Rehabilitation management method and device for patients with stroke and cognitive impairment

Abstract

The present invention relates to the field of rehabilitation medical technology, and discloses a rehabilitation management method and device for patients with stroke and cognitive impairment. The method comprises: performing character embedding on medical record data to obtain medical record features; calculating the attention score and semantic relationship of the medical record features to rehabilitation items, generating a first medical record expression and a second medical record expression based on the rehabilitation items according to the calculation results, predicting rehabilitation items based on the case expressions, and obtaining a rehabilitation plan for the patient; receiving rehabilitation data after the patient performs rehabilitation items in the rehabilitation plan, performing data correction and cluster optimization on the rehabilitation data to obtain a target matrix; calculating the distance between the target matrix and the matrix corresponding to the historical rehabilitation data, and determining the patient's rehabilitation status based on the distance value; adjusting the medical record data according to the rehabilitation status to obtain updated medical record data, and updating the rehabilitation items in the rehabilitation plan according to the updated medical record data. The present invention can improve the efficiency of rehabilitation management.

Description

Technical Field

[0001] The present invention relates to the field of rehabilitation medical technology, and in particular to a rehabilitation management method and device for patients with stroke or cognitive impairment. Background Art

[0002] With the development of artificial intelligence and modern medicine, intelligent medicine has gradually become a mainstream research trend. Within the field of rehabilitation therapy, rehabilitation treatment for common conditions such as stroke and cognitive impairment is particularly important. Stroke, also known as "stroke" or "cerebrovascular accident," is an acute cerebrovascular disease. It is a group of conditions, including ischemic and hemorrhagic strokes, that damage brain tissue due to the sudden rupture or blockage of a cerebral blood vessel, preventing blood flow to the brain. Cognitive impairment is a persistent decline in intellectual function while conscious, caused by various neurological and cerebrovascular lesions. It manifests as a decrease or loss of memory, judgment, language, and the ability to function independently. Different types of stroke or varying degrees of cognitive impairment require different rehabilitation treatment approaches, and the appropriate rehabilitation plan should be selected based on the individual patient's medical history.

[0003] Existing methods are all based on traditional means, manually evaluating large amounts of medical records and examination results to develop rehabilitation treatment plans. After the rehabilitation plan is implemented, the patient's recovery is evaluated and then the patient's rehabilitation treatment is managed. The entire process is time-consuming, labor-intensive, and relatively inefficient. To address this problem, machine learning methods and assessment tools are gradually being applied to the field of rehabilitation treatment. However, the machine learning process typically uses numerical and structured data and does not fully integrate the patient's medical records with the rehabilitation plan, resulting in low accuracy of the generated rehabilitation plan. When using assessment tools to evaluate rehabilitation data, they can only emphasize the results of individual assessments and cannot grasp the patient's overall rehabilitation plan feedback, resulting in inaccurate rehabilitation assessments and inflexible rehabilitation plan adjustments.

[0004] In summary, the existing technology has problems such as low accuracy in generating rehabilitation plans, inaccurate rehabilitation assessments, and inflexible adjustment of rehabilitation plans, which lead to low efficiency of rehabilitation management. Summary of the Invention

[0005] The present invention provides a rehabilitation management method and device for patients with stroke and cognitive impairment. Its main purpose is to solve the problem of low rehabilitation management efficiency caused by low accuracy of generated rehabilitation plans, inaccurate rehabilitation assessments and inflexible adjustment of rehabilitation plans.

[0006] To achieve the above objectives, the present invention provides a rehabilitation management method for patients with stroke and cognitive impairment, comprising:

[0007] Acquire the patient's medical record data, perform character embedding and text concatenation on the medical record data, and obtain medical record features;

[0008] Calculating an attention score of the medical record feature to a preset rehabilitation program, and generating a first medical record expression based on the rehabilitation program according to the attention score;

[0009] Calculating a semantic relationship between the rehabilitation project and the medical record feature, and generating a second medical record expression based on the rehabilitation project according to the semantic relationship;

[0010] Performing a rehabilitation project prediction based on the first medical record expression based on the rehabilitation project and the second medical record expression based on the rehabilitation project, obtaining a rehabilitation plan for the patient, and executing the rehabilitation projects in the rehabilitation plan;

[0011] Receiving rehabilitation data of the patient after performing the rehabilitation items in the rehabilitation program, performing data correction and cluster optimization on the rehabilitation data to obtain a target matrix;

[0012] Calculating a distance between the target matrix and a matrix corresponding to preset historical rehabilitation data, and determining the patient's rehabilitation status according to the distance;

[0013] The medical record data is adjusted according to the rehabilitation status to obtain updated medical record data, and the rehabilitation items in the rehabilitation plan are updated according to the updated medical record data.

[0014] Optionally, the performing character embedding and text concatenation on the medical record data to obtain medical record features includes:

[0015] Using a preset BERT model to perform character embedding on the medical record data to obtain multiple characters of the medical record data;

[0016] Utilizing a preset bidirectional recurrent neural network to capture contextual information corresponding to each of the characters;

[0017] The context information is string-concatenated to obtain concatenated characters, and a medical record feature is generated according to the concatenated characters corresponding to each character.

[0018] Optionally, before performing character embedding and text concatenation on the medical record data, the method further includes:

[0019] De-anonymizing sensitive data in the medical record data to obtain first cleansed data;

[0020] removing stop words from the first cleaned data to obtain second cleaned data;

[0021] Keyword screening is performed on the second cleaned data to obtain preprocessed data.

[0022] Optionally, calculating the attention score of the medical record feature to a preset rehabilitation program and generating a first medical record expression based on the rehabilitation program according to the attention score includes:

[0023] The attention score of the medical record feature to the preset rehabilitation items is calculated by the following formula:

[0024]

[0025] in, For the Character pair The attention score of each rehabilitation item, The medical record data The concatenation character corresponding to the characters, is the medical record feature, are the preset self-attention weight matrices, It is the preset self-attention mechanism;

[0026] The first medical record expression based on the rehabilitation project is generated by the following formula:

[0027]

[0028] in, The medical record data is based on The first medical record expression of a rehabilitation program.

[0029] Optionally, calculating the semantic relationship between the rehabilitation program and the medical record feature, and generating a second medical record expression based on the rehabilitation program according to the semantic relationship, includes:

[0030] The semantic relationship between the rehabilitation item and the medical record feature is calculated using the following formula:

[0031] ;

[0032] in, 、 Respectively characters and The contextual semantic relationship between rehabilitation projects, For the The vector matrix of rehabilitation items, are the forward expression and reverse expression of the medical record features respectively;

[0033] The second medical record expression based on the rehabilitation project is generated by the following formula:

[0034] ;

[0035] ;

[0036] in, The medical record data is based on The second medical record expression of a rehabilitation program.

[0037] Optionally, performing rehabilitation project prediction based on the first medical record expression based on the rehabilitation project and the second medical record expression based on the rehabilitation project to obtain the rehabilitation plan for the patient includes:

[0038] Setting weights according to the first medical record expression based on the rehabilitation program and the second medical record expression based on the rehabilitation program respectively;

[0039] generating a target medical record expression based on the rehabilitation program according to the weight, the first medical record expression based on the rehabilitation program, and the second medical record based on the rehabilitation program;

[0040] Utilizing a preset multi-layer perceptron to calculate the target medical record expression based on the rehabilitation project, and obtaining prediction results of multiple rehabilitation projects;

[0041] The rehabilitation items are screened according to the prediction results to obtain a rehabilitation plan.

[0042] Optionally, setting weights according to the first medical record expression based on the rehabilitation program and the second medical record expression based on the rehabilitation program respectively includes:

[0043] The weights of the first medical record expression based on the rehabilitation project and the second medical record expression based on the rehabilitation project are set by the following formula:

[0044]

[0045] in, Based on the The weight corresponding to the first medical record expression of each rehabilitation project, Based on the The weight corresponding to the second medical record expression of each rehabilitation project, The medical record data is based on The first medical record expression of a rehabilitation project, The medical record data is based on The second medical record expression of a rehabilitation project, are preset training parameters respectively.

[0046] Optionally, the calculating the target medical record expression based on the rehabilitation program by using a preset multi-layer perceptron includes:

[0047] The target medical record expression based on the rehabilitation project is calculated by the following formula:

[0048]

[0049] in, For the The predicted probability of a rehabilitation program, , 、 are the training parameters of the fully connected layer and the output layer in the multilayer perceptron, is the nonlinear activation function in the multilayer perceptron, Based on the The goal of a rehabilitation program is expressed in medical records.

[0050] Optionally, performing data correction and cluster optimization on the rehabilitation data to obtain a target matrix includes:

[0051] Normalizing the rehabilitation data, and performing data correction on the normalized rehabilitation data using a preset KNN algorithm to obtain a correction matrix;

[0052] Utilizing each network node in a preset convolutional network to calculate the correction matrix, a node matrix is obtained;

[0053] The node matrix is clustered using a preset clustering algorithm to obtain a target matrix.

[0054] In order to solve the above problems, the present invention further provides a rehabilitation management device for patients with stroke or cognitive impairment, the device comprising:

[0055] A medical record feature generation module is used to obtain the patient's medical record data, perform character embedding and text concatenation on the medical record data, and obtain medical record features;

[0056] A first medical record expression generating module is configured to calculate an attention score of the medical record feature to a preset rehabilitation program, and generate a first medical record expression based on the rehabilitation program according to the attention score;

[0057] A second medical record expression generating module is configured to calculate a semantic relationship between the rehabilitation project and the medical record feature, and generate a second medical record expression based on the rehabilitation project according to the semantic relationship;

[0058] a rehabilitation program generation module, configured to predict rehabilitation programs based on the first medical record expression based on rehabilitation programs and the second medical record expression based on rehabilitation programs, obtain a rehabilitation program for the patient, and execute rehabilitation programs in the rehabilitation program;

[0059] a rehabilitation data processing module, configured to receive rehabilitation data after the patient performs rehabilitation items in the rehabilitation program, perform data correction and cluster optimization on the rehabilitation data, and obtain a target matrix;

[0060] a rehabilitation status evaluation module, configured to calculate a distance between the target matrix and a matrix corresponding to preset historical rehabilitation data, and determine the patient's rehabilitation status based on the distance;

[0061] The rehabilitation item adjustment module is used to adjust the medical record data according to the rehabilitation situation to obtain updated medical record data, and update the rehabilitation items in the rehabilitation plan according to the updated medical record data.

[0062] Compared with the background technology, the present invention makes the generated medical record features more representative by embedding characters and connecting texts to medical record data; by calculating the attention score of medical record features to rehabilitation projects and calculating the semantic relationship between rehabilitation projects and medical record features, the relationship between patient medical records and rehabilitation projects is fully explored, and the data between patient medical records and rehabilitation projects is closely integrated, thereby improving the accuracy of rehabilitation project prediction; by performing data correction and clustering optimization processing on the rehabilitation data of patients performing rehabilitation projects, a target matrix is obtained, thereby achieving a global grasp of the overall implementation of the patient's rehabilitation program; by calculating the distance value between the target matrix and the matrix corresponding to the historical rehabilitation data, the patient's rehabilitation status is assessed, and then the rehabilitation program is adjusted to achieve efficient management of the patient's rehabilitation status. Therefore, the rehabilitation management method and device for stroke and cognitive impairment patients proposed in the present invention can solve the problem of low rehabilitation management efficiency caused by low accuracy of the generated rehabilitation program, inaccurate rehabilitation assessment, and inflexible adjustment of the rehabilitation program. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 A flowchart of a rehabilitation management method for stroke and cognitive impairment patients provided by one embodiment of the present invention;

[0064] Figure 2 A schematic diagram of a process for generating medical record features according to an embodiment of the present invention;

[0065] Figure 3 A schematic diagram of a process for generating a rehabilitation plan according to an embodiment of the present invention;

[0066] Figure 4 This is a functional module diagram of a rehabilitation management device for stroke and cognitive impairment patients provided by one embodiment of the present invention.

[0067] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0068] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0069] The embodiment of the present application provides a rehabilitation management method for patients with stroke or cognitive impairment. The execution subject of the rehabilitation management method for patients with stroke or cognitive impairment includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the rehabilitation management method for patients with stroke or cognitive impairment can be executed by software or hardware installed on a terminal device or a server device. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0070] Reference Figure 1 FIG. 1 is a flow chart of a rehabilitation management method for stroke and cognitive impairment patients according to an embodiment of the present invention. In this embodiment, the rehabilitation management method for stroke and cognitive impairment patients includes:

[0071] S1. Obtain the patient's medical record data, perform character embedding and text connection on the medical record data, and obtain medical record features.

[0072] In the embodiment of the present invention, the medical record data includes data on the patient's main symptoms, specialist physical examination, physical examination, diagnosis and treatment process, condition assessment, etc.

[0073] In an embodiment of the present invention, before embedding characters and concatenating texts on the medical record data, the method further includes:

[0074] De-anonymizing sensitive data in the medical record data to obtain first cleansed data;

[0075] removing stop words from the first cleaned data to obtain second cleaned data;

[0076] Keyword screening is performed on the second cleaned data to obtain preprocessed data.

[0077] In an embodiment of the present invention, the sensitive data may be personal information such as a user's name and address. By de-anonymizing, privacy leakage is prevented, the privacy security of the patient is guaranteed, and useless information such as the user's personal privacy information is eliminated in the first round.

[0078] In an embodiment of the present invention, a preset NLTK stop word list may be used to remove stop words in the first cleaned data. The stop words may be function words, conjunctions, etc. in the text.

[0079] In the embodiment of the present invention, the TF-IDF algorithm (Term Frequency-Inverse Document Frequency), the TextRank algorithm, or the PageRank algorithm may be used to perform keyword screening on the second cleaned data.

[0080] See also Figure 2 As shown, in the embodiment of the present invention, the character embedding and text connection of the medical record data to obtain medical record features includes:

[0081] S21. Using a preset BERT model to perform character embedding on the medical record data to obtain multiple characters of the medical record data;

[0082] S22, using a preset bidirectional recurrent neural network to capture context information corresponding to each of the characters;

[0083] S23. Perform string concatenation on the context information to obtain concatenated characters, and generate a medical record feature according to the concatenated characters corresponding to each character.

[0084] Specifically, the context information corresponding to the character is shown in the following formula:

[0085] ;

[0086] in, They are respectively Character context information for forward and reverse directions, The medical record data characters, Represented as the BERT bidirectional recurrent neural network;

[0087] The concatenation characters are as follows:

[0088] ;

[0089] in, The medical record data The concatenation character corresponding to the characters.

[0090] In the embodiment of the present invention, the medical record data contains multiple characters, each character corresponds to a serial character, and multiple serial characters constitute the medical record feature. The medical record feature can be expressed in the form of a vector matrix.

[0091] S2. Calculate the attention score of the medical record feature to the preset rehabilitation project, and generate a first medical record expression based on the rehabilitation project according to the attention score.

[0092] In an embodiment of the present invention, when the patient's medical record data contains "numbness in the right upper limb, the patient does not know the time of day, nor does he know where he is; the patient cannot pay attention continuously when performing an activity; can see objects but cannot recognize visual objects; cannot dress correctly, and cannot distinguish between inside and outside of clothes; memory confusion, cannot remember what he just said", the rehabilitation program prescribed for this patient includes: hand function training, orientation training, memory impairment training, attention training, perception training, and electroencephalography therapy. Among them, the content of "numbness in the right upper limb" is related to hand function training, which is used to improve hand function and enhance living ability; the content of "the patient cannot pay attention continuously when performing an activity" is related to attention training, which is used to improve the patient's concentration on things; the content of "can see objects but cannot recognize visual objects" is related to perception training, which is used to improve the ability to distinguish things. Therefore, the medical record data contains contextual information related to the rehabilitation items in the rehabilitation plan, wherein each character in the medical record data has a different contribution value to the rehabilitation item.

[0093] In the embodiment of the present invention, the attention score of the medical record feature to the preset rehabilitation items is calculated by the following formula:

[0094]

[0095] in, For the Character pair The attention score of each rehabilitation item, The medical record data The concatenation character corresponding to the characters, is the medical record feature, are the preset self-attention weight matrices, It is the preset self-attention mechanism;

[0096] The first medical record expression based on the rehabilitation project is generated by the following formula:

[0097]

[0098] in, The medical record data is based on The first medical record expression of a rehabilitation program.

[0099] S3. Calculate the semantic relationship between the rehabilitation project and the medical record feature, and generate a second medical record expression based on the rehabilitation project according to the semantic relationship.

[0100] In this embodiment of the present invention, the self-attention mechanism primarily focuses on medical record data. Different rehabilitation programs possess specific semantic information, implicitly describing the medical record data. To fully exploit this semantic information, rehabilitation programs can be represented as a trainable matrix. By capturing the semantic relevance between medical record data and rehabilitation programs through the attention mechanism, the prediction accuracy of medical record representations can be effectively improved.

[0101] In the embodiment of the present invention, the semantic relationship between the rehabilitation item and the medical record feature is calculated by the following formula:

[0102] ;

[0103] in, 、 Respectively characters and The contextual semantic relationship between rehabilitation projects, For the The vector matrix of rehabilitation items, are the forward expression and reverse expression of the medical record features respectively;

[0104] The second medical record expression based on the rehabilitation project is generated by the following formula:

[0105] ;

[0106] ;

[0107] in, The medical record data is based on The second medical record expression of a rehabilitation program.

[0108] S4. Predicting a rehabilitation project based on the first medical record expression based on the rehabilitation project and the second medical record expression based on the rehabilitation project, obtaining a rehabilitation plan for the patient, and executing the rehabilitation projects in the rehabilitation plan.

[0109] In the embodiment of the present invention, the first medical record expression based on rehabilitation items focuses on the specific content of the medical record data, and the second medical record expression based on rehabilitation items focuses on the semantic relationship between the medical record data and the rehabilitation items.

[0110] In an embodiment of the present invention, the rehabilitation program includes multiple rehabilitation items.

[0111] See also Figure 3 As shown, in an embodiment of the present invention, performing rehabilitation project prediction based on the first medical record expression based on the rehabilitation project and the second medical record expression based on the rehabilitation project to obtain the rehabilitation plan for the patient includes:

[0112] S31, setting weights according to the first medical record expression based on the rehabilitation program and the second medical record expression based on the rehabilitation program;

[0113] S32, generating a target medical record expression based on the rehabilitation program according to the weight, the first medical record expression based on the rehabilitation program, and the second medical record expression based on the rehabilitation program;

[0114] S33, using a preset multi-layer perceptron to calculate the target medical record expression based on the rehabilitation project to obtain prediction results of multiple rehabilitation projects;

[0115] S34. Screen the rehabilitation items according to the prediction results to obtain a rehabilitation plan.

[0116] In the embodiment of the present invention, in order to give full play to the advantages of the two medical record expressions, the importance between the two medical record expressions is represented by setting corresponding weights.

[0117] Specifically, the weights of the first medical record expression based on the rehabilitation project and the second medical record expression based on the rehabilitation project are set by the following formula:

[0118]

[0119] in, Based on the The weight corresponding to the first medical record expression of each rehabilitation project, Based on the The weight corresponding to the second medical record expression of each rehabilitation project, The medical record data is based on The first medical record expression of a rehabilitation project, The medical record data is based on The second medical record expression of a rehabilitation project, They are preset training parameters;

[0120] The target medical record based on the rehabilitation project is expressed as follows:

[0121]

[0122] in, Based on the The goal of each rehabilitation program is expressed in medical records;

[0123] The target medical record expression based on the rehabilitation project is calculated by the following formula:

[0124]

[0125] in, For the The predicted probability of a rehabilitation program, , 、 are the training parameters of the fully connected layer and the output layer in the multilayer perceptron, is the nonlinear activation function in the multilayer perceptron, Based on the The goal of a rehabilitation program is expressed in medical records.

[0126] In the embodiment of the present invention, after generating the predicted probability for each rehabilitation item, rehabilitation items with a predicted probability of 0 are filtered out, and rehabilitation items with a predicted probability of 1 are summarized to obtain a rehabilitation plan.

[0127] In the embodiment of the present invention, after the rehabilitation plan is generated, it is output to the user. In the embodiment of the present invention, interactive rehabilitation training can be achieved with the patient through means such as computers, artificial intelligence robots, VR (virtual reality technology), and remote rehabilitation diagnosis and treatment.

[0128] S5. Receiving rehabilitation data after the patient performs the rehabilitation items in the rehabilitation program, performing data correction and cluster optimization on the rehabilitation data, and obtaining a target matrix.

[0129] In an embodiment of the present invention, the rehabilitation data includes execution data generated during the execution of the rehabilitation items in the rehabilitation plan, data generated during condition testing after rehabilitation training, and CT images of body parts taken when the patient returns to the hospital after rehabilitation training.

[0130] For example, clinicians will use the Mini-Mental State Examination (MMSE), the Montreal Cognitive Assessment (MoCA), the Functional Independence Measurement (FIM), etc. to assess the cognitive function of patients with cognitive impairment. Doctors will judge the scale scores based on the test data of the scale and then determine the patient's recovery status.

[0131] In an embodiment of the present invention, performing data correction and cluster optimization on the rehabilitation data to obtain a target matrix includes:

[0132] Normalizing the rehabilitation data, and performing data correction on the normalized rehabilitation data using a preset KNN algorithm to obtain a correction matrix;

[0133] Utilizing each network node in a preset convolutional network to calculate the correction matrix, a node matrix is obtained;

[0134] The node matrix is clustered using a preset clustering algorithm to obtain a target matrix.

[0135] In an embodiment of the present invention, the rehabilitation data is normalized to a value between 0 and 1 through preprocessing to eliminate outliers. The KNN algorithm is used to construct a spatial matrix of the spatial information of the normalized rehabilitation data, and symmetry processing is performed to repair the relationship between adjacent points in the spatial matrix to obtain a correction matrix.

[0136] In an embodiment of the present invention, the convolutional network includes a multi-layer network and can be used for nonlinear feature aggregation. After the correction matrix is input into the convolutional network, a single embedding is output through each network node in the network, and the output of the previous network node is the input of the next network node.

[0137] The embodiment of the present invention can better capture the global information of the correction matrix through the convolutional network, better represent the characteristics of each network node, and realize the participation of all network nodes through node embedding.

[0138] In the embodiment of the present invention, the clustering algorithm includes but is not limited to the K-Means clustering algorithm, the AHC clustering algorithm, the BIRCH clustering algorithm, the GMM clustering algorithm, and the DPC clustering algorithm.

[0139] S6. Calculate a distance between the target matrix and a matrix corresponding to preset historical rehabilitation data, and determine the patient's rehabilitation status according to the distance.

[0140] In the embodiment of the present invention, the historical rehabilitation data includes historical rehabilitation data of patients and corresponding rehabilitation evaluation results.

[0141] In the embodiment of the present invention, the distance value between the target matrix and the matrix corresponding to the preset historical rehabilitation data is calculated using the following formula:

[0142]

[0143] in, is the distance value, is the dimension of the matrix corresponding to the target matrix and the preset historical rehabilitation data, is the target matrix, is the matrix corresponding to the historical rehabilitation data.

[0144] Another optional embodiment of the present invention may use Manhattan distance, Chebyshev distance, Minkowski distance, normalized Euclidean distance, cosine similarity, Jaccard distance, etc. to calculate the distance value between the target matrix and the matrix corresponding to the preset historical rehabilitation data.

[0145] In an embodiment of the present invention, the historical rehabilitation data corresponding to the matrix with the smallest distance value can be selected, and the corresponding rehabilitation evaluation result can be extracted based on the historical rehabilitation data. The rehabilitation evaluation result can be used as the evaluation result of the patient, and the evaluation result can be used to represent the patient's rehabilitation status.

[0146] S7. Adjust the medical record data according to the rehabilitation status to obtain updated medical record data, and update the rehabilitation items in the rehabilitation plan according to the updated medical record data.

[0147] In an embodiment of the present invention, different evaluation results of the recovery condition are compared with the patient's medical record data, so as to judge whether the recovery condition has improved and the degree of improvement. When the recovery condition has not changed, the relevant description content in the case data remains unchanged; when the recovery condition has improved and the degree of improvement is small, the relevant description content in the case data is updated according to the recovery condition.

[0148] In an embodiment of the present invention, after the medical record data is updated, rehabilitation project prediction can be re-performed on the updated case data. The process of updating the rehabilitation projects in the rehabilitation plan based on the updated medical record data is similar to the process in the above steps S1-S4, and will not be elaborated here.

[0149] Compared with the background technology, the present invention makes the generated medical record features more representative by embedding characters and connecting texts to medical record data; by calculating the attention score of medical record features to rehabilitation projects and calculating the semantic relationship between rehabilitation projects and medical record features, the relationship between patient medical records and rehabilitation projects is fully explored, and the data between patient medical records and rehabilitation projects is closely integrated, thereby improving the accuracy of rehabilitation project prediction; by performing data correction and clustering optimization processing on the rehabilitation data of patients performing rehabilitation projects, a target matrix is obtained, thereby achieving a global grasp of the overall implementation of the patient's rehabilitation program; by calculating the distance value between the target matrix and the matrix corresponding to the historical rehabilitation data, the patient's rehabilitation status is assessed, and then the rehabilitation project is adjusted to achieve efficient management of the patient's rehabilitation status. Therefore, the rehabilitation management method for stroke and cognitive impairment patients proposed in the present invention can solve the problem of low rehabilitation management efficiency caused by low accuracy of the generated rehabilitation program, inaccurate rehabilitation assessment, and inflexible adjustment of the rehabilitation program.

[0150] like Figure 4 1 is a functional module diagram of a rehabilitation management device for stroke and cognitive impairment patients provided by an embodiment of the present invention.

[0151] The rehabilitation management device 100 for stroke and cognitive impairment patients described in the present invention can be installed in an electronic device. Depending on the functionality implemented, the rehabilitation management device 100 for stroke and cognitive impairment patients can include a medical record feature generation module 101, a first medical record expression generation module 102, a second medical record expression generation module 103, a rehabilitation plan generation module 104, a rehabilitation data processing module 105, a rehabilitation status assessment module 106, and a rehabilitation program adjustment module 107. The modules described in the present invention, also referred to as units, refer to a series of computer program segments that can be executed by an electronic device processor and perform fixed functions, and are stored in the electronic device's memory.

[0152] The medical record feature generation module 101 is used to obtain the patient's medical record data, perform character embedding and text concatenation on the medical record data, and obtain medical record features;

[0153] The first medical record expression generating module 102 is configured to calculate the attention score of the medical record feature to the preset rehabilitation program, and generate a first medical record expression based on the rehabilitation program according to the attention score;

[0154] The second medical record expression generating module 103 is configured to calculate the semantic relationship between the rehabilitation program and the medical record feature, and generate a second medical record expression based on the rehabilitation program according to the semantic relationship;

[0155] The rehabilitation program generation module 104 is configured to predict rehabilitation programs based on the first medical record expression based on rehabilitation programs and the second medical record expression based on rehabilitation programs, obtain a rehabilitation program for the patient, and execute rehabilitation programs in the rehabilitation program;

[0156] The rehabilitation data processing module 105 is configured to receive rehabilitation data after the patient performs the rehabilitation items in the rehabilitation program, perform data correction and cluster optimization on the rehabilitation data, and obtain a target matrix;

[0157] The rehabilitation status evaluation module 106 is configured to calculate a distance between the target matrix and a matrix corresponding to preset historical rehabilitation data, and determine the patient's rehabilitation status based on the distance.

[0158] The rehabilitation item adjustment module 107 is configured to adjust the medical record data according to the rehabilitation status to obtain updated medical record data, and update the rehabilitation items in the rehabilitation plan according to the updated medical record data.

[0159] The electronic device may include a processor, a memory and a bus, and may also include a computer program stored in the memory and executable on the processor, such as a rehabilitation management method program based on use by patients with stroke or cognitive impairment.

[0160] The memory includes at least one type of readable storage medium, including flash memory, a mobile hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory may be an internal storage unit of the electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory may also be an external storage device of the electronic device, such as a plug-in mobile hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device. Furthermore, the memory may include both an internal storage unit of the electronic device and an external storage device. The memory can be used not only to store application software installed in the electronic device and various types of data, such as the code of a rehabilitation management method program for stroke and cognitive impairment patients, but also to temporarily store data that has been output or is about to be output.

[0161] In some embodiments, the processor may be comprised of an integrated circuit, such as a single packaged integrated circuit or multiple packaged integrated circuits with the same or different functions, including a combination of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor is the control core (control unit) of the electronic device, connecting the various components of the entire electronic device using various interfaces and circuits. It executes programs or modules stored in the memory (e.g., rehabilitation management programs for stroke and cognitive impairment patients) and accesses data stored in the memory to perform various functions of the electronic device and process data.

[0162] The bus may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to enable communication between the memory and at least one processor.

[0163] For example, although not shown, the electronic device may further include a power source (such as a battery) to power various components. Preferably, the power source may be logically connected to the at least one processor via a power management device, thereby enabling functions such as charge management, discharge management, and power consumption management via the power management device. The power source may further include any of one or more DC or AC power sources, a recharging device, a power failure detection circuit, a power converter or inverter, a power status indicator, and the like. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, and the like, which are not further detailed here.

[0164] Furthermore, the electronic device may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device and other electronic devices.

[0165] Optionally, the electronic device may further include a user interface, which may be a display or an input unit (such as a keyboard). Optionally, the user interface may also be a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display may also be appropriately referred to as a display screen or display unit, and is used to display information processed in the electronic device and to display a visual user interface.

[0166] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0167] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the module division is merely a logical function division, and other division methods may be used in actual implementation.

[0168] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network elements. Some or all of the modules may be selected to achieve the purpose of the solution of this embodiment according to actual needs.

[0169] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or hardware plus software functional modules.

[0170] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0171] Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the claims are intended to be embraced therein. Any reference to a figure in a claim should not be construed as limiting the claim to which it relates.

[0172] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a system claim may also be implemented by a single unit or device through software or hardware. Second-order terms are used to indicate names and do not imply any particular order.

[0173] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A rehabilitation management method for patients with stroke and cognitive impairment, characterized in that: The method comprises: Obtaining the patient's medical record data, performing character embedding and text concatenation on the medical record data, and obtaining medical record features; Calculating an attention score of the medical record feature to a preset rehabilitation program, and generating a first medical record expression based on the rehabilitation program according to the attention score; Calculating a semantic relationship between the rehabilitation project and the medical record feature, and generating a second medical record expression based on the rehabilitation project according to the semantic relationship; Predicting a rehabilitation project based on the first medical record expression based on the rehabilitation project and the second medical record expression based on the rehabilitation project, obtaining a rehabilitation plan for the patient, and executing the rehabilitation projects in the rehabilitation plan; Receiving rehabilitation data of the patient after performing the rehabilitation items in the rehabilitation program, performing data correction and cluster optimization on the rehabilitation data to obtain a target matrix; Calculating a distance between the target matrix and a matrix corresponding to preset historical rehabilitation data, and determining the patient's rehabilitation status according to the distance; Adjusting the medical record data according to the rehabilitation status to obtain updated medical record data, and updating the rehabilitation items in the rehabilitation plan according to the updated medical record data; The data correction and cluster optimization are performed on the rehabilitation data to obtain a target matrix, including: Normalizing the rehabilitation data, and performing data correction on the normalized rehabilitation data using a preset KNN algorithm to obtain a correction matrix; Utilizing each network node in a preset convolutional network to calculate the correction matrix, a node matrix is obtained; The node matrix is clustered using a preset clustering algorithm to obtain a target matrix.

2. The rehabilitation management method for stroke and cognitive impairment patients according to claim 1, characterized in that: The character embedding and text concatenation of the medical record data to obtain medical record features includes: Using a preset BERT model to perform character embedding on the medical record data to obtain multiple characters of the medical record data; Utilizing a preset bidirectional recurrent neural network to capture contextual information corresponding to each of the characters; The context information is string-concatenated to obtain concatenated characters, and a medical record feature is generated according to the concatenated characters corresponding to each character.

3. The rehabilitation management method for stroke and cognitive impairment patients according to claim 1, characterized in that: Before embedding characters and connecting texts to the medical record data, the method further includes: De-anonymizing sensitive data in the medical record data to obtain first cleansed data; removing stop words from the first cleaned data to obtain second cleaned data; Keyword screening is performed on the second cleaned data to obtain pre-processed data.

4. The rehabilitation management method for stroke and cognitive impairment patients according to claim 1, characterized in that: The step of calculating the attention score of the medical record feature to a preset rehabilitation program and generating a first medical record expression based on the rehabilitation program according to the attention score includes: The attention score of the medical record feature to the preset rehabilitation project is calculated by the following formula: in, For the Character pair The attention score of each rehabilitation project, The medical record data The concatenation character corresponding to the characters, is the medical record feature, are the preset self-attention weight matrices, It is the preset self-attention mechanism; The first medical record expression based on the rehabilitation project is generated by the following formula: in, The medical record data is based on The first medical record expression of a rehabilitation program.

5. The rehabilitation management method for stroke and cognitive impairment patients according to claim 1, characterized in that: The calculating the semantic relationship between the rehabilitation program and the medical record feature, and generating a second medical record expression based on the rehabilitation program according to the semantic relationship, includes: The semantic relationship between the rehabilitation item and the medical record feature is calculated using the following formula: ； in, 、 Respectively characters and The contextual semantic relationship between rehabilitation projects, For the The vector matrix of rehabilitation items, are the forward expression and reverse expression of the medical record features respectively; The second medical record expression based on the rehabilitation project is generated by the following formula: ； ； in, The medical record data is based on The second medical record expression of a rehabilitation program.

6. The rehabilitation management method for stroke and cognitive impairment patients according to claim 1, characterized in that: The performing of rehabilitation project prediction based on the first medical record expression based on the rehabilitation project and the second medical record expression based on the rehabilitation project to obtain the rehabilitation plan for the patient includes: Setting weights according to the first medical record expression based on the rehabilitation program and the second medical record expression based on the rehabilitation program respectively; generating a target medical record expression based on the rehabilitation program according to the weight, the first medical record expression based on the rehabilitation program, and the second medical record based on the rehabilitation program; Utilizing a preset multi-layer perceptron to calculate the target medical record expression based on the rehabilitation project, and obtaining prediction results of multiple rehabilitation projects; The rehabilitation items are screened according to the prediction results to obtain a rehabilitation plan.

7. The rehabilitation management method for stroke and cognitive impairment patients according to claim 6, characterized in that: The setting of weights according to the first medical record expression based on the rehabilitation project and the second medical record expression based on the rehabilitation project respectively includes: The weights of the first medical record expression based on the rehabilitation project and the second medical record expression based on the rehabilitation project are set by the following formula: in, Based on the The weight corresponding to the first medical record expression of each rehabilitation project, Based on the The weight corresponding to the second medical record expression of each rehabilitation project, The medical record data is based on The first medical record expression of a rehabilitation project, The medical record data is based on The second medical record expression of a rehabilitation project, are preset training parameters respectively.

8. The rehabilitation management method for stroke and cognitive impairment patients according to claim 6, characterized in that: The calculation of the target medical record expression based on the rehabilitation project by using a preset multi-layer perceptron includes: The target medical record expression based on the rehabilitation project is calculated by the following formula: in, For the The predicted probability of a rehabilitation program, , 、 are the training parameters of the fully connected layer and the output layer in the multilayer perceptron, is the nonlinear activation function in the multilayer perceptron, Based on the The goal of a rehabilitation program is expressed in medical records.

9. A rehabilitation management device for patients with stroke and cognitive impairment, characterized in that: The device is used for applying the rehabilitation management method for stroke and cognitive impairment patients according to any one of claims 1 to 8, comprising: A medical record feature generation module is used to obtain the patient's medical record data, perform character embedding and text concatenation on the medical record data, and obtain medical record features; A first medical record expression generating module is configured to calculate an attention score of the medical record feature to a preset rehabilitation program, and generate a first medical record expression based on the rehabilitation program according to the attention score; A second medical record expression generating module is configured to calculate a semantic relationship between the rehabilitation project and the medical record feature, and generate a second medical record expression based on the rehabilitation project according to the semantic relationship; a rehabilitation program generation module, configured to predict rehabilitation programs based on the first medical record expression based on rehabilitation programs and the second medical record expression based on rehabilitation programs, obtain a rehabilitation program for the patient, and execute rehabilitation programs in the rehabilitation program; a rehabilitation data processing module, configured to receive rehabilitation data after the patient performs rehabilitation items in the rehabilitation program, perform data correction and cluster optimization on the rehabilitation data, and obtain a target matrix; a rehabilitation status evaluation module, configured to calculate a distance between the target matrix and a matrix corresponding to preset historical rehabilitation data, and determine the patient's rehabilitation status based on the distance; The rehabilitation item adjustment module is used to adjust the medical record data according to the rehabilitation situation to obtain updated medical record data, and update the rehabilitation items in the rehabilitation plan according to the updated medical record data.

Images