A medical vulnerability event reporting method and device

Abstract

This application relates to a method and apparatus for reporting medical vulnerability events, belonging to the field of computer intelligent computing technology. It predicts whether a patient's condition will further develop based on currently recorded medical data, thus avoiding serious consequences caused by vulnerability events. The method includes: during the reading of system records, detecting key medical information related to a target event within a first medical event; inputting this key information into a pre-set medical event vulnerability assessment model; the target event includes the development of the first medical event into a second medical event; calculating the target event using the medical event vulnerability assessment model in a first dimension and a second dimension to obtain calculation results; the first dimension includes the probability of the first medical event developing into the second medical event; the second dimension includes the consequences of the first medical event developing into the second medical event; and reporting the target event to the corresponding processing port of the second medical event based on the calculation results.

Description

Technical Field

[0001] This application relates to the field of computer intelligent computing technology, specifically to a method and apparatus for reporting medical vulnerability events. Background Technology

[0002] To facilitate modern hospital information sharing (for medical information sharing) and enable departments and hospitals to fully utilize limited medical and health information, existing technologies have established information systems to support collaborative medical services. These systems are called Hospital Information Systems (HIS). However, current HIS systems can only record and analyze existing medical conditions, symptoms, diagnostic records, medication methods, and treatments. Their ability to promptly identify and predict further disease progression and take timely measures to prevent its deterioration needs improvement. Summary of the Invention

[0003] This application proposes a method for reporting medical vulnerability events. Based on currently recorded patient data, it predicts whether a patient's condition will worsen and calculates the potential consequences based on the possible progression of the condition. The resulting consequences are then reported to the relevant department with the capability to resolve the issue, thus preventing serious consequences from the occurrence of vulnerability events.

[0004] In a first aspect, one embodiment provides a method for reporting medical vulnerability events. The method includes: during the reading of system records, detecting key medical information related to a target event in a first medical event, and inputting the key medical information into a pre-set medical event vulnerability assessment model; the target event includes the development of the first medical event into a second medical event; calculating the target event using the medical event vulnerability assessment model in a first dimension and a second dimension to obtain calculation results; the first dimension includes the probability of the first medical event developing into the second medical event; the second dimension includes the consequences caused by the development of the first medical event into the second medical event; and reporting the target event to the processing port corresponding to the second medical event based on the calculation results.

[0005] In one embodiment, the target event is calculated using the medical event vulnerability assessment model in a first dimension and a second dimension to obtain the calculation results, including:

[0006] The probability of the target event occurring due to the key information about the patient's condition is determined based on the aforementioned medical event vulnerability assessment model.

[0007] The extent of loss caused by the second medical event is determined based on the aforementioned medical event vulnerability assessment model;

[0008] Based on the calculation results, the target event is reported to the processing port corresponding to the second medical event, including:

[0009] If the probability of the target event occurring due to the key information of the illness is greater than the first preset threshold, the target event will be reported to the processing port corresponding to the second medical event.

[0010] If the loss caused by the second medical event exceeds the second preset threshold, the target event will be reported to the processing port corresponding to the second medical event.

[0011] In one embodiment, determining the probability that the target event, caused by the critical information about the patient's condition, will occur within at least one recording period based on the medical event vulnerability assessment model includes:

[0012] If the average probability is found to be lower than the first preset threshold and greater than the third preset threshold in the medical event vulnerability assessment model, the average probability will be increased to exceed the first preset threshold.

[0013] In one embodiment, the target event is calculated using the medical event vulnerability assessment model in a first dimension and a second dimension to obtain the calculation results, including:

[0014] The probability of the target event occurring due to the key information about the patient's condition is determined based on the aforementioned medical event vulnerability assessment model.

[0015] The extent of loss caused by the second medical event is determined based on the aforementioned medical event vulnerability assessment model;

[0016] Based on the probability of the target event occurring and the extent of loss caused by the second medical event, the risk level of the target event is predicted;

[0017] Based on the calculation results, the target event is reported to the processing port corresponding to the second medical event, including:

[0018] Based on the risk level, the target event will be reported to the corresponding processing port of the second medical event.

[0019] In one embodiment, determining the extent of loss caused by the occurrence of the second medical event based on the medical event vulnerability assessment model includes:

[0020] Find the direct formula corresponding to the occurrence of the target event in the medical event vulnerability assessment model, or / and query the operation method corresponding to the second medical event;

[0021] Calculate the degree of loss corresponding to the aforementioned operational method.

[0022] In one embodiment, before predicting the risk level of the target event based on the probability of the target event occurring within at least one recording period and the extent of loss caused by the second medical event, the method further includes:

[0023] Based on the development records of different medical events stored in the medical event vulnerability assessment model, the first dimension and the second dimension are quantified.

[0024] Based on the probability of the target event occurring and the extent of loss caused by the second medical event, the risk level of the target event is predicted, including:

[0025] In the first and second dimensions after quantification, query the first quantified value corresponding to the target event and the second quantified value corresponding to the degree of loss caused by the second medical event;

[0026] The risk level of the target event is obtained by calculating the first quantification value and the second quantification value.

[0027] In one embodiment, the medical event vulnerability assessment model further includes a pre-set deep neural network, wherein the agent of the deep neural network is set as the medical event. During the interaction and decision-making process between the agent and the environment, the agent selects the direction of the medical event's development as an action. The decision-making process is a Markov decision based on the medical event development task model. The method further includes:

[0028] The key information about the patient's condition and the first medical event are input into the deep neural network, which then predicts whether the first medical event will develop into a second medical event.

[0029] In one embodiment, the step of training the deep neural network includes:

[0030] Collect samples of the development of different medical events; the medical event development sample records include a third medical event and the disease-related information of the third medical event;

[0031] The intelligent agent selects the change in the course of the third medical event and obtains the changed state;

[0032] If the distance between the changed state and the at least one final medical event is less than a first preset value, the agent is rewarded.

[0033] In one embodiment, after reporting the target event to the processing port corresponding to the second medical event, the method further includes:

[0034] The second medical event is displayed on the page corresponding to the second medical event.

[0035] Secondly, in one embodiment, a medical vulnerability event reporting device is provided, the device comprising:

[0036] The detection module is used to detect key information about the condition of a related target event in the first medical event during the process of reading system records, and input the key information about the condition into a pre-set medical event vulnerability assessment model; the target event includes the development of the first medical event into a second medical event;

[0037] The calculation module is used to calculate the target event using the medical event vulnerability assessment model in a first dimension and a second dimension to obtain the calculation results; the first dimension includes the probability of a first medical event developing into a second medical event; the second dimension includes the consequences of the first medical event developing into a second medical event.

[0038] The reporting module is used to report the target event to the processing port corresponding to the second medical event based on the calculation results.

[0039] The beneficial effects of this application are:

[0040] This application embodiment utilizes a model to automatically extract and analyze data recorded in a medical information system. Based on the basic information, treatment methods, examination data, and other information about a disease recorded in the medical information system, it calculates the probability that a disease will develop into another disease or develop other symptoms under the recorded objective conditions, as well as the severity of the consequences caused by the disease's development. Based on the probability of disease development and the severity of the consequences, it determines whether the reporting criteria for a medical vulnerability event are met. Because the development trend of the disease is predicted, if the medical vulnerability event is reported, it can be directly reported to the personnel capable of handling the medical vulnerability event, successfully intervening in the uncontrollable development of the disease, making early preparations, and avoiding economic losses caused by further deterioration of the patient's condition. Attached Figure Description

[0041] Figure 1 This is a flowchart of the steps for reporting medical vulnerability events according to an embodiment of this application;

[0042] Figure 2 This is a schematic diagram illustrating the quantization values ​​of the first and second dimensions in this application;

[0043] Figure 3 This is a functional block diagram of the medical vulnerability event reporting device proposed in the embodiments of this application. Detailed Implementation

[0044] The present application will now be described in further detail with reference to the accompanying drawings and specific embodiments. Similar elements in different embodiments are referred to by related similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the present application. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present application are not shown or described in the specification. This is to avoid obscuring the core parts of the present application with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0045] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0046] The serial numbers assigned to components in this document, such as "first" and "second," are used only to distinguish the described objects and have no sequential or technical meaning. The terms "connection" and "linkage" used in this application, unless otherwise specified, include both direct and indirect connections (linkages).

[0047] Figure 1 This is a flowchart of the steps for reporting medical vulnerability events according to an embodiment of this application, as shown below. Figure 1 As shown, the steps of the healthcare vulnerability event reporting method include:

[0048] S1: During the process of reading system records, if key information about the condition of the target event is detected in the first medical event, the key information about the condition will be input into the pre-set medical event vulnerability assessment model.

[0049] The target events include the progression from the first medical event to the second medical event.

[0050] The first medical event can be understood as the patient's primary condition, while the second medical event is a condition different from the first medical event, or a symptom different from the primary condition.

[0051] For example, the first medical event is "frozen shoulder". The relevant information recorded by the medical information system includes: "The patient received acupuncture treatment on August 1 and then developed pneumothorax on the spot". After data statistics and model analysis, it is found that "receiving acupuncture treatment" may be the key information of the condition related to the target event, "frozen shoulder" is the first medical event, and "pneumothorax" is the second medical event.

[0052] For example, the first medical event is "paroxysmal atrial fibrillation and dilated cardiomyopathy." The relevant information recorded by the medical information system includes: "The patient is suspected of having paroxysmal atrial fibrillation and dilated cardiomyopathy. During the infusion of amiodarone and dopamine into the left upper limb, swelling and pain gradually occurred." After data statistics and model analysis, it is concluded that "infusion of amiodarone and dopamine into the left upper limb, and dobutamine" may be the key information of the patient's condition related to the target event, and "swelling and pain" is the second medical event.

[0053] For example, the first medical event is "obese patient". The relevant information recorded by the medical information system includes: "stroke obese patient, bedridden, eventually developed bedsores". After data statistics and model analysis, it is found that "stroke and bedridden" may be the key information of the condition related to the target event, and "bedsores" is the second medical event.

[0054] This application embodiment reads medical records, outpatient records, examination data, medication information, treatment information, and other archives stored in the medical information system, and automatically extracts key information that may lead to the further development of the first medical event. Then, it uses a pre-set medical event vulnerability assessment model to analyze the key information, determining whether the first medical event, under the influence of objective factors, treatment methods, patient basic information, and other key information, will develop into other conditions, or exhibit other symptoms, or whether it will further deteriorate—these are all vulnerability events.

[0055] Patient basic information includes information such as age, gender, and underlying diseases.

[0056] S2: Calculate the target event using the medical event vulnerability assessment model in the first and second dimensions to obtain the calculation results.

[0057] The first dimension includes the probability of the first medical event developing into a second medical event; the second dimension includes the consequences of the first medical event developing into a second medical event.

[0058] One example of this application is to determine whether to report a vulnerability event based on the probability of a first medical event developing into a second medical event, the consequences of the first medical event developing into a second medical event, the severity of the first medical event developing into a second medical event, and the probability and consequences of the first medical event developing into a second medical event.

[0059] The greater the likelihood that the first medical event will develop into a second medical event, based on the records of the medical information system, the greater the risk that the first medical event will develop into a vulnerable event, the greater the consequences and losses caused by the development of the first medical event into a second medical event, and the greater the losses brought about by the development of the first medical event. Therefore, the development of the first medical event should be reported to the department that can handle the event.

[0060] In view of the above objective facts, this application provides a method for executing S2:

[0061] S21: Determine the probability of a target event occurring due to key information about the patient's condition based on a medical event vulnerability assessment model.

[0062] Before applying the medical event vulnerability assessment model, relevant data samples can be collected in advance to train the model. The model learns from the relevant data samples to determine the probability of a target event occurring due to key information about the patient's condition, or it can further determine the likelihood of a target event occurring due to key information about the patient's condition.

[0063] Alternatively, during the application of a medical event vulnerability assessment model, the vulnerability assessment model can continue to learn information about the development of the first medical event into the second medical event, adaptively adjusting the vulnerability assessment model to make it more robust.

[0064] For example, suppose relevant information records for N patients with "paroxysmal atrial fibrillation and dilated cardiomyopathy" are collected, resulting in N data samples. Among these, M samples record data on "infusion of amiodarone, dopamine, and dobutamine into the left upper limb," and M-1 of these M samples record patients experiencing "swelling and pain symptoms," where M is less than N, and NM is less than a constant C. Using these samples, a medical event vulnerability assessment model is trained. The model performs statistical analysis and learning on the sample data. During the application phase, when data on "paroxysmal atrial fibrillation, dilated cardiomyopathy, infusion of amiodarone, dopamine, and dobutamine into the left upper limb" is detected, the model outputs the probability that the key information about the patient's condition leads to the target event as (M-1) / N. Alternatively, the medical event vulnerability assessment model can further predict the likelihood of the key information causing the target event to occur as "very likely" based on the output probability of (M-1) / N.

[0065] In one example, performing a medical event vulnerability assessment model to determine the probability that a target event caused by critical information about a patient's condition will occur within at least one recording period further includes the steps of:

[0066] If the average probability found in the medical event vulnerability assessment model is lower than the first preset threshold but higher than the third preset threshold, the average probability will be increased to exceed the first preset threshold.

[0067] To more comprehensively prevent the occurrence of healthcare vulnerability events, the probability of a target event occurring can be increased based on the average probability of the target event calculated by the healthcare event vulnerability assessment model.

[0068] For example, suppose the healthcare event vulnerability assessment model outputs "stroke, obese patient, bedridden" and eventually develops bedsores.

[0069] S22: Determine the extent of loss caused by a second medical event based on a medical event vulnerability assessment model.

[0070] Before applying the medical event vulnerability assessment model, relevant data samples can be collected in advance. These data samples can further include labeled loss values, which can be manually confirmed. The medical event vulnerability assessment model is trained using these data samples, and the model learns the potential loss caused by the occurrence of a target event based on the data samples.

[0071] A deep neural network can be constructed in the medical event vulnerability assessment model. This network is trained using supervised learning based on the operational methods corresponding to the second medical event, the drugs used to treat the second medical event, the direct losses caused by the second medical event, and manually labeled loss values. After learning, the neural network can estimate the economic value of the target event based on the direct losses, the drugs used to treat the second medical event, and the operational methods corresponding to the second medical event. Then, based on the difference between the economic value and the labeled loss value, the deep neural network is adjusted, ultimately resulting in a loss level prediction model that can output the degree of loss based on the direct losses, the drugs used to treat the second medical event, and the operational methods corresponding to the second medical event.

[0072] The extent of loss resulting from a second medical event is determined based on a medical event vulnerability assessment model, including:

[0073] S201: Find the direct losses corresponding to the occurrence of the target event in the medical event vulnerability assessment model, and / or query the operational measures corresponding to the second medical event.

[0074] During the training phase, the direct losses corresponding to the occurrence of the target event can be obtained, and / or the corresponding operational methods for the second medical event can be queried. Relevant data can also be extracted from the medical information system: the operational methods corresponding to the second medical event, the drugs used to treat the second medical event, and the direct losses caused by the occurrence of the second medical event.

[0075] S202: Analyze the consequences to obtain the direct losses caused by the occurrence of the second medical event, obtain the degree of loss, and / or calculate the degree of loss corresponding to the operational measures.

[0076] For example, suppose relevant information records for N patients with "paroxysmal atrial fibrillation and dilated cardiomyopathy" are collected, resulting in N data samples. Among these, M samples record data on "infusion of amiodarone, dopamine, and dobutamine into the left upper limb." Of these M samples, M-1 records indicate that the patient experienced "swelling and pain symptoms," where M is less than N, and NM is less than a constant C. Each of the M-1 samples carries a label—a loss value of T. Using these samples, a medical event vulnerability assessment model is trained. The model performs statistical analysis and learning on the sample data, learning the relationship between the loss value and the first medical event, the second medical event, and key information. This allows it to predict the degree of loss caused by swelling and pain in patients with paroxysmal atrial fibrillation and dilated cardiomyopathy.

[0077] S23: If the critical information of the condition output by the vulnerability assessment model causes the probability of the target event to be greater than the first preset threshold, the target event will be reported to the corresponding processing port of the second medical event.

[0078] S24: If the loss caused by the second medical event output by the vulnerability assessment model is greater than the second preset threshold, the target event will be reported to the processing port corresponding to the second medical event.

[0079] S3: Based on the calculation results, report the target event to the corresponding processing port of the second medical event.

[0080] In one example, the processing port corresponding to the second medical event is connected to the storage area called by the operation interface of the second medical event handler during runtime. When the system detects that the second medical event handler has logged into the system and generates the operation interface, it calls the running file stored in a specific storage area. The processing port corresponding to the second medical event is connected to this specific storage area. Therefore, when the system detects that the second medical event handler has logged into the system, it can display the target event of the development from the first medical event to the second medical event. Since the second medical event handler can directly handle the second medical event, compared to notifying the personnel handling the first medical event or notifying the functional departments, this embodiment of the application reports the event to the processing port corresponding to the second medical event, directly reporting the event to the personnel who can handle the event, reducing the time spent in the message forwarding process and improving the efficiency of vulnerability event handling.

[0081] Therefore, after reporting the target event to the corresponding processing port of the second medical event, the method also includes:

[0082] The second medical event is displayed on the corresponding page to notify relevant personnel to handle the target event.

[0083] The methods for displaying secondary medical events include images, text, icons, etc.

[0084] This application embodiment utilizes a model to automatically extract and analyze data recorded in a medical information system. Based on the basic information, treatment methods, examination data, and other information about a disease recorded in the medical information system, it calculates the probability that a disease will develop into another disease or develop other symptoms under the recorded objective conditions, as well as the severity of the consequences caused by the disease's development. Based on the probability of disease development and the severity of the consequences, it determines whether the reporting criteria for a medical vulnerability event are met. Because the development trend of the disease is predicted, if the medical vulnerability event is reported, it can be directly reported to the personnel capable of handling the medical vulnerability event, successfully intervening in the uncontrollable development of the disease, making early preparations, and avoiding economic losses caused by further deterioration of the patient's condition.

[0085] In order to take into account the impact of the two factors on whether to report the target event, another embodiment of this application also proposes a way to perform S2; the two factors are the possibility of the first medical event developing into a second medical event and the degree of loss caused by the consequences of the first medical event developing into a second medical event.

[0086] After obtaining the probability of the target event occurring and the extent of loss caused by the second medical event, the risk level of the target event can be calculated based on the probability of the target event occurring and the extent of loss caused by the second medical event. Then, based on the risk level, the method of reporting the target event can be determined.

[0087] S2-1: Determine the probability of a target event occurring due to key information about the patient's condition based on a medical event vulnerability assessment model;

[0088] S2-2: Determine the extent of loss caused by a second medical event based on a medical event vulnerability assessment model;

[0089] S2-3: Based on the probability of the target event occurring and the extent of loss caused by the second medical event, predict the risk level of the target event.

[0090] One way to perform step S2-3, predicting the risk level of the target event, includes:

[0091] S301: Based on the development records of different medical events stored in the medical event vulnerability assessment model, quantify the first and second dimensions.

[0092] During the data collection phase, manual labeling can be used to quantify the first and second dimensions of a specific event.

[0093] For example, there are N records of the first medical event "stroke obese patient" and M records of "stroke obese patient, bedridden, eventually developed bedsores". The probability of stroke obese patient developing bedsores is calculated to be M / N. Based on this basic probability, the first dimension is quantified. The record of "stroke obese patient, bedridden, eventually developed bedsores" records the corresponding economic loss. Based on the average economic loss, the second dimension is quantified.

[0094] S302: Query the first quantitative value corresponding to the target event and the second quantitative value corresponding to the degree of loss caused by the occurrence of the second medical event in the risk consequence quantification matrix.

[0095] S303: Calculate the first and second quantification values ​​to obtain the risk level of the target event.

[0096] Figure 2 This is a schematic diagram illustrating the quantization values ​​of the first and second dimensions of this application, as shown below. Figure 2 As shown, in the application phase, the probability of the target event occurring can be considered based on the degree of loss and the probability of the target event occurring. The first dimension, the probability of the target event occurring, increases from the starting point, while the second dimension, the severity, also increases from the starting point. Figure 2 Find the corresponding value point and use the risk level corresponding to the value point as the risk level of the target event. For example, the probability of "a stroke patient who is obese, bedridden, and eventually develops bedsores" occurring is P, and the degree of loss is Q. Figure 2 If we find the point B corresponding to the probability P and the degree of loss Q, and the risk level corresponding to point B is H, then we can determine that the risk level of "stroke obese patient, bedridden, eventually developed bedsores" is H. Figure 2 E indicates extreme risk, requiring immediate reporting; H indicates high risk, requiring reporting to senior management; M indicates medium risk, requiring reporting to relevant personnel; and L indicates low risk, which can be temporarily left unreported.

[0097] This application also proposes a method for using a medical event vulnerability assessment model to predict target events:

[0098] Another deep neural network is constructed in the medical event vulnerability assessment model. This deep neural network is trained using reinforcement learning. The medical event is set as the agent, the possible development direction of the medical event is the action, and the state of the medical event is the state space. All sample data, action space, and state space uploaded to the medical event vulnerability assessment model constitute the learning environment of the agent. For example, the sample data may include, but is not limited to: "The patient was suspected of having paroxysmal atrial fibrillation and dilated cardiomyopathy. During the infusion of amiodarone and dopamine into the left upper limb, swelling and pain gradually appeared", "A stroke obese patient, bedridden, eventually developed bedsores", "A patient with frozen shoulder received acupuncture treatment on August 1 and developed pneumothorax on the spot", etc.

[0099] The agent learns that the current state is paroxysmal atrial fibrillation. Based on the decision function, it selects an action in the action space: to execute a step in the direction of developing pneumothorax. It calculates that the distance between the agent's current state and the final medical event "pneumothorax" has decreased to below a first preset value, and rewards the agent. Thus, the agent learns the direction of the disease's development.

[0100] Meanwhile, after the agent selects an action in the action space, it calculates the probability of selecting that action. After the agent develops to the state that is closest to the final medical event based on that action, it calculates the distance between the selection probability and the calibration probability. If the distance between the selection probability and the calibration probability is less than a second preset threshold, the agent is rewarded.

[0101] The calibration probability can be determined by collecting relevant information records of N patients with "paroxysmal atrial fibrillation and dilated cardiomyopathy" in accordance with other embodiments of this application, resulting in N data samples. Among them, M samples record data on "infusion of amiodarone, dopamine, and dobutamine into the left upper limb". Among the M samples, M-1 records show "swelling and pain symptoms" in the patients. M is less than N, and NM is less than the constant C. The calibration probability is then calculated.

[0102] After training, the agent in the deep neural network can automatically simulate the development of the disease based on the environment, thereby predicting whether the first medical event will develop into a second medical event.

[0103] The medical event vulnerability assessment model also includes a pre-set deep neural network. The agent within the deep neural network represents the medical event. During the interaction and decision-making process between the agent and the environment, the agent selects the direction of the medical event's development as its action. This decision-making process is a Markov decision based on the modeling of the medical event's development task. Other methods include:

[0104] Key information about the patient's condition and the first medical event are input into a deep neural network, which then predicts whether the first medical event will develop into a second medical event.

[0105] The key information about the condition of the first medical event and related target events is input into a medical event vulnerability assessment model that has undergone reinforcement learning. The medical event vulnerability assessment model can output the development result of the first medical event. If the first medical event develops into other conditions or develops other symptoms, it is determined that the first medical event has developed into the second medical event.

[0106] The process of training a healthcare event vulnerability assessment model using reinforcement learning includes:

[0107] S401: Collect samples of the development of different medical events; the medical event development sample records include third medical events and information related to the condition of the third medical events;

[0108] The third medical event refers to the final state of the development of the first medical event in the sample.

[0109] For example: "The patient was suspected of having paroxysmal atrial fibrillation and dilated cardiomyopathy. Amiodarone and dopamine were administered to the left upper limb via infusion. Swelling and pain gradually appeared during the dobutamine administration." "A stroke patient who was obese, bedridden, eventually developed bedsores."

[0110] S402: The agent selects the change in the course of the disease development of the third medical event and obtains the changed state;

[0111] S403: If the distance between the changed state and at least one final medical event is less than a first preset value, the agent is rewarded.

[0112] Figure 3 This is a functional block diagram of the medical vulnerability event reporting device proposed in the embodiments of this application, such as... Figure 3 As shown, the medical vulnerability event reporting device includes:

[0113] The detection module 31 is used to detect key information about the condition of a related target event in the first medical event during the process of reading system records, and input the key information about the condition into a pre-set medical event vulnerability assessment model; the target event includes the development of the first medical event into a second medical event;

[0114] The calculation module 32 is used to calculate the target event using the medical event vulnerability assessment model in the first and second dimensions to obtain the calculation results; the first dimension includes the probability of the first medical event developing into the second medical event; the second dimension includes the consequences of the first medical event developing into the second medical event.

[0115] The reporting module 33 is used to report the target event to the corresponding processing port of the second medical event based on the calculation results.

[0116] Figure 3 The implementation principle and technical effects of the medical vulnerability event reporting device provided in the illustrated embodiment can be further described in the relevant embodiments of the medical vulnerability event reporting method.

[0117] Optionally, the computing module includes:

[0118] The probability calculation submodule is used to determine the probability of a target event occurring due to key information about the patient's condition, based on a medical event vulnerability assessment model.

[0119] The loss calculation submodule is used to determine the extent of loss caused by the occurrence of a second medical event based on the medical event vulnerability assessment model.

[0120] The reporting module is specifically used to report the target event to the corresponding processing port of the second medical event if the probability of the target event caused by the key information of the patient's condition is greater than the first preset threshold; and to report the target event to the corresponding processing port of the second medical event if the degree of loss caused by the occurrence of the second medical event is greater than the second preset threshold.

[0121] Optionally, the probability calculation submodule is specifically used to increase the average probability to exceed the first preset threshold if the average probability found in the medical event vulnerability assessment model is lower than the first preset threshold and greater than the third preset threshold.

[0122] Optionally, the computing module includes:

[0123] The probability calculation submodule is used to determine the probability of a target event occurring due to key information about the patient's condition, based on a medical event vulnerability assessment model.

[0124] The loss calculation submodule is used to determine the extent of loss caused by the occurrence of a second medical event based on the medical event vulnerability assessment model.

[0125] The risk level calculation submodule is used to predict the risk level of the target event based on the probability of the target event occurring and the extent of loss caused by the second medical event.

[0126] The reporting module is specifically used to report the target event to the corresponding processing port of the second medical event according to the risk level.

[0127] Optionally, the risk level calculation submodule is specifically used to find the direct formula corresponding to the occurrence of the target event in the medical event vulnerability assessment model, or / and query the operation method corresponding to the second medical event;

[0128] Calculate the degree of loss corresponding to the operational methods.

[0129] Optionally, the healthcare vulnerability event reporting device also includes:

[0130] The quantification module is used to quantify the first and second dimensions based on the development records of different medical events stored in the medical event vulnerability assessment model.

[0131] The risk level calculation submodule is specifically used to query the first quantified value corresponding to the target event and the second quantified value corresponding to the degree of loss caused by the occurrence of the second medical event in the first and second dimensions after quantification.

[0132] Calculate the first and second quantification values ​​to obtain the risk level of the target event.

[0133] Optionally, the medical event vulnerability assessment model also includes a pre-set deep neural network, with the agent of the deep neural network representing the medical event. During the agent's interaction and decision-making process with the environment, it selects the direction of the medical event's development as its action. The decision-making process is a Markov decision based on the medical event development task model. The device also includes:

[0134] The prediction module is used to input key information about the patient's condition and the first medical event into a deep neural network for medical vulnerability events, and to use the neural network to predict whether the first medical event will develop into a second medical event.

[0135] Optionally, the medical vulnerability event reporting device also includes a module for training a deep neural network to collect different medical event development samples; the medical event development sample records include a third medical event and information on the disease association of the third medical event;

[0136] The agent selects the direction of the disease's progression in the third medical event and obtains the changed state;

[0137] If the distance between the changed state and at least one final medical event is less than a first preset value, the agent is rewarded.

[0138] Optionally, reporting of healthcare vulnerability events also includes:

[0139] The display module is used to display the second medical event on the corresponding page.

[0140] Regarding the modules / units included in the various devices described in the above embodiments, they can be software modules / units, hardware modules / units, or a combination of both. For example, for devices applied to or integrated into a chip, all modules / units can be implemented using hardware methods such as circuits, or at least some modules / units can be implemented using software programs running on a processor integrated within the chip, while the remaining modules / units can be implemented using hardware methods such as circuits. For devices applied to or integrated into a chip module, all modules / units can be implemented using hardware methods such as circuits. Different modules / units can be located in the same component (e.g., chip, circuit module, etc.) or different components of the chip module, or at least some modules / units can be implemented using software programs. The software program runs on the processor integrated inside the chip module, and the remaining modules / units can be implemented using hardware methods such as circuits. For each device applied to or integrated into an electronic terminal device, each of its modules / units can be implemented using hardware methods such as circuits. Different modules / units can be located in the same component (e.g., chip, circuit module, etc.) or different components within the electronic terminal device. Alternatively, at least some modules / units can be implemented using software programs that run on the processor integrated inside the electronic terminal device, and the remaining (if any) modules / units can be implemented using hardware methods such as circuits.

[0141] One embodiment of this application provides a computer-readable storage medium storing a program, the stored program including methods that can be loaded by a processor and processed in any of the above embodiments.

[0142] Those skilled in the art will understand that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the program can also be stored in a server, another computer, disk, optical disk, flash drive, or external hard drive, etc., and can be downloaded or copied to the memory of a local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be achieved.

[0143] The above examples illustrate this application only to aid understanding and are not intended to limit its scope. Those skilled in the art to which this application pertains can make various simple deductions, modifications, or substitutions based on the ideas presented.

Claims

1. A method for reporting medical vulnerability events, characterized in that, The method includes: During the reading of system records, key information about the patient's condition related to a target event is detected in the first medical event. This key information is then input into a pre-set medical event vulnerability assessment model. The target event includes the development of the first medical event into a second medical event. The first medical event is the patient's primary condition, and the second medical event is a condition different from the first medical event, or a symptom different from the primary condition. The related target event is the treatment method adopted for the first medical event, and this treatment method is key information that may lead to the further development of the first medical event, in order to determine whether the first medical event will develop into other conditions or cause other symptoms. The target event is calculated using the aforementioned medical event vulnerability assessment model in a first dimension and a second dimension to obtain the calculation results. The first dimension includes the probability that a first medical event will develop into a second medical event, which includes: determining the average probability within a recording period that the first medical event is caused by key information related to the target event leading to the second medical event based on the medical event vulnerability assessment model, and using this average probability as the probability that the first medical event will develop into the second medical event. The second dimension includes the consequences of the first medical event developing into the second medical event, including the degree of loss caused by the occurrence of the second medical event. Based on the calculation results, the target event is reported to the processing port corresponding to the second medical event, including: If the probability of the target event occurring due to the critical information of the condition is greater than a first preset threshold, the target event is reported to the processing port corresponding to the second medical event; if the degree of loss caused by the occurrence of the second medical event is greater than a second preset threshold, the target event is reported to the processing port corresponding to the second medical event. Alternatively, based on the probability of the target event occurring and the extent of loss caused by the second medical event, the risk level of the target event is predicted; based on the risk level, the target event is reported to the processing port corresponding to the second medical event.

2. The method for reporting medical vulnerability events according to claim 1, characterized in that, The extent of loss resulting from the second medical event is determined based on the aforementioned medical event vulnerability assessment model, including: Find the direct formula corresponding to the occurrence of the target event in the medical event vulnerability assessment model, or / and query the operation method corresponding to the second medical event; Calculate the degree of loss corresponding to the aforementioned operational method.

3. The method for reporting medical vulnerability events according to claim 1, characterized in that, Before predicting the risk level of the target event based on the probability of the target event occurring within at least one recording period and the extent of loss caused by the second medical event, the method further includes: Based on the development records of different medical events stored in the medical event vulnerability assessment model, the first dimension and the second dimension are quantified. Based on the probability of the target event occurring and the extent of loss caused by the second medical event, the risk level of the target event is predicted, including: In the quantified first and second dimensions, query the first quantified value corresponding to the target event and the second quantified value corresponding to the degree of loss caused by the second medical event; The risk level of the target event is obtained by calculating the first quantification value and the second quantification value.

4. The method for reporting medical vulnerability events according to claim 1, characterized in that, The medical event vulnerability assessment model also includes a pre-set deep neural network, where the agent of the deep neural network is designated as the medical event. During the interaction and decision-making process between the agent and the environment, the agent selects the direction of the medical event's development as its action. This decision-making process is a Markov decision based on the medical event's development task model. The method further includes: The key information about the patient's condition and the first medical event are input into the deep neural network, which then predicts whether the first medical event will develop into a second medical event.

5. The method for reporting medical vulnerability events according to claim 4, characterized in that, The steps for training the deep neural network include: Collect samples of the development of different medical events; the medical event development sample records include a third medical event and the disease-related information of the third medical event; The intelligent agent selects the change in the course of the third medical event and obtains the changed state; If the distance between the changed state and at least one final medical event is less than a first preset value, the agent is rewarded.

6. The method according to claim 1, characterized in that, After reporting the target event to the processing port corresponding to the second medical event, the method further includes: The second medical event is displayed on the page corresponding to the second medical event.

7. A medical vulnerability event reporting device, characterized in that, The medical vulnerability event reporting device includes: The detection module is used to detect key information about a patient's condition related to a target event within a first medical event during the reading of system records. This key information is then input into a pre-set medical event vulnerability assessment model. The target event includes the development of the first medical event into a second medical event. The first medical event is the patient's primary condition, and the second medical event is a condition different from the first medical event, or a symptom different from the primary condition. The related target event is the treatment method used for the first medical event, and this treatment method is key information that may lead to the further development of the first medical event, in order to determine whether the first medical event will develop into other conditions or present with other symptoms. The calculation module is used to calculate the target event using the medical event vulnerability assessment model in a first dimension and a second dimension to obtain calculation results. The first dimension includes the probability that a first medical event will develop into a second medical event, including: determining the average probability within a recording period that the first medical event is caused by key information related to the target event leading to the second medical event based on the medical event vulnerability assessment model, and using this average probability as the probability that the first medical event will develop into the second medical event. The second dimension includes the consequences of the first medical event developing into the second medical event, including the degree of loss caused by the occurrence of the second medical event. The reporting module, used to report the target event to the processing port corresponding to the second medical event based on the calculation result, includes: If the probability of the target event occurring due to the critical information of the condition is greater than a first preset threshold, the target event is reported to the processing port corresponding to the second medical event; if the degree of loss caused by the occurrence of the second medical event is greater than a second preset threshold, the target event is reported to the processing port corresponding to the second medical event. Alternatively, based on the probability of the target event occurring and the extent of loss caused by the second medical event, the risk level of the target event is predicted; based on the risk level, the target event is reported to the processing port corresponding to the second medical event.

Images