Pre-examination triage method and device, equipment and storage medium

Abstract

The invention discloses a pre-examination triage method, device and equipment and a storage medium, and belongs to the technical field of data analysis. The method comprises the following steps: acquiring chronic disease management data and current vital sign data of a patient; based on a linear weighting algorithm and the chronic disease management data, a chronic disease management score is obtained through calculation; and inputting the chronic disease management score and the current vital sign data into a triage grade prediction model to obtain a first triage grade of the patient. According to the method, the particularity of the vital signs of the chronic disease patient can be fully considered, misjudgment caused by a single vital sign is avoided, and the triage accuracy is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a pre-examination and triage method, device, equipment and storage medium. Background Art

[0002] With the aging of the population, elderly patients have gradually become a common population in the emergency department, and they are often accompanied by multiple chronic diseases. Due to the special health conditions of the elderly, they are easily misdiagnosed as critically ill patients during emergency triage, resulting in excessive intervention and waste of resources. Traditional triage methods are based only on single vital signs and symptoms, ignoring the impact of chronic disease management and multi-dimensional health data on triage accuracy, which leads to poor triage accuracy. Summary of the invention

[0003] The main purpose of the present invention is to provide a pre-examination triage method, device, equipment and storage medium, aiming to solve the technical problems of poor triage accuracy and easy misjudgment of traditional triage methods.

[0004] To achieve the above object, the present invention provides a triage pre-examination method, which comprises the following steps:

[0005] S1, obtain the patient's chronic disease management data and current vital signs data;

[0006] S2, based on the linear weighted algorithm and chronic disease management data, the chronic disease management score is calculated;

[0007] S3, input the chronic disease management score and current vital signs data into the triage level prediction model to obtain the patient's first triage level.

[0008] Optionally, the chronic disease management data includes the chronic disease type of the patient and the chronic disease characteristic value corresponding to the chronic disease type. Step S2 specifically includes:

[0009] Based on the patient's chronic disease type, query the preset weight coefficient table to determine the weight coefficient corresponding to the chronic disease characteristic value; the preset weight coefficient table has weight coefficients corresponding to different chronic disease characteristic values, and the weight coefficients are obtained based on one or more methods of regression analysis and machine learning;

[0010] Based on the weight coefficient, the chronic disease characteristic values ​​are weighted and summed to obtain the chronic disease management score.

[0011] Optionally, in step S3, the steps of constructing the triage level prediction model specifically include:

[0012] Obtain historical vital signs data and historical chronic disease management data of patients with multiple different types of chronic diseases;

[0013] Based on statistical methods, historical vital signs data are subjected to fluctuation analysis to obtain historical vital signs fluctuation parameters;

[0014] Based on the linear weighted algorithm and historical chronic disease management data, the historical chronic disease management score is calculated;

[0015] The Z-score standardization method was used to preprocess the historical chronic disease management scores and historical vital sign fluctuation parameters to obtain the first training set with unified dimensions;

[0016] Building a feature extraction layer based on a multi-layer perceptron model to extract features from the first training set;

[0017] Based on the self-attention mechanism, the features are weighted to obtain a weighted feature matrix; the attention weight calculation formula is:

[0018] ;

[0019] in, is the attention weight matrix, Q is the query matrix, K is the key matrix, V is the value matrix, d k is the dimension of the key vector;

[0020] Calculate the difference between the predicted triage grade probability and the actual triage grade, and construct the cross entropy loss function:

[0021] ;

[0022] in, is the loss value, is the actual triage level, To predict the probability of triage level, is the number of triage level categories;

[0023] Based on the cross entropy loss function and the weighted feature matrix, the deep learning model was trained to obtain the triage level prediction model.

[0024] Optionally, after step S3, the method further includes:

[0025] S4, obtain the patient's chronic disease control status, emotional state, and activity level;

[0026] S5, based on the preset quantitative standards, quantify the current vital signs, chronic disease control, emotional state and activity level respectively to obtain the vital signs score, chronic disease control score, emotional state score and activity level score;

[0027] S6, inputting the vital sign score, chronic disease control score, emotional state score, and activity level score into the score weight calculation model to obtain a dynamic score weight coefficient;

[0028] S7, based on the vital signs score, chronic disease control score, emotional state score, activity score and dynamic score weight coefficient, a weighted average calculation is performed to obtain a comprehensive score;

[0029] S8, adjusting the first triage level based on the comprehensive score to obtain a second triage level.

[0030] Optionally, in step S6, the step of constructing the scoring weight calculation model specifically includes:

[0031] Acquire historical vital sign data, historical emotional state data, historical chronic disease control data, historical activity data, and historical triage levels of multiple patients with different types of chronic diseases and perform preprocessing to obtain first preprocessing data;

[0032] Using L1 regularization, features related to the historical triage level were selected from the first preprocessed data to construct the second training set;

[0033] Based on the second training set and the random forest algorithm, model training is performed to obtain a scoring weight calculation model.

[0034] In addition, to achieve the above-mentioned purpose, the present invention also provides a pre-diagnosis and triage device, the device comprising:

[0035] Data acquisition module, used to obtain patients' chronic disease management data and current vital signs data;

[0036] Chronic disease analysis module, which is used to calculate the chronic disease management score based on the linear weighted algorithm and chronic disease management data;

[0037] The triage calculation module is used to input the chronic disease management score and current vital signs data into the triage level prediction model to obtain the patient's first triage level.

[0038] In addition, to achieve the above-mentioned purpose, the present invention also provides a pre-examination and triage device, the device comprising: a memory, a processor, and a pre-examination and triage program stored in the memory and executable on the processor, the pre-examination and triage program being configured to implement the steps of the pre-examination and triage method as described above.

[0039] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which a pre-examination and triage program is stored, and when the pre-examination and triage program is executed by a processor, the steps of the pre-examination and triage method as described above are implemented.

[0040] The present invention obtains the patient's chronic disease management data and current vital signs data to comprehensively evaluate the patient's health status. During the evaluation process, the chronic disease management data is quantitatively analyzed using a linear weighted algorithm to obtain a chronic disease management score that can accurately reflect the impact of the patient's chronic disease type on triage. The triage level prediction model is then used to comprehensively analyze the impact of the chronic disease management score and the current vital signs data on the triage level. This allows full consideration of the particularity of the vital signs of chronic disease patients, avoids misjudgment caused by a single vital sign, and effectively improves the accuracy of triage. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a structural diagram of a pre-diagnosis and triage device in a hardware operating environment involved in an embodiment of the present invention;

[0042] Figure 2 It is a schematic diagram of the flow chart of the first embodiment of the pre-examination and triage method of the present invention;

[0043] Figure 3 Schematic diagram of the flow chart of the second embodiment of the pre-examination and triage method of the present invention.

[0044] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0045] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0046] The inventive concept of the present application is further described below in conjunction with some specific embodiments and implementation methods.

[0047] Reference Figure 1 , Figure 1 The present invention is a schematic diagram of the structure of a pre-diagnosis and triage device in a hardware operating environment according to an embodiment of the present invention.

[0048] like Figure 1As shown, the pre-examination and triage device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (WIreless-FIdelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) memory, or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0049] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the pre-examination triage device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0050] like Figure 1 As shown, the memory 1005 as a computer-readable storage medium may include an operating system, a data storage module, a network communication module, a user interface module, and a pre-examination and triage program.

[0051] exist Figure 1 In the pre-examination and triage device shown, the network interface 1004 is mainly used for data communication with other devices; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the pre-examination and triage device of the present invention can be set in the pre-examination and triage device, and the pre-examination and triage device calls the pre-examination and triage program stored in the memory 1005 through the processor 1001, and executes the pre-examination and triage method provided by the embodiment of the present invention.

[0052] The present invention provides a pre-diagnosis method. Figure 2 , Figure 2 The figure is a flow chart of a first embodiment of a pre-examination and triage method of the present invention.

[0053] In this embodiment, the pre-examination and triage method includes:

[0054] Step S1: Obtain the patient's chronic disease management data and current vital signs data.

[0055] Specifically, in this embodiment, historical health data and actual historical triage levels can be obtained through the hospital's electronic medical record system. At the same time, the patient's current vital signs data, including but not limited to blood pressure, pulse, oxygen saturation, body temperature and respiratory rate, are obtained in real time through smart wearable devices.

[0056] Step S2: Based on the linear weighted algorithm and chronic disease management data, the chronic disease management score is calculated.

[0057] In the specific implementation process, the chronic disease management data includes the chronic disease type of the patient and the chronic disease characteristic value corresponding to the chronic disease type. Therefore, step S2 specifically includes:

[0058] Step S21: Based on the patient's chronic disease type, query the preset weight coefficient table to determine the weight coefficient corresponding to the chronic disease characteristic value.

[0059] Among them, the preset weight coefficient table has weight coefficients corresponding to different chronic disease characteristic values, and the weight coefficients are obtained based on one or more methods of regression analysis and machine learning.

[0060] Step S22: Based on the weight coefficient, the chronic disease characteristic values ​​are weighted and summed to obtain the chronic disease management score.

[0061] Specifically, chronic disease characteristic values ​​refer to the key indicators corresponding to different types of chronic diseases in chronic disease management. For example, the characteristic values ​​of diabetic patients include blood sugar level, body mass index, glycosylated hemoglobin and other indicators; while the characteristic values ​​of hypertensive patients include blood pressure value, heart rate and other indicators.

[0062] In the specific implementation process, the weight coefficient can be determined by regression analysis based on a large amount of historical health data or by applying machine learning methods, and then a preset weight coefficient table is constructed. When determining the weight coefficient, the chronic disease characteristic value corresponding to the patient's chronic disease type is used as the search condition, and the corresponding weight coefficient is obtained from the preset weight coefficient table. Finally, the chronic disease management score is calculated through a linear weighted algorithm.

[0063] Step S3: Input the chronic disease management score and current vital signs data into the triage level prediction model to obtain the patient's first triage level.

[0064] Specifically, the patient's chronic disease management score and current vital signs data are first input into the triage level prediction model, the probability values ​​of different triage levels are calculated, and the triage level with the largest probability value is selected as the patient's first triage level.

[0065] Furthermore, in the specific implementation process, in step S3, the steps of constructing the triage level prediction model specifically include:

[0066] Step S31: Obtain historical vital signs data and historical chronic disease management data of multiple patients with different chronic disease types.

[0067] Step S32: Based on statistical methods, perform fluctuation analysis on historical vital sign data to obtain historical vital sign fluctuation parameters.

[0068] The statistical method can be specifically used to calculate the standard deviation or the coefficient of variation. Specifically, for the blood pressure data in the historical vital signs, the standard deviation or the fluctuation range can be calculated as the fluctuation parameter; for the heart rate data in the historical vital signs, the standard deviation or the coefficient of variation can be calculated as the fluctuation parameter; for the blood oxygen saturation data, the fluctuation range can be calculated as the fluctuation parameter.

[0069] Step S33: Based on the linear weighted algorithm and historical chronic disease management data, the historical chronic disease management score is calculated.

[0070] Specifically, first identify the chronic disease types corresponding to different patients based on historical chronic disease management data. Then, for each patient, use the chronic disease characteristic value corresponding to the chronic disease type as the search condition, and query the corresponding weight coefficient from the preset weight coefficient table. Finally, calculate the historical chronic disease management score through a linear weighted algorithm.

[0071] Step S34: Use the Z-score standardization method to pre-process the historical chronic disease management scores and historical vital sign fluctuation parameters to obtain a first training set with unified dimensions.

[0072] Specifically, we first calculate the mean and standard deviation of the historical chronic disease management score and the historical vital signs fluctuation parameters, and then standardize each data point in the historical chronic disease data and historical vital signs according to the Z-score formula.

[0073] For each data point in the historical chronic disease management score To standardize:

[0074] ;

[0075] in, is the chronic disease management score of the i-th patient, is the mean of historical chronic disease management scores, is the standard deviation of historical chronic disease management scores, For data points The standardized results.

[0076] Normalize each data point Yi in the historical vital sign fluctuation parameters:

[0077] ;

[0078] in, is the historical vital sign fluctuation parameter of the i-th patient, is the mean of historical vital sign fluctuation parameters, is the standard deviation of historical vital sign fluctuation parameters, For data points The standardized results.

[0079] Step S35: construct a feature extraction layer based on the multi-layer perceptron model, extract features from the first training set, and obtain a feature matrix.

[0080] Specifically, the preprocessed historical chronic disease management data and historical vital sign fluctuation parameters in the first training set are input into the trained multi-layer perceptron model to obtain the activation output of each sample in the first training set when it passes through the hidden layer. The multi-layer perceptron model has multiple hidden layers, each of which extracts features at different levels. In order to avoid a large amount of information loss caused by deep feature extraction, the middle hidden output layer is selected as the output feature to obtain the feature matrix.

[0081] Step S36: Perform weighted processing on the features based on the self-attention mechanism to obtain a weighted feature matrix.

[0082] Among them, the attention weight calculation formula is:

[0083] ;

[0084] in, is the attention weight matrix, Q is the query matrix, K is the key matrix, V is the value matrix, d k is the dimension of the key vector.

[0085] Furthermore, the query matrix , the bond matrix , value matrix .

[0086] Wherein, X is the input feature matrix (N, D), which is extracted in step S35; N is the number of samples in the feature matrix; D is the feature dimension of each sample in the feature matrix; is the query parameter matrix, is the key parameter matrix, is the value parameter matrix, all of which are obtained through deep learning.

[0087] Finally, the attention weight matrix Multiply it with the value matrix V to obtain the weighted feature matrix.

[0088] Step S37: Calculate the difference between the predicted triage level probability and the actual triage level, and construct a cross entropy loss function:

[0089] ;

[0090] in, is the loss value, is the actual triage level, To predict the probability of triage level, is the triage level category number (level I, level II, level III, level IV, level V).

[0091] Step S38: Based on the cross entropy loss function and the weighted feature matrix, the deep learning model is trained to obtain a triage level prediction model.

[0092] Specifically, the cross entropy loss function is used to calculate the difference between the predicted probability and the actual triage level to obtain the loss value. Then, the back propagation algorithm is used to calculate the gradient of the loss function for the query, key, and value respectively, and the result is:

[0093] The gradient of the loss function with respect to the query matrix: ;

[0094] Gradient of the loss function with respect to the key matrix: ;

[0095] The gradient of the loss function with respect to the value matrix: .

[0096] Then, by using the chain rule, we can calculate the loss function for W Q , W K , W V The gradient is:

[0097] Gradient calculation formula:

[0098] ;

[0099] ;

[0100] ;

[0101] in, is the loss function The gradient of is the loss function The gradient of is the loss function The gradient of is the transposed matrix of the feature matrix X.

[0102] On this basis, the gradient descent algorithm is used according to the learning rate Update the parameter matrix:

[0103] Update formula:

[0104] ;

[0105] ;

[0106] ;

[0107] in, is the learning rate (used to control the magnitude of each update), is the updated query parameter matrix, is the updated key parameter matrix, is the updated value parameter matrix.

[0108] Finally, after multiple iterations, the triage level prediction model was trained.

[0109] It is worth mentioning that in the process of triage level prediction, some symptoms may have a significant impact on the triage level, while other irrelevant features may have little impact on the prediction results. The self-attention mechanism calculates the similarity between features and weights them, so that important symptoms or signs can receive higher attention, improving the prediction ability of the model. At the same time, the gradient descent algorithm is combined to continuously update the query parameter matrix, key parameter matrix, and value parameter matrix involved in the attention weight calculation formula, so that the model can gradually learn the optimal attention weight on the training set according to the feedback of the loss function, and thus can more accurately predict the triage level.

[0110] In this embodiment, the present invention obtains the patient's chronic disease management data and current vital signs data to comprehensively evaluate the patient's health status. During the evaluation process, the chronic disease management data is quantitatively analyzed using a linear weighted algorithm to obtain a chronic disease management score that can accurately reflect the impact of the patient's chronic disease type on triage. The triage level prediction model is then used to comprehensively analyze the impact of the chronic disease management score and the current vital signs data on the triage level. This allows full consideration of the particularity of the vital signs of chronic disease patients, avoids misjudgment caused by a single vital sign, and effectively improves the accuracy of triage.

[0111] Further, based on the first embodiment, a second embodiment is proposed, referring to Figure 3 , Figure 3 This is a flow chart of the second embodiment of the pre-examination and triage method of the present invention.

[0112] In some embodiments, factors such as chronic disease control, emotional state, and activity level can be taken into consideration to comprehensively analyze the patient's triage level. Therefore, after step S3, the method further includes:

[0113] Step S4: Obtain the patient's chronic disease control status, emotional state and activity level.

[0114] Step S5: Based on the preset quantitative standards, the current vital signs, chronic disease control status, emotional state and activity level are quantified respectively to obtain vital signs scores, chronic disease control scores, emotional state scores and activity level scores.

[0115] Specifically, first, the patient's chronic disease control status (for example, whether diabetes, hypertension, heart disease, etc. are effectively controlled) is obtained through the hospital's electronic medical record system; the patient's emotional state (for example, anxiety, depression, stress and other emotional problems) is obtained through psychological assessment tools, questionnaires or communication with patients; the patient's activity level (for example, amount of exercise, number of steps, exercise intensity, etc.) is obtained through the patient's smart wearable devices.

[0116] Furthermore, the acquired information such as current vital signs, chronic disease control, emotional state, and activity level are scored and quantified. Specifically, the preset quantification standards may include:

[0117] Vital signs scoring criteria:

[0118] Blood pressure score:

[0119] Systolic blood pressure (SBP): <90 mmHg, score 10; 90-139 mmHg, score 5; ≥140 mmHg, score 1.

[0120] Diastolic blood pressure (DBP): <60 mmHg, score 10; 60-89 mmHg, score 5; ≥90 mmHg, score 1.

[0121] Heart rate rating:

[0122] Heart rate (HR) <50 or >130 beats / minute, score 10; heart rate 50-100 beats / minute, score 5; heart rate 100-130 beats / minute, score 1.

[0123] Blood oxygen saturation score:

[0124] Less than 90%, score 10; between 90%-95%, score 5; greater than or equal to 95%, score 1.

[0125] Then, the obtained scores of each indicator are substituted into Formula 4 to calculate the final vital sign score. Formula 4 is:

[0126] ;

[0127] in, For systolic blood pressure score, The diastolic blood pressure score is Score your heart rate. Score for blood oxygen saturation. is the weight coefficient of systolic blood pressure, is the weight coefficient of diastolic pressure, is the weight coefficient of heart rate, is the weight coefficient of blood oxygen saturation, Score vital signs.

[0128] Chronic disease control scoring criteria:

[0129] Hypertension score: If blood pressure is higher than 140 / 90 mmHg for a long time, the score is 10; if blood pressure fluctuates greatly, the score is 5; if blood pressure is within the normal range, the score is 1.

[0130] Diabetes score: If fasting blood glucose is ≥8 mmol / L, the score is 10; if blood glucose fluctuates greatly, the score is 5; if fasting blood glucose is <6 mmol / L, the score is 1.

[0131] Other chronic diseases can also be assessed using similar scoring methods.

[0132] Then, the obtained chronic disease scores are calculated by using formula 5 to obtain the chronic disease control score. Formula 5 is:

[0133] ;

[0134] in, Score for chronic disease control, is the hypertension score weight, is the diabetes score weight, Score weights for other chronic diseases.

[0135] Emotional state scoring criteria: if anxiety or depression is severe (such as SADQ score ≥ 18), the score is 10; if anxiety or depression is mild (SADQ score 10-17), the score is 5; if the mood is stable (SADQ score <10), the score is 1.

[0136] Activity volume scoring criteria: If the number of steps per day is less than 1,000, the score is 10; if the number of steps per day is between 1,000 and 5,000, the score is 5; if the number of steps per day is more than 5,000, the score is 1.

[0137] Then, according to the above standards, calculate the scores corresponding to different activity intensities in the last few days, and then substitute the scores into Formula 6 to calculate the activity score. Formula 6 is:

[0138] ;

[0139] in, Rate the activity level. Weighting for very low activity scores, The weight of the moderate activity score is Score weight for higher activity levels.

[0140] Step S6: Input the vital sign score, chronic disease control score, emotional state score and activity level score into the score weight calculation model to obtain a dynamic score weight coefficient.

[0141] Step S7: Based on the vital signs score, chronic disease control score, emotional state score, activity score and dynamic score weight coefficient, a weighted average calculation is performed to obtain a comprehensive score.

[0142] Specifically, the weighted average calculation is performed on the scores of each dimension obtained above to obtain the final comprehensive score. Among them, the weighted average formula 7 is:

[0143] ;

[0144] in, For the comprehensive rating, Score vital signs, Score for chronic disease control, Rate your emotional state, Rate the activity level. is the weight of the vital sign score, is the weight of chronic disease control score, Score weights for emotional states, Activity score weight. The dynamic score weight coefficients include vital signs score weight, chronic disease control score weight, emotional state score weight and activity score weight.

[0145] In some specific real-time processes, the steps for constructing the scoring weight calculation model include:

[0146] Step S71: acquiring historical vital sign data, historical emotional state data, historical chronic disease control data, historical activity data, and historical triage levels of multiple patients with different chronic diseases and performing preprocessing to obtain first preprocessed data;

[0147] Step S72: using L1 regularization, filtering out features related to the historical triage level from the first preprocessed data, and constructing a second training set;

[0148] Step S73: Based on the second training set and the random forest algorithm, model training is performed to obtain a scoring weight calculation model.

[0149] Specifically, the acquired data is preprocessed, including removing invalid data, filling missing values, and standardizing all indicators to ensure that the data formats of different dimensions are consistent.

[0150] Next, the features related to triage decision-making were extracted, and the key features were screened again using L1 regularization to eliminate redundant or irrelevant features, retain features that had a significant impact on the triage level, and construct the second training set.

[0151] Then, the random forest model is trained based on the second training set to obtain the scoring weight calculation model. On this basis, the scoring weight calculation model can be used to calculate the contribution of each feature to the triage level according to the input patient real-time data, as the scoring weight coefficient of each feature. In this way, the weight coefficients of each dimension can be automatically adjusted according to the health status of patients with different types of chronic diseases to ensure the accuracy of the comprehensive score, making the comprehensive scoring system more personalized and adaptable.

[0152] Step S8: adjusting the first triage level based on the comprehensive score to obtain a second triage level.

[0153] Specifically, the higher the comprehensive score, the more serious the patient's condition, and the higher the triage level required by the patient. The comprehensive score is compared with the preset score of the first triage level to obtain the score difference. If the comprehensive score is higher than the preset score of the first triage level, and the score difference is greater than the preset difference, then the first triage level is increased by one level to obtain the second triage level. Alternatively, if the comprehensive score is lower than the preset score of the first triage level, and the score difference is greater than the preset difference, then the first triage level is reduced by one level to obtain the second triage level. Therefore, according to the triage level recommendation reflected by the comprehensive score, the first triage level is increased or decreased to obtain a second triage level with a more accurate triage strategy.

[0154] In this embodiment, by quantifying the patient's vital signs, chronic disease control, emotional state, and activity level, key features are screened out in combination with L1 regularization, and a model based on the random forest algorithm is constructed to calculate the weight coefficient, and then the patient's comprehensive score is calculated, and the patient's triage level is adjusted according to the comprehensive score. The patient's health status can be evaluated more accurately to improve the accuracy of triage.

[0155] Furthermore, in some specific implementations, a reasonable preset time interval is first set to periodically obtain the patient's vital sign data, and the preset time interval can be adjusted according to specific needs or the patient's condition. After each preset time interval, the patient's latest vital sign data is re-obtained through the smart wearable device, input into the triage level prediction model, and the latest triage level is calculated.

[0156] By regularly obtaining the patient's latest vital signs data and recalculating and updating the patient's triage level in combination with the chronic disease management score, the triage level can be adjusted dynamically and in a timely manner according to changes in the patient's health, ensuring that the triage level is more accurate.

[0157] Furthermore, in order to achieve the above-mentioned object, the present invention also provides a pre-examination and triage device, which may include:

[0158] Data acquisition module, used to obtain patients' chronic disease management data and current vital signs data;

[0159] Chronic disease analysis module, which is used to calculate the chronic disease management score based on the linear weighted algorithm and chronic disease management data;

[0160] The triage calculation module is used to input the chronic disease management score and current vital signs data into the triage level prediction model to obtain the patient's first triage level.

[0161] It should be noted that the functions that can be realized by each module in the pre-examination and triage device provided in this embodiment and the corresponding technical effects achieved can refer to the description of the specific implementation methods in each embodiment of the pre-examination and triage method of the present invention. For the sake of brevity of the specification, they will not be repeated here.

[0162] In addition, an embodiment of the present invention further proposes a computer-readable storage medium, on which a pre-examination and triage program is stored, and when the pre-examination and triage program is executed by a processor, the steps of the pre-examination and triage method as described above are implemented. Therefore, it will not be repeated here. In addition, the description of the beneficial effects of adopting the same method will not be repeated. For technical details not disclosed in the computer-readable storage medium embodiment involved in the present invention, please refer to the description of the method embodiment of the present invention. As an example, the program instructions can be deployed to be executed on one computing device, or on multiple computing devices located at one location, or on multiple computing devices distributed at multiple locations and interconnected by a communication network.

[0163] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a computer-readable storage medium (such as ROM / RAM, magnetic disk, optical disk) as above, and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods of various embodiments of the present invention.

[0164] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A pre-examination and triage method, characterized in that: The pre-examination and triage method comprises the following steps: S1, obtain the patient's chronic disease management data and current vital signs data; S2, calculating a chronic disease management score based on a linear weighted algorithm and the chronic disease management data; S3, inputting the chronic disease management score and the current vital sign data into a triage level prediction model to obtain the first triage level of the patient.

2. The pre-examination and triage method according to claim 1, characterized in that: The chronic disease management data includes the chronic disease type of the patient and the chronic disease characteristic value corresponding to the chronic disease type. S2 specifically includes: Based on the chronic disease type of the patient, query a preset weight coefficient table to determine the weight coefficient corresponding to the chronic disease characteristic value; the preset weight coefficient table has weight coefficients corresponding to different chronic disease characteristic values, and the weight coefficients are obtained based on one or more methods of regression analysis and machine learning; Based on the weight coefficient, the chronic disease characteristic values ​​are weighted and summed to obtain the chronic disease management score.

3. The pre-examination and triage method according to claim 1, characterized in that: In S3, the steps of constructing the triage level prediction model specifically include: Obtain historical vital signs data and historical chronic disease management data of multiple patients with different types of chronic diseases; Based on statistical methods, performing fluctuation analysis on the historical vital signs data to obtain historical vital signs fluctuation parameters; Based on the linear weighted algorithm and historical chronic disease management data, the historical chronic disease management score is calculated; Preprocessing the historical chronic disease management scores and the historical vital sign fluctuation parameters using a Z-score standardization method to obtain the first training set with unified dimensions; Building a feature extraction layer based on a multi-layer perceptron model to extract features from the first training set; The features are weighted based on the self-attention mechanism to obtain a weighted feature matrix; the attention weight calculation formula is: ； in, is the attention weight matrix, Q is the query matrix, K is the key matrix, V is the value matrix, d k is the dimension of the key vector; Calculate the difference between the predicted triage grade probability and the actual triage grade, and construct the cross entropy loss function: ； in, is the loss value, is the actual triage level, To predict the probability of triage level, is the number of triage level categories; Based on the cross entropy loss function and the weighted feature matrix, the deep learning model is trained to obtain the triage level prediction model.

4. The pre-examination and triage method according to claim 1, characterized in that: After S3, the method further includes: S4, obtaining the patient's chronic disease control status, emotional state and activity level; S5, based on preset quantitative standards, quantify the current vital signs, the chronic disease control status, the emotional state and the activity level respectively to obtain a vital sign score, a chronic disease control score, an emotional state score and an activity level score; S6, inputting the vital sign score, chronic disease control score, emotional state score, and activity level score into a score weight calculation model to obtain a dynamic score weight coefficient; S7, performing weighted average calculation based on the vital sign score, the chronic disease control score, the emotional state score, the activity score and the dynamic score weight coefficient to obtain a comprehensive score; S8: Adjust the first triage level based on the comprehensive score to obtain a second triage level.

5. The pre-examination and triage method according to claim 4, characterized in that: In S6, the step of constructing the scoring weight calculation model specifically includes: Acquire historical vital sign data, historical emotional state data, historical chronic disease control data, historical activity data, and historical triage levels of multiple patients with different types of chronic diseases and perform preprocessing to obtain first preprocessing data; Using L1 regularization, features related to the historical triage level are screened out from the first preprocessed data to construct a second training set; Based on the second training set and the random forest algorithm, model training is performed to obtain a scoring weight calculation model.

6. A pre-examination and triage device, characterized in that: The device comprises: Data acquisition module, used to obtain patients' chronic disease management data and current vital signs data; A chronic disease analysis module, used to calculate a chronic disease management score based on a linear weighted algorithm and the chronic disease management data; The triage calculation module is used to input the chronic disease management score and the current vital sign data into a triage level prediction model to obtain the first triage level of the patient.

7. A pre-examination and triage device, characterized in that: The device comprises: a memory, a processor, and a pre-examination and triage program stored in the memory and executable on the processor, wherein the pre-examination and triage program is configured to implement the steps of the pre-examination and triage method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a pre-examination and triage program, and when the pre-examination and triage program is executed by the processor, the steps of the pre-examination and triage method according to any one of claims 1 to 5 are implemented.

Images