Management system and method of central venous catheter for blood purification

Abstract

The invention belongs to the technical field of management, and discloses a management system and method for a central venous catheter for blood purification, and the system comprises an information management unit which is used for managing the patient information of a patient and the catheter information of a catheter disposed by the patient; the state management unit is used for managing the catheter use state and the physical sign state of the patient; the maintenance management unit is used for managing dressing change records, catheter treatment records and catheter drawing treatment records; the infection management unit is used for managing infection diagnosis records and complication diagnosis records; the infection early warning unit is used for carrying out infection early warning on the patient based on an early warning mechanism and generating early warning information; and the log management unit is used for generating a management log based on the patient information, the catheter information, the catheter use state, the patient sign state, the dressing change record, the catheter treatment record, the catheter drawing treatment record, the infection diagnosis record, the complication diagnosis record and the early warning information. Management of the whole use process of the central venous catheter is achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of management, and particularly relates to a management system and method for a central venous catheter for blood purification. Background Art

[0002] The central venous catheter for blood purification is the main treatment vascular access for patients undergoing blood purification treatment, and is used for continuous extracorporeal circulation for blood purification treatment. Currently, the central venous catheter for blood purification includes temporary catheters and long-term catheters. Compared with ordinary central venous catheters, this catheter has a larger lumen and can meet the blood flow requirements of 150 - 300 ml / min for blood purification treatment. Therefore, it is more likely to be infected compared with ordinary catheters.

[0003] Catheter-related bloodstream infection is an important quality management indicator in the medical industry. The occurrence of catheter-related bloodstream infection not only directly affects the service life of the catheter, but also has a serious impact on the prognosis of patients, especially critically ill patients. One of the main methods for preventing and controlling catheter-related bloodstream infection is to conduct regular and standardized catheter specialty evaluation and maintenance.

[0004] Regarding the specialty evaluation and maintenance of the central venous catheter for blood purification, some consensus has been reached in the industry: that is, the gauze dressing is changed every 48 hours, the transparent dressing is changed once a week, and it is changed immediately when there are signs of infection.

[0005] Due to the lack of a standardized evaluation and management process in the prior art, there are still problems of incomplete evaluation and untimely maintenance in the evaluation and maintenance of the central venous catheter, which may increase the incidence of catheter-related infections. Therefore, it is urgent to develop and formulate a management system for the central venous catheter for blood purification to further standardize the clinical evaluation and maintenance of the catheter. Summary of the Invention

[0006] The purpose of the present invention is to provide a management system and method for a central venous catheter for blood purification, so as to solve the problems that due to the lack of a standardized evaluation and management process in the prior art, there are still problems of incomplete evaluation and untimely maintenance in the evaluation and maintenance of the central venous catheter, which may increase the incidence of catheter-related infections.

[0007] To achieve the above purpose, the present invention adopts the following technical solutions: In the first aspect, the present invention provides a management system for a central venous catheter for blood purification, and the system includes: An information management unit, which is used to manage the patient information of the patient and the catheter information of the catheter placed by the patient; A status management unit, which is used to manage the catheter usage status and the patient physical sign status; A maintenance management unit, which is used to manage the dressing change records, catheter treatment records, and catheter removal treatment records; An infection management unit for managing infection diagnosis records and complication diagnosis records; An infection warning unit for building a warning mechanism, warning patients of infection based on the warning mechanism, and generating warning information; A log management unit for generating a management log based on patient information, catheter information, catheter usage status, patient physical sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, complication diagnosis records, and warning information.

[0008] Preferably, the warning mechanism is real-time warning based on multiple models, and the multiple models include: an early warning model, a local infection assessment model, and a function prediction model. The infection warning unit includes: A first prediction module for predicting the early infection risk of a patient based on the early warning model to obtain an early infection prediction result; A second prediction module for assessing the infection of the catheter insertion site of a patient based on the local infection assessment model to obtain a local infection prediction result; A third prediction module for predicting the functional abnormality of a catheter based on the function prediction model to obtain a catheter infection prediction result; An infection warning module for building an infection risk level based on the early infection prediction result, the local infection prediction result, and the catheter infection prediction result, and generating warning information based on the infection risk level.

[0009] Preferably, the system further includes: A prompt management unit for automatically generating operation prompts, and the operation prompts at least include: dressing change prompts and dressing replacement prompts; A storage management unit for structurally storing patient information, catheter information, catheter usage status, patient physical sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, and complication diagnosis records.

[0010] Preferably, the patient information at least includes: name, ID, diagnosis information, and underlying diseases; The catheter information at least includes: catheterization date, type, catheter insertion site, catheter length, and lumen specification; The patient physical sign status includes: body temperature, skin status around the catheter, and limb status; The catheter usage status at least includes: blood flow rate, dressing status, and fixation status; The dressing change records at least include: dressing change time, dressing change operator, dressing type, disinfectant type, and disinfection scope and method; The catheter handling records at least include: type of sealing solution, flushing frequency, and technique; The extubation process record at least includes: extubation reason, extubation operator, post-extubation complications, and catheter tip culture results; The infection diagnosis record at least includes: infection occurrence date, etiological results, and antibiotic usage plan; The complication diagnosis record at least includes: thrombosis, catheter displacement or dislodgment, and mechanical injury.

[0011] In a second aspect, the present invention provides a management method for a central venous catheter for blood purification, which is implemented based on the above-mentioned management system for a central venous catheter for blood purification. The method includes: Obtaining multi-dimensional management information of the patient, where the multi-dimensional management information includes: patient information, catheter information, catheter usage status, patient physical sign status, dressing change record, catheter treatment record, extubation treatment record, infection diagnosis record, and complication diagnosis record; Constructing an early warning mechanism to conduct infection early warning on the patient based on the early warning mechanism and generating early warning information; Generating a management log based on patient information, catheter information, catheter usage status, patient physical sign status, dressing change record, catheter treatment record, extubation treatment record, infection diagnosis record, complication diagnosis record, and early warning information.

[0012] Preferably, the early warning mechanism is real-time early warning based on multiple models. The multiple models include: an early warning model, a local infection assessment model, and a function prediction model; Conducting infection early warning on the patient based on the early warning mechanism and generating early warning information includes: Predicting the early infection risk of the patient based on the early warning model to obtain an early infection prediction result; Evaluating the infection of the catheter placement site of the patient based on the local infection assessment model to obtain a local infection prediction result; Predicting the functional abnormality of the catheter based on the function prediction model to obtain a catheter infection prediction result; Constructing an infection risk level based on the early infection prediction result, the local infection prediction result, and the catheter infection prediction result, and generating early warning information based on the infection risk level.

[0013] Preferably, the early warning model uses a random forest model or an XGBoost model, the local infection assessment model uses a convolutional neural network model, and the function prediction model uses an LSTM network; The early warning model, the local infection assessment model, and the function prediction model are trained using federated learning.

[0014] Preferably, the architecture of the federated learning includes a federated server and multiple participants. The early warning model, the local infection assessment model, and the function prediction model deployed on the federated server are used as the initial models. The method further includes: training the initial models, including: The federated server expands the type of any initial model to obtain several extended models, and the extended models have the same model parameters as the initial models. The federated server selects at least one extended model and randomly assigns it to multiple participants. Each participant trains the assigned extended model based on local data to obtain trained local model parameters. The federated server obtains the trained local model parameters of each participant and aggregates the trained local model parameters of each participant to obtain an aggregated model corresponding to at least one extended model. The federated server then randomly selects at least one aggregated model from the aggregated models corresponding to at least one extended model and re - randomly assigns it to each participant. Each participant updates and retrains the local model parameters based on the model parameters of the assigned aggregated model to obtain new local model parameters. The federated server obtains the new local model parameters and re - aggregates the new local model parameters until the aggregated models corresponding to each extended model converge, and uses the converged aggregated model as the trained initial model.

[0015] Preferably, each participant trains the assigned extended model based on local data to obtain trained local model parameters, including: The participant initializes the model parameters of the assigned extended model. Based on the adversarial algorithm, the local data is expanded to obtain a first sample and a second sample. The extended model is trained using local data to obtain a first training result, trained using the first sample to obtain a second training result, and trained using the second sample to obtain a third training result. A loss function is constructed based on the first training result, the second training result, and the third training result. Based on the loss function, the training loss value is calculated, and the model parameters of the extended model are iteratively updated based on the training loss value until a preset iteration termination condition is reached, and the trained local model parameters are obtained.

[0016] In a third aspect, the present invention provides a computer - readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the above - mentioned management method for the central venous catheter for blood purification.

[0017] Beneficial effects: Through the management system, the present invention can manage the whole process of using a central venous catheter, including patient information, catheter information, catheter usage status, patient vital signs status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, complication diagnosis records, etc., and use an early warning mechanism to give an infection early warning to the patient, conduct an infection early warning in a timely manner, and reduce the risk of the patient being infected; at the same time, a management log is generated to facilitate medical staff to view. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the embodiments of the present invention, but do not constitute a limitation to the embodiments of the present invention. In the drawings: Figure 1 is a block diagram of a management system for a central venous catheter for blood purification provided by an embodiment of the present invention; Figure 2 is a flowchart of a management method for a central venous catheter for blood purification provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the present invention in combination with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the structure of the drawings is only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. It should be noted here that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation to the present invention.

[0020] Embodiment 1 Figure 1 is a block diagram of a management system for a central venous catheter for blood purification provided by an embodiment of the present invention. As Figure 1 shown, this embodiment provides a management system for a central venous catheter for blood purification, and the system includes: an information management unit, a status management unit, a maintenance management unit, an infection management unit, an infection early warning unit, and a log management unit; The information management unit is used to manage the patient information of the patient and the catheter information of the catheter placed by the patient; the patient information includes but is not limited to: name, ID, diagnosis information (such as uremia) and underlying diseases (such as diabetes, immunosuppressive status, etc.); the catheter information at least includes: catheterization date, type (temporary / long-term), catheterization site (internal jugular vein, femoral vein, etc.), catheter length and lumen specification; at the same time, information such as catheterization operation records such as the operator, the use of ultrasonic guidance, and intraoperative complications is recorded.

[0021] The status management unit is used to manage the catheter usage status and the patient's physical sign status; the patient's physical sign status includes but is not limited to: body temperature, skin status around the catheter, and limb status; the catheter usage status includes but is not limited to: blood flow rate (whether it stably reaches 150 - 300 ml / min), dressing status (gauze / transparent dressing, presence of blood / fluid leakage, loosening or contamination), and fixation status (suture integrity, whether the catheter has shifted).

[0022] The maintenance management unit is used to manage dressing change records, catheter handling records, and catheter removal handling records; the dressing change records include but are not limited to: dressing change time, dressing change operator, dressing type, disinfectant type (such as chlorhexidine alcohol, povidone-iodine), and disinfection scope and method ((clockwise - counterclockwise alternation)); the catheter handling records include but are not limited to: type of catheter locking solution (heparin concentration, whether it contains antibiotics) and flushing frequency and technique (pulsatile positive pressure catheter locking); the catheter removal handling records include but are not limited to: reasons for catheter removal (treatment completion, infection, loss of function), catheter removal operator, post-removal complications, and catheter tip culture results.

[0023] The infection management unit is used to manage infection diagnosis records and complication diagnosis records; the infection diagnosis records include but are not limited to: date of infection occurrence, etiological results, and antibiotic usage regimen; the complication diagnosis records include but are not limited to: thrombosis, catheter displacement or extrusion, and mechanical injury.

[0024] The infection warning unit is used to construct a warning mechanism, perform infection warning on patients based on the warning mechanism, and generate warning information.

[0025] The log management unit is used to generate a management log based on patient information, catheter information, catheter usage status, patient physical sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, complication diagnosis records, and warning information; the management log of this embodiment is as follows in the table: Management Log Table

[0026] As a further optimization of this embodiment, the warning mechanism is real-time warning based on multiple models, the multiple models include: early warning model, local infection assessment model, and function prediction model, and the infection warning unit includes: first prediction module, second prediction module, third prediction module, and infection warning module; The first prediction module is used to predict the early infection risk of a patient based on an early warning model, and obtain an early infection prediction result. In this embodiment, the early warning model adopts a random forest model or an XGBoost model; by recording static data such as the patient's age, underlying diseases (diabetes, immunosuppression), catheter type / location, and dynamic data such as body temperature curve, dressing change interval, catheter blood flow trend, and changes in inflammatory markers, a random forest algorithm or an XGBoost algorithm is used to establish an early warning model to identify key risk factors (such as overdue dressing change, albumin < 30 g / L).

[0027] The second prediction module is used to evaluate the infection of the catheterization site of the patient based on a local infection assessment model, and obtain a local infection prediction result; in this embodiment, the local infection assessment model adopts a convolutional neural network model. By collecting skin photos at the catheter exit (which can be taken by a smartphone or bedside device), the convolutional neural network model is used to identify the red and swollen area, the nature of exudate (purulent / bloody), and the area with abnormal skin temperature.

[0028] The third prediction module is used to predict the functional abnormality of the catheter based on a function prediction model, and obtain a catheter infection prediction result; in this embodiment, the function prediction model adopts an LSTM (Long Short-Term Memory) network. By recording the real-time blood flow and pressure waveform data of the blood purification machine, the LSTM network is used to automatically identify abnormal hemodynamic patterns, such as abnormal waveforms caused by thrombus or biofilm (such as periodic flow decline).

[0029] The infection warning module is used to construct an infection risk level based on the early infection prediction result, the local infection prediction result, and the catheter infection prediction result, and generate a warning message based on the infection risk level.

[0030] In this embodiment, the infection risk level has three levels, including low risk, medium risk, and high risk. At the same time, corresponding treatment measures are matched according to the infection risk level. The warning message includes constructing a risk warning form using the infection risk level, treatment measures, and prediction results, and sending the risk warning form to medical staff to facilitate medical staff to take corresponding treatment measures in a timely manner. The risk warning form is shown in the following table: Risk Warning Form

[0031] Preferably, the system further includes: a prompt management unit and a storage management unit; The prompt management unit is used to automatically generate operation prompts, and the operation prompts include but are not limited to: dressing change prompt and dressing replacement prompt; The storage management unit is used for structurally storing patient information, catheter information, catheter usage status, patient vital sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, and complication diagnosis records.

[0032] Therefore, through the management system, the present invention can manage the entire process of using a central venous catheter, including patient information, catheter information, catheter usage status, patient vital sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, complication diagnosis records, etc., and use an early warning mechanism to give an infection warning to the patient, give an infection warning in a timely manner, and reduce the risk of the patient being infected; at the same time, a management log is generated to facilitate medical staff to view.

[0033] Embodiment 2 Figure 2 It is a flowchart of a management method for a central venous catheter for blood purification provided by an embodiment of the present invention. As Figure 2 shown, this embodiment provides a management method for a central venous catheter for blood purification, which is implemented based on the management system for a central venous catheter for blood purification in Embodiment 1. The method includes: Step S10: Obtain multi-dimensional management information of the patient. The multi-dimensional management information includes: patient information, catheter information, catheter usage status, patient vital sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, and complication diagnosis records.

[0034] Step S20: Construct an early warning mechanism, and give an infection warning to the patient based on the early warning mechanism to generate warning information.

[0035] Step S30: Generate a management log based on patient information, catheter information, catheter usage status, patient vital sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, complication diagnosis records, and warning information.

[0036] As a further optimization of this embodiment, the early warning mechanism is real-time early warning based on multiple models. The multiple models include: an early warning model, a local infection assessment model, and a function prediction model; giving an infection warning to the patient based on the early warning mechanism to generate warning information includes: Step a10: Predict the early infection risk of the patient based on the early warning model to obtain the early infection prediction result. In this embodiment, the early warning model uses a random forest model or an XGBoost model. By recording static data such as the patient's age, underlying diseases (diabetes, immunosuppression), catheter type / location, and dynamic data such as the body temperature curve, dressing change interval, catheter blood flow trend, and changes in inflammatory markers, a random forest algorithm or an XGBoost algorithm is used to establish an early warning model to identify key risk factors (such as overdue dressing change, albumin < 30 g / L).

[0037] Step a20: Evaluate the infection of the catheter insertion site of the patient based on the local infection assessment model to obtain the local infection prediction result. In this embodiment, the local infection assessment model uses a convolutional neural network model. By collecting skin photos at the catheter exit (which can be taken by a smartphone or bedside device), the convolutional neural network model is used to identify the range of redness and swelling, the nature of exudate (purulent / bloody), and the area of abnormal skin temperature.

[0038] Step a30: Predict the functional abnormality of the catheter based on the functional prediction model to obtain the catheter infection prediction result. In this embodiment, the functional prediction model uses an LSTM (Long Short-Term Memory) network. By recording the real-time blood flow rate and pressure waveform data of the blood purification machine, the LSTM network is used to automatically identify abnormal hemodynamic patterns, such as abnormal waveforms caused by thrombosis or biofilm (such as periodic flow decline).

[0039] Step a40: Construct an infection risk level based on the early infection prediction result, the local infection prediction result, and the catheter infection prediction result, and generate a warning message based on the infection risk level.

[0040] In this embodiment, the infection risk level has three levels, including low risk, medium risk, and high risk. At the same time, corresponding treatment measures are matched according to the infection risk level. The warning message includes constructing a risk warning table using the infection risk level, treatment measures, and prediction results, and sending the risk warning table to the medical staff to facilitate the medical staff to take corresponding treatment measures in a timely manner.

[0041] As a further optimization of this embodiment, the fragmentation of medical data and insufficient annotation (such as the lack of a high-quality infection picture library) result in low data quality and low prediction accuracy of the model; while multi-dimensional data is usually privacy data, it is difficult to obtain more sample data. To solve this technical problem, the early warning model, local infection assessment model, and function prediction model in this embodiment are trained using federated learning. The architecture of the federated learning includes a federated server and multiple participants. Multiple medical institutions are used as participants. The multiple participants collect local multi-dimensional data samples and store them in local databases. The multiple participants are all communicatively connected to the federated server. The early warning model, local infection assessment model, and function prediction model are deployed on the federated server. The federated server is responsible for allocating the corresponding models to the corresponding participants. The multiple participants jointly participate in the training of the model to improve the prediction accuracy of the model. Since the participants use local multi-dimensional data samples to perform training, the multi-dimensional data samples are not easily leaked, which can improve the security of the data.

[0042] In this embodiment, to further improve the security of the training process, the following training steps are performed on the federated server, that is, taking the early warning model, local infection assessment model, and function prediction model as initial models and training the initial models, including: Step b10: The federated server expands the type of any initial model to obtain several extended models, and the extended models have the same model parameters as the initial model. For example, for a convolutional neural network model, the extended convolutional neural network models include LeNet structure, AlexNet structure, VGGNet network (VGG16 and VGG19), and GoogLeNet structure.

[0043] Step b20: The federated server randomly selects at least one extended model and assigns it to multiple participants.

[0044] Step b30: Each participant trains the assigned extended model based on local data to obtain trained local model parameters, where the local data is multi-dimensional data samples.

[0045] Step b40: The federated server obtains the trained local model parameters of each participant and aggregates the trained local model parameters of each participant to obtain an aggregated model corresponding to at least one extended model.

[0046] Step b50: The federated server randomly selects at least one aggregated model from the aggregated models corresponding to at least one extended model and re-randomly assigns it to each participant.

[0047] Step b60: Each participant updates and retrains the local model parameters based on the model parameters of the assigned aggregated model to obtain new local model parameters.

[0048] Step b70: The federated server obtains the new local model parameters and re-aggregates the new local model parameters until the aggregated models corresponding to the extended models all converge, and uses the converged aggregated model as the trained initial model.

[0049] In this embodiment, the federated server randomly selects a certain number of extended models, and the number and type of the extended models selected each time are different. Then, the at least one selected extended model is randomly assigned to all participants. Each time during training, the type of the extended model assigned to the participants is also random and different, which can provide better guarantee for privacy protection in federated learning and provide a better privacy protection solution for management.

[0050] In this embodiment, although federated learning solves the problem of low data quality, when federated learning is under malicious attacks of adversarial samples, it cannot make accurate judgments, resulting in low robustness of the model. Therefore, in order to improve the robustness of federated learning, each participant trains the assigned extended model based on local data to obtain the trained local model parameters, including: Step b301: The participant initializes the model parameters of the assigned extended model.

[0051] Step b302: Extend the local data based on the adversarial algorithm to obtain the first sample and the second sample.

[0052] In this embodiment, the adversarial algorithm adopts the projected gradient descent adversarial algorithm and the fast gradient sign attack algorithm; the first sample is generated by the projected gradient descent adversarial algorithm and the second sample is generated by the fast gradient sign attack algorithm.

[0053] Among them, the functional expression of the first sample is: ; In the formula, is the first sample, is the local data, is the adversarial coefficient, is the projection function, is the loss function of the projected gradient descent adversarial algorithm, is the derivative of the loss function of the projected gradient descent adversarial algorithm.

[0054] Among them, the functional expression of the second sample is: ; In the formula, is the second sample, is the local data, is the attack coefficient, is the sign function, is the loss function of the fast gradient attack algorithm, is the derivative of the loss function of the fast gradient attack algorithm.

[0055] Step b303: Train the extended model using local data to obtain a first training result, train the extended model using a first sample to obtain a second training result, and train the extended model using a second sample to obtain a third training result.

[0056] Step b304: Construct a loss function based on the first training result, the second training result, and the third training result.

[0057] In this embodiment, an initial loss function is constructed through the first training result and the label of the local data, a loss function of the mapping gradient descent adversarial algorithm is constructed through the second training result and the first sample, a loss function of the fast gradient attack algorithm is constructed through the third training result and the second sample, and the three loss functions are weighted and averaged to obtain a final loss function.

[0058] Step b305: Calculate a training loss value based on the loss function, and iteratively update the model parameters of the extended model based on the training loss value until a preset iteration termination condition is reached, to obtain trained local model parameters.

[0059] Therefore, through the mapping gradient descent adversarial algorithm and the fast gradient attack algorithm, the present invention can resist various adversarial attacks and effectively improve the model robustness of federated learning.

[0060] The present invention can manage the entire process of using a central venous catheter through a management method, including patient information, catheter information, catheter usage status, patient vital sign status, dressing change records, catheter treatment records, catheter removal treatment records, infection diagnosis records, complication diagnosis records, etc., and use an early warning mechanism to give an infection warning to the patient, give an infection warning in a timely manner, and reduce the risk of the patient being infected; at the same time, a management log is generated for medical staff to view conveniently.

[0061] Embodiment III This embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the management method of the central venous catheter for blood purification in Embodiment II is implemented.

[0062] This embodiment also provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the management method of the central venous catheter for blood purification in Embodiment II is implemented.

[0063] The present invention can manage the entire process of using a central venous catheter, including patient information, catheter information, catheter usage status, patient vital sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, complication diagnosis records, etc., and use an early warning mechanism to give an infection warning to the patient, give an infection warning in a timely manner, and reduce the risk of the patient being infected; at the same time, a management log is generated to facilitate medical staff to view.

[0064] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0065] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0066] The above are only the embodiments of the present application and are not used to limit the present application. For those skilled in the art, the present application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A management system for a central venous catheter for blood purification, characterized in that, The system includes: An information management unit for managing patient information of patients and catheter information of the catheters placed by the patients; A status management unit for managing the usage status of catheters and the physical sign status of patients; A maintenance management unit for managing dressing change records, catheter handling records, and catheter removal handling records; An infection management unit for managing infection diagnosis records and complication diagnosis records; An infection warning unit for constructing a warning mechanism, performing infection warning on patients based on the warning mechanism, and generating warning information; A log management unit for generating a management log based on patient information, catheter information, catheter usage status, patient physical sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, complication diagnosis records, and warning information.

2. The management system of the central venous catheter for blood purification according to claim 1, characterized in that, The warning mechanism is real-time warning based on multiple models. The multiple models include: an early warning model, a local infection assessment model, and a function prediction model. The infection warning unit includes: A first prediction module for predicting the early infection risk of a patient based on the early warning model to obtain an early infection prediction result; A second prediction module for assessing the infection of the catheter placement site of a patient based on the local infection assessment model to obtain a local infection prediction result; A third prediction module for predicting the functional abnormality of a catheter based on the function prediction model to obtain a catheter infection prediction result; An infection warning module for constructing an infection risk level based on the early infection prediction result, the local infection prediction result, and the catheter infection prediction result, and generating warning information based on the infection risk level.

3. The management system of the central venous catheter for blood purification according to claim 1, characterized in that, The system further includes: A prompt management unit for automatically generating operation prompts. The operation prompts at least include: dressing change prompts and dressing replacement prompts; A storage management unit for structurally storing patient information, catheter information, catheter usage status, patient physical sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, and complication diagnosis records.

4. The management system of the central venous catheter for blood purification according to claim 1, characterized in that, The patient information at least includes: name, ID, diagnosis information, and underlying diseases; The catheter information at least includes: catheter placement date, type, catheter placement site, catheter length, and lumen specification; The patient physical sign status includes: body temperature, skin status around the catheter, and limb status; The catheter usage status at least includes: blood flow rate, dressing status, and fixation status; The dressing change records at least include: dressing change time, dressing change operator, dressing type, disinfectant type, and disinfection scope and method; The catheter handling records at least include: type of sealing solution, flushing frequency, and technique; The catheter removal handling records at least include: catheter removal reason, catheter removal operator, complications after catheter removal, and catheter tip culture result; The infection diagnosis records at least include: infection occurrence date, etiological result, and antibiotic usage plan; The complication diagnosis records at least include: thrombosis, catheter displacement or prolapse, and mechanical injury.

5. A management method for a central venous catheter for blood purification, the method being implemented based on the management system for a central venous catheter for blood purification according to any one of claims 1-4, characterized in that, The method includes: Obtain the multi-dimensional management information of the patient, where the multi-dimensional management information includes: patient information, catheter information, catheter usage status, patient physical sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, and complication diagnosis records; Construct an early warning mechanism, and based on the early warning mechanism, conduct infection early warning for the patient to generate early warning information; Generate a management log based on the patient information, catheter information, catheter usage status, patient physical sign status, dressing change records, catheter handling records, catheter removal handling records, infection diagnosis records, complication diagnosis records, and early warning information.

6. The management method of the central venous catheter for blood purification according to claim 5, characterized in that The early warning mechanism is a real-time early warning based on multiple models, and the multiple models include: an early warning model, a local infection assessment model, and a function prediction model; conducting infection early warning for the patient based on the early warning mechanism to generate early warning information, including: Predict the early infection risk of the patient based on the early warning model to obtain an early infection prediction result; Conduct an infection assessment on the catheter insertion site of the patient based on the local infection assessment model to obtain a local infection prediction result; Predict the functional abnormality of the catheter based on the function prediction model to obtain a catheter infection prediction result; Construct an infection risk level based on the early infection prediction result, local infection prediction result, and catheter infection prediction result, and generate early warning information based on the infection risk level.

7. The management method of the central venous catheter for blood purification according to claim 6, characterized in that, The early warning model uses a random forest model or an XGBoost model, the local infection assessment model uses a convolutional neural network model, and the function prediction model uses an LSTM network; the early warning model, local infection assessment model, and function prediction model are trained using federated learning.

8. The management method of the central venous catheter for blood purification according to claim 7, wherein The framework of the federated learning includes a federated server and multiple participants, and the early warning model, local infection assessment model, and function prediction model deployed on the federated server are used as initial models; The method further includes: training the initial models, including: The federated server expands the type of any initial model to obtain several extended models, and the extended models have the same model parameters as the initial models; The federated server selects at least one extended model and randomly assigns it to multiple participants; Each participant trains the assigned extended model based on local data to obtain trained local model parameters; The federated server obtains the trained local model parameters of each participant and aggregates the trained local model parameters of each participant to obtain an aggregated model corresponding to at least one extended model; The federated server then randomly selects at least one aggregated model from the aggregated models corresponding to at least one extended model and re-randomly assigns it to each participant; Each participant updates and retrains the local model parameters based on the model parameters of the assigned aggregated model to obtain new local model parameters; The federated server obtains the new local model parameters and re-aggregates the new local model parameters until the aggregated models corresponding to each extended model converge, and uses the converged aggregated model as the trained initial model.

9. The management method of the central venous catheter for blood purification according to claim 1, characterized in that Each participant trains the assigned extended model based on local data to obtain trained local model parameters, including: The participant initializes the model parameters of the assigned extended model; Expand the local data based on the adversarial algorithm to obtain the first sample and the second sample; Use the local data to train the extended model to obtain the first training result, use the first sample to train the extended model to obtain the second training result, and use the second sample to train the extended model to obtain the third training result; Construct a loss function based on the first training result, the second training result, and the third training result; Based on the loss function, calculate the training loss value, and iteratively update the model parameters of the extended model based on the training loss value until the preset iteration termination condition is reached to obtain the trained local model parameters.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the management method of the central venous catheter for blood purification described in any one of claims 5-9.

Images