Hospital information intercommunication system and method

Abstract

The invention relates to the technical field of hospital information automation, and discloses a hospital information intercommunication system and method. The hospital information intercommunication system comprises a grading model training module used for acquiring historical patient information and training a grading model based on a CNN neural network model; the application number acquisition module is used for acquiring an application number at the mobile terminal; the priority acquisition module is used for acquiring the priority of the patient; the transfer duration prediction module is used for predicting the transfer time required by the patient to arrive at the window position; the dynamic scheduling module is used for acquiring all the application numbers in the transfer time, and optimizing by taking the queuing window and the queuing order of each application number as optimization targets through a grey wolf optimization algorithm to obtain an optimal queuing scheme; and the number calling module is used for generating a queuing number of the patient according to the optimal queuing scheme when the patient arrives at the window position, and sending the queuing number to the mobile terminal and the window management end, so that intelligent intercommunication of hospital information is realized, the patient queuing process is optimized, and the medical service efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of hospital information automation, and in particular to a hospital information intercommunication system and method. Background Art

[0002] The medical and health industry has always been an industry closely related to people's lives. In recent years, the medical system has developed rapidly, the scale of hospitals has expanded rapidly, the number of people seeking medical treatment has increased day by day, and the workload of windows of various functional types is very large. Problems such as disorderly queues in hospitals, long queues, long waiting times for examinations, and inconvenient medical processes are particularly prominent. These problems have, to a certain extent, become the main influencing factors for the inability of hospitals to operate healthily and the low patient satisfaction.

[0003] In order to solve the above problems, some automatic triage solutions divide patients into two levels: high priority and low priority, and then jump the queue for high priority patients; some solutions divide the windows into special population windows and general population windows, and different categories of patients queue up for treatment at different windows; no matter which solution is used, there are problems of poor fairness, waste of resources and low efficiency. Summary of the invention

[0004] The present invention aims to provide a hospital information intercommunication system and method, which can automatically perform priority classification according to patient information after queuing application on a mobile terminal, obtain the optimal queuing plan based on the Grey Wolf Optimization Algorithm, obtain the queuing number according to the optimal queuing plan, and send the queuing number to the window management end for calling, so as to realize the intelligent intercommunication of hospital information, optimize the patient queuing process, and improve the efficiency of medical services.

[0005] In order to achieve the above object, the present invention adopts the following technical scheme:

[0006] A hospital information intercommunication system, comprising:

[0007] The classification model training module is used to obtain historical patient information and train a classification model based on the CNN neural network model, taking the patient information as input and the priority as output;

[0008] The application number acquisition module is used to make a virtual queue application on a mobile terminal and obtain an application number;

[0009] A priority acquisition module, used for acquiring the priority of the patient through a grading model according to the patient information;

[0010] The transfer time prediction module predicts the transfer time required for the patient to reach the window location based on the patient's video information and navigation information;

[0011] The dynamic scheduling module is used to obtain all application numbers within the transfer time, and optimize the queue window and queue position of each application number through the gray wolf optimization algorithm to obtain the optimal queue plan;

[0012] The calling module is used to generate the patient's queue number according to the optimal queuing plan when the patient arrives at the window position, and send it to the mobile terminal and the window management end.

[0013] The principle and advantage of this solution are: in actual application, by training the hierarchical model, it is convenient to apply to various types of windows and obtain the priority more accurately. Obtaining the application number on the mobile terminal can enter the queue sequence in advance on the one hand, and facilitate the global optimization of the system on the other hand. In the existing scheme, there is a situation where the patient is not on the scene when the number is called. Repeated calling will affect the overall work efficiency. After skipping the number, the patient needs to queue up again or jump in line, which affects the overall fairness and the patient's experience; through the transfer time prediction, it is convenient to ensure that the patient's queue number is effective after the patient arrives at the window, improve the global efficiency, and optimize the queuing scheme of all application numbers within the transfer time. When the patient arrives at the window position, the queue number is regenerated, which can improve the global optimization efficiency while ensuring fairness. In this application, based on the window status information and the patient's priority, the gray wolf optimization model is used to optimize the queue window and queue position of each application number, which can further improve the global efficiency and fairness.

[0014] Preferably, as an improvement, the hierarchical model training module further includes:

[0015] The text preprocessing submodule is used to obtain patient information text and perform denoising and word segmentation processing;

[0016] The vocabulary building submodule is used to assign a unique vocabulary ID to each vocabulary and build a vocabulary;

[0017] The sequence encoding submodule is used to create a vector for each word in the vocabulary. The length of the vector is equal to the size of the vocabulary, and all positions except the corresponding word are 1 and all other positions are 0;

[0018] Input data construction submodule: used to convert each sentence in the patient information text into a sequence consisting of vocabulary IDs.

[0019] Technical effect: It is easy to complete the encoding process of scattered and complex patient information text, so that the patient information text can be processed by the CNN neural network model.

[0020] Preferably, as an improvement, the hierarchical model training module further includes:

[0021] Window monitoring submodule, used to obtain window status information in real time;

[0022] The window utilization acquisition submodule is used to obtain the window utilization according to the analysis of the window status information;

[0023] The priority level determination submodule is used to dynamically adjust the priority level according to the window utilization.

[0024] Technical effect: It is easy to determine a reasonable priority level, making the CNN neural network model training more accurate.

[0025] Preferably, as an improvement, the dynamic scheduling module further includes:

[0026] The queuing scheme optimization model creation submodule is used to obtain window feature information, match the model parameters required according to the window feature information, create a queuing scheme optimization model in combination with the gray wolf optimization algorithm, and iterate;

[0027] The queuing scheme optimization model triggering submodule is used to obtain virtual queuing applications in real time and trigger the creation and iteration of the queuing scheme optimization model based on the virtual queuing applications.

[0028] Technical effect: Windows of different functional types have different corresponding priority urgency and queuing requirements. Model parameters are matched according to window characteristics to improve the adaptability of the queuing scheme optimization model. Taking into account the dynamic changes of the queue over time, the creation and iteration of the queuing scheme optimization model is triggered when a new application number is generated, so as to improve the timeliness of the scheme update.

[0029] Preferably, as an improvement, the queuing scheme optimization model creation submodule includes an initialization molecule module, which is used to define the wolf pack search space, set parameters, randomly generate wolf packs and initialize leaders α, β, and δ wolves; the definition of the wolf pack search space includes defining the value range of the queuing scheme, the queuing scheme corresponds to gray wolf individuals, and the position vector of each gray wolf individual includes a window number and a queue position; the setting parameters includes setting the wolf pack size, number of iterations, convergence factor and coordination coefficient vector.

[0030] Technical effect: The wolf pack individuals correspond to the queuing scheme, which makes it easier to determine the search starting point through initialization and promotes global search.

[0031] Preferably, as an improvement, the queuing scheme optimization model creation submodule further includes a fitness function determination numerator module, which is used to determine the fitness function according to the revenue, and the fitness function is:

[0032] T all =w1Σ(y ix *μ i )+w2*H t +w3*F t

[0033] Among them, y ix is the waiting time of patient x with priority i, μ i is the waiting time influence constant of patients with priority i, w1, w2, w3 are weight values, H t represents the total number of service visits at each window within time period t, F t is the service time difference value.

[0034] Technical effect: It is easy to minimize the overall waiting time and maximize the window utilization, improving the overall patient satisfaction and window service efficiency.

[0035] Preferably, as an improvement, the service time difference value calculation model is:

[0036]

[0037] Among them, μ is the fitness adjustment constant value, h it is the number of patients with priority i, y ix is the waiting time of patient x with priority i, is the average waiting time for patients with priority i, N i is the priority number.

[0038] Technical effect: It is easy to balance the waiting time of patients of different priorities and further improve fairness.

[0039] Preferably, as an improvement, the queuing scheme optimization model creation submodule further includes a boundary determination molecular module for determining a boundary function, wherein the boundary function includes:

[0040] Position mi ≤Position m(i-1)

[0041] Nu m ≤M

[0042] t x ∈t g

[0043] Among them, Position mi is the rank of the i-priority patient in window m, Nu m is the service load of window m, M is the upper limit of the service load of window m; t x Time to complete service for patient x, t g The prescribed completion time for patient x.

[0044] Technical effect: It is convenient to ensure that patients with high priority in this round are ahead of patients with low priority, and at the same time ensure that patients can complete the service within the specified time and the window load does not exceed the upper limit.

[0045] Preferably, as an improvement, it also includes:

[0046] Information exchange module, used to update and share patient information, doctor schedule information, and drug inventory information in real time;

[0047] The intelligent management module is used to synchronously update the training classification model and the queuing scheme optimization model when the patient information, doctor scheduling information, and drug inventory information are updated; it is also used to send reminder information to the mobile terminal.

[0048] Technical effect: easy to improve management efficiency.

[0049] Also included is a hospital information intercommunication method, which is used in a hospital information intercommunication system. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 The figure is a schematic diagram of the structure of a hospital information communication system. DETAILED DESCRIPTION

[0051] The following is further described in detail through specific implementation methods:

[0052] The embodiment is basically as shown in the attached Figure 1 As shown:

[0053] A hospital information intercommunication system, comprising:

[0054] The classification model training module is used to obtain historical patient information and train a classification model based on the CNN neural network model, taking patient information as input and priority as output. The classification model can help medical staff quickly identify which patients need immediate treatment and which patients can be treated later.

[0055] The historical patient information that needs to be collected for training the classification model includes basic personal information, symptom information, vital signs information, test results information, past medical history information, socioeconomic factors and psychological status information. Specifically, considering that patients of different age groups face different health risks, some diseases have different incidences or manifestations in different genders; the body mass index (BMI) needs to be calculated to understand the nutritional status of the patient. Basic personal information includes: age, gender, weight and height. Symptom information includes the patient's main symptom description, such as chest pain, dyspnea, etc.; symptom duration; accompanying symptoms, that is, whether there are other related symptoms, such as fever, nausea, etc. Vital sign information includes: body temperature, blood pressure, heart rate, respiratory rate, oxygen saturation, abnormal body temperature indicates infection or other disease states, high blood pressure or low blood pressure is a sign of serious illness, abnormal heart rate indicates heart problems, shortness of breath indicates lung problems, and low oxygen saturation indicates respiratory problems. Test result information includes: blood tests, such as white blood cell count, hemoglobin level, etc.; urine analysis, used to detect kidney function or urinary system problems; imaging test results, such as X-rays, CT scans, etc. Past medical history includes chronic diseases, allergies, and surgical history. Chronic diseases include diabetes and heart disease. Allergy history makes it easy to understand whether the patient is allergic to certain drugs or substances. Surgical history makes it easy to understand whether the patient has undergone surgery and its impact. Socioeconomic factors include living environment, occupation, and insurance coverage. Living environment makes it easy to judge the impact in combination with the patient's living conditions; certain occupations increase the risk of illness; insurance coverage makes it easy to understand whether the patient can afford the cost of treatment. Psychological state includes anxiety level and depression symptoms. Anxiety can affect the patient's subjective feelings and treatment compliance; depression can affect the patient's physical health. The above information can comprehensively and accurately identify key features, thereby providing personalized solutions and rapid responses.

[0056] The hierarchical model training module also includes: a text preprocessing submodule, which is used to obtain patient information text and perform denoising and word segmentation processing; specifically, denoising processing is to remove noise, delete irrelevant characters (such as punctuation marks), numbers or HTML tags, etc.; word segmentation processing: split sentences into words or tags; also includes case conversion processing: converting text to lowercase to reduce the size of the vocabulary; stop word removal processing: removing common meaningless words. Vocabulary construction submodule, used to assign a unique vocabulary ID to each vocabulary and build a vocabulary. Sequence encoding submodule, used to create a vector for each word in the vocabulary, the vector length is equal to the vocabulary size, except for the position of the corresponding word is 1, the rest of the positions are 0. Input data construction submodule: used to convert each sentence in the patient information text into a sequence composed of vocabulary IDs, and combine multiple sentences into a batch for neural network processing.

[0057] The hierarchical model training module also includes: a window monitoring submodule, which is used to obtain window status information in real time. The window status information includes the window opening status, such as opening time, closing time, and opening duration, and also includes service records, such as the time when each application number starts to receive service, the time when each application number completes service, and the time spent by each application number during the service; it also includes queuing information, such as queue length, average waiting time, and idle time; a window utilization acquisition submodule, which is used to obtain window utilization based on window status information analysis; window utilization is the busyness of each service window over a period of time. The level of window utilization reflects the efficiency of window resource utilization. The window utilization is the proportion of the total working time of the window to the available time. The total working time of the window is the sum of the time when the window is in the open state and is providing services. The available time of the window is the sum of the time when the window is in the open state, including service time and idle time. The priority level determination submodule is used to dynamically adjust the priority level according to the window utilization rate, pre-store dynamic adjustment rules, and trigger dynamic adjustment when the window utilization rate changes beyond a threshold. Reasonable allocation of priority levels facilitates significant improvement in patient waiting time and service efficiency. In this embodiment, when the window utilization rate is low, the priority level is reduced to simplify the queuing process. When the window utilization rate is high, the priority level is increased to subdivide the patient group and improve service efficiency.

[0058] The application number acquisition module is used to apply for virtual queuing on the mobile terminal and obtain the application number. By obtaining the application number, it is convenient to enter the queue sequence in advance and improve the global optimization ability of the system. The priority acquisition module is used to obtain the patient's priority through the grading model according to the patient information.

[0059] The transfer time prediction module predicts the transfer time required for the patient to reach the window position based on the patient's video information and navigation information; obtains the patient's moving speed based on the patient's moving video information in the hospital, obtains the patient's total moving path length based on the navigation information, and predicts the transfer time required for the patient to reach the window position based on the speed and path length. In this embodiment, in order to make the prediction more accurate, the transfer time can also be predicted in sections based on the hospital layout such as corridors, elevators, stairs, etc., and finally the sum of the time is calculated as the required transfer time.

[0060] The dynamic scheduling module is used to obtain all application numbers within the transfer time, and optimize the queuing window and queuing position of each application number by the gray wolf optimization algorithm to obtain the optimal queuing solution; the dynamic scheduling module also includes:

[0061] The queuing scheme optimization model creation submodule is used to obtain window feature information, which includes window function and window quantity, match the model parameters required in the pre-stored parameter query table according to the window feature information, create a queuing scheme optimization model in combination with the Gray Wolf Optimization Algorithm, and iterate.

[0062] The queuing scheme optimization model creation submodule includes an initialization molecule module, which is used to define the wolf pack search space, set parameters, randomly generate wolf packs and initialize leaders α, β, and δ wolves; the definition of the wolf pack search space includes defining the value range of the queuing scheme, the queuing scheme corresponds to the gray wolf individual, and the position vector of each gray wolf individual includes the window number and the queue position. The window number is the window number assigned to each application number, and the queue position is the queue order in the specified window. If there are (x) application numbers that need to be queued in (i) windows, then the position vector of each "wolf" is a (2x)-dimensional vector, in which the first (x) elements represent the window number corresponding to each application number, and the second (x) elements represent the queue position of each application number in the specified window; the setting parameters includes setting the wolf pack size, the number of iterations, the convergence factor and the coordination coefficient vector.

[0063] The queuing scheme optimization model creation submodule also includes a fitness function determination numerator module, which is used to determine the fitness function according to the income. The fitness function is:

[0064] T all =w1∑(y ix *μ i )+w2*H t +w3*F t

[0065] Among them, y ix is the waiting time of patient x with priority i, μ i is the waiting time influence constant of patients with priority i, w1, w2, w3 are weight values, H t represents the total number of service visits at each window within time period t, F t is the service time difference value. Through the above fitness function, it is convenient to minimize the overall waiting time and maximize the window utilization, thereby improving the overall patient satisfaction and window service efficiency.

[0066] The service time difference value calculation model is:

[0067]

[0068] Among them, μ is the fitness adjustment constant value, h it is the number of patients with priority i, y ix is the waiting time of patient x with priority i, is the average waiting time for patients with priority i, N i is the priority number.

[0069] The queuing scheme optimization model creation submodule also includes a boundary determination molecular module for determining a boundary function, and the boundary function includes:

[0070] Position mi ≤Position m(i-1)

[0071] Nu m ≤M

[0072] t x ∈t g

[0073] Among them, Position ni Nu is the rank of the i-priority patient in window m, where i-priority is greater than i-1. Through the constraint condition, it is easy to ensure that in the current configuration, the high-priority patient ranks before the low-priority patient; Nu m is the service load of window m, M is the upper limit of the service load of window m, and defines the upper limit of the service capacity of each window to ensure that the service request of the window does not exceed its capacity at any time; t x Time to complete service for patient x, t g The specified completion time for patient x gives users with strict time requirements a higher priority to ensure that they can complete the service within the specified time.

[0074] The queuing plan optimization model triggering submodule is used to obtain virtual queuing applications in real time, and trigger the creation and iteration of the queuing plan optimization model based on the virtual queuing application. Taking into account the dynamic changes of the queue over time, the creation and iteration of the queuing plan optimization model is triggered when a new application number is generated, so as to improve the timeliness of the plan update.

[0075] The calling module is used to generate the patient's queue number according to the optimal queuing plan when the patient arrives at the window position, and send it to the mobile terminal and the window management end. In order to further improve the information exchange capability, it also includes: an information exchange module for real-time update and sharing of patient information, doctor scheduling information, and drug inventory information; an intelligent management module for synchronously updating the training classification model and the queuing plan optimization model when the patient information, doctor scheduling information, and drug inventory information are updated; and it is also used to send reminder information to the mobile terminal.

[0076] Also included is a hospital information intercommunication method, which is used in a hospital information intercommunication system.

[0077] The above is only an embodiment of the present invention, and the common knowledge such as the known specific technical solutions and / or characteristics in the solution is not described in detail here. It should be pointed out that for those skilled in the art, without departing from the technical solution of the present invention, several modifications and improvements can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A hospital information communication system, characterized in that: include: The classification model training module is used to obtain historical patient information and train a classification model based on the CNN neural network model, taking the patient information as input and the priority as output; The application number acquisition module is used to make a virtual queue application on a mobile terminal and obtain an application number; A priority acquisition module, used for acquiring the priority of the patient through a grading model according to the patient information; The transfer time prediction module predicts the transfer time required for the patient to reach the window location based on the patient's video information and navigation information; The dynamic scheduling module is used to obtain all application numbers within the transfer time, and optimize the queue window and queue position of each application number through the gray wolf optimization algorithm to obtain the optimal queue plan; The calling module is used to generate the patient's queue number according to the optimal queuing plan when the patient arrives at the window position, and send it to the mobile terminal and the window management end.

2. A hospital information intercommunication system according to claim 1, characterized in that: The hierarchical model training module also includes: The text preprocessing submodule is used to obtain patient information text and perform denoising and word segmentation processing; The vocabulary building submodule is used to assign a unique vocabulary ID to each vocabulary and build a vocabulary; The sequence encoding submodule is used to create a vector for each word in the vocabulary. The length of the vector is equal to the size of the vocabulary, and all positions except the corresponding word are 1 and all other positions are 0; Input data construction submodule: used to convert each sentence in the patient information text into a sequence consisting of vocabulary IDs.

3. A hospital information intercommunication system according to claim 1, characterized in that: The hierarchical model training module also includes: Window monitoring submodule, used to obtain window status information in real time; The window utilization acquisition submodule is used to obtain the window utilization according to the analysis of the window status information; The priority level determination submodule is used to dynamically adjust the priority level according to the window utilization.

4. A hospital information intercommunication system according to claim 1, characterized in that: The dynamic scheduling module also includes: The queuing scheme optimization model creation submodule is used to obtain window feature information, match the model parameters required according to the window feature information, create a queuing scheme optimization model in combination with the gray wolf optimization algorithm, and iterate; The queuing scheme optimization model triggering submodule is used to obtain virtual queuing applications in real time and trigger the creation and iteration of the queuing scheme optimization model based on the virtual queuing applications.

5. A hospital information intercommunication system according to claim 4, characterized in that: The queuing scheme optimization model creation submodule includes an initialization molecule module, which is used to define the wolf pack search space, set parameters, randomly generate wolf packs and initialize leaders α, β, and δ wolves; the definition of the wolf pack search space includes defining the value range of the queuing scheme, the queuing scheme corresponds to gray wolf individuals, and the position vector of each gray wolf individual includes a window number and a queue position; the setting parameters includes setting the wolf pack size, number of iterations, convergence factor and coordination coefficient vector.

6. A hospital information intercommunication system according to claim 4, characterized in that: The queuing scheme optimization model creation submodule also includes a fitness function determination numerator module, which is used to determine the fitness function according to the income. The fitness function is: T all =w1∑(y ix *μ i )+w2*H t +w3*F t Among them, y ix is the waiting time of patient x with priority i, μ i is the waiting time influence constant of patients with priority i, w1, w2, w3 are weight values, H t represents the total number of service visits at each window within time period t, F t is the service time difference value.

7. A hospital information intercommunication system according to claim 6, characterized in that: The service time difference value calculation model is: Among them, μ is the fitness adjustment constant value, h it is the number of patients with priority i, y ix is the waiting time of patient x with priority i, is the average waiting time for patients with priority i, N i is the priority number.

8. A hospital information intercommunication system according to claim 4, characterized in that: The queuing scheme optimization model creation submodule also includes a boundary determination molecular module for determining a boundary function, and the boundary function includes: Position mi ≤Position m(i-1) Not m ≤M t x ∈t g Among them, Position mi is the rank of the i-priority patient in window m, Nu m is the service load of window m, M is the upper limit of the service load of window m; t x Time to complete service for patient x, t g The prescribed completion time for patient x.

9. A hospital information intercommunication system according to claim 8, characterized in that: Also includes: Information exchange module, used to update and share patient information, doctor schedule information, and drug inventory information in real time; The intelligent management module is used to synchronously update the training classification model and the queuing scheme optimization model when the patient information, doctor scheduling information, and drug inventory information are updated; it is also used to send reminder information to the mobile terminal.

10. A hospital information exchange method, characterized by: Used in a hospital information communication system as described in claims 1-9.

Images