Risk early warning system for adjusting patient monitoring levels

Abstract

The disclosed embodiments disclose a risk warning system for adjusting patient monitoring levels. A specific implementation of the system includes: obtaining historical and current patient monitoring data for a patient to be evaluated, as well as retrieving basic patient information for the patient to be evaluated; generating pre- and post-operative vital sign comparison data; selecting a complication structure tree that matches the patient to be evaluated to obtain a target complication structure tree; and modifying the target complication structure tree to obtain a modified complication structure tree; obtaining case data that matches the pre- and post-operative vital sign comparison data from a pre-connected case data server, and processing each acquired case data using a second data processing device to obtain a processed case data set; generating a personalized patient prediction model; and performing a risk level assessment on the patient to be evaluated, and controlling a human-computer interaction device to execute corresponding risk warning operations. This implementation can improve patient safety.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the medical field and the computer technology field, and particularly to a risk warning system for adjusting a patient monitoring level. Background Art

[0002] Studies have shown that the probability of unplanned return to the ICU (Intensive Care Unit) is high for critically ill neurosurgery patients after they are transferred out of the ICU. Once an unplanned return to the ICU occurs, a series of problems will arise, including increased hospitalization costs and length of stay, increased delirium, and reduced prognosis. Therefore, it is particularly important to identify high-risk groups of patients transferred out of neurological critical care early so that they can be given precise care and reduce the incidence of unplanned return to the ICU. Currently, in clinical practice, the risk prediction of unplanned return to the ICU mostly uses manual judgment, and the prediction results are highly subjective and vary greatly due to the different professional abilities of each nurse. Therefore, there is an urgent need for a risk warning system that can accurately and objectively predict high-risk groups of unplanned return to the ICU, improve risk monitoring of patients, and thereby improve patient safety.

[0003] The above information disclosed in this Background section is only for enhancement of understanding of the background of the inventive concept and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0004] The content of this disclosure is used to briefly introduce concepts that will be described in detail in the detailed description section below. The content of this disclosure is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0005] Some embodiments of the present disclosure propose a risk warning system for adjusting a patient monitoring level to solve the technical problems mentioned in the above background technology section.

[0006] In a first aspect, some embodiments of the present disclosure provide a risk warning system for adjusting a patient monitoring level, the system comprising: obtaining historical monitoring data and current monitoring data of a patient to be evaluated from a first data preprocessing device, and retrieving basic patient information of the patient to be evaluated from a multi-physiological parameter database; generating pre- and post-operative physical sign comparison data based on the current monitoring data, historical monitoring data and basic patient information, wherein the pre- and post-operative physical sign comparison data is used to characterize changes in physical sign data of the patient to be evaluated before and after the operation; selecting a complication tree structure that matches the patient to be evaluated from a set of complication tree structures pre-established in the multi-physiological parameter database to obtain a target complication tree structure, and performing a comparison of the target complication structure based on the pre- and post-operative physical sign comparison data. The complication tree is modified to obtain a modified complication tree; case data matching the pre- and post-operative physical sign comparison data are obtained from a pre-connected case data server, and each acquired case data is processed by a second data processing device to obtain a processed case data set; the processed case data set is input into a risk assessment module to obtain a patient personalized prediction model, wherein the risk assessment module performs online additional training on the risk prediction basic model through the processed case data set to generate a patient personalized prediction model; the risk level of the patient to be evaluated is assessed through the patient personalized prediction model and the modified complication tree to obtain risk level assessment information, and according to the risk level assessment information, the human-computer interaction device is controlled to perform corresponding risk warning operations.

[0007] In a second aspect, some embodiments of the present disclosure provide a risk warning device for adjusting a patient monitoring level, the device comprising: an acquisition unit, configured to acquire historical patient monitoring data and current monitoring data of a patient to be evaluated from a first data preprocessing device, and retrieve basic patient information of the patient to be evaluated from a multi-physiological parameter database; a first generation unit, configured to generate pre- and post-operative physical sign comparison data based on the current monitoring data, historical monitoring data and basic patient information, wherein the pre- and post-operative physical sign comparison data is used to characterize changes in physical sign data of the patient to be evaluated before and after the operation; a selection and correction unit, configured to select a complication structure tree matching the patient to be evaluated from a set of complication structure trees pre-established in the multi-physiological parameter database, obtain a target complication structure tree, and perform correction on the target complication structure tree based on the pre- and post-operative physical sign comparison data. The tree is modified to obtain a modified complication structure tree; the acquisition and processing unit is configured to obtain case data matching the above-mentioned pre- and post-operative physical sign comparison data from a pre-connected case data server, and perform data processing on each acquired case data through a second data processing device to obtain a processed case data set; the model training unit is configured to input the above-mentioned processed case data set into the risk assessment module to obtain a patient personalized prediction model, wherein the above-mentioned risk assessment module performs online additional training on the risk prediction basic model through the above-mentioned processed case data set to generate a patient personalized prediction model; the assessment and risk warning unit is configured to perform risk level assessment on the patient to be assessed through the above-mentioned patient personalized prediction model and the above-mentioned modified complication structure tree to obtain risk level assessment information, and control the human-computer interaction device to perform corresponding risk warning operations based on the above-mentioned risk level assessment information.

[0008] In a third aspect, some embodiments of the present disclosure provide an electronic device comprising: one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by one or more processors, the one or more processors implement the system described in any implementation method of the above-mentioned first aspect.

[0009] In a fourth aspect, some embodiments of the present disclosure provide a computer-readable medium having a computer program stored thereon, wherein when the program is executed by a processor, the system described in any implementation of the first aspect is implemented.

[0010] The above-mentioned various embodiments of the present disclosure have the following beneficial effects: Through the risk warning system for adjusting patient monitoring levels of some embodiments of the present disclosure, the monitoring accuracy of patient monitoring can be improved and patient safety can be enhanced. Specifically, the reason for insufficient patient monitoring and a high probability of unplanned ICU return is that manual judgment is often used, the prediction results are highly subjective, and there is a large difference due to the professional ability of each nursing staff. Based on this, some embodiments of the risk warning system for adjusting patient monitoring levels of the present disclosure firstly realize the automation of patient monitoring by setting up a risk warning system and including a first data preprocessing device, a multi-physiological parameter database, a risk assessment module, a second data processing device, and a human-computer interaction device. Specifically, first, historical patient monitoring data and current patient monitoring data of the patient to be evaluated are obtained from the first data preprocessing device, and the patient basic information of the patient to be evaluated is retrieved from the multi-physiological parameter database. Based on the current monitoring data, historical monitoring data, and the patient basic information, pre- and post-operative vital sign comparison data are generated, wherein the pre- and post-operative vital sign comparison data is used to characterize the changes in the vital sign data of the patient to be evaluated before and after surgery. Considering the impact of surgery on the patient's body, a multi-physiological parameter database is established to retain pre- and post-operative data for comparison. This allows for identification of changes in the patient's physical condition before and after surgery. A complication structure tree matching the patient to be evaluated is then selected from the pre-established set of complication structure trees in the multi-physiological parameter database to generate a target complication structure tree. This target complication structure tree is then modified based on the pre- and post-operative physical sign comparison data to generate a revised complication structure tree. Considering that different surgeries correspond to different complications, as well as varying degrees and probabilities of complications, a tree structure is established to associate corresponding complications with different surgeries. This classification and refinement of the tree structure further enhances the risk warning system's ability to predict a patient's post-operative condition. Furthermore, the introduction of the tree structure allows for rapid identification of associated complications and their probabilities within a vast amount of data. Furthermore, considering the diverse physical conditions of different patients, the complication structure tree is further adjusted based on pre- and post-operative physical sign comparison data, i.e., utilizing the patient's own physical changes, to generate a revised complication structure tree. This makes the revised complication structure tree more suitable for the patient to be evaluated. Then, case data matching the pre- and post-operative physical sign comparison data is obtained from a pre-connected case data server. Each acquired case data is processed by a second data processing device to generate a processed case data set. This processed case data set is input into a risk assessment module to generate a patient-specific prediction model. The risk assessment module then performs online training on a basic risk prediction model using the processed case data set to generate the patient-specific prediction model.Considering that significant coupling still exists when only referring to changes in the patient's own vital signs, the pre-trained risk prediction basic model is supplemented online with processed case data that matches the patient's vital signs and case data. This results in a personalized patient prediction model. The inclusion of case data similar to the patient's allows the model training to not only largely preserve the patient's own changes but also reduce coupling by supplementing the model with similar case data. Finally, the risk level of the patient being evaluated is assessed using this personalized patient prediction model and the revised complication structure tree. Risk level assessment information is obtained, and based on this risk level assessment information, the human-computer interaction device is controlled to execute the corresponding risk warning operation. This allows for timely adjustment of the patient's monitoring level, reducing the risk of unplanned ICU return and improving patient safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The above and other features, advantages, and aspects of the various embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0012] Figure 1 is a flow chart of some embodiments of a risk warning system for adjusting a patient monitoring level according to the present disclosure;

[0013] Figure 2 It is a schematic diagram of the complication tree;

[0014] Figure 3 This is a partial schematic diagram of the revised complication tree;

[0015] Figure 4 is a schematic structural diagram of some embodiments of a risk warning device for adjusting a patient monitoring level according to the present disclosure;

[0016] Figure 5 It is a structural diagram of an electronic device suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0017] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of protection of the present disclosure.

[0018] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other.

[0019] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0020] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, they should be understood as "one or more".

[0021] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0022] With regard to the collection, storage, and use of user personal information (such as user vital signs data, user basic data, etc.) involved in this disclosure, before performing the corresponding operations, the relevant organizations or individuals shall fulfill their obligations, including conducting personal information security impact assessments, fulfilling the obligation to inform the personal information subjects, and obtaining the authorization and consent of the personal information subjects in advance.

[0023] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0024] Figure 1 A process 100 of some embodiments of a risk warning system for adjusting a patient monitoring level according to the present disclosure is shown. The above-mentioned risk warning system includes: a first data preprocessing device, a multi-physiological parameter database, a risk assessment module, a second data processing device, and a human-computer interaction device. Among them, the first data preprocessing device can be a device for processing personal data of patients to be evaluated. The second data processing device can be a device for processing other case data. The multi-physiological parameter database can be used to store patient data. The risk assessment module can be a module for performing risk assessment on patients to be evaluated to determine whether they have the risk of unplanned return to the ICU and adjust the monitoring strategy. The human-computer interaction device can be a device for notifying the corresponding medical terminal to issue a risk warning.

[0025] The risk warning system for adjusting the patient monitoring level includes the following steps:

[0026] Step 101 : Acquire historical patient monitoring data and current patient monitoring data of a patient to be evaluated from a first data preprocessing device, and retrieve basic patient information of the patient to be evaluated from a multi-physiological parameter database.

[0027] In some embodiments, the risk warning system (i.e., the executing entity) for adjusting the patient monitoring level can obtain the patient historical monitoring data and current monitoring data of the patient to be evaluated from the first data preprocessing device in a wired or wireless manner, and retrieve the patient basic information of the above-mentioned patient to be evaluated from the multi-physiological parameter database.

[0028] It should be noted that the computing device (executing entity) described above can be either hardware or software. When the computing device is hardware, it can be implemented as a distributed cluster consisting of multiple servers or terminal devices, or as a single server or terminal device. When the computing device is software, it can be installed in the hardware devices listed above. It can be implemented as multiple software programs or software modules, for example, to provide distributed services, or as a single software program or software module. This is not specifically limited here. It should be understood that the number of computing devices can be any number depending on implementation needs.

[0029] In some optional implementations of some embodiments, the execution subject obtains historical patient monitoring data and current patient monitoring data of the patient to be evaluated from the first data preprocessing device, and retrieves basic patient information of the patient to be evaluated from the multi-physiological parameter database, including:

[0030] The first step is to control the first data preprocessing device to obtain initial historical vital sign data and current monitoring data from the patient's vital sign monitoring device. The current monitoring data includes: postoperative baseline data, postoperative muscle strength test values, postoperative metabolic score values, postoperative mental cognition score values, disease identification, surgical identification, and indwelling drainage tube information.

[0031] Here, postoperative baseline data may be basic data representative of the patient's physical condition. For example, blood pressure, blood oxygen concentration, body temperature, heart rate, arterial pressure, respiratory rate, hematocrit, and white blood cell count. The postoperative muscle strength test value may be the patient's pre-tested muscle strength level after surgery. For example, the muscle strength level may be measured using Manual Muscle Testing (MMT). The postoperative metabolic score may be used to represent the patient's postoperative metabolic capacity. For example, it may be measured using the Mini-Nutrition Assessment or the Nutrition Risk in Critically Ill Score (NUTRIC). The postoperative mental cognition score may represent the patient's mental state as measured by mental health assessment methods. For example, mental health assessment methods may include the Mini-Mental State Examination (MMSE), the Montreal Cognitive Assessment (MoCA), the Hamilton Depression Rating Scale (HAMD), or the Brief Psychiatric Rating Scale (BPRS). Furthermore, the disease identifier may be a unique identifier of the patient's condition. The surgical identifier may be an identifier of the surgical procedure performed on the patient. The indwelling drainage tube information may include the number of indwelling drainage tubes and the drainage tube identifier corresponding to each indwelling drainage tube. The indwelling drainage tube information indicates the status of the drainage tubes remaining in the patient's body after surgery.

[0032] The second step is to control the first data preprocessing device to perform data standardization on the initial historical vital sign data to obtain the patient historical monitoring data. The patient historical monitoring data includes preoperative muscle strength test values, preoperative metabolic score values, preoperative basic data, and preoperative vital sign ability values. Secondly, data standardization can be performed by arranging the various data items in the initial historical vital sign data according to a preset field order to obtain the patient historical monitoring data. This facilitates subsequent data comparison. Here, the patient historical monitoring data can represent various data items measured before the patient undergoes surgery.

[0033] The third step is to retrieve the basic information of the patient to be evaluated from the multi-physiological parameter database, wherein the basic information of the patient includes: past medical history information, smoking duration, and drinking duration.

[0034] Step 102: Generate pre- and post-operative vital sign comparison data based on current monitoring data, historical monitoring data, and basic patient information.

[0035] In some embodiments, the execution entity may generate pre- and post-operative physical sign comparison data based on the current monitoring data, historical monitoring data, and the patient's basic information. The pre- and post-operative physical sign comparison data is used to characterize changes in the physical sign data of the patient to be evaluated before and after the operation.

[0036] In practice, when predicting risk for patients before and after surgery, the pre- and post-operative data can be compared. Alternatively, the two sets of data can be compared: obtaining the patient's physical sign data before and after ICU discharge to predict the possibility of an unplanned return to the ICU. Alternatively, obtaining the patient's physical sign data before and after ICU discharge to predict whether the patient will need to return to the ICU after surgery. Therefore, no specific limitations are given.

[0037] In some optional implementations of some embodiments, the execution entity generates pre- and post-operative physical sign comparison data based on the current monitoring data, historical monitoring data, and the patient's basic information, including:

[0038] In the first step, the postoperative baseline data in the current monitoring data is evaluated using a preset scoring scale to generate a current baseline data score. This current baseline data score is then compared with the health score in the preoperative baseline data to generate an acute physiology and chronic health score. The scoring scale may be the Apache2 (Acute Physiology and Chronic Health) scoring scale. The scoring scale can be used to evaluate the patient's postoperative baseline data (e.g., body temperature, mean arterial pressure, heart rate, respiratory rate, oxygenation, blood pH, hematocrit, white blood cell count, etc.) to generate a current baseline data score. The current baseline data score can represent the overall score of the patient's postoperative baseline data. The difference between the current baseline data score and the health score in the preoperative baseline data can then be determined as the acute physiology and chronic health score.

[0039] The second step is to generate a muscle strength change grade based on the absolute value of the change between the preoperative muscle strength test value and the postoperative muscle strength test value. The muscle strength change grade can be determined as the difference between the muscle strength grades corresponding to the preoperative muscle strength test value and the postoperative muscle strength test value.

[0040] For example, the absolute difference in muscle strength change is two levels (from a level 5 strength test value to a level 3 strength test value after surgery). In practice, a decrease in muscle strength of two levels or more after surgery can be considered a significant change.

[0041] The third step is to generate a metabolic change coefficient based on the postoperative metabolic score and the preoperative metabolic score. The metabolic difference between the postoperative metabolic score and the preoperative metabolic score can be determined as the metabolic change coefficient.

[0042] The fourth step is to generate a mental cognition change coefficient based on the postoperative mental cognition score and the preoperative mental cognition score in the preoperative basic data. The mental cognition change coefficient can be determined as the difference between the postoperative mental cognition score and the preoperative mental cognition score in the preoperative basic data.

[0043] In the fifth step, the above-mentioned patient basic information, the above-mentioned acute physiological and chronic health scores, the above-mentioned muscle strength change level, the above-mentioned metabolic change coefficient, the above-mentioned mental cognitive change coefficient, the disease mark, the surgical mark and the indwelling drainage tube information are determined as the physical sign comparison data before and after the operation.

[0044] Step 103, selecting a complication tree structure that matches the patient to be evaluated from a set of complication tree structures pre-established in the multi-physiological parameter database to obtain a target complication tree structure, and modifying the target complication tree structure based on the pre- and post-operative physical sign comparison data to obtain a modified complication tree structure.

[0045] In some embodiments, the above-mentioned execution entity can select a complication tree structure that matches the above-mentioned patient to be evaluated from a set of complication tree structures pre-established in a multi-physiological parameter database to obtain a target complication tree structure, and modify the above-mentioned target complication tree structure based on the above-mentioned pre- and post-operative physical sign comparison data to obtain a modified complication tree structure.

[0046] In some optional implementations of some embodiments, the execution entity selects a complication tree structure that matches the patient to be evaluated from a set of complication tree structures pre-established in a multi-physiological parameter database to obtain a target complication tree structure, including:

[0047] A complication tree structure that matches the current monitoring data, including the surgical identifier and the disease identifier, is selected from the multi-physiological parameter database as the target complication tree structure. If no match is found, a new single-item match is performed using the surgical identifier or the disease identifier. If a single-item match is found for one complication tree structure, it is used as the target complication tree structure. If two single-item matches are found for two complication tree structures, the two complication tree structures are merged into a target complication tree structure, which includes the associated complication identifiers and corresponding complication probabilities.

[0048] In practice, given the complexity and sheer number of complications, manual verification can easily overlook certain items or inaccurately evaluate them. Therefore, by introducing a complication structure tree, a corresponding complication tree can be established for each procedure and condition. This allows for intuitive extraction of the corresponding complication and determination of its probability of occurrence. Furthermore, the complication structure tree is constructed using a single primary node—that is, it is constructed with each procedure identifier or condition identifier as the primary node. Therefore, when matching both the procedure identifier and the condition identifier simultaneously, it is easy to match two complication structures, each corresponding to a non-overlapping set of complications. Consequently, when two complication structures are matched, the subnodes (i.e., the complication identifier nodes) under the two primary nodes (the procedure identifier and the condition identifier) ​​can be merged into a single structure tree to obtain the target complication structure tree. This not only reduces the time and computing resources required for data matching but also provides technical support for subsequent condition matching.

[0049] As an example, Figure 2 The diagram shows a complication structure tree. The main node is a neurosurgery procedure (e.g., ventricular shunt surgery). Each corresponding complication category is a child node, namely, intracranial complications, infection-related complications, metabolic / electrolyte disorders, and other complications. Each child node can also be divided into at least one complication as a leaf node. For example, the leaf nodes of intracranial complications include leaf nodes such as recurrence of hydrocephalus, subdural hematoma, and intracranial infection. Each leaf node also includes a corresponding complication probability. The complication probability of a child node is the sum of the leaf nodes under it.

[0050] In some optional implementations of some embodiments, the execution subject modifies the target complication structure tree based on the pre- and post-operative physical sign comparison data to obtain a modified complication structure tree, including:

[0051] The first step is to determine the probability correction ratio and probability correction identifier corresponding to the above-mentioned patient basic information, the above-mentioned acute physiological and chronic health scores, the above-mentioned muscle strength change level, the above-mentioned metabolic change coefficient, the above-mentioned mental cognitive change coefficient, the disease identifier, the surgical identifier, and the indwelling drainage tube information in the above-mentioned pre- and post-operative physical sign comparison data according to the pre-selected probability correction table.

[0052] Among them, the probability correction table may include various fields and corresponding probability correction ratios. Here, each field in the above-mentioned probability correction table may correspond to multiple field value intervals, and each field value may correspond to a probability correction ratio. Specifically, each field may be a field value interval and a corresponding probability correction ratio preselected by a piecewise function. Thus, the corresponding probability correction ratio can be extracted for each parameter in the above-mentioned patient basic information, the above-mentioned acute physiology and chronic health score, the above-mentioned muscle strength change level, the above-mentioned metabolic change coefficient, the above-mentioned mental cognitive change coefficient, the disease identification, the surgical identification, and the indwelling drainage tube information.

[0053] In the second step, the concurrent probability values ​​corresponding to the respective probability correction identifiers in the target complication structure tree are modified by using the respective probability correction ratios to obtain a modified complication structure tree.

[0054] For example, for every five-point increase in the acute physiology and chronic health score, the probability of inflammatory complications increases by 5% (up to a maximum of 50%). A muscle strength change of grade 2 or higher increases the probability of neuromuscular complications by 20%. Each drainage tube indwelling for more than seven days increases the probability of intracranial complications by 8%. A postoperative temperature exceeding 38.5°C for three consecutive days increases the probability of infection-related complications by multiplying by 1.5. Furthermore, patient information, including past medical history, smoking history, and alcohol consumption, can be modified to reflect these probabilities. For example, a history of cardiovascular disease increases the risk of deep vein thrombosis by 25%. A recent smoking history of less than six months increases the risk of lung infection by 35% and the risk of delayed wound healing by 25%. Chronic alcohol abuse increases the risk of electrolyte imbalance by 42%. Alcohol consumption within three days of surgery increases the probability of coagulation abnormalities by 30%.

[0055] like Figure 3 As shown, a partial diagram of the revised complication structure tree. Figure 2 For different patients, the probability values ​​after correction are all improved. This further shows that for different patients, it is necessary to introduce the changes in their individual physical characteristics before and after surgery to modify the complication structure tree to make it more consistent with the patient's individual disease situation.

[0056] Step 104 , obtaining case data that matches the pre- and post-operative physical sign comparison data from the pre-connected case data server, and performing data processing on each acquired case data by a second data processing device to obtain a processed case data set.

[0057] In some embodiments, the execution entity may obtain case data that matches the pre- and post-operative physical sign comparison data from a pre-connected case data server, and process the obtained case data through a second data processing device to obtain a processed case data set.

[0058] In some optional implementations of some embodiments, the execution subject obtains case data matching the pre- and post-operative physical sign comparison data from a pre-connected case data server, and processes each acquired case data through a second data processing device to obtain a processed case data set, including:

[0059] The first step is to obtain case data from the case data server that matches the pre- and post-operative physical sign comparison data and the revised complication structure tree, thereby obtaining a case dataset. The case data server can include data from a multi-physiological parameter database and case data from other shared platforms. Secondly, a match can be achieved by matching the surgical and disease identifiers in the case data. This means extracting case data corresponding to the same case and surgical procedure.

[0060] In the second step, the second data processing device is used to filter each case data item in the case data set to obtain a processed case data set. Each processed case data item in the processed case data set matches the age group, gender, surgical location identifier, surgical identifier, and disease identifier of the patient to be evaluated. Here, the second data processing device can be used to filter the case data so that the processed case data item matches the age group, gender, surgical location identifier, surgical identifier, and disease identifier of the patient to be evaluated. This allows for the selection of similar case data items. This avoids model prediction errors caused by age group and gender differences, thereby improving the model's prediction accuracy for complication probability.

[0061] Step 105: Input the processed case data set into the risk assessment module to obtain a patient-personalized prediction model.

[0062] In some embodiments, the execution entity may input the processed case dataset into a risk assessment module to generate a patient-specific prediction model. The risk assessment module may perform online additional training on the risk prediction basic model using the processed case dataset to generate the patient-specific prediction model.

[0063] In practice, the above-mentioned basic risk prediction model can be a pre-set teacher model. Here, the teacher model structure may include an input layer, a first hidden layer, a second hidden layer, and an output layer. Specifically, the input dimension of the input layer can be N×D (i.e., the number of input batches × the feature dimension), for example, 1×20. The output dimension can be N×128. The first hidden layer can utilize a bidirectional LSTM (Long Short-Term Memory) and a fully connected layer for feature processing. The second hidden layer can perform feature fusion based on the feature dimension through feature concatenation and an attention mechanism. The output layer can utilize a fully connected layer and a softmax normalization layer as the prediction output, outputting a sequence of probability prediction values ​​of the same length as the pre-set complication sequence. Thus, the basic risk prediction model is obtained by pre-training the teacher model. Furthermore, the dataset used in the pre-training process includes data from various case studies. This enables the teacher model to have the basic ability to predict the probability of various complications based on the patient's physical condition.

[0064] Here, through knowledge distillation, the aforementioned basic risk prediction model can be used as a teacher model, and the processed case dataset can be used as training samples to train a corresponding student model, resulting in a personalized patient prediction model. Specifically, the student model architecture can include a personalized input layer (for example, processing the processed case data into an N×5 dimension as input and outputting an N×64-dimensional data matrix), a nonlinear activation layer, a regularization layer, a fully connected layer, a concatenation layer, a fused fully connected layer, and an output layer. During student model training, the weights of the teacher model can be frozen. This allows for a personalized complication prediction model to be tailored to each patient while also relying on the teacher model to reduce coupling. Furthermore, different student models can be trained at different postoperative time points, or the same student model can be trained cumulatively.

[0065] Step 106 , performing risk level assessment on the patient to be assessed using the patient personalized prediction model and the modified complication structure tree, obtaining risk level assessment information, and controlling the human-computer interaction device to perform corresponding risk warning operations based on the risk level assessment information.

[0066] In some embodiments, the above-mentioned execution entity can perform risk level assessment on the patient to be evaluated through the above-mentioned patient personalized prediction model and the above-mentioned revised complication structure tree, obtain risk level assessment information, and control the human-computer interaction device to perform corresponding risk warning operations based on the above-mentioned risk level assessment information.

[0067] In some optional implementations of some embodiments, the execution subject performs a risk level assessment on the patient to be evaluated using the patient personalized prediction model and the modified complication structure tree to obtain risk level assessment information, and controls the human-computer interaction device to perform corresponding risk warning operations based on the risk level assessment information, including:

[0068] The first step is to assess the risk level of the patient to be evaluated using the personalized patient prediction model to obtain a sequence of risk level assessment probability values. Each risk level assessment probability value in the sequence corresponds to a node in the revised complication structure tree. The current monitoring data, historical monitoring data, and basic patient information of the patient to be evaluated can be combined into model input data according to a preset field order, which is then fed into the personalized patient prediction model to obtain the sequence of risk level assessment probability values.

[0069] The second step is to perform a secondary adjustment on the revised complication structure tree using the risk level assessment probability value sequence to obtain a secondary adjusted complication structure tree. For each risk level assessment probability value in the risk level assessment probability value sequence and the probability value of the corresponding node in the revised complication structure tree, the weighted sum of the two probability values ​​can be used as the secondary adjusted probability value. This yields the secondary adjusted complication structure tree.

[0070] In the third step, in response to the secondary adjusted complication structure tree satisfying a preset first maintenance adjustment condition, maintenance strategy information is generated based on the secondary adjusted complication structure tree, and the maintenance strategy information is transmitted to the maintenance processing terminal via the human-computer interaction device. The first maintenance adjustment condition may be that there are no nodes in the secondary adjusted complication structure tree that exceed the warning probability value (i.e., the complication probability value does not exceed the corresponding warning probability value). Thus, the patient data corresponding to the complication identifier whose complication probability value exceeds the low-risk warning probability value can be used as a maintenance target, and corresponding maintenance strategy information can be generated. For example, if the maintenance target is a postoperative muscle strength test value, then the maintenance strategy information is information targeted at muscle strength training recovery. For another example, if the maintenance target is a postoperative metabolic score value, then the maintenance strategy information is information targeted at metabolic recovery. Here, the maintenance processing terminal may be a medical terminal.

[0071] In the fourth step, in response to the second adjusted complication structure tree meeting the preset second maintenance adjustment condition, the human-computer interaction device is controlled to execute a corresponding risk warning operation to adjust the patient's monitoring level. The second maintenance adjustment condition can be the presence of a node in the second adjusted complication structure tree exceeding a warning probability value (i.e., a complication probability value exceeding the corresponding warning probability value). For example, the probability of an intracranial complication exceeding the warning probability value. Consequently, the human-computer interaction device can be controlled to issue a warning prompt to the medical terminal, allowing for adjustment of the patient's monitoring level, such as returning the patient to the ICU monitoring room.

[0072] In practice, the modified complication tree structure, adjusted for a single condition or procedure, still exhibits a high degree of coupling. Therefore, a model prediction approach was introduced to train a basic risk prediction model from a large number of cases. This makes the model's predictions of complication probability more universal. However, given the hidden differences between different procedures or cases, using the same model for predictions would be difficult to accurately identify individual cases. Therefore, a knowledge distillation approach was introduced to train a student model using similar case data. This further improves the student model's accuracy in predicting complication probabilities associated with a single procedure or case. Finally, by performing a secondary adjustment on the modified complication tree, the results of the two approaches are combined. This results in predictions that are not only universal but also more tailored to the actual circumstances of the patients being evaluated. This allows for accurate and objective predictions of whether a patient will need to return to the ICU. This enhances patient risk monitoring and ensures patient safety.

[0073] The above-mentioned various embodiments of the present disclosure have the following beneficial effects: Through the risk warning system for adjusting patient monitoring levels of some embodiments of the present disclosure, the monitoring accuracy of patient monitoring can be improved and patient safety can be enhanced. Specifically, the reason for insufficient patient monitoring and a high probability of unplanned ICU return is that manual judgment is often used, the prediction results are highly subjective, and there is a large difference due to the professional ability of each nursing staff. Based on this, some embodiments of the risk warning system for adjusting patient monitoring levels of the present disclosure firstly realize the automation of patient monitoring by setting up a risk warning system and including a first data preprocessing device, a multi-physiological parameter database, a risk assessment module, a second data processing device, and a human-computer interaction device. Specifically, first, historical patient monitoring data and current patient monitoring data of the patient to be evaluated are obtained from the first data preprocessing device, and the patient basic information of the patient to be evaluated is retrieved from the multi-physiological parameter database. Based on the current monitoring data, historical monitoring data, and the patient basic information, pre- and post-operative vital sign comparison data are generated, wherein the pre- and post-operative vital sign comparison data is used to characterize the changes in the vital sign data of the patient to be evaluated before and after surgery. Considering the impact of surgery on the patient's body, a multi-physiological parameter database is established to retain pre- and post-operative data for comparison. This allows for identification of changes in the patient's physical condition before and after surgery. A complication structure tree matching the patient to be evaluated is then selected from the pre-established set of complication structure trees in the multi-physiological parameter database to generate a target complication structure tree. This target complication structure tree is then modified based on the pre- and post-operative physical sign comparison data to generate a revised complication structure tree. Considering that different surgeries correspond to different complications, as well as varying degrees and probabilities of complications, a tree structure is established to associate corresponding complications with different surgeries. This classification and refinement of the tree structure further enhances the risk warning system's ability to predict a patient's post-operative condition. Furthermore, the introduction of the tree structure allows for rapid identification of associated complications and their probabilities within a vast amount of data. Furthermore, considering the diverse physical conditions of different patients, the complication structure tree is further adjusted based on pre- and post-operative physical sign comparison data, i.e., utilizing the patient's own physical changes, to generate a revised complication structure tree. This makes the revised complication structure tree more suitable for the patient to be evaluated. Then, case data matching the pre- and post-operative physical sign comparison data is obtained from a pre-connected case data server. Each acquired case data is processed by a second data processing device to generate a processed case data set. This processed case data set is input into a risk assessment module to generate a patient-specific prediction model. The risk assessment module then performs online training on a basic risk prediction model using the processed case data set to generate the patient-specific prediction model.Considering that significant coupling still exists when only referring to changes in the patient's own vital signs, the pre-trained risk prediction basic model is supplemented online with processed case data that matches the patient's vital signs and case data. This results in a personalized patient prediction model. The inclusion of case data similar to the patient's allows the model training to not only largely preserve the patient's own changes but also reduce coupling by supplementing the model with similar case data. Finally, the risk level of the patient being evaluated is assessed using this personalized patient prediction model and the revised complication structure tree. Risk level assessment information is obtained, and based on this risk level assessment information, the human-computer interaction device is controlled to execute the corresponding risk warning operation. This allows for timely adjustment of the patient's monitoring level, reducing the risk of unplanned ICU return and improving patient safety.

[0074] Further references Figure 4 As an implementation of the systems shown in the above figures, the present disclosure provides some embodiments of a risk warning device for adjusting a patient monitoring level. These device embodiments are similar to Figure 1 Corresponding to the system embodiments shown, the risk warning device for adjusting the patient monitoring level can be specifically applied to various electronic devices.

[0075] like Figure 4As shown, some embodiments of the risk warning device 400 for adjusting the patient monitoring level include: an acquisition unit 401, a first generation unit 402, a selection and correction unit 403, an acquisition and processing unit 404, a model training unit 405, and an estimation and risk warning unit 406. The acquisition unit 401 is configured to acquire the patient's historical monitoring data and current monitoring data of the patient to be evaluated from the first data pre-processing device, and retrieve the patient's basic information of the patient to be evaluated from the multi-physiological parameter database; the first generation unit 402 is configured to generate pre- and post-operative physical sign comparison data based on the current monitoring data, historical monitoring data, and the patient's basic information, wherein the pre- and post-operative physical sign comparison data is used to characterize the changes in the physical sign data of the patient to be evaluated before and after the operation; the selection and correction unit 403 is configured to select a complication tree structure that matches the patient to be evaluated from a set of complication tree structures pre-established in the multi-physiological parameter database to obtain a target complication tree structure, and correct the target complication tree structure based on the pre- and post-operative physical sign comparison data to obtain a corrected complication tree structure; the acquisition and correction unit 404 is configured to obtain the target complication tree structure and correct the target complication tree structure based on the pre- and post-operative physical sign comparison data to obtain a corrected complication tree structure; the acquisition and correction unit 405 is configured to obtain the target complication tree structure and correct the target complication tree structure based on the pre- and post-operative physical sign comparison data to obtain a corrected complication tree structure; the acquisition and correction unit 406 is configured to obtain the target complication tree structure and correct the target complication tree structure based on the pre- and post-operative physical sign comparison data to obtain a corrected complication tree structure. The processing unit 404 is configured to obtain case data that matches the above-mentioned before-and-after physical sign comparison data from a pre-connected case data server, and to perform data processing on each acquired case data through a second data processing device to obtain a processed case data set; the model training unit 405 is configured to input the above-mentioned processed case data set into the risk assessment module to obtain a patient personalized prediction model, wherein the above-mentioned risk assessment module performs online additional training on the risk prediction basic model through the above-mentioned processed case data set to generate a patient personalized prediction model; the assessment and risk warning unit 406 is configured to perform risk level assessment on the patient to be evaluated through the above-mentioned patient personalized prediction model and the above-mentioned revised complication structure tree, obtain risk level assessment information, and control the human-computer interaction device to perform corresponding risk warning operations based on the above-mentioned risk level assessment information.

[0076] It is understood that the various units recorded in the risk warning device 400 for adjusting the patient monitoring level are similar to those in the reference Figure 1 Therefore, the operations, features and beneficial effects described above for the system are also applicable to the device 400 and the units included therein, and will not be repeated here.

[0077] Reference below Figure 5 , which shows a structural diagram of an electronic device (such as a risk warning system) suitable for implementing some embodiments of the present disclosure. Figure 5 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure. Figure 5As shown, the computer device includes a processor, a memory and a network interface connected via a system bus, wherein the memory may include a non-volatile storage medium and an internal memory. The non-volatile storage medium may store an operating system and a computer program. The computer program includes program instructions, which, when executed, enable the processor to execute any risk warning system for adjusting the patient monitoring level. The processor is used to provide computing and control capabilities to support the operation of the entire computer device. The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium, which, when executed by the processor, enables the processor to execute any risk warning system for adjusting the patient monitoring level. The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will understand that Figure 5 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present disclosure, and does not constitute a limitation on the computer device to which the solution of the present disclosure is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0078] It should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0079] Wherein, in one embodiment, the processor is used to run a computer program stored in a memory to implement the following steps: obtaining the patient's historical monitoring data and current monitoring data of the patient to be evaluated from a first data preprocessing device, and retrieving the patient's basic information of the patient to be evaluated from a multi-physiological parameter database; generating before-and-after physical sign comparison data based on the current monitoring data, historical monitoring data and the patient's basic information, wherein the before-and-after physical sign comparison data is used to characterize the changes in the physical sign data of the patient to be evaluated before and after the operation; selecting a complication structure tree that matches the patient to be evaluated from a set of complication structure trees pre-established in the multi-physiological parameter database to obtain a target complication structure tree, and performing a comparison of the target complication structure tree based on the before-and-after physical sign comparison data. The tree is modified to obtain a modified complication structure tree; case data matching the before-and-after physical sign comparison data are obtained from a pre-connected case data server, and each acquired case data is processed by a second data processing device to obtain a processed case data set; the processed case data set is input into a risk assessment module to obtain a patient personalized prediction model, wherein the risk assessment module performs online additional training on the risk prediction basic model through the processed case data set to generate a patient personalized prediction model; risk level assessment is performed on the patient to be evaluated through the patient personalized prediction model and the modified complication structure tree to obtain risk level assessment information, and according to the risk level assessment information, the human-computer interaction device is controlled to perform corresponding risk warning operations.

[0080] An embodiment of the present disclosure also provides a computer-readable storage medium, on which a computer program is stored. The computer program includes program instructions. The method implemented when the program instructions are executed can refer to the various embodiments of the risk warning system for adjusting the patient monitoring level disclosed in the present disclosure.

[0081] The computer-readable storage medium may be an internal storage unit of the computer device described in the aforementioned embodiment, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a SmartMediaCard (SMC), a Secure Digital (SD) card, a flash memory card, etc., provided on the computer device.

[0082] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or system comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or system comprising the element.

[0083] The above descriptions are merely some preferred embodiments of the present disclosure and illustrate the underlying technical principles. Those skilled in the art should understand that the scope of the invention encompassed by the embodiments of the present disclosure is not limited to technical solutions formed by specific combinations of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned inventive concept. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure.

Claims

1. A risk warning system for adjusting patient monitoring levels, characterized in that: The risk warning system includes: a first data preprocessing device, a multi-physiological parameter database, a risk assessment module, a second data processing device and a human-computer interaction device, wherein: Acquire patient historical monitoring data and current monitoring data of the patient to be evaluated from a first data preprocessing device, and retrieve patient basic information of the patient to be evaluated from a multi-physiological parameter database, wherein the first data preprocessing device is controlled to acquire initial historical vital sign data and current monitoring data from a patient vital sign monitoring device, wherein the current monitoring data includes: postoperative basic data, postoperative muscle strength test value, postoperative metabolic score value, postoperative mental cognition score value, disease identification, surgical identification, and indwelling drainage tube information; control the first data preprocessing device to perform data standardization on the initial historical vital sign data to obtain the patient historical monitoring data, wherein the patient historical monitoring data includes: preoperative muscle strength test value, preoperative metabolic score value, and preoperative basic data; retrieve the patient basic information of the patient to be evaluated from the multi-physiological parameter database, wherein the patient basic information includes: past medical history information, smoking duration, and drinking duration; Generate pre- and post-operative physical sign comparison data based on the current monitoring data, historical monitoring data, and the patient's basic information, wherein the pre- and post-operative physical sign comparison data is used to characterize changes in the physical sign data of the patient to be evaluated before and after the operation; Selecting a complication tree structure that matches the patient to be evaluated from a set of complication tree structures pre-established in a multi-physiological parameter database to obtain a target complication tree structure, and revising the target complication tree structure based on the pre- and post-operative physical sign comparison data to obtain a revised complication tree structure; Acquiring case data that matches the pre- and post-operative physical sign comparison data from a pre-connected case data server, and performing data processing on each acquired case data by a second data processing device to obtain a processed case data set, wherein case data that matches the pre- and post-operative physical sign comparison data and the revised complication structure tree is acquired from the case data server to obtain the case data set; screening each case data in the case data set by the second data processing device to obtain a processed case data set, wherein each processed case data in the processed case data set matches the age group, gender, surgical location identifier, surgical identifier, and disease identifier of the patient to be evaluated; Inputting the processed case data set into a risk assessment module to obtain a patient-personalized prediction model, wherein the risk assessment module performs online additional training on the risk prediction basic model using the processed case data set to generate a patient-personalized prediction model, wherein, through knowledge distillation, the risk prediction basic model is used as a teacher model, and the processed case data set is used as a training sample to train the corresponding student model to obtain the patient-personalized prediction model; The risk level of the patient to be evaluated is assessed using the patient personalized prediction model and the modified complication structure tree to obtain risk level assessment information, and the human-computer interaction device is controlled to perform corresponding risk warning operations based on the risk level assessment information.

2. The system according to claim 1, wherein: The generating of pre- and post-operative physical sign comparison data based on the current monitoring data, historical monitoring data, and the patient's basic information includes: Evaluate the postoperative basic data in the current monitoring data using a preset scoring scale to generate a current basic data scoring value, and compare the current basic data scoring value with the health scoring value in the preoperative basic data to generate an acute physiological and chronic health score; generating a muscle strength change grade according to an absolute value of a change between the preoperative muscle strength test value and the postoperative muscle strength test value; generating a metabolic change coefficient according to the postoperative metabolic score value and the preoperative metabolic score value; generating a mental cognition change coefficient according to the postoperative mental cognition score value and the preoperative mental cognition score value in the preoperative basic data; The patient's basic information, the acute physiological and chronic health scores, the muscle strength change level, the metabolic change coefficient, the mental cognitive change coefficient, the disease mark, the surgery mark and the indwelling drainage tube information are determined as the pre- and post-operative physical sign comparison data.

3. The system according to claim 2, characterized in that The step of selecting a complication tree structure that matches the patient to be evaluated from a set of complication tree structures pre-established in the multi-physiological parameter database to obtain a target complication tree structure includes: A complication tree structure that matches the current monitoring data, including the surgical identifier and the disease identifier, is selected from the multi-physiological parameter database as the target complication tree structure, wherein, in response to no match, a single match is re-performed using the surgical identifier or the disease identifier, and in response to a single match to one complication tree structure, it is used as the target complication tree structure, and in response to a single match to two complication tree structures, the two complication tree structures are fused into a target complication tree structure, wherein the target complication tree structure includes the associated complication identifiers and the corresponding concurrency probabilities.

4. The system according to claim 3, characterized in that The target complication tree structure is modified based on the pre- and post-operative physical sign comparison data to obtain a modified complication tree structure, including: According to a pre-selected probability correction table, the probability correction ratio and probability correction indicator corresponding to the patient's basic information, the acute physiological and chronic health score, the muscle strength change level, the metabolic change coefficient, the mental cognitive change coefficient, the disease indicator, the surgical indicator, and the indwelling drainage tube information in the pre- and post-operative physical sign comparison data are respectively determined; By using each probability correction ratio, the concurrent probability value corresponding to each probability correction identifier in the target complication structure tree is obtained to obtain a corrected complication structure tree.

5. The system according to claim 4, characterized in that The step of performing risk level assessment on the patient to be evaluated by using the patient personalized prediction model and the modified complication structure tree to obtain risk level assessment information, and controlling the human-computer interaction device to perform corresponding risk warning operations based on the risk level assessment information, includes: Performing risk level assessment on the patient to be evaluated using the patient personalized prediction model to obtain a risk level assessment probability value sequence, wherein each risk level assessment probability value in the risk level assessment probability value sequence corresponds to a node in the revised complication structure tree; Performing a secondary adjustment on the revised complication tree structure using the risk level assessment probability value sequence to obtain a secondary adjusted complication tree structure; In response to the secondarily adjusted complication structure tree satisfying a preset first maintenance adjustment condition, generating maintenance strategy information according to the secondarily adjusted complication structure tree, and sending the maintenance strategy information to a maintenance processing terminal via the human-computer interaction device; In response to the complication structure tree after the secondary adjustment satisfying the preset second maintenance adjustment condition, the human-computer interaction device is controlled to perform a corresponding risk warning operation to adjust the patient monitoring level.

6. A risk warning device for adjusting a patient monitoring level, comprising: an acquisition unit configured to acquire patient historical monitoring data and current monitoring data of a patient to be evaluated from a first data preprocessing device, and to retrieve patient basic information of the patient to be evaluated from a multi-physiological parameter database, wherein the first data preprocessing device is controlled to acquire initial historical vital sign data and current monitoring data from a patient vital sign monitoring device, wherein the current monitoring data includes: postoperative basic data, postoperative muscle strength test value, postoperative metabolic score value, postoperative mental cognition score value, disease identification, surgical identification, and indwelling drainage tube information; the first data preprocessing device is controlled to perform data standardization on the initial historical vital sign data to obtain the patient historical monitoring data, wherein the patient historical monitoring data includes: preoperative muscle strength test value, preoperative metabolic score value, and preoperative basic data; and the patient basic information of the patient to be evaluated is retrieved from the multi-physiological parameter database, wherein the patient basic information includes: past medical history information, smoking duration, and drinking duration; A first generating unit is configured to generate pre- and post-operative vital sign comparison data based on the current monitoring data, the historical monitoring data, and the patient's basic information, wherein the pre- and post-operative vital sign comparison data is used to characterize changes in vital sign data of the patient to be evaluated before and after the operation; a selection and correction unit configured to select a complication tree structure that matches the patient to be evaluated from a set of complication tree structures pre-established in a multi-physiological parameter database to obtain a target complication tree structure, and to correct the target complication tree structure based on the pre- and post-operative physical sign comparison data to obtain a corrected complication tree structure; an acquisition and processing unit configured to acquire case data matching the pre- and post-operative vital sign comparison data from a pre-connected case data server, and to perform data processing on each acquired case data through a second data processing device to obtain a processed case data set, wherein case data matching the pre- and post-operative vital sign comparison data and the revised complication structure tree are acquired from the case data server to obtain the case data set; and each case data in the case data set is screened through the second data processing device to obtain a processed case data set, wherein each processed case data in the processed case data set matches the age group, gender, surgical location identifier, surgical identifier, and disease identifier of the patient to be evaluated; a model training unit configured to input the processed case data set into a risk assessment module to obtain a patient-personalized prediction model, wherein the risk assessment module performs online additional training on a risk prediction basic model using the processed case data set to generate a patient-personalized prediction model, wherein the risk prediction basic model is used as a teacher model and the processed case data set is used as a training sample to train a corresponding student model through knowledge distillation to obtain the patient-personalized prediction model; The evaluation and risk warning unit is configured to perform risk level evaluation on the patient to be evaluated through the patient personalized prediction model and the revised complication structure tree, obtain risk level evaluation information, and control the human-computer interaction device to perform corresponding risk warning operations based on the risk level evaluation information.

7. An electronic device comprising: one or more processors; a storage device having one or more programs stored thereon, When the one or more programs are executed by the one or more processors, the one or more processors implement the system according to any one of claims 1 to 5.

8. A computer-readable medium having a computer program stored thereon, wherein: When the program is executed by a processor, the system according to any one of claims 1 to 5 is implemented.

Images