Elderly patient falling event data prediction method based on common disease index

Abstract

The invention provides an old patient falling event data prediction method based on a common disease index. According to the method, health record data of a target elderly patient is obtained, a target associated disease data set is determined from the health record data according to a preset tumble associated disease map, a target co-disease index is determined according to the target associated disease data set, and then the tumble associated disease map is determined only when the target co-disease index is smaller than a preset index threshold value. And extracting a patient feature set according to the health record data, and inputting the patient feature set into a preset tumble event prediction model to output a tumble risk level of the target elderly patient, thereby realizing automatic prediction and evaluation of the tumble risk level of the target elderly patient.

Description

Technical Field

[0001] The present application relates to data processing technology, and in particular to a method for predicting fall event data in elderly patients based on a comorbidity index. Background Art

[0002] As the global population ages, the health issues of the elderly population are receiving increasing attention. Frequent falls in elderly patients not only seriously affect their quality of life, but may also cause a series of serious health problems, such as fractures, craniocerebral injuries, and even endanger their lives. Therefore, how to accurately predict the risk of falls in elderly patients and take effective preventive measures has become an urgent problem to be solved in the current medical and nursing fields.

[0003] Traditionally, the assessment of fall risk for elderly patients mainly relies on the clinical experience and intuitive judgment of medical staff. This method has problems such as strong subjectivity and inconsistent assessment standards, making it difficult to achieve automatic predictive assessment of fall risk for elderly patients. Summary of the invention

[0004] The present application provides a method for predicting fall event data of elderly patients based on a comorbidity index, so as to realize automatic prediction and evaluation of the fall risk level of target elderly patients.

[0005] In a first aspect, the present application provides a method for predicting fall event data of elderly patients based on a comorbidity index, comprising:

[0006] Acquire health record data of a target elderly patient, and determine a target associated disease data set from the health record data according to a preset fall-associated disease atlas, so as to determine a target comorbidity index according to the target associated disease data set, wherein the target associated disease data in the target associated disease data set include target associated disease types and weight coefficients corresponding to the target associated disease types;

[0007] Only when the target comorbidity index is less than a preset index threshold, extracting a patient feature set according to the health record data, the patient feature set including the target comorbidity index, physiological feature evaluation parameters, and spatial perception ability evaluation parameters of the target elderly patient;

[0008] The patient feature set is input into a preset fall event prediction model to output the fall risk level of the target elderly patient.

[0009] In the above scheme, by combining the health record data of elderly patients and the preset fall-related disease map, the types of diseases that are highly correlated with the risk of falls and their weight coefficients can be systematically identified, thereby constructing a target comorbidity index. This not only takes into account the impact of a single disease, but also integrates the interactions and associations between multiple diseases, making the assessment of the risk of falls in elderly patients more comprehensive and scientific.

[0010] When the target comorbidity index is less than the preset index threshold, the patient's physiological characteristics evaluation parameters and spatial perception ability evaluation parameters are further extracted to form a personalized patient feature set. This prediction method based on the comorbidity index and individual characteristics can conduct risk assessment based on the specific situation of each elderly patient, improving the accuracy and practicality of the prediction.

[0011] Then, by inputting the patient feature set into the preset fall event prediction model to output the fall risk level of the target elderly patient, it can provide a clear reference for medical institutions, nursing staff and patients' families, helping them to formulate targeted preventive measures and intervention plans, thereby effectively reducing the incidence of falls in elderly patients and mitigating their consequences. In turn, by quickly identifying high-risk patients, it helps medical institutions to rationally allocate limited medical resources.

[0012] In addition, as the patient's health status changes, the health record data also changes accordingly, so it is also possible to dynamically monitor and update the fall risk of elderly patients in real time, ensuring the timeliness and accuracy of the prediction results.

[0013] Optionally, after determining the target comorbidity index according to the target-associated disease data set, the method further includes:

[0014] If the target comorbidity index is greater than or equal to the preset index threshold, the fall risk level of the target elderly patient is determined to be a first risk level, wherein the first risk level is the highest risk level among the fall risk levels.

[0015] In the above scheme, a clear preset index threshold is set to quickly determine the fall risk level of elderly patients. When the target comorbidity index reaches or exceeds this threshold, the fall risk level is directly determined to be the highest level (first risk level), which greatly simplifies the risk assessment process and enables high-risk patients to be quickly identified.

[0016] Moreover, when the target comorbidity index reaches or exceeds this threshold, there is no need to extract the patient feature set and input it into the preset fall event prediction model for complex data extraction and calculation. This can not only ensure the reliability of the prediction results, but also avoid the occupation of excessive computing resources and ensure the speed of result output.

[0017] Optionally, the preset fall-associated disease map includes a node set and an edge set, each node in the node set is used to represent a corresponding associated disease type, and each edge in the edge set is used to connect two different nodes, and is used to represent the association relationship between the associated disease types corresponding to the two connected nodes; correspondingly, determining a target associated disease data set from the health record data according to the preset fall-associated disease map, so as to determine a target comorbidity index according to the target associated disease data set, includes:

[0018] Traversing each node in the node set corresponding to the preset fall-related disease map, and adding the nodes corresponding to the associated disease types existing in the health record data to the target node set;

[0019] Determine a target edge set according to the target node set P and the corresponding edge set in the preset fall-related disease graph, wherein a target edge in the target edge set is used to connect two target nodes in the target node set;

[0020] Determine a comorbidity impact path set according to the target node set and the target edge set;

[0021] Determine the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set according to the preset node weight coefficient and the preset edge weight coefficient of each node in the preset fall-related disease graph;

[0022] The target comorbidity index is determined according to the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set.

[0023] In the above scheme, by constructing a preset fall-related disease map containing a set of nodes and a set of edges, various types of related diseases in the health record data of elderly patients and the complex relationships between them can be systematically identified and characterized. This systematic analysis method ensures the comprehensiveness and accuracy of fall risk prediction and can more comprehensively consider the comprehensive impact of multiple disease factors on the fall risk of elderly patients.

[0024] After determining the target associated disease data set, the comorbidity impact path set is further determined based on the target node set and edge set. These path sets intuitively show the mutual influence and transmission path between different diseases. By dynamically analyzing these paths, the potential impact of comorbid conditions on the risk of falls in elderly patients can be more accurately assessed, improving the accuracy and timeliness of predictions.

[0025] By using the preset node weight coefficients and edge weight coefficients, the path comorbidity impact index of each comorbidity impact path can be quantitatively calculated, and then the target comorbidity index can be determined. This quantitative evaluation method not only provides a scientific prediction basis, but also makes the prediction results more intuitive and easy to understand. At the same time, accurate prediction based on the target comorbidity index can provide medical institutions and patients' families with more specific fall risk level information, which is helpful to formulate targeted preventive measures and intervention strategies.

[0026] In addition, the construction process of the preset fall-related disease map and the setting of the weight coefficients of nodes and edges are flexible and scalable. With the continuous deepening of medical research and the continuous discovery of disease associations, the map can be updated and improved to meet the prediction needs of different elderly patient groups. This flexibility and scalability make the above solution have a long life cycle and broad application prospects.

[0027] Optionally, the preset fall-related disease atlas also includes a hierarchical architecture, each layer in the hierarchical architecture is used to accommodate nodes in the node set, each node in the node set is configured with a corresponding level in the hierarchical architecture, and the level of each node in the node set is negatively correlated with the corresponding point weight coefficient; correspondingly, determining the comorbidity impact path set according to the target node set P and the target edge set includes:

[0028] Sorting the target node set in ascending order according to the level information of each node in the node set to generate an updated target node set, wherein the updated target node set includes a subset of target nodes corresponding to each level of the hierarchical architecture;

[0029] Traversing each target node in the target node subset to determine a starting target node, the starting target node being one of the end nodes on the target edge in the starting target edge subset of the target edge set, the starting target edge subset being a set consisting of all target edges in the target edge set that include the starting target node, and the other end nodes on the target edges in the starting target edge subset corresponding to a level in the hierarchical architecture that is greater than the level corresponding to the starting target node in the hierarchical architecture;

[0030] Traversing each target node in the target node subset to determine a termination target node, the termination target node being one of the end nodes on the target edge in the termination target edge subset of the target edge set, the termination target edge subset being a set consisting of all target edges in the target edge set L that include the termination target node, and the levels corresponding to the other end nodes on the target edges in the termination target edge subset in the hierarchical architecture are all less than or equal to the levels corresponding to the termination target edge subset in the hierarchical architecture;

[0031] Generate a starting target node set according to all starting target nodes determined at each level of the hierarchical architecture, and generate a terminating target node set according to all terminating target nodes;

[0032] The comorbidity impact path set is determined according to the starting target node set and the terminating target node set, wherein the starting node of the target node sequence corresponding to the comorbidity impact path in the comorbidity impact path set belongs to the starting target node set, the terminating node belongs to the terminating target node set, and the nodes in the target node sequence are connected in sequence through the target edges in the target edge set.

[0033] In the above scheme, a hierarchical architecture is introduced into the preset fall-related disease map, so that the disease nodes are logically organized into different levels, which helps to realize the hierarchical analysis of the comorbidity of elderly patients. Each level represents a different disease complexity and risk level, which helps to understand and evaluate the potential impact of different diseases on the risk of falls in more detail. By setting the negative correlation between the node level and the weight coefficient, the contribution of nodes with higher levels in the comorbidity impact path is relatively low. The method for determining the starting target node and the ending target node based on the hierarchical architecture effectively reduces the redundant paths in the set of comorbidity impact paths, and only retains those paths that transfer influences from high levels to low levels or between the same levels. This optimization not only reduces the computational complexity, but also improves the efficiency and stability of the prediction model.

[0034] Optionally, determining the target comorbidity index according to the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set S includes:

[0035] Determine a corresponding comorbidity impact path node sequence according to each comorbidity impact path in the comorbidity impact path set;

[0036] The target comorbidity index is determined according to the comorbidity influence path node sequence corresponding to each comorbidity influence path in the comorbidity influence path set and the preset node weight coefficient and the preset edge weight coefficient of each node in the preset fall-related disease map.

[0037] In the above scheme, the node sequence of the comorbidity impact path of each path is determined through detailed analysis of each path in the comorbidity impact path set. This step enables the prediction method to accurately capture the complex interactions and associations between different diseases, thus providing a solid foundation for the quantitative evaluation of comorbidity impacts. The preset node weight coefficient and the preset edge weight coefficient are introduced, and the target comorbidity index is calculated in combination with the comorbidity impact path node sequence. The setting of this weight coefficient reflects the importance of different diseases and their interrelationships in the prediction of fall risk, ensuring the accuracy and scientificity of the evaluation results.

[0038] Taking into account the diversity and complexity of the impact pathways of comorbidities, a comprehensive evaluation of each pathway can more comprehensively reflect the impact of comorbidities on the risk of falls in elderly patients. This method significantly improves the accuracy and reliability of predictions compared to single disease evaluation or simple summation. By setting weight coefficients, the subjectivity and bias in the evaluation process are reduced, making the prediction results more objective and credible.

[0039] Optionally, the physiological characteristic assessment parameter includes at least one of a muscle strength risk assessment index, a balance ability risk assessment index, a hearing risk assessment index, and a vision risk assessment index;

[0040] The spatial perception capability assessment parameter includes at least one of a direction discrimination capability risk assessment index, a spatial positioning capability risk assessment index, and a depth perception risk assessment index.

[0041] In the above scheme, by incorporating physiological characteristic assessment parameters such as muscle strength risk assessment index, balance ability risk assessment index, hearing risk assessment index and vision risk assessment index, the physiological function status of elderly patients can be comprehensively assessed. These parameters are directly related to the physical stability, coordination ability and perception of external stimuli of elderly patients, which are crucial for accurately predicting the risk of falls. Taking into account multiple physiological characteristic assessment parameters, the prediction model can capture subtle changes in the physiological functions of elderly patients, thereby more accurately assessing their risk of falling. This multi-factor comprehensive evaluation method can significantly improve the accuracy and reliability of the prediction compared to the evaluation of a single factor.

[0042] In addition, the evaluation dimensions of the prediction model have been further expanded by adding spatial perception ability assessment parameters (such as direction discrimination ability, spatial positioning ability, and depth perception ability). These abilities are crucial for elderly patients to walk, turn, avoid obstacles, and other actions in their daily lives, and are directly related to their risk of falling. By evaluating these abilities, we can have a more comprehensive understanding of the risk factors for falling in elderly patients.

[0043] Optionally, the step of inputting the patient feature set into a preset fall event prediction model to output a fall risk level of the target elderly patient includes:

[0044] Determining the fall risk assessment score according to the patient feature set;

[0045] The fall risk level of the target elderly patient is output according to the fall risk assessment score and a preset fall risk assessment threshold interval distribution.

[0046] In the above scheme, by calculating the fall risk assessment score, a quantitative assessment of the patient's fall risk is achieved. This quantitative assessment method is more objective and accurate than subjective judgment, and provides a more accurate basis for risk assessment for medical institutions and nursing staff. Then, according to the preset fall risk assessment threshold interval distribution, the fall risk assessment score is converted into a specific fall risk level. This grading method makes the risk assessment results more intuitive and easy to understand, which is convenient for medical staff to apply in actual work.

[0047] In a second aspect, the present application provides a device for predicting fall event data of elderly patients based on a comorbidity index, comprising:

[0048] An acquisition module is used to acquire health record data of a target elderly patient, and determine a target-associated disease data set from the health record data according to a preset fall-associated disease atlas, wherein the target-associated disease data in the target-associated disease data set include target-associated disease types and weight coefficients corresponding to the target-associated disease types;

[0049] a processing module, configured to determine a target comorbidity index according to the target associated disease data set,

[0050] An extraction module, configured to extract a patient feature set according to the health record data only when the target comorbidity index is less than a preset index threshold, the patient feature set including the target comorbidity index, physiological feature evaluation parameters, and spatial perception ability evaluation parameters of the target elderly patient;

[0051] The processing module is further used to input the patient feature set into a preset fall event prediction model to output the fall risk level of the target elderly patient.

[0052] Optionally, the processing module is further used to determine that the fall risk level of the target elderly patient is a first risk level, wherein the first risk level is the highest risk level among the fall risk levels.

[0053] Optionally, the preset fall-associated disease graph includes a node set and an edge set, each node in the node set is used to represent a corresponding associated disease type, and each edge in the edge set is used to connect two different nodes to represent the association relationship between the associated disease types corresponding to the two connected nodes; the processing module is specifically used to:

[0054] Traversing each node in the node set corresponding to the preset fall-related disease map, and adding the nodes corresponding to the associated disease types existing in the health record data to the target node set;

[0055] Determine a target edge set according to the target node set P and the corresponding edge set in the preset fall-related disease graph, wherein a target edge in the target edge set is used to connect two target nodes in the target node set;

[0056] Determine a comorbidity impact path set according to the target node set and the target edge set;

[0057] Determine the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set according to the preset node weight coefficient and the preset edge weight coefficient of each node in the preset fall-related disease graph;

[0058] The target comorbidity index is determined according to the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set.

[0059] Optionally, the preset fall-related disease atlas further includes a hierarchical structure, each layer in the hierarchical structure is used to accommodate nodes in the node set, each node in the node set is configured with a corresponding level in the hierarchical structure, and the level of each node in the node set is negatively correlated with the corresponding point weight coefficient; the processing module is specifically used to:

[0060] Sorting the target node set in ascending order according to the level information of each node in the node set to generate an updated target node set, wherein the updated target node set includes a subset of target nodes corresponding to each level of the hierarchical architecture;

[0061] Traversing each target node in the target node subset to determine a starting target node, the starting target node being one of the end nodes on the target edge in the starting target edge subset of the target edge set, the starting target edge subset being a set consisting of all target edges in the target edge set that include the starting target node, and the other end nodes on the target edges in the starting target edge subset corresponding to a level in the hierarchical architecture that is greater than the level corresponding to the starting target node in the hierarchical architecture;

[0062] Traversing each target node in the target node subset to determine a termination target node, the termination target node being one of the end nodes on the target edge in the termination target edge subset of the target edge set, the termination target edge subset being a set consisting of all target edges in the target edge set L that include the termination target node, and the levels corresponding to the other end nodes on the target edges in the termination target edge subset in the hierarchical architecture are all less than or equal to the levels corresponding to the termination target edge subset in the hierarchical architecture;

[0063] Generate a starting target node set according to all starting target nodes determined at each level of the hierarchical architecture, and generate a terminating target node set according to all terminating target nodes;

[0064] The comorbidity impact path set is determined according to the starting target node set and the terminating target node set, wherein the starting node of the target node sequence corresponding to the comorbidity impact path in the comorbidity impact path set belongs to the starting target node set, the terminating node belongs to the terminating target node set, and the nodes in the target node sequence are connected in sequence through the target edges in the target edge set.

[0065] Optionally, the processing module is specifically used to:

[0066] Determine a corresponding comorbidity impact path node sequence according to each comorbidity impact path in the comorbidity impact path set;

[0067] The target comorbidity index is determined according to the comorbidity influence path node sequence corresponding to each comorbidity influence path in the comorbidity influence path set and the preset node weight coefficient and the preset edge weight coefficient of each node in the preset fall-related disease map.

[0068] Optionally, the physiological characteristic assessment parameter includes at least one of a muscle strength risk assessment index, a balance ability risk assessment index, a hearing risk assessment index, and a vision risk assessment index;

[0069] The spatial perception capability assessment parameter includes at least one of a direction discrimination capability risk assessment index, a spatial positioning capability risk assessment index, and a depth perception risk assessment index.

[0070] Optionally, the processing module is specifically used to:

[0071] Determining the fall risk assessment score according to the patient feature set;

[0072] The fall risk level of the target elderly patient is output according to the fall risk assessment score and a preset fall risk assessment threshold interval distribution.

[0073] In a third aspect, the present application provides an electronic device, including:

[0074] processor; and,

[0075] A memory, configured to store executable instructions of the processor;

[0076] The processor is configured to perform any possible method described in the first aspect by executing the executable instructions.

[0077] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement any possible method described in the first aspect.

[0078] The present application provides a method for predicting fall event data for elderly patients based on a comorbidity index, which obtains health record data of a target elderly patient, and determines a target associated disease data set from the health record data according to a preset fall-associated disease map, so as to determine a target comorbidity index according to the target associated disease data set. Then, only when the target comorbidity index is less than a preset index threshold, a patient feature set is extracted according to the health record data, and the patient feature set is input into a preset fall event prediction model to output the fall risk level of the target elderly patient, thereby realizing automatic prediction and evaluation of the fall risk level of the target elderly patient. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0080] Figure 1 is a flow chart of a method for predicting fall event data of elderly patients based on a comorbidity index according to an exemplary embodiment of the present application;

[0081] Figure 2 is a flow chart of a method for predicting fall event data of elderly patients based on a comorbidity index according to another exemplary embodiment of the present application;

[0082] Figure 3 is a structural schematic diagram of a device for predicting fall event data of elderly patients based on a comorbidity index according to an exemplary embodiment of the present application;

[0083] Figure 4 It is a schematic diagram of the structure of an electronic device according to an exemplary embodiment of the present application.

[0084] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0085] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0086] In order to solve the above problems, the embodiment provided by the present application obtains the health record data of the target elderly patient and determines the target associated disease data set from it according to the preset fall-associated disease map. The set includes the target associated disease types and their corresponding weight coefficients, which are used to calculate the target comorbidity index. The comorbidity index not only considers the impact of a single disease, but also integrates the interactions and associations between multiple diseases, providing a more comprehensive and scientific basis for the assessment of fall risk.

[0087] When the target comorbidity index is less than the preset index threshold, the patient's physiological characteristics evaluation parameters and spatial perception ability evaluation parameters are further extracted from the health record data to construct a patient feature set. This step ensures that a more detailed individual feature evaluation is performed only when the risk of comorbidity is relatively low, thereby optimizing resource utilization and improving evaluation efficiency. Then, the constructed patient feature set is input into the preset fall event prediction model, which processes the data through an internal algorithm and outputs the fall risk level of the target elderly patient. This step realizes the quantitative assessment of fall risk and provides a scientific basis for medical decision-making.

[0088] When the target comorbidity index is greater than or equal to the preset index threshold, the patient's fall risk is directly determined to be the highest level (first risk level). This design simplifies the assessment process, allowing high-risk patients to be quickly identified and appropriate preventive measures to be taken.

[0089] In addition, the preset fall-related disease graph (including node sets, edge sets and hierarchical architecture) is used to determine the target-related disease data set, and the comorbidity impact path set and target comorbidity index can be further calculated, thereby ensuring the comprehensiveness and accuracy of fall risk assessment and fully considering the complex correlation between diseases.

[0090] By combining health record data and fall-related disease maps, the system identifies the types of diseases that are highly correlated with the risk of falls and their weight coefficients, and constructs a scientific and comprehensive comorbidity index. Based on the comprehensive evaluation of the comorbidity index and individual physiological characteristics and spatial perception ability, personalized prediction of the risk of falls in elderly patients is achieved, and the accuracy of the prediction is improved. The preset fall event prediction model can automatically process data and output the fall risk level, which improves the evaluation efficiency. At the same time, the evaluation results provide a clear reference for medical institutions and nursing staff, which helps to formulate targeted preventive measures and intervention plans.

[0091] In summary, the embodiments provided in this application achieve the goal of comprehensive, scientific and personalized prediction of the fall risk of elderly patients through the steps of comorbidity index construction, conditional risk assessment execution, prediction model application, rapid identification of high comorbidity risk, and detailed physiological characteristics and spatial perception ability assessment.

[0092] Figure 1 FIG. 1 is a flow chart of a method for predicting fall event data of elderly patients based on a comorbidity index according to an exemplary embodiment of the present application. Figure 1 As shown, the method for predicting fall event data of elderly patients based on comorbidity index provided in this embodiment includes:

[0093] S101. Obtain health record data of a target elderly patient, and determine a target associated disease data set from the health record data according to a preset fall-associated disease map, so as to determine a target comorbidity index according to the target associated disease data set.

[0094] In this step, the health record data of the target elderly patients are obtained, and a target associated disease data set is determined from the health record data according to a preset fall-associated disease map, so as to determine a target comorbidity index according to the target associated disease data set. The target associated disease data in the target associated disease data set include target associated disease types and weight coefficients corresponding to the target associated disease types.

[0095] Specifically, first, the health record data of the target elderly patients is collected through the electronic health record system or the database of the medical institution. These data include but are not limited to the basic information of the patients (such as age, gender, living habits, etc.), past medical history, current diagnosis results, treatment records and physical examination results.

[0096] In addition, the preset fall-related disease atlas is a network model that contains a variety of disease types related to fall risk and their associations. The atlas consists of a set of nodes and a set of edges. Each node represents an associated disease, and the edge represents the association between diseases. According to this atlas, the patient's health record data is traversed to identify all the disease types that exist in the atlas and use them as part of the target associated disease data set. At the same time, according to the weight coefficient preset for each disease node in the atlas, the weight value corresponding to each target associated disease is recorded.

[0097] Based on the target-related disease data set and the weight coefficients of the nodes and edges in the graph, the target comorbidity index is calculated through a series of complex algorithms. This index comprehensively reflects the impact of comorbid conditions on the risk of falls in elderly patients.

[0098] S102. Only when the target comorbidity index is less than a preset index threshold, extract a patient feature set based on the health record data.

[0099] In this step, only when the target comorbidity index is less than the preset index threshold, a patient feature set is extracted based on the health record data, and the patient feature set includes the target comorbidity index, physiological characteristic evaluation parameters, and spatial perception ability evaluation parameters of the target elderly patient.

[0100] Specifically, after calculating the target comorbidity index, it is first determined whether the index is less than a preset index threshold. This threshold is set based on a large amount of historical data and expert experience to distinguish between high and low fall risks.

[0101] If the target comorbidity index is less than the preset threshold, the patient's physiological characteristic assessment parameters and spatial perception ability assessment parameters are further extracted from the health record data. Physiological characteristic parameters may include muscle strength risk assessment index, balance ability risk assessment index, hearing risk assessment index and vision risk assessment index, etc.; spatial perception ability parameters may include direction discrimination ability risk assessment index, spatial positioning ability risk assessment index and depth perception risk assessment index, etc. It is worth noting that the above-mentioned physiological characteristic assessment parameters and spatial perception ability assessment parameters can be obtained through physical examination or related tests.

[0102] Then, the target comorbidity index, the extracted physiological characteristic assessment parameters, and the spatial perception ability assessment parameters were integrated into a patient feature set for subsequent fall risk assessment.

[0103] Further, if the target comorbidity index is greater than or equal to a preset index threshold, the fall risk level of the target elderly patient is determined to be a first risk level, wherein the first risk level is the highest risk level among the fall risk levels.

[0104] In the above scheme, a clear preset index threshold is set to quickly determine the fall risk level of elderly patients. When the target comorbidity index reaches or exceeds this threshold, the fall risk level is directly determined to be the highest level (first risk level), which greatly simplifies the risk assessment process and enables high-risk patients to be quickly identified.

[0105] Moreover, when the target comorbidity index reaches or exceeds this threshold, there is no need to extract the patient feature set and input it into the preset fall event prediction model for complex data extraction and calculation. This can not only ensure the reliability of the prediction results, but also avoid the occupation of excessive computing resources and ensure the speed of result output.

[0106] S103: Input the patient feature set into a preset fall event prediction model to output the fall risk level of the target elderly patient.

[0107] Specifically, a pre-trained fall event prediction model can be selected, which can receive a patient feature set as input and output the corresponding fall risk level. Among them, the training of the model is usually based on a large amount of fall event data and non-fall event data, using machine learning or deep learning algorithms. The constructed patient feature set is input into the preset fall event prediction model. After the model processes the input data, it outputs the fall risk level of the target elderly patient. This level is usually divided based on certain scoring criteria or threshold intervals, providing an intuitive reference for medical institutions and nursing staff.

[0108] In one possible design, the selected prediction model can be an integrated learning model based on the Gradient Boosting Decision Tree (GBDT). The GBDT model is suitable for dealing with complex problems such as fall risk prediction because of its powerful ability to handle nonlinear relationships, its ability to naturally handle missing values, and its robustness to outliers. The configuration of the model can be as follows:

[0109] Input layer: Receives a set of patient features as input, including target comorbidity index, physiological characteristic assessment parameters (such as muscle strength risk assessment index, balance ability risk assessment index, etc.) and spatial perception ability assessment parameters (such as direction discrimination ability risk assessment index, spatial positioning ability risk assessment index, etc.).

[0110] Hidden layer: Depending on the complexity of the specific problem, multiple hidden layers can be set. Each layer contains a certain number of decision trees, and the prediction results are gradually optimized by improving layer by layer.

[0111] Output layer: Outputs the fall risk level of the target elderly patient. The output layer usually uses a softmax activation function to convert the model's prediction results into a probability distribution and selects the category with the highest probability as the fall risk level.

[0112] The training process of the model is as follows:

[0113] Dataset division: A large amount of data on falls and non-falls of elderly patients is collected as training and test sets. The training set is used to train the model, and the test set is used to evaluate the model performance.

[0114] Feature engineering: preprocessing of the collected data, including data cleaning (handling missing values, outliers, etc.), feature selection (selecting features that have a significant impact on the risk of falls based on business understanding and statistical tests), and feature transformation (such as normalization, standardization, etc.).

[0115] Initialize the model: set the basic parameters of the model, such as learning rate, number of trees, tree depth, etc.

[0116] Iterative training: In each iteration, the model first calculates the residual between the current prediction result and the true label.

[0117] Based on the current residual, train a new decision tree to fit this part of the residual.

[0118] Add the newly trained decision tree to the model and update the model's predictions.

[0119] When the preset number of iterations is reached or the model's prediction performance on the validation set is no longer significantly improved, training is stopped.

[0120] Performance evaluation: Use the test set to evaluate the performance of the model. Common evaluation indicators include accuracy, recall, F1 score, etc. Based on the evaluation results, you can adjust the model parameters or reselect features to further optimize the model performance.

[0121] Then, the constructed patient feature set is organized in the format required by the model and input into the pre-trained GBDT model. The model performs complex operations and logical judgments on the input features through a layer-by-layer decision tree to comprehensively evaluate the impact of various physiological characteristics, spatial perception ability, and comorbid conditions on the risk of falls.

[0122] After the model processes the input data, it outputs the fall risk level of the target elderly patient. The level is divided based on the model's predicted probability and the preset scoring criteria or threshold range. For example, different threshold ranges can be set to correspond to different fall risk levels (such as low risk, medium risk, high risk, etc.), providing an intuitive reference for medical institutions and nursing staff.

[0123] In this embodiment, the health record data of the target elderly patient is obtained, and a target associated disease data set is determined from the health record data according to a preset fall-associated disease map, so as to determine a target comorbidity index according to the target associated disease data set. Then, only when the target comorbidity index is less than a preset index threshold, a patient feature set is extracted according to the health record data, and the patient feature set is input into a preset fall event prediction model to output the fall risk level of the target elderly patient, thereby realizing automatic prediction and evaluation of the fall risk level of the target elderly patient.

[0124] Figure 2 FIG. 1 is a flow chart of a method for predicting fall event data of elderly patients based on a comorbidity index according to another exemplary embodiment of the present application. Figure 2 As shown, the method for predicting fall event data of elderly patients based on comorbidity index provided in this embodiment includes:

[0125] S201. Obtain health record data of target elderly patients.

[0126] Specifically, the health record data of the target elderly patients can be collected through the electronic health record system or the database of the medical institution. Such data includes but is not limited to the basic information of the patient (such as age, gender, living habits, etc.), past medical history, current diagnosis results, treatment records and physical examination results.

[0127] S202: Add the nodes corresponding to the associated disease types existing in the health record data to the target node set.

[0128] After obtaining the health record data of the target elderly patient, each node in the corresponding node set in the preset fall-related disease map is traversed, and the nodes corresponding to the associated disease types existing in the health record data are added to the target node set.

[0129] The preset fall-related disease atlas is a network model that contains a variety of disease types related to fall risk and their associations. The atlas consists of a set of nodes and a set of edges. Each node represents an associated disease, and the edge represents the association between diseases. According to this atlas, the patient's health record data is traversed to identify all the disease types that exist in the atlas and use them as part of the target associated disease data set. At the same time, according to the preset weight coefficient for each disease node in the atlas, the weight value corresponding to each target associated disease is recorded.

[0130] In this step, each node in the corresponding node set in the preset fall-related disease map is traversed, and the nodes corresponding to the associated disease types in the health record data are added to the target node set P = {p1, p2, ..., p i ,…,p n}, where p i is the i-th target node in the target node set P, and n is the number of target nodes in the target node set P.

[0131] S203: Determine a target edge set according to the target node set and a corresponding edge set in a preset fall-related disease graph.

[0132] In this step, the target edge set L = {l1,l2,…,l j ,…,l m}, the target edge in the target edge set L is used to connect two target nodes in the target node set P, where l j is the j-th target edge in the target edge set L, and m is the number of target edges in the target edge set L.

[0133] S204: Determine a set of comorbidity impact paths according to the target node set and the target edge set.

[0134] In this step, the comorbidity impact path set S = {S 1 ,S 2 ,…,S k , …,S t}, where S k is the kth comorbidity impact path in the comorbidity impact path set S, t is the number of comorbidity impact paths in the comorbidity impact path set S, and the kth comorbidity impact path s k It includes at least two target nodes in the target node set P and target edges existing in the target edge set L for sequentially connecting adjacent target nodes.

[0135] In a possible implementation, the preset fall-related disease map also includes a hierarchical architecture, each layer in the hierarchical architecture is used to accommodate nodes in the node set, and each node in the node set is configured with a corresponding level in the hierarchical architecture, and the level of each node in the node set is negatively correlated with the corresponding point weight coefficient.

[0136] Specifically, the comorbidity impact path set S is determined based on the target node set P and the target edge set L = {S 1 ,S 2 ,…,S k ,…,S t}, which may include:

[0137] According to the hierarchical information of each node in the node set, the target node set P = p1, p2, ..., p i ,…,p n} are sorted in ascending order to generate the update target node set P' = {P'1, P'2, ..., P' g , …,P' h}, where P' g To update the target node subset consisting of the target nodes located at the g-th level of the hierarchical structure in the target node set P', where h is the number of levels of the hierarchical structure;

[0138] Traverse the target node subset P' g Each target node in the target edge set L is selected to determine the starting target node, the starting target node is one of the end nodes on the target edge in the starting target edge subset of the target edge set L, the starting target edge subset is a set of all target edges in the target edge set L including the starting target node, and the other end nodes on the target edges in the starting target edge subset have a level in the hierarchical architecture greater than the level in the hierarchical architecture corresponding to the starting target node;

[0139] Traverse the target node subset P' g Each target node in the target edge set L is selected to determine the termination target node, the termination target node is one of the end nodes on the target edge in the termination target edge subset of the target edge set L, the termination target edge subset is a set of all target edges in the target edge set L that include the termination target node, and the level corresponding to the other end node on the target edge in the termination target edge subset in the hierarchical architecture is less than or equal to the level corresponding to the termination target edge subset in the hierarchical architecture;

[0140] Generate a start target node set according to all start target nodes determined at each level of the hierarchical architecture, and generate a stop target node set according to all stop target nodes;

[0141] The comorbidity impact path set S is determined according to the starting target node set and the ending target node set, wherein the kth comorbidity impact path s in the comorbidity impact path set S is k The starting node of the corresponding target node sequence belongs to the starting target node set, the ending node belongs to the ending target node set, and the nodes in the target node sequence are connected in sequence through the target edges in the target edge set L.

[0142] S205. Determine a target comorbidity index according to the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set.

[0143] Specifically, according to the kth comorbidity impact path S in the comorbidity impact path set S k Determine the sequence of nodes in the pathway affected by comorbidity in, is the node sequence S of the comorbidity impact path kThe xth node in , y is the node sequence S of the co-morbidity impact path k The number of nodes in the

[0144] Using formula 1, the target comorbidity index C is determined according to the comorbidity impact path node sequence corresponding to each comorbidity impact path in the comorbidity impact path set S and the preset fall-related disease map. Formula 1 is:

[0145]

[0146] in, is the node sequence S of the comorbidity impact path k The xth node in The node weight coefficients preset in the preset fall-related disease map, is the node sequence S of the comorbidity impact path k The xth node in With the x-1th node The edge weight coefficients between the edges are preset in the preset fall-related disease graph.

[0147] To further illustrate, when x=3, When x=4, And so on.

[0148] It is worth noting that the above formula 1 realizes the quantitative assessment of the comorbidity of elderly patients, thereby calculating the target comorbidity index. This index is an important basis for assessing the risk of falls in elderly patients. Specifically, formula 1 achieves its purpose through the following technical effects:

[0149] Comprehensive consideration of the complex relationships between diseases: Formula 1 not only considers the weight of each associated disease (node), but also comprehensively considers the interactions and associations between diseases through edge weight coefficients. This processing method can fully capture the complex network relationships between diseases and ensure accurate assessment of comorbid conditions.

[0150] Quantifying the impact of paths: In the set of comorbidity impact paths, each path may have different degrees of impact on the risk of falls. Formula 1 quantifies the impact of these paths by accumulating the product of the node weight and edge weight on each path. This method can distinguish the contribution of different paths to the risk of falls and improve the accuracy of the assessment.

[0151] Support for dynamic assessment: As the health status of elderly patients may change over time and during treatment, their comorbidities may also change accordingly. Formula 1 supports dynamic calculation of the target comorbidity index based on the latest health record data, ensuring the timeliness and accuracy of the assessment results.

[0152] Improved prediction accuracy: By comprehensively considering the combined impact of multiple diseases and their interrelationships on fall risk, Formula 1 significantly improves the accuracy of fall risk assessment. Compared with single disease assessment or simple summation methods, it can more comprehensively reflect the actual risk status of elderly patients.

[0153] S206. Only when the target comorbidity index is less than a preset index threshold, extract a patient feature set based on the health record data.

[0154] In this step, only when the target comorbidity index is less than the preset index threshold, a patient feature set is extracted based on the health record data, and the patient feature set includes the target comorbidity index, physiological characteristic evaluation parameters, and spatial perception ability evaluation parameters of the target elderly patient.

[0155] Specifically, after calculating the target comorbidity index, it is first determined whether the index is less than a preset index threshold. This threshold is set based on a large amount of historical data and expert experience to distinguish between high and low fall risks.

[0156] If the target comorbidity index is less than the preset threshold, the patient's physiological characteristic assessment parameters and spatial perception ability assessment parameters are further extracted from the health record data. Physiological characteristic parameters may include muscle strength risk assessment index, balance ability risk assessment index, hearing risk assessment index and vision risk assessment index, etc.; spatial perception ability parameters may include direction discrimination ability risk assessment index, spatial positioning ability risk assessment index and depth perception risk assessment index, etc. It is worth noting that the above-mentioned physiological characteristic assessment parameters and spatial perception ability assessment parameters can be obtained through physical examination or related tests.

[0157] Then, the target comorbidity index, the extracted physiological characteristic assessment parameters, and the spatial perception ability assessment parameters were integrated into a patient feature set for subsequent fall risk assessment.

[0158] Further, if the target comorbidity index is greater than or equal to a preset index threshold, the fall risk level of the target elderly patient is determined to be a first risk level, wherein the first risk level is the highest risk level among the fall risk levels.

[0159] In the above scheme, a clear preset index threshold is set to quickly determine the fall risk level of elderly patients. When the target comorbidity index reaches or exceeds this threshold, the fall risk level is directly determined to be the highest level (first risk level), which greatly simplifies the risk assessment process and enables high-risk patients to be quickly identified.

[0160] Moreover, when the target comorbidity index reaches or exceeds this threshold, there is no need to extract the patient feature set and input it into the preset fall event prediction model for complex data extraction and calculation. This can not only ensure the reliability of the prediction results, but also avoid the occupation of excessive computing resources and ensure the speed of result output.

[0161] S207: Input the patient feature set into a preset fall event prediction model to output the fall risk level of the target elderly patient.

[0162] In a possible implementation, Formula 2 may be used to determine the fall risk assessment score F according to a patient feature set. Formula 2 is:

[0163]

[0164] Among them, γ i is the preset weight value corresponding to the i-th patient feature in the patient feature set, v i is the evaluation index corresponding to the i-th patient feature in the patient feature set, and u is the total number of patient features in the patient feature set;

[0165] The fall risk level of the target elderly patient is output according to the fall risk assessment score F and the preset fall risk assessment threshold interval distribution.

[0166] It is worth noting that Formula 2 comprehensively considers multiple characteristic factors in the patient feature set (such as comorbidity index, physiological characteristic evaluation parameters, spatial perception ability evaluation parameters, etc.) and their impact on the risk of falling through weighted summation. Each characteristic factor is assigned a different preset weight value according to its importance, and is multiplied with the corresponding evaluation index and summed up, thereby achieving a comprehensive and quantitative assessment of the patient's fall risk. Based on the calculated fall risk assessment score, the score can be converted into a specific fall risk level through the preset fall risk assessment threshold interval distribution, thereby providing a clear reference for medical institutions, nursing staff and patients' families, helping them to formulate corresponding preventive measures and intervention plans according to the risk level.

[0167] Figure 3 FIG. 1 is a schematic diagram of a device for predicting fall events in elderly patients based on a comorbidity index according to an exemplary embodiment of the present application. Figure 3 As shown, the device 300 for predicting fall event data of elderly patients based on comorbidity index provided in this embodiment includes:

[0168] The acquisition module 310 is used to acquire the health record data of the target elderly patient, and determine a target-related disease data set from the health record data according to a preset fall-related disease atlas, wherein the target-related disease data in the target-related disease data set includes target-related disease types and weight coefficients corresponding to the target-related disease types;

[0169] A processing module 320 is used to determine a target comorbidity index according to the target associated disease data set,

[0170] An extraction module 330, configured to extract a patient feature set according to the health record data only when the target comorbidity index is less than a preset index threshold, the patient feature set including the target comorbidity index, physiological feature evaluation parameters, and spatial perception ability evaluation parameters of the target elderly patient;

[0171] The processing module 320 is further configured to input the patient feature set into a preset fall event prediction model to output a fall risk level of the target elderly patient.

[0172] Optionally, the processing module 320 is further used to determine that the fall risk level of the target elderly patient is a first risk level, wherein the first risk level is the highest risk level among the fall risk levels.

[0173] Optionally, the preset fall-associated disease graph includes a node set and an edge set, each node in the node set is used to represent a corresponding associated disease type, and each edge in the edge set is used to connect two different nodes to represent the association relationship between the associated disease types corresponding to the two connected nodes; the processing module 320 is specifically used to:

[0174] Traversing each node in the node set corresponding to the preset fall-related disease map, and adding the nodes corresponding to the associated disease types existing in the health record data to the target node set;

[0175] Determine a target edge set according to the target node set and the corresponding edge set in the preset fall-related disease atlas, wherein a target edge in the target edge set is used to connect two target nodes in the target node set;

[0176] Determine a comorbidity impact path set according to the target node set and the target edge set;

[0177] Determine the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set according to the preset node weight coefficient and the preset edge weight coefficient of each node in the preset fall-related disease graph;

[0178] The target comorbidity index is determined according to the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set.

[0179] Optionally, the preset fall-related disease atlas further includes a hierarchical structure, each layer in the hierarchical structure is used to accommodate nodes in the node set, each node in the node set is configured with a corresponding level in the hierarchical structure, and the level of each node in the node set is negatively correlated with the corresponding point weight coefficient; the processing module 320 is specifically used to:

[0180] Sorting the target node set in ascending order according to the level information of each node in the node set to generate an updated target node set, wherein the updated target node set includes a subset of target nodes corresponding to each level of the hierarchical architecture;

[0181] Traversing each target node in the target node subset to determine a starting target node, the starting target node being one of the end nodes on the target edge in the starting target edge subset of the target edge set, the starting target edge subset being a set consisting of all target edges in the target edge set that include the starting target node, and the other end nodes on the target edges in the starting target edge subset corresponding to a level in the hierarchical architecture that is greater than the level corresponding to the starting target node in the hierarchical architecture;

[0182] Traversing each target node in the target node subset to determine a termination target node, the termination target node being one of the end nodes on the target edge in the termination target edge subset of the target edge set, the termination target edge subset being a set consisting of all target edges in the target edge set L that include the termination target node, and the levels corresponding to the other end nodes on the target edges in the termination target edge subset in the hierarchical architecture are all less than or equal to the levels corresponding to the termination target edge subset in the hierarchical architecture;

[0183] Generate a starting target node set according to all starting target nodes determined at each level of the hierarchical architecture, and generate a terminating target node set according to all terminating target nodes;

[0184] The comorbidity impact path set is determined according to the starting target node set and the terminating target node set, wherein the starting node of the target node sequence corresponding to the comorbidity impact path in the comorbidity impact path set belongs to the starting target node set, the terminating node belongs to the terminating target node set, and the nodes in the target node sequence are connected in sequence through the target edges in the target edge set.

[0185] Optionally, the processing module 320 is specifically configured to:

[0186] Determine a corresponding comorbidity impact path node sequence according to each comorbidity impact path in the comorbidity impact path set;

[0187] The target comorbidity index is determined according to the comorbidity influence path node sequence corresponding to each comorbidity influence path in the comorbidity influence path set and the preset node weight coefficient and the preset edge weight coefficient of each node in the preset fall-related disease map.

[0188] Optionally, the physiological characteristic assessment parameter includes at least one of a muscle strength risk assessment index, a balance ability risk assessment index, a hearing risk assessment index, and a vision risk assessment index;

[0189] The spatial perception capability assessment parameter includes at least one of a direction discrimination capability risk assessment index, a spatial positioning capability risk assessment index, and a depth perception risk assessment index.

[0190] Optionally, the processing module 320 is specifically configured to:

[0191] Determining the fall risk assessment score according to the patient feature set;

[0192] The fall risk level of the target elderly patient is output according to the fall risk assessment score and a preset fall risk assessment threshold interval distribution.

[0193] Figure 4 is a schematic diagram of the structure of an electronic device according to an exemplary embodiment of the present application. Figure 4 As shown, an electronic device 400 provided in this embodiment includes: a processor 401 and a memory 402; wherein:

[0194] The memory 402 is used to store computer programs, and the memory may also be a flash memory.

[0195] The processor 401 is used to execute the execution instructions stored in the memory to implement each step in the above method. For details, please refer to the relevant description in the above method embodiment.

[0196] Optionally, the memory 402 may be independent or integrated with the processor 401 .

[0197] When the memory 402 is a device independent of the processor 401, the electronic device 400 may further include:

[0198] The bus 403 is used to connect the memory 402 and the processor 401 .

[0199] This embodiment further provides a readable storage medium, in which a computer program is stored. When at least one processor of an electronic device executes the computer program, the electronic device executes the methods provided in the above-mentioned various implementation modes.

[0200] This embodiment also provides a program product, which includes a computer program stored in a readable storage medium. At least one processor of the electronic device can read the computer program from the readable storage medium, and at least one processor executes the computer program so that the electronic device implements the methods provided in the above various embodiments.

[0201] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the claims.

[0202] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for predicting fall events in elderly patients based on comorbidity index, characterized in that: include: Acquire health record data of a target elderly patient, and determine a target associated disease data set from the health record data according to a preset fall-associated disease atlas, so as to determine a target comorbidity index according to the target associated disease data set, wherein the target associated disease data in the target associated disease data set include target associated disease types and weight coefficients corresponding to the target associated disease types; Only when the target comorbidity index is less than a preset index threshold, extracting a patient feature set according to the health record data, the patient feature set including the target comorbidity index, physiological feature evaluation parameters, and spatial perception ability evaluation parameters of the target elderly patient; The patient feature set is input into a preset fall event prediction model to output the fall risk level of the target elderly patient.

2. The method for predicting fall events in elderly patients based on comorbidity index according to claim 1, characterized in that: After determining the target comorbidity index according to the target-associated disease data set, the method further includes: If the target comorbidity index is greater than or equal to the preset index threshold, the fall risk level of the target elderly patient is determined to be a first risk level, wherein the first risk level is the highest risk level among the fall risk levels.

3. The method for predicting fall events in elderly patients based on comorbidity index according to claim 2, characterized in that: The preset fall-associated disease graph includes a node set and an edge set, each node in the node set is used to represent a corresponding associated disease type, and each edge in the edge set is used to connect two different nodes, and is used to represent the association relationship between the associated disease types corresponding to the two connected nodes; correspondingly, determining a target associated disease data set from the health record data according to the preset fall-associated disease graph, so as to determine a target comorbidity index according to the target associated disease data set, includes: Traversing each node in the node set corresponding to the preset fall-related disease map, and adding the nodes corresponding to the associated disease types existing in the health record data to the target node set; Determine a target edge set according to the target node set and the corresponding edge set in the preset fall-related disease atlas, wherein a target edge in the target edge set is used to connect two target nodes in the target node set; Determine a comorbidity impact path set according to the target node set and the target edge set; Determine the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set according to the preset node weight coefficient and the preset edge weight coefficient of each node in the preset fall-related disease graph; The target comorbidity index is determined according to the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set.

4. The method for predicting fall events in elderly patients based on comorbidity index according to claim 3, characterized in that: The preset fall-related disease atlas also includes a hierarchical structure, each layer in the hierarchical structure is used to accommodate the nodes in the node set, each node in the node set is configured with a corresponding level in the hierarchical structure, and the level of each node in the node set is negatively correlated with the corresponding point weight coefficient; Correspondingly, determining the comorbidity impact path set according to the target node set P and the target edge set includes: Sorting the target node set in ascending order according to the level information of each node in the node set to generate an updated target node set, wherein the updated target node set includes a subset of target nodes corresponding to each level of the hierarchical architecture; Traversing each target node in the target node subset to determine a starting target node, the starting target node being one of the end nodes on the target edge in the starting target edge subset of the target edge set, the starting target edge subset being a set consisting of all target edges in the target edge set that include the starting target node, and the other end nodes on the target edges in the starting target edge subset corresponding to a level in the hierarchical architecture that is greater than the level corresponding to the starting target node in the hierarchical architecture; Traversing each target node in the target node subset to determine a termination target node, the termination target node being one of the end nodes on the target edge in the termination target edge subset of the target edge set, the termination target edge subset being a set consisting of all target edges in the target edge set L that include the termination target node, and the levels corresponding to the other end nodes on the target edges in the termination target edge subset in the hierarchical architecture are all less than or equal to the levels corresponding to the termination target edge subset in the hierarchical architecture; Generate a starting target node set according to all starting target nodes determined at each level of the hierarchical architecture, and generate a terminating target node set according to all terminating target nodes; The comorbidity impact path set is determined according to the starting target node set and the terminating target node set, wherein the starting node of the target node sequence corresponding to the comorbidity impact path in the comorbidity impact path set belongs to the starting target node set, the terminating node belongs to the terminating target node set, and the nodes in the target node sequence are connected in sequence through the target edges in the target edge set.

5. The method for predicting fall events in elderly patients based on comorbidity index according to claim 4, characterized in that: The determining the target comorbidity index according to the path comorbidity impact index corresponding to each comorbidity impact path in the comorbidity impact path set S includes: Determine a corresponding comorbidity impact path node sequence according to each comorbidity impact path in the comorbidity impact path set; The target comorbidity index is determined according to the comorbidity influence path node sequence corresponding to each comorbidity influence path in the comorbidity influence path set and the preset node weight coefficient and the preset edge weight coefficient of each node in the preset fall-related disease map.

6. The method for predicting fall events in elderly patients based on comorbidity index according to any one of claims 1 to 5, characterized in that: The physiological characteristic assessment parameter includes at least one of a muscle strength risk assessment index, a balance ability risk assessment index, a hearing risk assessment index, and a vision risk assessment index; The spatial perception capability assessment parameter includes at least one of a direction discrimination capability risk assessment index, a spatial positioning capability risk assessment index, and a depth perception risk assessment index.

7. The method for predicting fall events in elderly patients based on comorbidity index according to claim 6, characterized in that: The step of inputting the patient feature set into a preset fall event prediction model to output the fall risk level of the target elderly patient includes: Determining the fall risk assessment score according to the patient feature set; The fall risk level of the target elderly patient is output according to the fall risk assessment score and a preset fall risk assessment threshold interval distribution.

8. A device for predicting fall events in elderly patients based on comorbidity index, characterized in that: include: An acquisition module is used to acquire health record data of a target elderly patient, and determine a target-associated disease data set from the health record data according to a preset fall-associated disease atlas, wherein the target-associated disease data in the target-associated disease data set include target-associated disease types and weight coefficients corresponding to the target-associated disease types; a processing module, configured to determine a target comorbidity index according to the target associated disease data set, An extraction module, configured to extract a patient feature set according to the health record data only when the target comorbidity index is less than a preset index threshold, the patient feature set including the target comorbidity index, physiological feature evaluation parameters, and spatial perception ability evaluation parameters of the target elderly patient; The processing module is further used to input the patient feature set into a preset fall event prediction model to output the fall risk level of the target elderly patient.

9. An electronic device, characterized in that: include: processor; as well as, A memory, configured to store executable instructions of the processor; The processor is configured to perform the method of any one of claims 1 to 7 by executing the executable instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

Images