Nursing management system and method for multiple myeloma patients

Abstract

The present invention discloses a nursing management system and method for multiple myeloma patients, which relates to the field of medical care technology, including obtaining medical data of multiple myeloma patients through a data fusion platform, integrating medical data through multi-source heterogeneous data sets; formulating a medical plan for multiple myeloma patients based on the preprocessing results of medical data; and making intervention suggestions for the medical plan according to a medical treatment decision model to complete the nursing management of multiple myeloma patients. The present invention realizes efficient integration and high-quality processing of medical data, can timely identify and repair abnormal data, ensure the accuracy and safety of the medical plan, and significantly improves the nursing quality and treatment effect of multiple myeloma patients in combination with the intervention suggestions of the medical treatment decision model.

Description

Technical Field

[0001] The present invention relates to the field of medical care technology, in particular to a multiple myeloma patient care management system and method. Background Art

[0002] With the rapid development of information technology and the increasing attention paid to health management by society, healthcare informatics plays an increasingly important role in improving the efficiency and quality of medical services. In particular, with the support of advanced technologies such as big data and cloud computing, the medical industry is undergoing unprecedented changes. For the management of complex diseases such as multiple myeloma, although the existing medical information system can provide a certain degree of help, it is still insufficient in terms of accurate treatment plan recommendations, long-term patient follow-up management, and doctor-patient communication. In order to better meet the nursing needs of multiple myeloma patients and improve the level of disease management, it is particularly urgent to develop a more intelligent and personalized patient care management system.

[0003] The existing nursing management method for multiple myeloma patients mainly relies on traditional paper records and manual follow-up. This method is not only inefficient, but also prone to information loss or untimely updates, which seriously affects the evaluation of patient treatment effects and the formulation of subsequent treatment plans. In addition, due to the lack of effective data integration and analysis tools, it is difficult for doctors to fully grasp the changes in patients' conditions, resulting in restrictions on the formulation of personalized treatment plans. In order to solve the above problems, the present invention uses a multiple myeloma patient nursing management system and method, and uses advanced information and communication technologies to achieve efficient collection, storage, analysis and sharing of patient data, thereby providing patients with more accurate and continuous nursing services. Summary of the invention

[0004] In view of the problems existing in the existing multiple myeloma patient care management system and method, the present invention is proposed.

[0005] Therefore, the purpose of the present invention is to provide a multiple myeloma patient care management system and method. In view of the problems of data silos and insufficient standardization in the existing multiple myeloma patient care management system, the present invention achieves efficient integration and standardized processing of data by constructing a multi-source heterogeneous data fusion platform and federated learning technology, thereby improving the quality and availability of medical data.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a nursing management method for multiple myeloma patients, which includes building a data fusion platform for multiple myeloma patients, obtaining medical data of multiple myeloma patients through the data fusion platform, and integrating the medical data through a multi-source heterogeneous data set, wherein the multi-source heterogeneous data set includes completing networking of hospitals in different regions based on the Internet and obtaining a data protocol for medical information records from different regions;

[0008] The construction of the data fusion platform for multiple myeloma patients includes connecting the physical characteristic data of the multiple myeloma patients to the multiple myeloma patient care management system through an API interface, and performing standardization processing on the physical characteristic data of the multiple myeloma patients, wherein the standardization processing includes format conversion of the physical characteristic data of the multiple myeloma patients;

[0009] A treatment prediction model is used to pre-process the integrated medical data, and a medical plan is formulated for multiple myeloma patients based on the pre-processing results of the medical data. An anomaly detection algorithm is constructed in the multiple myeloma patient care management system to identify the medical plans formulated by multiple myeloma patients, and intervention recommendations are made on the medical plans based on the medical treatment decision model to complete the nursing management of multiple myeloma patients.

[0010] As a preferred embodiment of the nursing management method for multiple myeloma patients of the present invention, the construction of a data fusion platform for multiple myeloma patients includes connecting the physical characteristic data of multiple myeloma patients to the nursing management system for multiple myeloma patients through an API interface, and performing standardization processing on the physical characteristic data of multiple myeloma patients, wherein the standardization processing includes format conversion of the physical characteristic data of multiple myeloma patients;

[0011] The format conversion includes performing process decomposition on the body feature data and calculating and obtaining a standardized function. The formula for calculating and obtaining the standardized function is: ;in, represents the normalization function, represents the body feature data cleaning function, represents standardized data, and Indicates the field encoding. Represents the encoding normalization function.

[0012] As a preferred embodiment of the nursing management method for multiple myeloma patients of the present invention, the step of obtaining the medical data of multiple myeloma patients includes training the standardized function using a federated learning technique, wherein the training includes training a global model of the standardized function without exchanging the standardized data with external data;

[0013] Combining the process of standardized data and federated learning technology, constructing a standardized data processing flow, using the standardized data as the input of the federated learning technology, and uploading it to the central server in the processing flow, wherein the central server aggregates the standardized data, generates a local model, and sends the trained global model to each local model, thereby completing the acquisition of medical data of multiple myeloma patients;

[0014] The formula for sending the trained global model to each local model is: ;in, Indicates The global model parameters after rounds of iterations, Indicates The local model is The local model parameters after rounds of iterations, represents the weight coefficient, represents the learning rate, represents the gradient of the loss function, Indicates the number of times.

[0015] As a preferred solution of the nursing management method for multiple myeloma patients of the present invention, the medical data integration includes repairing the medical data of multiple myeloma patients using a Gaussian mixture model, and the repair includes calculating the probability density value of the data points of each local model using a Gaussian mixture model GMM, and comparing it with the adaptive threshold Make comparisons;

[0016] When the probability density value is less than the threshold When , the local model at this time is abnormal;

[0017] Otherwise, when the probability density value is greater than or equal to the threshold When , the local model is normal;

[0018] The calculation formula of the probability density value is: ;in, represents the probability density value, represents the Gaussian component, represents the mean medical data of multiple myeloma patients, represents the data points of each local model, represents the mixing weight, Indicates the number of times;

[0019] When the local model is abnormal, the abnormal local model is optimized by using the interpolation method. The optimization includes linear interpolation of the data points of the local model before and after the normal local model. The calculation formula of the linear interpolation is: ;in, and represents the adjacent normal data points of each local model, and Represents the timestamps corresponding to the adjacent normal data points of each local model, represents linear interpolation, Indicates time.

[0020] As a preferred solution of the nursing management method for multiple myeloma patients of the present invention, the preprocessing of the integrated medical data includes training the integrated medical data to generate a treatment prediction model, and the formula for training the integrated medical data is: ;in, represents the treatment prediction model, represents the actual observed value, Indicates The feature vector of the samples, Indicates the number of times, Represents model parameters Next pair of samples predictions, represents the loss function, Indicates The actual observed values, represents the training network, represents the model parameters;

[0021] The formulation of a medical plan for a multiple myeloma patient includes generating a treatment plan for the patient based on the output of the treatment prediction model, wherein the treatment plan includes a treatment effect assessment, a risk assessment, and a comprehensive assessment;

[0022] when When , the treatment effect evaluation is valid; otherwise, when When , the treatment effect evaluation is invalid, among which, It represents the treatment effect assessment threshold in the multiple myeloma patient care management system;

[0023] when When , the risk assessment is valid; otherwise, The risk assessment is invalid when represents the risk assessment threshold in the multiple myeloma patient care management system;

[0024] when and When , the patient receives treatment according to the treatment plan generated; otherwise, when and When the patient is given a treatment plan, the treatment is abandoned.

[0025] As a preferred solution of the method for nursing management of multiple myeloma patients described in the present invention, wherein: the abnormality detection algorithm constructed includes: Identifying the medical plan developed for a patient with multiple myeloma, the formula for constructing the anomaly detection algorithm is: ;in, represents the probability density value of each data point, Indicates the constructed anomaly identification period, Indicates critical value;

[0026] When a local model is abnormal and the timestamps corresponding to the adjacent normal data points of each local model are on weekends or holidays, it is marked as a first-level medical abnormality;

[0027] When a local model is abnormal and the timestamps corresponding to the adjacent normal data points of each local model are on a weekday, it is marked as a secondary medical abnormality;

[0028] The making intervention suggestion for the medical plan includes notifying the medical plan according to the medical abnormality.

[0029] As a preferred solution of the nursing management method for multiple myeloma patients of the present invention, wherein: the notification of the medical plan includes analyzing according to the primary medical abnormality and the secondary medical abnormality;

[0030] When the multiple myeloma patient care management system detects a level one medical anomaly, the patient is flagged for care through institutional care;

[0031] When the multiple myeloma patient care management system detects a secondary care anomaly, the patient is flagged for care through institutional care or secondary care.

[0032] In a second aspect, the present invention provides a multiple myeloma patient care management system, comprising:

[0033] An integration module, which builds a data fusion platform for multiple myeloma patients, obtains the medical data of multiple myeloma patients through the data fusion platform, and integrates the medical data through multi-source heterogeneous data sets;

[0034] A preprocessing module, which uses a treatment prediction model to preprocess the integrated medical data and formulates a medical plan for multiple myeloma patients based on the preprocessing results of the medical data;

[0035] The nursing management module builds an anomaly detection algorithm in the multiple myeloma patient nursing management system, identifies the medical plans formulated by multiple myeloma patients, makes intervention recommendations on the medical plans based on the medical treatment decision model, and completes the nursing management of multiple myeloma patients.

[0036] The beneficial effects of the present invention are:

[0037] The present invention solves the problems of data silos and insufficient standardization by constructing a multi-source heterogeneous data fusion platform and federated learning technology, and realizes efficient integration and high-quality processing of medical data. Through Gaussian mixture models and anomaly detection algorithms, it can timely identify and repair abnormal data to ensure the accuracy and safety of medical plans. The personalized treatment plan generated based on the treatment prediction model, combined with the intervention recommendations of the medical treatment decision model, significantly improves the nursing quality and treatment effect of multiple myeloma patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0039] Figure 1 A specific flow chart of a multiple myeloma patient care management method provided by one embodiment of the present invention.

[0040] Figure 2 A nursing management schematic diagram of a nursing management method for multiple myeloma patients provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0042] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0043] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0044] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0045] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0046] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0047] Embodiment 1 of the present invention: Referring to Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a nursing management method for multiple myeloma patients, comprising:

[0048] S1: Constructing a data fusion platform for multiple myeloma patients, obtaining the medical data of multiple myeloma patients through the data fusion platform, and integrating the medical data through multi-source heterogeneous data sets, wherein the multi-source heterogeneous data sets include completing networking of hospitals in different regions based on the Internet and obtaining data protocols for medical information records from different regions; constructing the data fusion platform for multiple myeloma patients includes connecting the physical characteristic data of multiple myeloma patients to the multiple myeloma patient care management system through an API interface, and standardizing the physical characteristic data of multiple myeloma patients, wherein the standardization includes format conversion of the physical characteristic data of multiple myeloma patients.

[0049] Among them, building a data fusion platform for multiple myeloma patients includes connecting the physical characteristic data of multiple myeloma patients to the multiple myeloma patient care management system through an API interface, and standardizing the physical characteristic data of multiple myeloma patients, and the standardization includes format conversion of the physical characteristic data of multiple myeloma patients;

[0050] The format conversion includes process decomposition of the body feature data and calculation of the standardized function. The formula for calculating the standardized function is: ;in, represents the normalization function, represents the body feature data cleaning function, represents standardized data, and Indicates the field encoding. Represents the encoding normalization function.

[0051] S1.1: Acquiring medical data of multiple myeloma patients includes training a standardized function using a federated learning technique, wherein the training includes training a global model of the standardized function without exchanging the standardized data with external data;

[0052] Combining the process of standardized data and federated learning technology, a standardized data processing flow is constructed, standardized data is used as the input of the federated learning technology, and uploaded to the central server in the processing flow. The central server includes aggregating standardized data, generating local models, and sending the trained global model to each local model to complete the acquisition of medical data of multiple myeloma patients;

[0053] The formula for sending the trained global model to each local model is: ;in, Indicates The global model parameters after rounds of iterations, Indicates The local model is The local model parameters after rounds of iterations, represents the weight coefficient, represents the learning rate, represents the gradient of the loss function, Indicates the number of times.

[0054] Furthermore, during the training process, Hospital A and Hospital B each used local standardized data to train local models. After the training was completed, the parameters of these local models were uploaded to the central server. The central server aggregated the uploaded local model parameters to generate a global model. In the first round of iterations, the central server received the local model parameters of Hospital A and Hospital B, which were recorded as 0.85 for Hospital A and 0.78 for Hospital B, respectively. These parameters were aggregated according to the weight coefficient and learning rate to generate a global model parameter value of 0.82.

[0055] The central server sends the trained global model parameter value of 0.82 to each medical institution. Hospital A and Hospital B continue to train their local models using the new global model parameters. This process is repeated for multiple rounds until the model converges. For example, after the fifth round of iterations, the global model parameter value generated by the central server stabilizes at 0.84, indicating that the model has converged. In this way, each medical institution can jointly improve the performance of the model without sharing the original data, ensuring the efficiency and security of the acquisition and processing of medical data of multiple myeloma patients. The training data is shown in Table 1 below:

[0056] Table 1 Training data display table

[0057] Iteration round Hospital A local model parameters Hospital B local model parameters Central server global model parameters Round 1 0.85 0.78 0.82 Round 2 0.83 0.80 0.82 Round 3 0.84 0.81 0.83 Round 4 0.84 0.82 0.83 Round 5 0.84 0.83 0.84

[0058] As can be seen from Table 1, with the increase in iteration rounds, the local model parameters of Hospital A and Hospital B gradually approach each other, and the global model parameters generated by the central server gradually stabilize. After the fifth round of iteration, the global model parameter value stabilized at 0.84, indicating that the model has converged. This approach enables various medical institutions to jointly improve the performance of the model without sharing the original data, ensuring the efficiency and security of the acquisition and processing of medical data for multiple myeloma patients.

[0059] S1.2: The integration of medical data includes the use of Gaussian mixture models to repair the medical data of multiple myeloma patients. The repair includes the use of Gaussian mixture models GMM to calculate the probability density values ​​of the data points of each local model and the adaptive threshold Make comparisons;

[0060] When the probability density value is less than the threshold When , the local model at this time is abnormal;

[0061] Otherwise, when the probability density value is greater than or equal to the threshold When , the local model is normal;

[0062] The calculation formula for the probability density value is: ;in, represents the probability density value, represents the Gaussian component, represents the mean medical data of multiple myeloma patients, represents the data points of each local model, represents the mixing weight, Indicates the number of times;

[0063] When the local model is abnormal, the interpolation method is used to optimize the abnormal local model. The optimization includes linear interpolation of the data points of the local model before and after the normal local model. The calculation formula of linear interpolation is: ;in, and represents the adjacent normal data points of each local model, and Represents the timestamps corresponding to the adjacent normal data points of each local model, represents linear interpolation, Indicates time.

[0064] Furthermore, the Gaussian mixture model is used to repair the medical data of multiple myeloma patients. Specifically, the probability density value of the data points of each local model is calculated and compared with the adaptive threshold. Assuming that the adaptive threshold is set to 0.05, for a data point of a local model, its probability density value is calculated. For a heart rate data point of patient A of 75 times / minute, the calculated probability density value is 0.04. The calculated probability density value is compared with the adaptive threshold. If the probability density value is less than the threshold, the data point is considered to be an outlier. Since the probability density value 0.04 is less than 0.05, the heart rate data point is determined to be Abnormal value; if the probability density value is greater than or equal to the threshold, the data point is considered to be a normal value. For a heart rate data point of patient B, the heart rate data point is 78 times / minute, and the calculated probability density value is 0.06, so the heart rate data point is determined to be a normal value; when the local model is abnormal, the abnormal local model is optimized by interpolation. For example, for the abnormal heart rate data point of patient A of 75 times / minute, the adjacent normal data points are used for linear interpolation. Assuming that the adjacent normal data points are 78 times / minute and 80 times / minute respectively, the corresponding timestamps are 10:00 and 10:10 respectively, and the new data point obtained by interpolation is 79 times / minute.

[0065] S2: Use the treatment prediction model to pre-process the integrated medical data, and formulate a medical plan for multiple myeloma patients based on the pre-processing results of the medical data.

[0066] Among them, preprocessing the integrated medical data includes training the integrated medical data to generate a treatment prediction model. The formula for training the integrated medical data is: ;in, represents the treatment prediction model, represents the actual observed value, Indicates The feature vector of the samples, Indicates the number of times, Represents model parameters Next pair of samples predictions, represents the loss function, Indicates The actual observed values, represents the training network, represents the model parameters;

[0067] Formulating a medical plan for multiple myeloma patients includes generating a treatment plan for the patient based on the output of the treatment prediction model, and the treatment plan includes treatment effect evaluation, risk assessment and comprehensive assessment;

[0068] when When , the treatment effect evaluation is valid; otherwise, when When , the treatment effect evaluation is invalid, among which, It represents the treatment effect assessment threshold in the multiple myeloma patient care management system;

[0069] when When , the risk assessment is valid; otherwise, The risk assessment is invalid when represents the risk assessment threshold in the multiple myeloma patient care management system;

[0070] when and When , the patient receives treatment according to the treatment plan generated; otherwise, when and When the patient is given a treatment plan, the treatment is abandoned.

[0071] S3: Build an anomaly detection algorithm in the multiple myeloma patient care management system to identify the medical plans formulated by multiple myeloma patients, make intervention recommendations on the medical plans based on the medical treatment decision model, and complete the care management of multiple myeloma patients.

[0072] Among them, building anomaly detection algorithms includes adaptive threshold based Identify the medical plans made by patients with multiple myeloma. The formula for building an anomaly detection algorithm is: ;in, represents the probability density value of each data point, Indicates the constructed anomaly identification period, Indicates critical value;

[0073] When a local model is abnormal and the timestamps corresponding to the adjacent normal data points of each local model are on weekends or holidays, it is marked as a first-level medical abnormality;

[0074] When a local model is abnormal and the timestamps corresponding to the adjacent normal data points of each local model are on a weekday, it is marked as a secondary medical abnormality;

[0075] Recommendations for intervention in the medical plan include notification of the medical plan based on medical anomalies.

[0076] S3.1: Notification of the medical plan includes analysis based on primary and secondary medical anomalies;

[0077] When the multiple myeloma patient care management system detects a level one medical anomaly, the patient is flagged for care through institutional care;

[0078] When the multiple myeloma patient care management system detects a secondary care anomaly, the patient is flagged for care through institutional care or secondary care.

[0079] In a preferred embodiment, a multiple myeloma patient care management system includes an integration module, which constructs a data fusion platform for multiple myeloma patients, obtains medical data of multiple myeloma patients through the data fusion platform, and integrates the medical data through multi-source heterogeneous data sets; a preprocessing module, which uses a treatment prediction model to preprocess the integrated medical data, and formulates a medical plan for multiple myeloma patients based on the preprocessing results of the medical data; a nursing management module, which constructs an anomaly detection algorithm in the multiple myeloma patient care management system, identifies the medical plan formulated by the multiple myeloma patient, makes intervention recommendations on the medical plan according to the medical treatment decision model, and completes the nursing management of multiple myeloma patients.

[0080] The above-mentioned unit modules may be embedded in or independent of a processor in a computer device in the form of hardware, or may be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to the above-mentioned modules.

[0081] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0082] In summary, the present invention solves the problems of data silos and insufficient standardization by constructing a multi-source heterogeneous data fusion platform and federated learning technology, and realizes efficient integration and high-quality processing of medical data. Through the Gaussian mixture model and anomaly detection algorithm, it can timely identify and repair abnormal data to ensure the accuracy and safety of the medical plan. The personalized treatment plan generated based on the treatment prediction model, combined with the intervention recommendations of the medical treatment decision model, significantly improves the nursing quality and treatment effect of patients with multiple myeloma.

[0083] Embodiment 2 of the present invention: Referring to Figure 1 and Figure 2 , which is the second embodiment of the present invention, provides a multiple myeloma patient care management system and method. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0084] The present invention constructs a data fusion platform for multiple myeloma patients, integrates and repairs medical data using federated learning technology and Gaussian mixture models, generates a treatment prediction model, and formulates a personalized treatment plan based on the model output. Physical characteristic data of multiple myeloma patients, such as hemoglobin level, white blood cell count and heart rate, are obtained from multiple data sources such as hospital electronic medical record systems, laboratory test result systems and wearable devices through API interfaces. These data are standardized, including data cleaning, format conversion and coding normalization. For example, hemoglobin levels are uniformly converted to g / dL, and heart rate data are uniformly converted to times / minute. Federated learning technology is used to train the standardized data to generate a local model. Hospital A and Hospital B use local standardized data to train local models, respectively. The generated local model parameters are 0.85 and 0.78, respectively. The central server aggregates these parameters to generate a global model parameter value of 0.82, and the global model parameters are sent to various medical institutions. After the fifth round of iteration, the global model parameter value is stabilized at 0.84. The Gaussian mixture model is used to train the local model of multiple myeloma. The medical data of myeloma patients were repaired, and the probability density values ​​of the data points of each local model were calculated and compared with the adaptive threshold of 0.05. For a heart rate data point of 75 beats / minute of patient A, the calculated probability density value was 0.04, which was determined to be an abnormal value. The adjacent normal data points of 78 beats / minute and 80 beats / minute were used for linear interpolation, and the new data point obtained by interpolation was 79 beats / minute. The generated treatment prediction model was used to preprocess the integrated medical data, and a treatment plan was generated for the patient based on the model output. For patient A, the treatment effect value predicted by the model was 0.85, and the treatment effect evaluation threshold was 0.7, which determined that the treatment effect evaluation was effective. The treatment risk value predicted by the model was 0.3, and the risk evaluation threshold was 0.5, which determined that the risk evaluation was invalid. Therefore, it was recommended that patient A accept the current treatment plan; on the contrary, for patient B, the treatment effect value predicted by the model was 0.65, which was less than the threshold of 0.7, and the treatment effect evaluation was invalid. The treatment risk value predicted by the model was 0.6, which was greater than the threshold of 0.5, and the risk evaluation was effective. Therefore, it was recommended that patient B consider giving up treatment. The experimental data are shown in Table 2 below:

[0085] Table 2 Experimental data table

[0086] Data Types Specific data value Remark Hemoglobin level 12.5 g / dL Standardized data Heart rate 75 times / minute Standardized data White blood cell count 7.2 × 10^3 / μL Standardized data Hospital A local model parameters 0.85 First iteration Hospital B local model parameters 0.78 First iteration Central server global model parameters 0.82 First iteration Central server global model parameters 0.84 Fifth iteration Patient A's heart rate data points 75 times / minute Raw data Probability density value of patient A's heart rate data point 0.04 Calculation results Adaptive Threshold 0.05 Anomaly detection threshold Interpolated heart rate data points 79 times / minute Interpolation results of adjacent normal data points before and after Prediction value of treatment effect of patient A 0.85 Model Output Treatment effect assessment threshold 0.7 Preset Threshold Treatment risk prediction value for patient A 0.3 Model Output Treatment risk assessment thresholds 0.5 Preset Threshold Prediction value of treatment effect for patient B 0.65 Model Output Treatment risk prediction value for patient B 0.6 Model Output

[0087] Table 2 clearly shows the specific data involved in the technical steps and their uses, ensuring the scientificity and accuracy of the medical data management and treatment plan formulation for multiple myeloma patients. The comparison of the prior art of the present invention is shown in Table 3 below:

[0088] Table 3 Comparison table of the prior art of the present invention

[0089] Compare Projects Prior art Technical solution of the present invention Advantages Data silo problem It is difficult to share data between medical institutions, resulting in data fragmentation Build a multi-source heterogeneous data fusion platform and obtain data through API interfaces Solve data silo problems and improve data integrity Data Standardization Inconsistent data formats make effective integration difficult Data cleaning, format conversion and coding standardization Ensure data consistency and comparability Privacy Protection There may be risks of privacy leakage during data sharing Federated learning technology is used to keep data local Protect patient privacy and enhance data security Model Training Requires a large amount of centralized data, and training efficiency is low Federated learning technology is used to train local models locally and aggregate global models Improve training efficiency and reduce data transmission costs Anomaly Detection Lack of effective anomaly detection mechanism may lead to misjudgment Anomaly detection using Gaussian mixture models Timely discover and repair abnormal data to improve data quality Evaluation of treatment efficacy Reliance on manual experience, highly subjective Objective assessment based on treatment prediction models Improve the accuracy and reliability of assessments risk assessment Lack of quantitative indicators, the evaluation results are not accurate enough Quantitative assessment based on risk value predicted by the model Provide more accurate risk assessment to guide clinical decision making Personalized treatment plan Lack of personalized treatment plans and unsatisfactory results Generate personalized treatment plans based on model output Improve treatment outcomes and patient outcomes Data repair The method of handling abnormal data is single and the effect is limited Data repair using interpolation and Kalman filtering Improve the accuracy and robustness of data repair System Integration Each system operates independently and lacks a coordination mechanism Integrate modular design to achieve seamless integration of data and models Improve overall system performance and scalability

[0090] From Table 3, it can be seen that the present invention has obvious advantages in data island problems, data standardization, privacy protection, model training, anomaly detection, treatment effect evaluation, risk assessment, personalized treatment plan, data repair and system integration, and can better meet the needs of nursing management of multiple myeloma patients.

[0091] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for nursing management of patients with multiple myeloma, characterized in that: include: S1: constructing a data fusion platform for multiple myeloma patients, obtaining medical data of multiple myeloma patients through the data fusion platform, and integrating the medical data through multi-source heterogeneous data sets, wherein the multi-source heterogeneous data sets include completing networking of hospitals in different regions based on the Internet and obtaining data protocols for medical information records from different regions; constructing the data fusion platform for multiple myeloma patients includes connecting the physical characteristic data of multiple myeloma patients to the multiple myeloma patient care management system through an API interface, and standardizing the physical characteristic data of multiple myeloma patients, wherein the standardization includes format conversion of the physical characteristic data of multiple myeloma patients; S1.1: Acquiring medical data of multiple myeloma patients includes training a standardized function using a federated learning technique, wherein the training includes training a global model of the standardized function without exchanging the standardized data with external data; Combining the process of standardized data and federated learning technology, a standardized data processing flow is constructed, standardized data is used as the input of the federated learning technology, and uploaded to the central server in the processing flow. The central server includes aggregating standardized data, generating a global model, and sending the trained global model to each local model to complete the acquisition of medical data of multiple myeloma patients; The formula for sending the trained global model to each local model is: ； in, Indicates The global model parameters after rounds of iterations, Indicates The local model is The local model parameters after rounds of iterations, represents the weight coefficient, represents the learning rate, represents the gradient of the loss function, Indicates the number of models; S1.2: The integration of medical data includes the use of Gaussian mixture models to repair the medical data of multiple myeloma patients. The repair includes the use of Gaussian mixture models GMM to calculate the probability density values ​​of the data points of each local model and the adaptive threshold Make comparisons; When the probability density value is less than the threshold When , the local model at this time is abnormal; Otherwise, when the probability density value is greater than or equal to the threshold When , the local model is normal; The calculation formula for the probability density value is: ； in, represents the probability density value, represents the Gaussian component, represents the mean medical data of multiple myeloma patients, represents the data points of each local model, represents the mixing weight; When the local model is abnormal, the interpolation method is used to optimize the abnormal local model. The optimization includes linear interpolation of the data points of the local model before and after the normal local model. The calculation formula of linear interpolation is: ； in, and represents the adjacent normal data points of each local model, and Represents the timestamps corresponding to the adjacent normal data points of each local model, represents linear interpolation, Indicates time; S2: Use the treatment prediction model to pre-process the integrated medical data, and formulate a medical plan for multiple myeloma patients based on the pre-processing results of the medical data; Among them, preprocessing the integrated medical data includes training the integrated medical data to generate a treatment prediction model. The formula for training the integrated medical data is: ； in, represents the treatment prediction model, represents the actual observed value, Indicates The feature vector of the samples, Indicates the number of times, Represents model parameters Next pair of samples predictions, represents the loss function, Indicates The actual observed values, represents the training network, represents the model parameters; Generate a treatment plan for the patient based on the output of the treatment prediction model; S3: Build an anomaly detection algorithm in the multiple myeloma patient care management system to identify the medical plans formulated by multiple myeloma patients, make intervention recommendations on the medical plans based on the medical treatment decision model, and complete the care management of multiple myeloma patients.

2. The method for nursing and managing multiple myeloma patients according to claim 1, characterized in that: The format conversion includes performing process decomposition on the body feature data and calculating and obtaining a standardized function. The formula for calculating and obtaining the standardized function is: ； in, represents the normalization function, represents the body feature data cleaning function, represents standardized data, and Indicates the field encoding. Represents the encoding normalization function.

3. A multiple myeloma patient care management system, based on the multiple myeloma patient care management method according to any one of claims 1 to 2, characterized in that: include: An integration module, which builds a data fusion platform for multiple myeloma patients, obtains the medical data of multiple myeloma patients through the data fusion platform, and integrates the medical data through multi-source heterogeneous data sets; A preprocessing module, which uses a treatment prediction model to preprocess the integrated medical data and formulates a medical plan for multiple myeloma patients based on the preprocessing results of the medical data; The nursing management module builds an anomaly detection algorithm in the multiple myeloma patient nursing management system, identifies the medical plans formulated by multiple myeloma patients, makes intervention recommendations on the medical plans based on the medical treatment decision model, and completes the nursing management of multiple myeloma patients.

Images