interface 605 connected via a device bus 601. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium 606 and internal memory 607. The non-volatile storage medium 606 stores operating devices, computer programs, and a database. The internal memory 607 provides an environment for the operation of the operating devices and computer programs in the non-volatile storage medium. When the processor 602 executes the computer program stored in the memory, it implements the pediatric sepsis risk warning method of Embodiment 1 described above, as follows:
[0112] The system acquires time-series white blood cell count data for N consecutive days after admission or diagnosis of sepsis in the target pediatric patient, forming a raw data sequence. The raw data sequence is preprocessed to eliminate baseline differences between individuals. The preprocessed data sequence is then input into a pre-trained white blood cell dynamic trajectory classification model to output the probability of the target pediatric patient belonging to each category, and the category with the highest probability is determined as the current trajectory category of the target pediatric patient. Based on a pre-defined trajectory category-risk level mapping rule, the current trajectory category is mapped to the corresponding sepsis risk level. Finally, based on the sepsis risk level, a decision is made on whether to generate an early warning message.
[0113] Furthermore, it may also include: pushing the early warning information and corresponding supporting data to at least one of the doctor's workstation, nurse's station, and mobile terminal, and highlighting it, wherein the supporting data includes trajectory map and risk level.
[0114] Example 5:
[0115] This embodiment provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the pediatric sepsis risk warning method of Embodiment 1 above, as follows:
[0116] The system acquires time-series white blood cell count data for N consecutive days after admission or diagnosis of sepsis in the target pediatric patient, forming a raw data sequence. The raw data sequence is preprocessed to eliminate baseline differences between individuals. The preprocessed data sequence is then input into a pre-trained white blood cell dynamic trajectory classification model to output the probability of the target pediatric patient belonging to each category, and the category with the highest probability is determined as the current trajectory category of the target pediatric patient. Based on a pre-defined trajectory category-risk level mapping rule, the current trajectory category is mapped to the corresponding sepsis risk level. Finally, based on the sepsis risk level, a decision is made on whether to generate an early warning message.
[0117] Furthermore, it may also include: pushing the early warning information and corresponding supporting data to at least one of the doctor's workstation, nurse's station, and mobile terminal, and highlighting it, wherein the supporting data includes trajectory map and risk level.
[0118] It should be noted that the computer-readable storage medium in this embodiment can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.
[0119] In this embodiment, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this embodiment, the computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable storage medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable storage medium can be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.
[0120] The computer-readable storage medium described above can be used to write computer programs for executing this embodiment in one or more programming languages or combinations thereof. These programming languages include object-oriented programming languages—such as Java, Python, and C++—and conventional procedural programming languages—such as C or similar programming languages. The program can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).
[0121] In summary, this invention can automatically analyze continuous white blood cell count data after a patient is admitted to the hospital, identify its characteristic dynamic change trajectory, and automatically determine the risk level based on the trajectory pattern, and issue timely warnings.
[0122] The above description is merely a preferred embodiment of the present invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects. The scope of the present invention is defined by the appended claims rather than the foregoing description, and thus all variations falling within the meaning and scope of the equivalents of the claims are intended to be included within the present invention.
Claims
1. A pediatric sepsis risk early warning method based on white blood cell dynamic trajectory, characterized in that, The method includes: Obtain time-series white blood cell count data of the target pediatric patient for N consecutive days after admission or diagnosis of sepsis to form the raw data sequence; The original data sequences are preprocessed to eliminate baseline differences between individuals; The preprocessed data sequence is input into a pre-trained leukocyte dynamic trajectory classification model to output the probability of the target pediatric patient belonging to each category, and the category with the highest probability is determined as the current trajectory category of the target pediatric patient; According to the preset trajectory category-risk level mapping rules, the current trajectory category is mapped to the corresponding sepsis risk level; Whether to generate an early warning message depends on the sepsis risk level.
2. The pediatric sepsis risk warning method of claim 1, wherein, The preprocessing of the original data sequence includes: The original data sequence is standardized or normalized.
3. The pediatric sepsis risk alerting method of claim 1, wherein, The leukocyte dynamic trajectory classification model is constructed using a group basis trajectory modeling method. Based on time-series leukocyte count data from a historical cohort of pediatric sepsis patients, it identifies K significantly different typical trajectory categories. The training process of the leukocyte dynamic trajectory classification model includes: Obtain time-series white blood cell count data from a historical cohort of pediatric sepsis patients; Time-series white blood cell count data of historical pediatric sepsis patients were cleaned to remove outliers; for missing data, linear interpolation or carry-over from the last observation was used to fill in the gaps. Using the cleaned time-series white blood cell count data as input features, the maximum likelihood estimation method was used to fit the group basis trajectory model. Simultaneously, we examine the number of groups and the order of the trajectory function to identify potential heterogeneous trajectory subgroups in the data, and characterize the changing trends of different forms through the order of the polynomial. Candidate models were screened by comprehensively using three indicators: Bayesian information criterion, average posterior probability, and clinical interpretability, in order to determine the optimal model. The parameters of the trained optimal model were then solidified to obtain a trained white blood cell dynamic trajectory classification model.
4. The pediatric sepsis risk warning method of claim 3, wherein, The typical trajectory categories are three: persistently high value, gradually decreasing value, and persistently low value. The term "persistently high value" refers to a white blood cell count that remains at a high level throughout the entire observation period. The term "gradual decline" refers to a white blood cell count that starts from a mid-to-high baseline level and shows a gradual downward trend over time. The term "persistently low value" refers to a white blood cell count that remains at a low level throughout the observation period.
5. The pediatric sepsis risk alerting method of claim 1, wherein, The process of determining whether to generate an early warning message based on the sepsis risk level includes: If the sepsis risk level is intermediate or low, continue monitoring; If the sepsis risk level is high, an early warning message will be generated.
6. The pediatric sepsis risk warning method according to any one of claims 1-5, wherein, The method further includes: The warning information and corresponding supporting data are pushed to at least one of the doctor's workstation, nurse's station, and mobile terminal and highlighted. The supporting data includes trajectory map and risk level.
7. A pediatric sepsis risk warning system based on white blood cell dynamic trajectory, characterized in that, The system includes: The data interface module is used to acquire time-series white blood cell count data of the target pediatric patient for N consecutive days after admission or diagnosis of sepsis, forming the raw data sequence; The preprocessing module is used to preprocess the raw data sequence to eliminate baseline differences between individuals; The core analysis module is used to input the preprocessed data sequence into the pre-trained white blood cell dynamic trajectory classification model to output the probability of the target pediatric patient belonging to each category, and to determine the category with the highest probability as the current trajectory category of the target pediatric patient. The risk level determination module is used to map the current trajectory category to the corresponding sepsis risk level according to the preset trajectory category-risk level mapping rules; The early warning generation module is used to determine whether to generate early warning information based on the sepsis risk level.
8. The pediatric sepsis risk warning system of claim 7, wherein, The system also includes: The push module is used to push early warning information and corresponding supporting data to at least one of the doctor's workstation, nurse's station, and mobile terminal, and to highlight it. The supporting data includes trajectory map and risk level.
9. A computer device comprising a processor and a memory for storing a processor-executable program, characterized in that, When the processor executes the program stored in the memory, it implements the pediatric sepsis risk warning method according to any one of claims 1-7.
10. A computer-readable storage medium storing a program, characterized in that, When the program is executed by the processor, it implements the pediatric sepsis risk warning method according to any one of claims 1-7.